D B Dengue - apps.searo.who.intapps.searo.who.int/pds_docs/B4325.pdf · attachment to the Editor at...

Dengue Bulletin – Volume 32, 2008 iii

Acknowledgements

Editor, Dengue Bulletin, WHO/SEARO, gratefully thanks the following for peer reviewing manuscriptssubmitted for publication.

In-house Review

Nand L. Kalra: Reviewed the manuscripts in respect of format check, content, conclusions drawn,including condensation of tabular and illustrative materials for clear, concise and focused presentationand bibliographic references. He was also involved in the final stages of printing of the Bulletin.

Peer reviewers

1. Linda S. LloydPublic Health Consultant3443 Whittier St.San Diego, CA 92106, USAE-mail: [email protected]

2. Philip McCallVector Research GroupLiverpool Tropical School of MedicinePembroke PlaceLiverpool L3 5QA, UKE-mail: [email protected]

3. Jeffrey L. LennonLiberty UniversityDepartment of Health Sciences and KinesiologyLynchburg, VA 24502, USAE-mail: [email protected]

4. Alaka SinghHealth System FinancingDepartment of Health Systems DevelopmentRegional Office for South-East AsiaWorld Health OrganizationI.P. EstateNew Delhi – 110002, IndiaE-mail: [email protected]

5. Khanchit LimpakarnjanaratCommunicable Diseases Surveillance and ResponseCommunicable Diseases DepartmentRegional Office for South-East AsiaWorld Health OrganizationI.P. Estate, New Delhi – 110002, IndiaE-mail: [email protected]

6. Olaf HorstickSpecial Programme for Research and Training in Tropical Diseases (TDR)World Health OrganizationAvenue Appia 201211 Geneva 27, SwitzerlandE-mail: [email protected]

7. Tassanee SilawanThe Rural Health Training & Research CenterFaculty of Public HealthMahidol University30/1 Mitsampan Road, Soong Noen DistrictNakhonratchasima Province, Thailand 30170E-mail: [email protected]

8. John AaskovWHO Collaborating Centre for Arbovirus Reference and ResearchQueensland University of TechnologyBrisbane, AustraliaE-mail: [email protected]

9. Roberto BarreraDengue Branch, DVBIDCenters for Disease Control and Prevention1324 Calle CañadaSan Juan, Puerto RicoE-mail: [email protected]

10. Tai Ji ChoongEnvironmental Health DepartmentNational Environment Agency40 Scotts RoadEnvironment Building #13-00Singapore 228231E-mail: [email protected]

iv Dengue Bulletin – Volume 32, 2008

11. Andrew Keith FalconarDepartment of Infectious & Tropical DiseasesLondon School of Hygiene & Tropical MedicineKeppel StreetLondon WC1E 7HT, UKE-mail: [email protected]

12. Anuja MathewCenter for Infectious Disease & Vaccine ResearchUniversity of Massachusetts Medical School55 Lake Avenue NorthWorcester, MA 01655, USAE-mail: [email protected]

13. Sander KoenraadtLaboratory of EntomologyWageningen UniversityP.O. Box 80316700 EH Wageningen, The NetherlandsE-mail: [email protected]

14. Vincent CorbelCentre de Recherches Entomologiques de Cotonou (CREC)Institut de Recherche pour leDéveloppement (IRD)01 BP 4414 RP, Cotonou Benin, FranceE-mail: [email protected]

15. Denise ValleLaboratório de Fisiologia e Controle de Artrópodes VetoresInstituto Oswaldo Cruz, FIOCRUZRio de Janeiro, RJ, BrazilE-mail: [email protected]

16. To SethaNational Dengue Control ProgrammeMinistry of HealthPhnom Penh, CambodiaE-mail: [email protected]

17. Sokrin KhunNational Centre for Health PromotionMinistry of HealthPhnom Penh, CambodiaE-mail: [email protected]

18. Dave ChadeeDepartment of Life SciencesUniversity of the West IndiesSt. Augustine, Trinidad, West IndiesE-mail: [email protected]

19. Polly LeungDepartment of Health Technology and InformaticsThe Hong Kong Polytechnic UniversityHung Hom, KowloonHong KongE-mail: [email protected]

20. Asif M. KhanDepartment of MicrobiologyYong Loo Lin School of MedicineNational University of Singapore5 Science Drive 2, Singapore 117597E-mail: [email protected]

21. Pei-Yun ShuResearch and Diagnostic CenterCenters for Disease ControlDepartment of HealthTaipei, Taiwan, Republic of ChinaE-mail: [email protected]

22. Duncan R. SmithMolecular Pathology LaboratoryInstitute of Molecular Biology and GeneticsMahidol University, Salaya Campus25/25 Phuttamontol Sai 4, SalayaNakorn Pathom, 73170, ThailandE-mail: [email protected]

23. Sushil Kumar KabraPediatric Pulmonology DivisionDepartment of PediatricsAll India Institute of Medical SciencesNew Delhi 110029, IndiaE-mail: [email protected]

24. Siripen KalayanaroojWHO Collaborating Centre for Case Management of Dengue/DHF/DSSQueen Sirikit National Institute of Child Health (Children’s Hospital)Bangkok, ThailandE-mail: [email protected]

25. Guey Chuen PerngDepartment of Pathology and Laboratory MedicineEmory Vaccine CenterSchool of MedicineEmory UniversityAtlantaGeorgia 30322, USAE-mail: [email protected]

Dengue Bulletin – Volume 32, 2008 v

26. Brian KayQueensland Institute of Medical ResearchPost Office Royal Brisbane HospitalBrisbane, QLD 4029AustraliaE-mail: [email protected]

27. Craig WilliamsSansom InstituteSchool of Pharmacy & Medical SciencesUniversity of South AustraliaGPO Box 2471 AdelaideSouth Australia 5000E-mail: [email protected]

28. Jennifer KyleDivision of Infectious DiseasesSchool of Public HealthUniversity of CaliforniaBerkeley, 1 Barker Hall #424Berkeley, CA 94720-7354, USAE-mail: [email protected]

29. Michael A. JohanssonDengue BranchDivision of Vector-Borne Infectious DiseasesCenters for Disease Control and PreventionSan Juan, Puerto RicoE-mail: [email protected]

30. Han-Chung WuInstitute of Cellular and Organismic BiologyAcademia Sinica128 Academia Road, Section 2, NankangTaipei, 115, TaiwanE-mail: [email protected]

31. Scott HalsteadUniformed Services University of the Health SciencesBethesda, Maryland, USAE-mail: [email protected]

32. Jackie DeenInternational Vaccine InstituteSan 4-8 Bongcheon-7-dong, Kwanak-guSeoul 151-818, KoreaE-mail: [email protected]

33. Nguyen Thanh HungChildren’s Hospital Number OneHo Chi Minh City, Viet NamE-mail: [email protected]

34. Eric MartínezPedro Kouri Institute of Tropical MedicineHavana City, CubaE-mail: [email protected]

35. Lucy Lum Chai SeeUniversity of Malaya Medical CenterKuala Lumpur, MalaysiaE-mail: [email protected]

36. V.K. SaxenaCentre for Medical Entomology & Vector ManagementandDivision of Malariology and CoordinationNational Institute of Communicable Diseases22 Shamnath MargDelhi 110054, IndiaE-mail: [email protected]

37. Martin GrobuschInfectious Diseases UnitDivision of Clinical Microbiology and Infectious DiseasesNational Health Laboratory ServiceandSchool of PathologyFaculty of Health SciencesUniversity of the Witwatersrand7 York Road, Parktown 2196Johannesburg, South AfricaE-mail: [email protected]

38. Jane CardosaInstitute of Health and Community MedicineUniversiti Malaysia SarawakKota SamarahanSarawak, MalaysiaE-mail: [email protected]

39. Allison ImrieUniversity of Western AustraliaDiscipline of Microbiology and Immunology M502School of Biomedical, Biomolecular and Chemical Sciences35 Stirling HighwayCrawley WA 6009AustraliaE-mail: [email protected]

vi Dengue Bulletin – Volume 32, 2008

40. I-Ming TangE-mail: [email protected]

41. Lourdes EstevaDepartamento de MatemáticasFacultad de Ciencias, UNAMCircuito ExteriorMéxico, D.F. 04510E-mail: [email protected]

42. Philippe DussartLaboratoire de VirologieCNR arbovirus et virus influenza pour la région Antilles GuyaneInstitut Pasteur de la Guyane23, Avenue Pasteur - BP 601097306 Cayenne CedexE-mail: [email protected]

43. Rita NogueiraLaboratory of FlavivirusInstitute Oswaldo Cruz, FIOCRUZRio de Janeiro, BrazilE-mail: [email protected]

44. Ananda AmarasinghePediatric Dengue Vaccine InitiativeInternational Vaccine InstituteSan 4-8 Bongcheon-7-dong, Kwanak-guSeoul 151-818, KoreaE-mail: [email protected]

45. Christophe LagneauDirecteur recherche et développementEID Méditerranée165, Avenue Paul RimbaudF-34184 Montpellier Cedex 4FranceE-mail: [email protected]

46. Zairi JaalVector Control Research UnitSchool of Biological SciencesUniversiti Sains Malaysia11800 PenangMalaysiaE-mail: [email protected]

47. Michael NathanE-mail: [email protected]

48. Morteza ZaimWHO Pesticide Evaluation Scheme (WHOPES)Vector Ecology & ManagementDepartment of Control of Neglected Tropical DiseasesWorld Health Organization20 Avenue AppiaCH-1211 Geneva 27, SwitzerlandE-mail: [email protected]

49. Raman VelayudhanVector Ecology and ManagementDepartment of Control of Neglected Tropical DiseasesWorld Health Organization20 Avenue AppiaCH-1211 Geneva 27SwitzerlandE-mail: [email protected]

50. Dr John EhrenbergMalaria, Other Vectorborne and Parasitic DiseasesRegional Office for the Western PacificP.O. Box No. 293212115 ManilaPhilippinesEmail: [email protected]

The quality and scientific stature of the Dengue Bulletin is largely due to the conscientiousefforts of the experts and also due to the positive response of contributors to comments andsuggestions.

Dengue Bulletin – Volume 32, 2008 vii

Contents

1. WHO's efforts for the development of a dengue vaccine ................................ 1Sutee Yoksan

2. Socioeconomic determinants of dengue incidence in Singapore .................. 17Stefan Ma, Eng Eong Ooi, Kee Tai Goh

3. Forecasting dengue incidence in Dhaka, Bangladesh:A time series analysis ..................................................................................... 29M.A.H. Zamil Choudhury, Shahera Banu, M. Amirul Islam

4. Dengue vector surveillance in Hong Kong – 2007 ........................................ 38M.W. Lee, M.Y. Fok

5. Re-emergence of dengue in Argentina: Historical developmentand future challenges .................................................................................... 44Héctor Masuh

6. Duration of short-lived cross-protective immunity against aclinical attack of dengue: A preliminary estimate .......................................... 55Hiroshi Nishiura

7. Discrimination between primary and secondary dengue virus infectionby using an immunoglobulin G avidity test .................................................... 67A. Chakravarti, M. Matlani, A. Kumar

8. DENV-3 genotype III circulating in São Paulo, Brazil,from 2003 to 2008 is not associated with denguehaemorrhagic fever/dengue shock syndrome ................................................ 73Luiza Antunes de Castro-Jorge, Daniel Macedo de Melo Jorge,Benedito Antônio Lopes da Fonseca

9. Application of monoclonal antibody DSSC7 for detectingdengue infection in Aedes aegypti based on immunocytochemicalstreptavidin biotin peroxidase complex assay (ISBPC) .................................... 83Umniyati SR, Sutaryo, Wahyono D, Artama WT, Mardihusodo SJ,Soeyoko, Mulyaningsih B, Utoro T

viii Dengue Bulletin – Volume 32, 2008

Contents

10. Enhancement of MHC class I binding and immunogenic propertiesof the CTL epitope peptides derived from dengue virus NS3protein by anchor residue replacement ........................................................ 99Hideyuki Masaki, Yoshiki Fujii, Kiyohiro Irimajiri, Takanori T. Tomura, Ichiro Kurane

11. Liver function tests in patients with dengue viral infection .......................... 110Rajoo Singh Chhina, Omesh Goyal, Deepinder Kaur Chhina, Prerna Goyal,Raj Kumar, Sandeep Puri

12. Changing clinical manifestations of dengue infection in north India ............ 118Chandrakanta, Rashmi Kumar, Garima, Jyotsana Agarwal, Amita Jain, Rachna Nagar

13. Dengue virus serotype 3 (genotype III) from Colombia: A perspectiveof its pathogenic potential ........................................................................... 126Sergio Yebrail Gómez Rangel, Christian Julián Villabona-Arenas,Flor Angela Torres Pimiento, Daniel Rafael Miranda-Esquivel,Raquel Elvira Ocazionez Jimenez

14. Applied informatics manipulation for fight against dengue ........................... 138Viroj Wiwanitkit

15. Community participation and social engagement in theprevention and control of dengue fever in rural Cambodia ......................... 145Sokrin Khun, Lenore Manderson

16. Dengue in the National Capital Territory (NCT) ofDelhi (India): Epidemiological and entomological profilefor the period 2003 to 2008 ........................................................................ 156J. Nandi, R.S. Sharma, P.K. Dutta, G.P.S. Dhillon

17. Increased utilization of treatment centre facilities duringa dengue fever outbreak in Kolkata, India ................................................... 162Shanta Dutta, Jacqueline L. Deen, Dipika Sur, Byomkesh Manna,Suman Kanungo, Barnali Bhaduri, Anna Lena Lopez, Lorenz von Seidlein,John D. Clemens, Sujit K. Bhattacharya

18. Entomological survey of dengue vectors as basis fordeveloping vector control measures in Barangay Poblacion,Muntinlupa City, Philippines, 2008 .............................................................. 167Estrella Irlandez Cruz, Ferdinand V. Salazar, Elizabeth Porras,Remigio Mercado, Virginia Orais, Juancho Bunyi

Contents

Dengue Bulletin – Volume 32, 2008 ix

19. Aedes survey of selected public hospitals admitting denguepatients in Metro Manila, Philippines .......................................................... 171Estrella Irlandez Cruz, Ferdinand V. Salazar, Wilfredo E. Aure, Elizabeth P. Torres

20. Epidemiological and entomological aspects of an outbreak ofchikungunya in Lakshadweep Islands, India, during 2007 ............................ 178R.S. Sharma, M.K. Showkath Ali, G.P.S. Dhillon

21. Effect of pyriproxyfen in Aedes aegypti populations with differentlevels of susceptibility to the organophosphate temephos ............................ 186Maria Teresa Macoris Andrighetti, Fernanda Cerone, Marcelo Rigueti,Karen Cristina Galvani, Maria de Lourdes da Graça Macoris

22. Effectiveness of pyriproxyfen-controlled release block againstlarvae of Aedes (Stegomyia) aegypti in Kuala Lumpur, Malaysia ................... 199C.D. Chen, W.A. Andy-Tan, S.R. Loke, H.L. Lee, A.R. Yasmin, M. Sofian-Azirun

23. Laboratory evaluation of Mesocyclops aspericornis as abiocontrol agent of Aedes aegypti ............................................................... 207R. Ramanibai, Kanniga S.

Viewpoint

24. Management dilemmas in the treatment of dengue fever ........................... 211Kolitha H Sellahewa

Short notes

25. An outbreak of dengue in Moreh: A small rural town inManipur near Indo-Myanmar border .......................................................... 219Kalpana Baruah, K. Indra Singh, C.M. Agrawal, G.P.S. Dhillon

26. Imported dengue fever cases in Gunma prefecture, Japan .......................... 222Yukio Morita, Tomoyuki Suzuki, Kunihisa Kozawa, Masahiro Noda,Nobuhiko Okabe, Hirokazu Kimura

27 Concurrent dengue fever and bacterial septicemiaduring the 2008 dengue outbreak in Delhi ................................................. 226Subhash C. Arya, Nirmala Agarwal

x Dengue Bulletin – Volume 32, 2008

Contents

28. Fibre-glass drums as the key containers of Aedes aegypti breedingin apartments occupied by expatriates in Jeddah, Saudi Arabia ................... 228A.A.N. Aljawi, T. Mariappan, A. Abo-Khatwa, Khalid M. Al-Ghamdi, Hani M. Aburas

Book reviews

29. Asia-Pacific Dengue Strategic Plan (2008–2015) ......................................... 232

30. Dengue prevention and control .................................................................. 233

31. Six countries test dengue interventions ....................................................... 235

32. Dengue in Africa: Emergence of DENV-3,Côte d'Ivoire, 2008 ..................................................................................... 237

33. Instructions for contributors ......................................................................... 241

Dengue Bulletin – Volume 32, 2008 1

WHO's efforts for the development of a dengue vaccine

Sutee Yoksan#

Centre for Vaccine Development, Institute of Molecular Biosciences, Mahidol University, Thailand

Abstract

Background: Dengue fever and dengue haemorrhagic fever (DF/DHF) are caused by dengue (DENV)viruses. There are four antigenically related, but distinct, DENV serotypes (DENV-1 through DENV-4).Humans are the amplifying vertebrate hosts and Aedes mosquitoes are the primary mosquito vectors aswell as the reservoir of infection.

DENV infections cause a spectrum of diseases, ranging from asymptomatic infections to infectionscomplicated by haemorrhage, shock and death. Infection with DENV of one serotype results in apparentlife-long monotypic immunity against that serotype but not against any other serotype. Thus, separateinfections with all four DENV serotypes are theoretically possible in a single host. It should be noted thatin Thailand, all the four DENV serotypes co-circulate, thereby resulting in multiple exposures and thepotential for re-infection with different serotypes.

Initial vaccine development: In 1980, Mahidol University committed to develop a live-attenuatedtetravalent DENV vaccine. The DENV vaccine development project was supported by a grant from theWHO Regional Office for South-East Asia (ICP RPD 002/DHF). DENV-1 and -2 obtained from DHFpatients and DENV-4 obtained from a DF patient were serially passaged in primary dog kidney (PDK)cells certified to be free from human and canine infectious agents. DENV-3 obtained from DHF patientswas first passaged in primary green monkey kidney (PGMK) cells and then in certified Fetal Rhesus Lung(FRhL) cells. The degree of attenuation was empirically based on certain biological markers.

Bulk seed productions were eventually prepared in pilot production facilities at Mahidol University’sCentre for Vaccine Development at the Institute of Molecular Biosciences. They were subjected togeneral safety tests and monkey neurovirulence tests in accordance with the US FDA requirements.These pre-clinical tested candidate DENV viruses were approved for proceeding to the clinical evaluationphase by a WHO-appointed Scientific Steering Committee and by the Ethical Review Committee of theThai Ministry of Public Health.

The monovalent live-attenuated viruses – DENV-1 PDK-13, DENV-2 PDK-53, DENV-3 PGMK-30/FRhL-3 and DENV-4 PDK-48 – were first tested in flavivirus non-immune adult subjects, followed by bivalent,trivalent and tetravalent vaccine clinical trials. All vaccine recipients developed either a mild or noadverse reaction to the vaccine. The immunogenicity data were discussed.

#E-mail: [email protected]

2 Dengue Bulletin – Volume 32, 2008


Due to viral interference of each DENV components in the combinations, 12 DENV formulations wereevaluated for confirmation of safety and immunoginecity profiles in 155 Thai children aged 3–15years. Preliminary data were analysed and processed for further development.

Collaboration with Sanofi Pasteur: In order to make productive use of this research, Mahidol Universityentered into a collaborative licensing agreement in DENV vaccine production in 1993 with France-based Sanofi Pasteur, the vaccine division of Sanofi-Aventis Group and the largest company in theworld devoted entirely to human vaccines. DENV vaccine based on this approach was prepared forproduction on an industrial scale in France using specific-pathogen-free (SPF) dog colony and FRhLcells. The vaccine was presented in a lyophilized (freeze-dried) form and reconstituted with water forinjection in order to deliver a 0.5 ml specified dose. Multiple dose presentations were planned for atarget population of children and adults living in or travelling to DENV-endemic areas.

The current strategy of creating tetravalent DENV vaccine formulations can lead to an unbalancedimmune response. This is attributed to viral interference that apparently comes into play when threemonovalent vaccine viruses DENV-1, DENV-2 and DENV-4 are mixed with DENV-3 to create a tetravalentformulation.

More research is needed on a priority basis to work out the viral interference factor in order to make theproduction of a tetravalent vaccine out of our attenuated DENV-3 candidate vaccine strain a success.

Keywords: Live attenuated tetravalent dengue vaccine; WHO/SEARO; PDK cells.

Historical development

Dengue haemorrhagic fever (DHF) was firstrecognized as a new disease in Manila in 1954[1].The disease affected mainly children and wascharacterized by the acute onset of high fever,petechial haemorrhage and shock. The secondlarge outbreak occurred in Manila again in 1956which resulted in more than 1200 cases, with10 to 15 per cent case-fatality rate[2]. In 1958, anoutbreak of DHF occurred in Bangkok and itsnearby areas. Almost 2500 cases with 10 percent case-fatality rate were recorded[3]. Since then,DHF has become a serious public health problem,causing large-scale morbidity and mortality amongchildren in the South-East Asia and the WesternPacific regions of WHO. Well-establishedepidemics have also been reported fromMyanmar, China, Cambodia, Indonesia, Laos,Malaysia, Philippines, Thailand and Viet Nam.

In the WHO South-East Asia Region, DHFis a major public health problem in Indonesia,Myanmar and Thailand[4].

The first meeting of the South-East AsiaRegional Advisory Committee on MedicalResearch (SEA/ACMR), held in New Delhi, 5–9 January 1976, recommended that researchon DHF should be given a high priority. Duringthe second session of the SEA/ACMR held inNew Delhi, 23–27 August 1976, a review wasmade of the history of the spread of this diseasethrough several countries of the Region, withan evaluation of the current state of knowledgeon its epidemiology, virology, pathogenesis andthe related problems of clinical management.

A meeting of the Research Study Groupon DHF was held in New Delhi on 24-25February 1977. Several measures with potentialfor the prevention and control of this diseasewere considered. After detailed discussions,the group made its recommendations, of whichthe two important ones were: (i) vaccineresearch; and (ii) control of Aedes aegypti.

The first plan of vaccine research wasdeveloped, which, inter alia, proposed that



virologists from the South-East Asia and theWestern Pacific regions be trained in researchand development of the vaccine at the Schoolof Tropical Medicine and Medical Microbiology,University of Hawaii. On the completion ofthe training (1980), the participants on returnto their respective countries were impressedupon to get directly involved in the nationalDENV vaccine development programme. Thetime frame needed for the development ofthe DENV vaccine programme was proposedto be 3–5 years.

It was understood that most countrieswith DHF problem would like to participate inthe field trials of DENV vaccine at a later stagewhen the vaccines would be ready.

In 1978, a research steering committeerecommended to WHO to take positive stepstowards DENV vaccine development bydesignating the then Department of Pathology,Faculty of Medicine, Ramathibodi Hospital,Mahidol University, now known as the Centrefor Vaccine Development, Mahidol Universityat Salaya, Thailand, to undertake the researchfor the development of the vaccine.

Funding of this project began in April 1980.Three laboratories were equipped for DENVvaccine research and development. A virologistwas recruited and sent to the University ofHawaii for the initial phase of research as wellas for advanced training while equipping of thelaboratory continued. The laboratory was readyfor operation in early November 1980. Detailedand comprehensive standard operatingprocedures (SOPs) for vaccine developmentwere prepared. Protocols were available. Testswere signed by operators and were checkedand signed by the supervisor.

In 1987, the site for DENV vaccinedevelopment moved to the Centre for VaccineDevelopment of Mahidol University at Salaya,Nakhonpathom, Thailand. Equipments for the

centre were donated by the ItalianGovernment. Another four additional buildingswere constructed between 1988–1990, whichincluded an experimental animal house andvaccine pilot plant buildings. The entire vaccinecompound was designated for DENV vaccinedevelopment. Airlocks and a hepafiltered airsupply were generated to control potentialcross-contamination.

Rationale for dengue vaccinedevelopment

The scientific hypothesis behind thedevelopment of a tetravalent DENV vaccineagainst DHF can be summarized as follows:

(1) Adults developed a higher rate ofseroconversion of antibody responseagainst DENV viruses and appeared tobe less susceptible to DHF. Thenaturally-acquired immunity appearedto protect the individuals against theinfection. The immunization of targetpopulations could result in thedevelopment of a protective antibodyresponse in individuals and could helpin protection against the disease.

(2) It had also been shown that a monoor bitypic antibody response could bea risk factor for DHF if sequentialinfection by other serotypes of DENVviruses occurred. It was imperativethat the DENV vaccine should be ableto confer the protective immunityagainst all four serotypes of DENVinfection and provide life-longimmunity. This called for thedevelopment of a live-attenuatedtetravalent DENV vaccine.

(3) The target population for immunizationagainst DHF should be toddlers 1–3years old.



Technical consideration ondengue vaccine developmentat Mahidol University

The objectives of this programme were toselect strains of DENV-1, -2, -3 and -4 whichshowed promise of being attenuated for humanuse and produced in cell substrates. All thefour DENV virus serotypes being developed inThailand were passaged serially in cell culturewithout specific selection.

Dengue virus strains selected forattenuation attempts

DENV-1 (16007-TC-10 2/14/74)

Isolated from a DHF patient in Thailand in 1964had been passaged in tissue culture beforeinoculation into Toxorhynchites amboinensis.The first intrathoracic passage was No. 167164and the second was No. 167376 (received fromDr. Robert E. Tesh on 17 June 1980).

DENV-2 (16681 LLC-1 1/22/73)

Isolated from a DHF case in 1964 fromThailand had been passaged in tissue culturebefore inoculation into Toxorhynchitesamboinensis. The first intrathoracic passage wasNo. 167165 and the second passage was No.167377 and were received on 17 June 1980.The parent culture strain was virulent for man,having produced typical DF in a laboratoryworker who was exposed accidentally(unpublished observations, Dr S.B. Halstead).

DENV-3 (16562 TC-7 1/31/72)

Virus was isolated in 1964 from a DHF case inthe Philippines. It had been passaged in cellculture before inoculation into Toxorhynchites

amboinensis. The first intrathoracic passage wasNo. 167166; the second passage was No.167378 (received for PDK cell passage on 17June 1980). The parent culture passaged viruswas demonstrated to be virulent for man havingproduced DF in an accidentally infectedlaboratory worker (unpublished observations,Dr S.B. Halstead).

DENV-4 (1036)

Virus was isolated from a DF case from Indonesiain 1976 using Aedes aegypti and kindly furnishedby Dr Duane G. Gubler. The fourth passagewas used to initiate the vaccine studies.

Mosquito inoculation

At the University of Hawaii (Pacific ResearchUnit), five adult laboratory-rearedToxorhynchites amboinensis were inoculatedintrathoracically with strains of DENV-1 to -4[5].The inoculum was approximately 0.0003 ml.Mosquitoes were maintained on 10% sucrosesolution at 28 °C for 12 days. At the end of theincubation period, each group of insects waskilled by freezing, their heads removed andtriturated in 5.0 ml phosphate buffer saline,containing 0.5% gelatin, 30% heat-inactivatedcalf serum and penicillin and streptomycin.After centrifugation at 5 °C for 30 minutes, eachsupernatant fluid was inoculated into anothergroup of five Toxorhynchites amboinensis. Theseinsects were also held at 28 °C for 13 days. Atthe end of the incubation period, they werekilled by freezing.

Preparation of mosquitosuspensions

The head was removed from the infectedmosquitoes with a sterile surgical blade andplaced in a mortar. Body parts were kept in a



sterile vial and frozen at –70 °C. The virusdiluent, with 30% heat-inactivated calf serumin phosphate-buffered saline, pH 7.5,penicillin/streptomycin, was added to groundmosquito heads, 2.5 ml/5 mosquitoes. Aftercentrifugation at 10 000 rpm for 30 minutes at5 °C, the supernatant fluids were filteredthrough at 0.45 micron millipore filter. Filtrateswere used to inoculate PDK and PGMK cells.

Preparation of primary dogkidney cells

The work was done in the laboratories of theDepartment of Tropical Medicine and MedicalMicrobiology, University of Hawaii, and wassupported, in part, by a grant from theRockefeller Foundation to Dr S.B. Halstead.

Each lot and sub-lot of dog kidney cellswere subjected to safety tests to assure thatthe cells and supernatant fluid were free ofinfectious agents. Tests included for exclusionof bacterial, fungi, mycoplasma and cytopathicand haemadsorbing agents.

Development of attenuated strainsof DENV 1-4 viruses

Mosquito suspensions of the DENV-1 (16007),DENV-2 (16681) and DENV-4 (1036) wereattenuated by serial passages in PDK cell cultureat 32°C without cloning or deliberate selection.The procedure relied on the spontaneousappearance of variants and selection forattenuated variants by the biological pressureof the abnormal host cell. This general methodhad been successful with several other live virusvaccines, e.g. rubella and mumps.

The DENV-3 (16562) virus was attenuatedby serial passages in PGMK cell; however,attempts to adapt it to PDK cells had failed.

DENV viruses were serially passaged inPDK cells as illustrated in Figure 1.

Figure 1: Schema for attempted attenuationof dengue viruses

At every fifth passage level, a moderate-sized virus seed was prepared. This virus wasstudied for plaque size morphology in LLC-MK2cells, temperature sensitivity to replication shut-off, suckling mouse neurovirulence and growthin human monocytes, viraemia and antibodyresponse in primates.

When the passaged virus presented areduction in plaque size, temperature sensitivityfor replication and absence of viraemia andreduced antibody response in monkeys,“Masters seed”, “Production seed” and“Candidate vaccine” were prepared. Safety testson the Production seed and Candidate vaccineincluded inoculation of neutralized virus intoadult and suckling mice, guinea pigs, rabbits andseveral tissue culture systems. Furthermore, itwas also assured that the candidate vaccineproduced no neurovirulence followingintracerebral inoculation in monkeys. Theattenuated strains thus developed could helptowards worldwide stock of candidate denguevaccines[6].

What constitutes a satisfactory level ofattenuation remains uncertain. Hypothetically,we would like to have a vaccine which was

DENV VIRUS

Mosquito (2 passages)

PGMK(parental virus)

PDK 1

PDK 5

PDK 10

etc.



avirulent, i.e. viruses which did not have thecapability to cause direct cell injury, but theprotective antigenic epitopes of these avirulentviruses should still be preserved and effectivelypresented to both the B and T lymphocytes ofthe vaccines to confer both humoral andcellular immunities. It was very difficult todefine the attenuation of the dengue virusesby a specific series of the biological markers.These observations remained unsubstantiateddue to the fact that there was no known animalmodel for DHF. Man represented the onlyalternative testing model of vaccine efficacy.

Markers of attenuation

To define the level of attenuation of the virusesat the present, it could at best be empirical.The assessment was based on the findings ofa combination of markers.

Evidence for attenuation was based on acomparison of the high passage viruses withthe parent virus in several in vivo and in vitrotests: plaque size, temperature sensitivity,replication in human monocytes, and monkeyviraemia. These characteristics had been shownto be related to human virulence with otherexperimental DENV vaccines.

The DENV-1 PDK 43, DENV-2 PDK 53,DENV-3 PGMK 33 and DENV-4 PDK 48candidate viruses produced a uniform smallplaque size when assayed in LLC-MK2 cells.They revealed temperature sensitivity by theplaquing efficiency test. High PDK or PGMKpassages had significantly reduced virulence forsuckling mice by the intracerebral route. Allthe DENV candidate viruses produced low orno ability to replicate in human monocytes invitro. All of them showed low or no viraemiaafter inoculation with 104–105 plaque formingunit (pfu) of each candidate viruses withmoderate specific neutralizing antibodyresponses. Reduced neurovirulence for mice

was observed with DENV-1, DENV-2 andDENV-3 candidate viruses. However, theDENV-4 PDK 48 candidate viruses still revealedmodulate neurovirulence in suckling mice.

Safety test

Safety tests of the cell substrate, the candidateviruses and the candidate vaccines weredesigned according to the United States FDAregulations as applied to live attenuatedvaccines produced in the United States. Testsincluded microbial sterility; and search foradventitious agents in PGMK cells, adult andinfant mice, guinea pigs, and rabbits.Haemadsorption agents were sought in cell-culture experiments. A second tier of testsrequired for additional safety were performedat the virology laboratory of the Departmentof Tropical Medicine and Medical Microbiology,University of Hawaii.

The team could establish the capability toperform monkey neurovirulence test inThailand and slides of monkey tissue wereindependently reviewed by an experiencedneuropathologist.

Peer review of the vaccinedevelopment project

Candidate DENV vaccines considered to besufficiently attenuated were submitted to aninternational panel of experts in DENV forreview. This panel had met annually once ayear in Bangkok for twelve times from 1983 to1994. The function of the panel of experts wasto review the scientific work, including visit tothe site of vaccine development, in order topursue and examine the facilities, and to auditthe raw data. The record books were reviewedby two of the peer reviewers in detail forcompleteness and for the accuracy of summary



data presented. The peer group gaverecommendations to the Ministry of PublicHealth of Thailand and to WHO-SEARO basedon their assessment whether the candidatevaccines were suitable for vaccine trials inhuman beings or not. This process was uniquefor WHO programmes[7-18] (Table).

The DENV-1 (16007) PDK 43, DENV-2(16681) PDK 53, DENV-3 (16562) PGMK 30FRhL 3 and DENV-4 (1036) PDK 48 met theUS FDA requirement for microbial safety andmonkey neurovirulence test for live attenuatedviral vaccine conducted by laboratory at MahidolUniversity as well as by an independentlaboratory at the Walter Reed Army Instituteof Research, USA. They were approved by aninternational peer review group based on theexamination of the result of safety test and byan on-site examination of the facilities,laboratory record and log books. Comfirmatoryhistopathological examination was done at theethical review conducted by a committee forhuman experimentation of the Mahidol

University and by a similar committee of theMinistry of Public Health, which was satisfactoryand these candidate vaccines were approvedfor clinical trials.

Clinical trials of monovalentdengue vaccines[19]

The site for the small-scale experimental clinicaltrial was Lampoon and Loei provinces, an areawhere there was low prevalence of Ae. aegyptimosquitoes. The trials were conducted duringthe cold season so as to minimize the risk ofother arbovirus infections and possibletransmission of vaccine viruses. The populationchosen to conduct the trials consisted offlavivirus non-immune young male adults. Theinitial trial was conducted in two phases usingfirst two and then eight volunteers to increasethe safety factor. The protocol called for closeobservation during the first 21 days.

DENV-1 (16007) candidate vaccines

The candidate DENV-1 (16007) PDK 43vaccine was passed 43 times in PDK cell. Theevolution of the biological markers tested wasas follows: plaques in LLC-MK2 cells becameof small size (= 1 mm) after passages 10–15.Temperature sensitivity at 39 °C was achievedat passage 30. Ability to grow in humanperipheral blood lymphocytes (PBL) was lostat passage 20. Suckling mouse neurovirulencewas reduced to minimal level at passage 15.After 43 passages, all monkeys that receivedthe DENV-3 PDK 43 showed no or lowviraemia. Based on these results, passage 43was selected for phase 1 trial.

Six flavivirus non-immune subjects wereinoculated with a dose of 2.1 to 3.5 × 104 pfu.Clinical symptoms were mild in all volunteersand only one of them showed very minimalnose bleeding, without other haemorrhagic

Table: Dengue Virus DevelopmentProgramme WHO Peer Review Meeting,

Bangkok, ThailandOrganized by WHO/SEARO,WHO Project: ICP RPD/002

Serial No. Date

1st 1-5 August 19832nd 23-25 August 19843rd 31 July - 2 August 19854th 20-22 August 19865th 27-30 July 19876th 1-5 August 19887th 7-11 August 19898th 29-30 September 19909th 26-28 August 199110th 24-26 August 199211th 23-25 August 199312th 29-31 August 1994



manifestations. In no case was absoluteleukopenia observed. Clinical chemistry wasnormal in all volunteers. Immune response, asmeasured by plaque reduction neutralization test(PRNT), was detected in only 1 out of 6 flavivirusnon-immune volunteers. Two of theseronegative subjects were challenged againwith the same dose of DENV-1 PDK 43 at 3months. Again, they failed to develop anyneutralizing antibody. The conclusion was thatDENV-1 PDK 43 had been over-attenuated andthat it was necessary to try lower passage levels.A candidate vaccine was then prepared fromDENV-1 PDK 30, and after the usual safety tests,was infected into five adult male volunteers whowere seronegative for both Japaneseencephalitis (JE) and DENV viruses. Two of thefive seroconverted, but the responses were low.The exercise was thus repeated using DENV-1PDK 20 but again the antibody responses werelow and only three of the five seroconverted.The conclusion was thus reached that DENV-1,PDK 20, PDK 30 and PDK 43 were over-attenuated to be useful as candidate vaccinesfor DENV-1 in people that were seronegativefor previous exposure to flaviviruses. The DENV-1 PDK 13 virus that was used in the next trialshowed evidence of lesser attenuation than PDK20, PDK 30 or PDK 43 in that it still replicatedin human monocytes. When seven DENV andJE antibody negative male volunteers wereinjected with DENV-1 PDK 13, fiveseroconverted to DENV-1 within 30 days. Therewas some evidence of rhinitis but this may havebeen coincidental and the significance of theobservation could not be assessed. There wasalso a slight fall in leukocyte counts on day 10,but the virus was not isolated from blood at anystage.

DENV-2 (16681) candidate vaccine[19]

The trial of the DENV-2 candidate vaccine,DENV-2 (16681) PDK 53, was carried out asphase 1a and 1b trial.

The initial phase 1a of DENV-2 PDK 53candidate vaccine in ten 18–30-years-old malehuman subjects showed encouraging results.None of the 10 persons vaccinated were febrileor incapacitated; side-reactions possiblyattributable to the candidate vaccine werelimited to slight leukopenia, occasionallyabnormally large platelets and a few transientcomplaints such as mild aches and pains.

All vaccinated persons developed DENV-2 neutralizing antibody. Those subjects withpreexisting antibody to JE virus respondedserologically more rapidly than those subjectswithout preexisting flavivirus antibody beforevaccination.

Serological tests carried out one year afterthe vaccination showed that neutralizingantibodies were present in 100 per cent ofthe volunteers.

The phase 1b trial of DENV-2 PDK 53candidate vaccine involved sixteen 15–30-years-old male volunteers, 15 of whom wereflavivirus non-immune. Four doses of varyingvirus dilutions were given to groups of fourvolunteers each, and every person developedDENV-2 neutralizing antibodies, regardless ofvaccine virus dilutions. Abnormal lymphocytesand a slight decrease in lymphocyte numberswere consistently observed between days 6and 10. As in the phase 1a trial, no adversereactions to the vaccine were observed.

Viraemia was detected in one volunteerand virus isolated in C6/36 cells from day 6serum. The virus had growth characteristicssimilar to those of the candidate DENV vaccinevirus. On the basis of 1a and 1b studies it wasrevealed that viraemia occurred between days6 and 10. It was unlikely to occur after day 14because of the onset of neutralizing antibodies.It is possible that viraemia may precede thetime of the lowest white cell counts, whichfrequently occurred on day 6[19].



A dose response study in adults, based on5-fold dilutions of the vaccine, showed anestimated 50 per cent infectious dose of 5–7pfu.

DENV-3 (16562) candidate vaccines[19]

The DENV-3 (16562) parent virus did not growin PDK cells and was passaged in PGMK cells.At passage PGMK 30, the virus was still able toreplicate in PBL and produced plaques ofvarying sizes. After passage 34, plaques wereuniformly small. Two PGMK passages wereselected for adaptation to FRhL cells: 30 and35. In FRhL cells, DENV-3 attained titres onelog higher than in PGMK cells. With bothpassages (PGMK 30/F2) and PGMK 35/F2),biological markers were considered to besatisfactory: plaques were of small size, no CPEin LLC-MK2 cells, temperature sensitive at38.3 °C, no growth in human PBL, and reducedneurovirulence for suckling mice.

No adventitious agents were found inPGMK cells analysed by electron microscopy.Safety tests of PGMK cells were beingcompleted at the National Institute forBiological Standards and Control (London) andthe National Biological Standards Laboratory(Canberra). The cells had been found to befree of mycoplasma, mycobacteria and otheradventitious agents. Tests to detect simianretroviruses, SV5 and SV40, were negative.

Biological characteristics of DENV-3candidate vaccine viruses

The DENV-3 (16562) PGMK 30 passage virushad mixed plaque morphology (medium andsmall), a restrictive temperature of 40 °Ccaused CPE in LLC-MK2 cells and grew inhuman PBL. DENV-3 PGMK 30, FRhL-3 virushad small and pinpoint plaque morphology,restricted growth at 38 °C, and did not causeCPE in LLC-MK2 cells. Considerable change,

presumably selection, had occurred with FRhLpassage. The virus recovered from a volunteerwho received PGMK 30, FRhL-3 vaccine hadbiological markers (medium) plaque size, CPEin LLC-MK2 similar to earlier passage vaccinewas either genetically prone to reversion orcontained an undetected subpopulation ofmore virulent virus.

Three passage levels of DENV-3 (16562)were given to volunteers; PGMK 33, PGMK30-FRhL-2, and PGMK 30 FRhL-3. The FRhL-passaged viruses differed from the PGMK 33in being more temperature-sensitive, less ableto produce CPE in LLC-MK2 cells, and havinguniform small plaque morphology.

Four volunteers received the PGMK 30FRhL-3 virus at doses of 1 × 104 to 6.5 × 104

pfu. One of two volunteers seroconverted atthe lower dose. The volunteer who failed toconvert at the lower dose was revaccinated atthe higher dose and seroconverted. Twovolunteers seroconverted at a higher dose.Only minor symptoms and no fever wereobserved. Satisfactory primary immuneresponses were observed in three volunteers;the fourth, who was JE immune, had asecondary-type serological response. The virusisolated from the serum of one volunteerexhibited medium-sized plaque morphologyand its characteristics of earlier passage virus.

In other trials, two volunteers receivedPGMK 33 vaccine and four volunteers receivedPGMK 30 FRhL-2 vaccine. Both of thosevaccines contained both medium and smallplaque sizes and were less temperature-sensitive than the PGMK 30 FRhL-3 vaccine.Both vaccines immunized satisfactorily at dosesof 104; however, brief febrile responses andmild symptoms were observed.

The PGMK-30, FRhL-3 vaccine appearedto be less reactogenic than the other twoDENV-3 candidate vaccines and wasimmunogenic at a dose of 5 × 104 pfu.



DENV-4 (1036) PDK 48 candidatevaccine

DENV-4 (1036) virus was passaged in PDK cellsto passage 48. Biological markers of PDK 48included small plaque and no cytopathic effectin LLC-MK2 cells, temperature replicative shut-off at 39 °C, and average survival of 12 days insuckling mouse. PDK 48 virus replicated inperipheral blood mononuclear cells. Rhesusmonkeys inoculated with PDK 48 virus did notdevelop viraemia but seroconverted. Onevaluation of the monkey neurovirulence tests,the panel of experts concluded that there wasno significant difference between the parentalDENV-4 virus and DENV-4 PDK 48 candidatevaccine, and it was thus acceptable to proceedwith phase 1a clinical trial. An additional fourrhesus monkeys had been tested with PDK 48virus; enhanced neurovirluence was not found.The monkey neurovirulence test result wasconsidered satisfactory, and it appeared feasibleto proceed with PDK 48 as a candidate vaccine.

The phase 1a clinical trial was theinoculation of five flavivirus seronegativevolunteers with 1-2 × 104 pfu of DENV-4 PDK48. All volunteers developed specificneutralizing antibodies which first appearedfrom days 13–16, and peaked in titre at day30 post-inoculation. Clinical signs wereunremarkable in all volunteers, and novolunteer developed fever. Clinical symptomswere generally absent, although two volunteersreported eye pain and headache. In one ofthese volunteers the headache re-occurred fora period of about 2 weeks. All volunteersshowed normal blood chemistry profiles.Haematological studies revealed a transientincrease in atypical lymphocytes in threevolunteers. All volunteers showed a temporarydepression in total white blood cell counts;however, there was no absolute leukopenia.Virus was recovered only from plasma, and theviraemia appeared to be low in titre, since novirus could be detected by direct plaque assay.

The recovered virus strains shared all biologicalmarkers with the vaccine candidate, exceptthat one strain showed an extended mean dayto death in suckling mice of 20 days.

The phase 1b trial was designed todetermine the minimum infective dose of theDENV-4 vaccine candidate. The 1b trial wastemporarily divided into two phases with groupsof seven and five volunteers. The vaccine wasdiluted from 1:5 (3700 pfu) to 1:1000 (12-15pfu) and each dilution was inoculated into 1–3volunteers. None of the volunteers showedfever or rash, and clinical signs and symptomswere mild, although eight of the twelvevolunteers reported transient headache and eyepain. Blood chemistries were normal, andhaematological findings were similar to thoseseen in the phase 1a trial.

All groups of volunteers inoculated with adilution of 7 × 102 pfu or greater developedspecific neutralizing antibody. Two of the twovolunteers at 7 × 102 pfu seroconverted, oneof the two at 1.5 × 102 pfu seroconverted,and none of the three volunteers at 63–77 pfuseroconverted. In total, combining the phase1a and phase 1b results, 10 of the 10volunteers inoculated with vaccine doses of7 × 102 pfu or greater seroconverted.

Polyvalent vaccine clinicaltrials[20]

Bivalent vaccine DENV-2 (16681)PDK 53 and DENV-4 (1036) PDK 48clinical trial

The aim of the DENV vaccine developmentprogramme was to develop and administer avaccine containing a mixture of multiple DENVserotypes. The rationale was based on theprovision of providing protection to all serotypesthat would minimize any chances of futureDENV infection enhancement and adverse hostreaction. This trial was designed to conduct in



humans in support of the hypothetical conceptthat multiple simultaneous infections withcandidate vaccines were possible, safe, andeffective. Eleven male flavivirus non-immunesubjects aged 16 to 31 years received bivalentDENV-2 and DENV-4 candidate vaccines.

The neutralizing antibody responses toDENV-2 at 6 months ranged from 1:52 to 1:440and that of DENV-4 ranged from 1:44 to 1:310.A low titre of neutralizing antibodies to DENV-1, DENV-3 and JE viruses were detected early,but this disappeared by 6 months.

The bivalent DENV-2–DENV-4 candidatevaccine was both immunogenic and withoutunacceptable reactions. Moreover, the dose ofDENV-2 and DENV-4 viruses was acceptableand formed to be the basis for future trials[20].

Bivalent vaccine DENV-1 (16007)PDK 13 and DENV-4 (1036) PDK 48clinical trial[20]

The bivalent DENV-1 and DENV-4 vaccine’shuman clinical trial was conducted in Loeiprovince, Thailand. Seven male flavivirus non-immune subjects, aged 16 to 30 years,received bivalent DENV-1 and DENV-4candidate vaccine.

All seven subjects seroconverted to bothDENV-1 and DENV-4 since the presence ofneutralization antibodies were detected byPRNT to both DENV-1 and DENV-4. There wasno difference in response between thosereceiving candidate vaccines in separate armsand among those receiving mixed vaccine inanother arm.

The bivalent DENV-1 and DENV-4candidate vaccine induced specific responseto both DENV-1 and DENV-4 but low titres ofheterologous neutralizing antibody were foundand the vaccine, was without any adversereactions, among the recipients[20].


Seven male subjects aged between 17 and 32years received bivalent DENV-1 and DENV-2candidate vaccine.

All seven subjects seroconverted to bothDENV-1 and DENV-4 since the presence ofneutralization antibodies were detected byPRNT to both DENV-1 and DENV-4. There wasno difference in response between thosereceiving candidate vaccines in separate armsand among those receiving mixed vaccine inanother arm.

The bivalent DENV-1 and DENV-4candidate vaccine induced specific responseto both DENV-1 and DENV-4 but low titres ofheterologous neutralizing antibody were foundand the vaccine was without any adversereactions among the recipients[25].


Seven male subjects aged between 17 and 32years received bivalent DENV-1 and DENV-2candidate vaccine.

All subjects seroconverted to DENV-1 andDENV-2 by 30 days. Titres of DENV-1neutralizing antibody ranged between 1:27 and1:70 in the six subjects who were non-immunebefore vaccination. Titres of DENV-2 rangedfrom 1:26 to 1:120 and at 30 days no cross-reaction with other flaviviruses was observed.

The bivalent DENV-1 and DENV-2candidate vaccine was both immunogenic andwithout any adverse reactions[25].



Trivalent vaccine DENV-1 (16007)PDK 13, DENV-2 (16681) PDK 53 andDENV-4 (1036) PDK 48 clinical trial[21]

The human clinical trial comprised of a mixtureof three monovalent DENV vaccines (DENV-1, -2 and -4) which was inoculated into eachof 12 male adults. The study was performedin a subdistrict of Loei province in north-eastThailand, where the prevalence of Ae. aegyptior Ae. albopictus mosquitoes was low. Theobjective of this study was to determine thesafety and feasibility of simultaneousadministration of three DENV candidatevaccines.

Of the 12 recipients, nine were flavivirusnon-immune; they all developed serumneutralizing antibodies to all the three DENVviruses.

The results of this study showed that itwas possible to infect humans safely with threeattenuated DENV viruses. The median doseof DENV-1 virus was close to optimum whereasthe dose of DENV-4 was too low. Thesuccessful administration of a trivalent DENVvaccine indicated that Mahidol University hadachieved another important milestone on theroad to the development of a tetravalentdengue vaccine[21].

Trials of tetravalent vaccine inchildren aged 5–12 years[22]

The tetravalent DENV vaccine candidateappeared to be safe when administered tochildren aged 5–12 years. Children becamejust febrile, and this usually did not last for morethan a day. One volunteer had a rash thatpersisted for three days.

Two trends of serological response to thetetravalent vaccine were observed among thevolunteers. First, the infectious dose that was

calculated for adults was not equivalent for thechildren in the age groups studied. A trend ofan increasing rate of seroconversion amongchildren was noted with a decreasing vaccinedosage; however, an optimum dose forchildren 5–12-years-old still had to bedetermined.

Second, children with preexistingantibodies to either DENV or JEV appeared torespond better to the tetravalent vaccine thandid children who were completely non-immune. Even so, not all volunteers withpreexisting flavivirus antibodies responded toall four DENV serotypes.

Collaboration withmanufacturer[23]

The Mahidol vaccine was licensed to PasteurMérieux (now Sanofi Pasteur) in France forlarge-scale production under GoodManufacturing Practice (GMP) conditions.

The master seed, the production seed andthe candidate DENV vaccines were sent toPasteur Merieux in February 1993, shortly afterthe agreement was signed in January 1993.Three technical meetings at Pasteur Merieuxwere held in Lyon, France, between 1994–1996. Industrial production of the fourmonovalent vaccines was achieved by 1995.All biological studies, including monkeyneurovirulence studies were repeated.

The first clinical trial carried out using theMahidol/PMsv tetravalent vaccine in USvolunteers suggested that the combination offour attenuated strains appeared to result inincreased reactogenicity and diminishedtolerability. Antibody responses werepredominantly directed against DENV-3 withlow or undetectable titres against the remainingthree serotypes[23]. This outcome appeared tobe the result of preferential replication of



DENV-3 in the tetravalent vaccine. Themechanism of such viral interference was notknown. But it had been suggested that the ratioof the four attenuated viruses in the tetravalentformulation may be an important factor. Asubsequent clinical study in Thailand[24] showedthat varying and reducing the concentrationsof the DENV-3 strain resulted in an improvedclinical safety profile of the tetravalent vaccine.About 71% seroconversion (against all 4serotypes) was observed after a two-dosevaccination schedule in this study. Severaldifferent reformulations of the tetravalentvaccine were being evaluated in order toprovide a more balanced immune response toeach serotype[23].

Second generationrecombinant vaccines

A Cooperative Research and DevelopmentAgreement (CRADA) was entered into withthe Division of Vector-Borne InfectiousDiseases, National Center for InfectiousDisease, Centers for Disease Control andPrevention (CDC), Fort Collins, Colorado,USA, to develop the second-generationrecombinant vaccines using complementaryDNA (cDNA) technology. As per thememorandum of understanding (MoU), the firstshipment of candidate DENV vaccines (DENV-1 and DENV-2) was sent to CDC in August1994. The DENV-3 and DENV-4 vaccines weresent shortly thereafter. The agreement calledfor Mahidol University to provide support forone locally-trained technician and for CDC toprovide support to one Thai investigatorengaged in research and capacity-buildingactivities at Fort Collins. Vaccine developmentstudies were realized at CDC while biologicalmarker testing was partially done in Thailand.

Conclusion on present statusof PDK-based live-attenuateddengue vaccine

The importance of DENV vaccine developmentwas imperative in order to improve publichealth throughout the world and was highlydesirable for WHO to provide financial supportfor this programme. The peer group summarizedthe progress as follows:

(1) Monovalent candidate vaccines

(a) DENV-1: A usable candidatevaccine

(b) DENV-2: A near-perfectcandidate vaccine

(c) DENV-3: The most recentlydeveloped candidate vaccine,somewhat more reactogenic thanthe other candidate vaccines. Asearch for a better vaccine shouldproceed.

(d) DENV-4: A very good product.

(2) Bivalent and trivalent combinationsusing DENV-1 PDK 13, DENV-2 PDK53 and DENV-4 PDK 48 hadundergone phase 1 trials in adults withsatisfactory results.

(3) Tetravalent vaccine was acceptablysafe. Interference was noticed aftermixing of the DENV-3 PGMK-30/F3 inthe combination.

Lessons learned

There was a general consensus that vaccinationcan be one of the most cost-effective ways toprevent DF and DHF. The aim of this projectwas to develop a safe and immunogenic



vaccine against the four DENV serotypes. Eachof the four monovalent vaccines as well as thebivalent and the trivalent vaccines weredeveloped and tested step by step in thelaboratory and in human volunteers. By 1992,the attenuated, tetravalent vaccine was beingtested for immunogenicity and safety in humanvolunteers. Formal phase 1 and phase 2 clinicaltrials had proven the vaccines to be both safeand immunogenic in humans. Human trials ofthe tetravalent vaccine were successfullyconcluded.

In November 1992, WHO headquartersand WHO/SEARO announced the attainmentof the objective of the dengue vaccinedevelopment project at Mahidol as follows:“Vaccine for Dengue Haemorrhagic Fever”.From this study, it was proved that PDK cellscould be used successfully for attenuationattempts. The DENV-2 PDK 53, which was oneof the important outcomes of this study, hasbeen further used as a backbone to constructlive molecular DENV vaccines in the USA.

Research as well as relevant capability-building activities at Mahidol University wereestablished with the advice of the internationalpeer group which met annually. However, theinitial expectation in 1985 that DENV vaccinedevelopment would be completed withinthree years proved too optimistic.

Considerable research capacity buildingtook place as part of the research project supportduring that decade. The various technologiesrequired for vaccine development andlaboratory-scale production were transferred.They included continuous tissue culture,development of PDK and other cell lines,monkey tests for neurovirulene, etc. The annualmeetings of the peer group itself providedvaluable scientific advice to the project. Inaddition, Mahidol University scientists weresupported for visits and contacts with variousscientists and institutions in other countries.

Meanwhile, Mahidol University expandedthe physical and other infrastructure requiredfor vaccine development and pilot scale up. Avaccine development centre building and alaboratory animal centre were completed atthe new Salaya campus. Equipment forupscaling was received as donation from theItalian Government.

The DENV vaccine development project wasacknowledged to be a worthy scientificachievement in the area of health. Suchachievements could occur due to the long-termcommitment of scientists in Thailand, thecontinuous support of the Government ofThailand, and the initial impetus and sustainedcommitment and support provided by WHO/SEARO. The Government of Thailand andMahidol University provided the major resources.WHO provided about US$ 2.5 million during aperiod of 15 years. Other donors contributedsubstantial amounts at various stages of theproject for specific components of theprogramme. Success was due to scientificcorrectness of the research, the outstandingleadership of late Prof. N. Bhamarapravati ofMahidol University, sound research managementby several parties, and the sustained commitmentand technical support through the years of theWHO Regional Office for South-East Asia.

Acknowledgements

The author is highly grateful to Dr Scott B.Halstead, to all the members of the Peer ReviewGroup and the participating scientists. The authoris also thankful to all the staff of the secretariatfor their full cooperation for the success of thepeer review meetings. The contributions madeby the scientists of the Centre for VaccineDevelopment, Mahidol University, includingProf. S. Angsubhakorn, Drs Mi Mi Khin, N.Nitatpattana, K. Tubthong, W. Kanitwithayanun,N. Jirakanjanakit, T. Sanohsomnieng, S.Palabodeewat, and S. Phunyahathaikul aregreatly appreciated.



[1] Quintos FN, Lim LE, Juliano L, Reyes A, LacsonP. Haemorrhagic fever observed amongchildren in the Philippines. Philipp J Pediatr1954; 3:1-19.

[2] Hammon WM, Rudnick A, Sather GE, RogersKD, Chan V, Dizon JJ, Basaco-Sevilla V. Studieson Philippine haemorrhagic fever: relation todengue viruses. In: Proceedings of the ninthpacific science congress of the Pacific scienceassociation, 1957 (Vol. 17, Public Health andMedical Sciences), Bangkok, 1962.

[3] Jatanasen S, Skuntanaga P, Dhanasiri C. Someaspects of epidemiology of Thai haemorrhagicfever 1958-1961. In: Symposium onhaemorhagic fever. SEATO medical researchmonograph no. 2. Bangkok: Post Publishing,1962. pp. 6-21.

[4] World Health Organization, Regional Officefor South-East Asia. From the editor’s desk.Dengue Bulletin. 2007; 31.

[5] Rosen L, Gubler D. The use of mosquitoes todetect and propagate dengue viruses. Am J TropMed Hyg. 1974; 23(6):1153-60.

[6] Yoksan S, Bhamarapravati N, Halstead S B.Dengue virus vaccine development: study onbiological markers of uncloned dengue 1-4viruses serially passaged in primary kidneycells. In: St. George T D, Kay B H, Blok J, editors.Arbovirus research in Australia. Proceedings ofthe fourth symposium. Brisbane, Australia:CSIRO/QIMR; 1986. pp. 35–38.

[7] Dengue virus vaccine development, PeerReview Meeting, Bangkok, 1-5 August 1983,Report to the Regional Director.

[8] Development of dengue virus vaccine, SecondPeer Review Meeting, Bangkok, 23-25 August1984, Report to the Regional Director.

[9] Development of dengue virus vaccine, ThirdPeer Review Meeting, Bangkok, 31 July-2August 1985, Report to the Regional Director.

[10] Development of dengue virus vaccine, FourthPeer Review Meeting, Bangkok, 20-22 August1986, Report to the Regional Director.

[11] Development of dengue virus vaccine, FifthPeer Review Meeting, Bangkok, 27-30 July1987, Report to the Regional Director.

[12] Development of dengue virus vaccine, SixthPeer Review Meeting, Bangkok, 1-5 August1988, Report to the Regional Director.

[13] Development of dengue virus vaccine, SeventhPeer Review Meeting, Bangkok, 7-11 August1989, Report to the Regional Director.

[14] Development of dengue virus vaccine, EighthPeer Review Meeting , Bangkok, 29-30September 1990, Report to the RegionalDirector.

[15] Development of dengue virus vaccine, NinthPeer Review Meeting, Bangkok, 26-28 August1991, Report to the Regional Director.

[16] Development of dengue virus vaccine, TenthPeer Review Meeting, Bangkok, 24-26 August1992, Report to the Regional Director.

[17] Dengue vaccine development, Report of theEleventh WHO Peer Review Meeting, Bangkok,23-25 August 1993.

[18] Dengue vaccine development, Report of theTwelfth WHO Peer Review Meeting, Bangkok,29-31 August 1994.

[19] Bhamarapravati N, Yoksan S, ChayaniyayothinT, Angsubphakorn S, Bunyaratvej A.Immunization with a live attenuated dengue-2-virus candidate vaccine (16681-PDK 53):clinical, immunological and biologicalresponses in adult volunteers. Bull World HealthOrgan 1987;65(2):189-95.

[20] Bhamarapravati N, Yoksan S. Study of bivalentdengue vaccine in volunteers. Lancet1989;1(8646):1077.

References



[21] Bhamarapravati N, Yoksan S. Immunisation inhumans with live attenuated trivalent vaccines.South East Asian J Trop Med and Pub Health.1993; 24 Supp 1: 246-249.

[22] Bhamarapravati N, Yoksan S. Live attenuateddengue tetravalent vaccine. Gubler DJ, KunoG, eds. Dengue and dengue hemorrhagic fever.New York: CAB International, 1997. pp. 367–377.

[23] Kanesa-thasan N, Sun W, Kim-Ahn G, VanAlbert S, Putnak J.R, King A, RaengsakulsrachB, Christ-Schmidt H, Gilson K, Zahradnik J.M,Vaughn D.W, Innis B.L, Saluzzo JF, Hoke Jr C.H.

Safety and immunogenicity of attenuateddengue virus vaccines (Aventis Psteur) inhuman volunteers. Vaccine 19; 2001: 3179-3188.

[24] Sabchareon A, Lang J, Chanthavanich P, YoksanS, Forrat R, Attanath P, Sirivichayakul C, PengsaaK, Pojjaroen-Anant C, Chokejindachai W,Jagsudee A, Saluzzo JF, Bhamarapravati N.Safety and immunogenicity of tetravalent live-attenuated dengue vaccines in Thai adultvolunteers: role of serotype concentration,ratio, and multiple doses. Am J Trop Med Hyg.2002; 66(3): 264-72.


Socioeconomic determinants of dengue incidence inSingapore

Stefan Maa#, Eng Eong Ooib, Kee Tai Gohc

aEpidemiology and Disease Control Division, Ministry of Health, Singapore

bDefence Medical and Environmental Research Institute, DSO National Laboratories, Singapore

cDirector of Medical Services’ Office, Ministry of Health, Singapore

Abstract

Community participation is critical in sustaining vector population control activities in order to preventdengue transmission. However, disease exposure in a community is often not uniform across the entirepopulation and the identification of “at-risk” groups would enable the disease prevention effort to befocused and thus cost-effective. We performed ecological data analyses to study the association betweensocioeconomic variables and dengue incidence in Singapore from 1998 to 2002. Our results indicatedthat the DF/DHF incidence was ecologically associated with some socioeconomic/demographiccharacteristics of the population. Areas with a high proportion of socioeconomically disadvantagedresidents had also a significantly higher DF/DHF incidence. The Aedes population density of larvae wasnot related to this difference in the DF/DHF incidence, indicating that additional risk factors werepresent in these population sub-groups, and that dengue control in Singapore could benefit from amore focused effort in outreach to the population of relatively lower socioeconomic levels.

Keywords: Dengue; Socioeconomic; Geographical; Singapore.

#E-mail: [email protected]; Tel: (65) 6325 1165; Fax: (65) 6325 9194

Introduction

Dengue fever/dengue haemorrhagic fever (DF/DHF) remains a major health problem in manyareas of the world, especially in south-eastAsia[1-3]. Much effort has been focused on theprevention and control of dengue infection.The only effective strategy to control a DF/DHF outbreak, in the absence of a vaccine, isto eliminate Aedes mosquitoes and its larvalbreeding habitats[4].

In Singapore, DHF was first recognized asa public health problem during the early 1960sand a nationwide Aedes control programme,which incorporated source reduction, publiceducation and law enforcement, wasimplemented in 1969. The Aedes house index(HI) (% of premises positive for Aedes breeding)was markedly reduced from more than 25%before 1970 to 1–2% for the entire countrysince 1982. The significant decline of the DF/


Socioeconomic determinants of dengue incidence in Singapore

DHF incidence (about 60 per 100 000population in 1973 to below 10 per 100 000in 1982) was the result of an effective Aedessurveillance and control programme[5]. Althoughthe Aedes mosquito population density hasbeen reduced and maintained to a relativelylow level, as indicated by the overall houseindex, there was a progressive resurgence ofDF (but proportionately less DHF) with aperiodicity of about 5 to 6 years from 1992onwards[6-8].

In the last two decades, several studieshave investigated the risk factors for DF/DHFin affected communities, including those withpoor living conditions, social inequalities andilliteracy[3]. Identified DF/DHF risk factors varygreatly depending on the location, populationdensity, previous exposure to specific serotypesand availability of oviposition sites. Seasonaldistribution has also been reported with theAedes aegypti population density and DF/DHFincidence being associated with elevatedtemperature and rainfall in certain regions[9].However, not much is known about howsocioeconomic or demographic variables couldinfluence the occurrence of DF/DHF in urbancentres, such as Singapore. This may haveparticular importance since DF/DHF outbreaksare likely to initiate from urban centres[10,11].Geographical correlation analysis may help toanswer this question so as to shed some lighton providing various perspectives for publichealth policy-makers when designing controlmeasures for DF/DHF.

The aim of this study was to examinewhether or not there was any correlationbetween socioeconomic/demographic variablesand DF/DHF incidence by geographical areasusing the Development Guide Plan (DGP)zones in Singapore. Our findings have directand immediate implications for dengueprevention.

Materials and methods

Units of analysis

The analyses were based on the geographicalunits, namely, Development Guide Plan areas.The DGP is a detailed urban plan used for eachof the 55 planning areas in Singaporedesignated by the Urban RedevelopmentAuthority of Singapore, the nation’s planningand conservation authority. Each DGP covers aplanning area with a population of around150 000 served by a town centre. In order toobtain stability and reduce sampling variabilityof disease incidence, 4 DGP areas with lessthan 10 000 persons (ranging from 1085 to 9403persons) have been merged with the adjacentzones in our analysis. We also excluded 23DGP zones from our analyses that arecomposed of rural areas with very smallpopulation size in which no censusenumeration was conducted. In addition, onlya few cases of dengue (ranging from 0 to 11cases per year) were notified from these 23excluded DGP zones during the study period.Hence, only 28 of the DGP groups withenumerated population denominators wereincluded in the final analyses.

Population census

The socioeconomic/demographic (SED)variables by DGP were extracted from theSingapore Population Census 2000 reports[12]

and are described in Table 1. The proportionswith individual SED characteristics computedfor each DGP group were used as SED variablesfor the analysis.

Data analyses

The residential address of each notified case ofDF/DHF was first geo-coded into 28 DGP groups



Table 1: Spearman’s rank correlation coefficient (r) between dengue incidence and proportionswith individual socioeconomic/demographic characteristics (SED1-18), and average densities of

Aedes aegypti and Aedes albopictus (%) (AED1-2)[based on 28 Development Guide Plan (DGP) groups in Singapore, 1998–2002]

Socioeconomic/demographic variable (denominator) r (p value)SED1: Landed properties* and others (resident population) 0.645 (0.001)SED2: Services production industries(working residents aged 15 years and over) 0.592 (0.002)SED3: 'Other' ethnic group (resident population) 0.560 (0.004)SED4: Aged 65 years and above (resident population) 0.559 (0.004)SED5: Non-owner (resident private households) 0.533 (0.007)SED6: Financial & business services industries(working residents aged 15 years and over) 0.500 (0.010)SED7: Widowed female(resident population aged 15 years and over) 0.475 (0.014)SED8: Economically inactive(resident population aged 15 years and over) 0.464 (0.016)SED9: Attending upper secondary education(resident students aged 5 years and over) 0.453 (0.019)SED10: Monthly gross income from work below SGD1000(resident private households) 0.430 (0.026)SED11: Female Chinese (resident population) 0.429 (0.026)SED12: No family nucleus** (resident private households) 0.427 (0.027)SED13: Female living alone(resident population aged 15 years and over) 0.403 (0.037)AED1: Average density of Ae. aegypti (%) 0.395 (0.040)SED14: Household size: 8 or above(resident private households) 0.394 (0.041)SED15: Household size: 1 (resident private households) 0.394 (0.041)SED16: Attending university education(resident students aged 5 years and over) 0.365 (0.058)AED2: Average density of Ae. albopictus (%) 0.273 (0.156)SED17: Indian (resident population) 0.260 (0.177)SED18: Workers: agriculture & fishery, craftsmen, etc.(working residents aged 15 years and over) -0.450 (0.019)

* Refers to residents who are living at residential unit with individual ground contact which do not include multi-level apartment buildings. Types of landed properties include bungalow/detached house, linked house, semi-detached, terrace house, town house, and cluster housing, etc. For the latter housing type, it is a hybrid betweenconventional landed housing and condominium housing and all these units have ground contact but with sharedfacilities similar to those found in condominiums. Most of these landed property buildings usually are less than 4floors per building.

** Refers to a household formed by a person living alone or living with others but which does not constitute anyfamily nucleus. Thus it can refer to individuals, but not necessarily to people living alone.



using ArcView software version 9.1 (ESRI,Redlands, CA). Crude DGP-specific DF/DHFannual incidence was calculated based on thetotal number of cases reported from 1998 to2002 inclusive, with the person-yearsdenominator being the sum of the annualpopulation estimates for each of the years. Thepopulation estimates were obtained byinterpolating or extrapolating linear trend of eachDGP zone from the Population Census 1990 tothe Population Census 2000 to account forpopulation changes over the study period.

Three levels of data analysis, namely,correlation analysis, factor analysis and linearregression analysis, were conducted in thisstudy.

Firstly, the possible geographical correlationbetween the crude DF/DHF incidence (on logscale) and each SED variable was assessed byusing the Spearman’s rank correlationcoefficient.

Secondly, the potential SED variablesidentified by the correlation analysis were thensummarized into factor scores obtained byusing the exploratory factor approach, whichexamine how underlying constructs influencethe response on a number of measuredvariables[13]. In the factor analysis, the maximumlikelihood estimation was used to determinethe number of factors to retain, followed byorthogonal (varimax) rotation to assist in theinterpretation of the factors and to ensure thatthe factors were uncorrelated. SED variableswith rotated factor loadings (equivalent toPearson’s correlation coefficients between eachvariable and each factor) having absolute valuesof 0.6 or greater were used in interpreting thefactors and considered “dominant” as thedefining SED variables for the identification ofspecific factors[13]. Scores were computed forrotated factors as the sum of products ofobserved variables multiplied by their factorloading.

Thirdly, six weighted linear regressionmodels were then used to study theassociations between each of the respectivefactor score-based variables and the DGP group-specific DF/DHF incidence, with or withoutadjustment for mosquito indices, as a sensitivityanalysis (namely, models 1–6 in Figure 1). Thisanalysis allowed for the investigation of theassociation between the socioeconomicinfluence and DF/DHF incidence because factorscore-based variables minimized the multi-co-linearity problem present in conventionalregression analyses. It, therefore, allowed usto obtain a stable estimation. Log-transformation of the DGP group-specific DF/DHF incidence was taken as a dependentvariable. The regression was weighted by theDGP group’s population size.

The average densities of the two Aedesmosquitoes were defined as the total numberof Ae. aegypti and Ae. albopictus larvaeobserved, divided by the number of premiseswhere larval breeding was found, respectively.These two mosquito variables were alsosubsequently included together with 2-factorsscore-based variables in the weighted linearregression analyses.

All analyses were performed with S-Plussoftware version 6.0 (Insightful Corporation,Seattle, Washington) and Stata software version8.0 (Stata Corporation, College Station, TX,USA).

Results

Geographical variations in DF/DHFincidence

During the study period (1998–2002), therewere 16.4 million person-years of observationwith 11 888 reported cases of DF/DHF (afterexcluding 210 cases that occurred in the non-



sampled DGP zones). Figure 2 shows theincidence of DF/DHF in each of the 28 DGPgroups studied in Singapore. The overallincidence of DF/DHF was 72.7 per 100 000person-years. The incidence of DF/DHF rangedfrom 18.8 to 271.2 per 100 000 and all theincidence rates were significantly different(p<0.01) from the DGP average, except fortwo DGP groups (DGP groups 16 and 17)(Figure 2).

Association with socioeconomic/demographic factors

From the pair-wise ecological correlationanalyses (Table 1), the following variables weremost significantly associated (p<0.01) with theincidence of DF/DHF: landed properties andothers (r=0.65); services production industries(0.59); and ‘other’ ethnic group (0.56); aged65+ (0.56); and non-owner tenancy (0.53).

Figure 1: Schematic diagrams of respective weighted linear regression model used

Model 1: Scores of Factor 1 DF/DHF incidence

Model 2:

Model 3:

Ae. aegypti

DF/DHF incidence

Model 4:

Ae. albopictus

DF/DHF incidence

Model 5: DF/DHF incidence

Scores of Factor 2 DF/DHF incidence

Scores of Factor 1

Scores of Factor 2

Scores of Factor 1

Scores of Factor 2

DF/DHF incidence

Ae. aegypti

Ae. albopictus

Model 6:



In the factor analysis, two factors wereextracted based on 16 SED variables that couldexplain 80% of the total variations (Table 2).The first factor with the first six greatest inabsolute value of factor loadings wereconsidered “dominant”: low gross monthlyincome from work (<Singapore dollars,SGD1000), no family nucleus, aged 65+, livingalone, female widowed and economicallyinactive. This is consistent with “retiredelderly”. Likewise, for the second factoridentified had: workers in the businessindustries, ‘other’ ethnic group (minority livingin Singapore), household size 8 or above,landed property residents, and attending uppersecondary school as dominant variables (Table2). This appears to be consistent with “middle-class adults”.

These 2-factors score-based variables werethen included in the weighted linear regressionanalysis. The two indices for the average density

of Aedes mosquitoes were also included in theregression analysis for further adjustment. Fourvariables together could explain 32% of thevariations (R2=0.32) in the regression model(model 6 in Table 3). However, only scores inthe first factor consistent with “retired elderly”remained significantly and positively associated(p=0.022) with DF/DHF incidence.

Discussion

Singapore is a modern and highly urbanizedtropical island city state with one of the highesturban population densities in the world (6004residents per square kilometre). In this study,substantial geographical variations in theincidence of DF/DHF were observed, and thesevariations have been shown to be associatedwith differences in the socioeconomic/demographic characteristics of the population.

Figure 2: Incidence rates per 100 000 person-years of dengue fever by Development GuidePlan (DGP) group in Singapore, 1998–2002

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

050

100

150

200

250

DGP group

DGP Group-wide incidenceInci

den

cep

er1

00

00

0



Exploratory factor analysis is a data-reduction statistical method that has beenwidely used to identify and summarize manyinter-relationships that exist among individualvariables[13]. In a factor analysis used in thisstudy (schematic diagram shown in Figure 3),inter-correlated variables (variables SED1-16)are grouped into smaller numbers of newvariables (2 factors). Such an approach allows

for the simplification of the data set analysedin order to gain practically relevant insights intothe underlying risk factors and true exposuresthat are linked to adverse health effects.

Our study had several methodologicallimitations. Firstly, the actual extent of dengueinfection is probably an underestimate since itis based on the notified cases. However, the

Table 2: Results of factor analysis estimated using maximum likelihood estimation:rotated factors and factor loadings

[based on 28 Development Guide Plan (DGP) groups in Singapore, 1998-2002]

Socioeconomic/demographic variable (denominator) Factor 1 Factor 2(Retired elderly) (Middle-class adults)

SED1: Landed properties and others (resident population) 0.826SED2: Services production industries(working residents aged 15 years and over) 0.724SED3: 'Other' ethnic group (resident population) 0.938SED4: Aged 65 years and above (resident population) 0.962SED5: Non-owner (resident private households) 0.906SED6: Financial & business services industries(working residents aged 15 years and over) 0.956SED7: Widowed female(resident population aged 15 years and over) 0.945SED8: Economically inactive(resident population aged 15 years and over) 0.904SED9: Attending upper secondary education(resident students aged 5 years and over) 0.747SED10: Monthly gross income from work below SGD1000(resident private households) 0.995SED11: Female Chinese (resident population) 0.622SED12: No family nucleus (resident private households) 0.963SED13: Female living alone(resident population aged 15 years and over) 0.784SED14: Household size: 8 or above(resident private households) 0.877SED15: Household size: 1 (resident private households) 0.938SED16: Attending university education(resident students aged 5 years and over) 0.709Percentage total variance 46.3% 34.1%Percentage cumulative variance 46.3% 80.4%

Data are factor loadings, the correlation between the individual variable and each factor. Only variables withloadings ≥ ±0.60 are shown. SED17 and SED18 are not shown because their loadings < ±0.60.



Table 3: Summary of multiple linear regression analysis weighted by the DGP group’spopulation size

Model 1: DF/DHF incidence = Scores of 0.391 – – – 20.7%Factor 1 (0.015)

Model 2: DF/DHF incidence = Scores of Factor 2 – 0.313 – – 8.0%(0.145)

Model 3: DF/DHF incidence = Av. density of – – 0.041 – 11.7%Ae. aegypti % (0.075)Model 4: DF/DHF incidence = Av. density of – – – 0.035 8.8%Ae. albopictus % (0.126)

Model 5: DF/DHF incidence = Scores of 0.383 0.293 – – 27.7%Factor 1 + Scores of Factor 2 (0.015) (0.132)

Model 6: DF/DHF incidence = Scores of 0.399 0.092 0.036 –0.007 32.0%Factor 1 + Scores of Factor 2 + Av. density (0.022) (0.724) (0.256) (0.785)of Ae. aegypti % + Av. density of Ae. albopictus %

Reg. Coef.† Reg. Coef.† Reg. Coef.† Reg. Coef.†

(p value) (p value) (p value) (p value)

Model

Variable(s) included in the regression analysis

R2‡

Factor 1*(Retiredelderly)

Factor 2*(Middle-

classadults)

Averagedensity of

Ae. aegypti(%)

Averagedensity of

Ae.albopictus

(%)

proportion of sub-clinical to clinical cases maynot be as high as previously reported[14] inSingapore since almost 90% of the total caseshappen in adults[7,15]. If this discount factoroccurred randomly across all geographical areas,the magnitude of ecological correlation analysesbased on relative geographical variations wouldonly tend to bias toward the null[16]. In addition,regarding the size and choice of a geographicalunit, a recent study showed that for a giventime point and deprivation score, thedeprivation gap in crude survival was some 25times smaller when estimated with largergeographical units than with small ones. Thissuggests that our estimates are conservative[17].

Secondly, like other dengue-endemiccountries, the DF/DHF incidence may beaffected by various environmental factors, such

* Scores of Factor 1 and Factor 2 obtained from factor analysis and socioeconomic variables contributed to eachfactor referred to in Table 2.† Regression coefficient estimated from weighted linear regression.‡ R-squared interpreted as percentage of variation explained by respective model.

as rainfall, humidity and temperature. A localstudy showed that these meteorologicalconditions preceded the DF/DHF incidence by8–20 weeks[18]. While there may be differencesin such environmental factors in different partsof Singapore, these may not be significant giventhe very small geographical size of the island.Furthermore, as the DF/DHF incidence wasaggregated over a 5-year period in our analysis,the short- and medium-term fluctuations ofmeteorological conditions are less likely to havean impact on our analysis.

Thirdly, the association betweensocioeconomic/demographic factors and DF/DHF incidence was assessed based onecological data. Therefore, as with anyecological analysis, interpretation of thesefindings must be done with caution. Individual-



based studies are needed to validate thehypothesis generated from these findings[19].

In a cancer study conducted in Thailand,the socio-demographic characteristics thatinfluenced the decision-making of the patient’scaretaker to receive alternative therapy includedthe level of education, occupation, residentialareas and lay symptom assessment[20]. For theeconomic factors, the capability to reimbursethe cost of treatment, the family income andthe financial resources were also important[20].In our study, the DF/DHF incidence wasecologically associated with somesocioeconomic/demographic characteristics of

the population, such as those with low income(economically inactive; Spearman’s r=0.46 ormonthly gross income from work belowSGD1000; 0.43), living alone (household size1; 0.39), gender (female living alone; 0.40 orwidowed female; 0.48), and less attention paidto environmental hygiene (aged 65+; 0.56, nofamily nucleus; 0.43, household size 1; 0.39,and widowed female; 0.48), a group we havecollectively referred to as “retired elderly”. Insuch a population group, the Aedes larvalbreeding sites in the domestic and peri-domesticenvironment could increase due to poorhygiene[21] and failure to check for breeding andreluctance to have their homes fogged with

Figure 3: Schematic diagram showing inter-correlated variables (please see variablessocioeconomic/demographic variables: SED1-16 in Table 1) grouped into smaller numbers of

new variables (2 uncorrelated factors: retired elderly and middle-class adults) using data-reduction statistical method, factor analysis

Factor 1:Retiredelderly

Factor 2:Middle -class

adults

SED4

SED1

SED7

SED8

SED10

SED5

SED15

SED16

SED13

SED12

SED2

SED3

SED6

SED11SED14

SED9

Factor 1:Retiredelderly

Factor 2:Middle -class

adults

SED4

SED1

SED7

SED8

SED5

SED16

SED13

SED2

SED3

SED6

SED11SED14

SED9



insecticide[22]. In addition, non-owner tenancyhouseholders (r=0.53) could be less responsiblein cleaning up their premises. Living in landedproperties was also associated with a higher DF/DHF incidence (r=0.65) as it has beenconsistently observed that there are morebreeding habitats in these premises[5].

We found that areas with a high proportionof socioeconomically disadvantaged residentshave significantly higher incidence rates of DF/DHF. Ae. aegypti and Ae. albopictus populationdensities, taken individually without theinclusion of other factors, are insufficient toaccount for the observed difference in the DF/DHF incidence rates (Models 3 and 4 in Table3). It is interesting to note that in the samemultivariable regression analysis, the regressioncoefficient of the average density of Ae.albopictus on DF/DHF incidence wassignificantly reduced (regression coefficients:0.035 in Model 4 vs –0.007 in Model 6), butthe Ae. aegypti variable had less reductionalthough both variables did not reach statisticalsignificance (regression coefficients: 0.041 inModel 3 vs 0.036 in Model 6, Table 3). Thissuggests that Ae. aegypti remains the principalvector for dengue virus transmission, despitethe greater abundance of Ae. albopictus.

Because this finding is based on ecologicaldata, we cannot conclude that persons frompoor families have a higher risk of DF/DHFwithout further prospective studies. However,it is consistent with the hypothesis thatsusceptibility to infection is associated with lowsocioeconomic status[23]. The higher DF/DHFincidence in the socioeconomicallydisadvantaged residents could also likely be dueto socio-behavioral barriers in seeking healthcare[24,25] or some other behavioural orenvironmental processes operating at householdor individual levels that supported breeding ofAedes mosquitoes (e.g. monthly gross incomefrom work below SGD1000 was correlated with

average densities of Ae. aegypti, r=0.703, datanot shown) and transmission of dengueviruses[26].

Until a safe and effective vaccine isavailable, controlling Aedes mosquitoes, activelaboratory-based case as well as entomologicalsurveillance[27,28], and improved casemanagement are the principal options availablefor reducing the burden of DF/DHF inSingapore. More cost-effective integratedcontrol measures such as public healtheducational campaign targeting ‘hot-spot’ areas,in which both DF/DHF incidence and factorscores are high, could be a logical approach tominimize the impact of the disease[26-30]. [The‘high-high’ areas defined as the first one-thirdof the DGP groups with high DF/DHFincidence accounted for 28.3% of the totalannual cases but only 10.4% of the totalpopulation size, and the first one-third of thefactor scores (data not shown)].

In conclusion, the results of this studysuggest that dengue control in Singapore couldbenefit with a more focused effort in outreachto the population in relatively lowersocioeconomic regions. Further efforts shouldbe directed at addressing the barriers tobehavioural change, correcting misconceptionon the spread of dengue by social and closecontact, and educating them and the illiterateon measures to prevent dengue[22].

Acknowledgements

We thank Ms J.K.Y. Wong, Ms E. Loh, Mr C.T.Heng and Ms P.Y. Soh of the NationalEnvironmental Agency for their assistance ingeo-coding of data to the DGP zones. We alsothank Ms L. Kurupatham and Dr P.L. Ooi ofthe Communicable Diseases Division and DrY.B. Cheung for their helpful comments on anearly version of the manuscript.



References

[1] Monath TP. Dengue: the risk to developed anddeveloping countries. Proc Natl Acad Sci.1994; 91: 2395-400.

[2] Gubler DJ. The global pandemic of dengue/dengue haemorrhagic fever: current status andprospects for the future. Ann Aca MedSingapore. 1998; 27: 227-34.

[3] Guzman MG, Kouri G. Dengue: an update.Lancet Infect Dis. 2002; 2: 33-42.

[4] Kay B, Vu SN. New strategy against Aedesaegypti in Vietnam. Lancet 2005; 365: 613-7.

[5] Goh KT. Dengue – re-emerging infectiousdisease in Singapore. In: Goh KT (ed).Dengue in Singapore. Technical MonographSeries 2. Inst itute of EnvironmentalEpidemiology. Ministry of the Environment.1998. pp. 33-49.

[6] Goh KT. Changing epidemiology of dengue inSingapore. Lancet. 1995; 346: 1098.

[7] Ooi EE, Hart TJ, Tan HC, Chan SH. Dengueseroepidemiology in Singapore. Lancet. 2001;357: 685-6.

[8] Ooi EE, Goh KT, Gubler DJ. Dengueprevention and 35 years of vector control inSingapore. Emerg Infect Dis. 2006; 12: 887-93.

[9] Hales S, Weinstein P, Woodward A. Denguefever epidemics in the South Pacific: drivenby El Nino Southern Oscillation? Lancet. 1996;346: 1664-5.

[10] Cummings DA, Irizarry RA, Huang NE et al.Travelling waves in the occurrence of denguehaemorrhagic fever in Thailand. Nature.2004; 427: 344-7.

[11] Gubler DJ. Cities spawn epidemic dengueviruses. Nature Medicine. 2004; 10: 129-130.

[12] Singapore Census of Population 2000.Statistical release 4: geographic distribution andtravel. Singapore: Department of Statistics.2001.

[13] Kleinbaum DG, Kupper LL, Muller DC.Applied regression analysis and othermultivariable methods. Boston: KentPublishing Company, 1988.

[14] Burke DS, Nisalak A, Johnson DE, Scott RM. Aprospective study of dengue infections inBangkok. Am J Trop Med Hyg. 1988; 38:172-80.

[15] Koh BKW, Ng LC, Kita Y et al. The 2005 dengueepidemic in Singapore: epidemiology,prevention and control. Ann Acad MedSingapore. 2008. 37: 538-45.

[16] Brenner H, Savitz DA, Jöckel KH, GreenlandS. Effects of nondifferential exposuremisclassification in ecologic studies. Am JEpidemiol. 1992; 135: 85-95.

[17] Woods LM, Rachet B, Coleman MP. Choiceof geographic unit influences socioinequalities in breast cancer survival. Brit J Can.2005; 92: 1279-82.

[18] Heng BH, Goh KT, Neo KS. Environmentaltemperature, Aedes aegypti house index andrainfall as predictors of annual epidemics ofdengue fever and dengue haemorrhagic feverin Singapore. In: Goh KT (ed). Dengue inSingapore. Institute of EnvironmentalEpidemiology. Singapore: Ministry of theEnvironment. Technical Monograph Series 2,1998. pp.138-49.

[19] Hofer TP, Wolfe RA, Tedeschi PJ, McMahonLF, Griffith JR. Use of community versusindividual socioeconomic data in predictingvariation in hospital use. Health Sev Res. 1998;33: 243-59.

[20] Okanurak K, Sornmani S, Mas-ngammueng Ret al. Treatment seeking behavior of DHFpatients in Thailand. Southeast Asian J TropMed Pub Hlth. 1997; 28: 351-8.

[21] Danis-Lozano R, Rodriguez MH, Hernandez-Avila M. Gender-related family head schoolingand Aedes aegypti larval breeding risk inSouthern Mexico. Salud Publica Mex. 2002;44: 237-42.



[22] Heng BH, Goh KT, How ST et al. Knowledge,attitude, belief and practice on dengue andAedes mosquito. In: Goh KT (ed). Dengue inSingapore. Technical Monograph Series 2.Singapore: Institute of EnvironmentalEpidemiology, Ministry of the Environment.1998. pp.167-83.

[23] Deolalikar AB, Laxminarayan R.Socioeconomic determinants of diseasetransmission in Cambodia. Resources for theFuture July 2000: Discussion paper 00-32.http://www.rff.org, 2002.

[24] ICDDR Bangladesh. Dengue illnesses inhospitalized patients in Dhaka, 2001. Healthand Science Bulletin. 2002; 1: 2-6.

[25] Renganathan E, Parks W, Lloyd L et al.Towards sustaining behavioural impact indengue prevention and control. DengueBulletin. 2003; 27: 6-12.

[26] Ali M, Wagatsuma Y, Emch M et al. Use of ageographic information system for definingspatial risk for dengue transmission in

Bangladesh: role of Aedes albopictus in anurban outbreak. Am J Trop Med Hyg. 2003;69: 634-40.

[27] Gubler DJ. Aedes aegypti and Aedes aegypti-borne disease control in the 1990s: Top Downor Bottom Up. Am J Trop Med Hyg. 1989; 40:571-8.

[28] Carstairs V. Deprivation indices: theirinterpretation and use in relation to health. JEpidemiol Com Hlth. 1995; 49 Suppl 2: S3-8.

[29] Gubler DJ, Casta-Valez A. A program forprevention and control of epidemic dengueand dengue hemorrhagic fever in Puerto Ricoand the U.S. Virgin Islands. Bull Pan Am HealthOrgan. 1991; 25: 237-47.

[30] Ho SC, Nam AC. Factor influencing theoutcome of health campaigns: a case study inSingapore. Int J Hlth Educ. 1980; 23: 247-52.


Forecasting dengue incidence in Dhaka, Bangladesh:A time series analysis

M.A.H. Zamil Choudhurya,b#, Shahera Banub, M. Amirul Islamc,d

aInternational Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), Mohakhali,Dhaka-1212, Bangladesh

bDepartment of Microbiology and Hygiene, Bangladesh Agricultural University,Mymensingh 2202, Bangladesh

cDepartment of Agricultural Statistics, Bangladesh Agricultural University, Mymensingh 2202, Bangladesh

dSouthampton Statistical Sciences Research Institute, University of Southampton, UK

Abstract

This article attempts to model the monthly number of dengue fever (DF) cases in Dhaka, Bangladesh,and forecast the dengue incidence using time series analysis. Seasonal Autoregressive Integrated MovingAverage (SARIMA) models have been developed on the monthly data collected from January 2000 toOctober 2007 and validated using the data from September 2006 to October 2007. The resultsshowed that the predicted values were consistent with the upturns and downturns of the observedseries. The SARIMA (1,0,0)(1,1,1)12 model has been found as the most suitable model with leastNormalized Bayesian Information Criteria (BIC) of 11.918 and Mean Absolute Percent Error (MAPE) of595.346. The model was further validated by Ljung-Box test (Q18=15.266 and p>.10) with nosignificant autocorrelation between residuals at different lag times. Finally, a forecast for the periodNovember 2007 to December 2008 was made, which showed a pick in the incidence of DF duringJuly 2008, with estimated cases as 689.

Keywords: Dengue; Time series analysis; SARIMA; Disease prediction; Dhaka, Bangladesh.

#E-mail: [email protected]; Phone: +610430504183

Introduction

Dengue is one of the most important emergingviral diseases of major public health concernin Bangladesh. The disease is transmittedthrough the bite of the Aedes aegypti and Ae.albopictus mosquitoes[1]. It causes a broadspectrum of clinical manifestations in humans

ranging from the acute febrile illness, denguefever (DF), to the life-threatening denguehaemorrhagic fever/dengue shock syndrome(DHF/DSS)[2].

Dengue was first reported as “Dacca fever”in Bangladesh in 1964 by Aziz and hiscolleagues[3]. Subsequent reports suggested that


Forecasting dengue incidence in Dhaka, Bangladesh: A time series analysis

dengue fever may have been occurringsporadically in Bangladesh from 1964 to 1999[3-

9]. The first epidemic of dengue was reportedin the capital city, Dhaka in the year 2000[10,11].Since then the disease has shown an annualoccurrence in all major cities of the country.During January 2000–December 2007,Bangladesh recorded a total of 22 245 casesand 233 deaths (1.04%). Of these, Dhakaaccounted for 20 115 cases and 181 deaths(0.9%).

In the absence of a vaccine and specifictreatment available for dengue, vector controlremains the only option. Early warning aboutthe disease based on forecasting, therefore,becomes crucial for the prevention and controlof dengue in Bangladesh. The time seriesanalyses methodology has been increasinglyused in the field of epidemiological researchon infectious diseases, particularly in theassessment of health services[12-16]. In healthscience research, Autoregressive IntegratedMoving Average (ARIMA) models[12-18] as wellas Seasonal Autoregressive Integrated MovingAverage (SARIMA)[19-20] models are useful toolsfor analysing time series data containing ordinaryor seasonal trends to develop a predictiveforecasting model. There have been efforts inforecasting dengue incidence in different partsof the world using both ARIMA[21,22] andSARIMA[19] modelling. This study is aimed atdeveloping univarite time series models toforecast the monthly dengue incidence inDhaka based on reported monthly casesavailable from 2000–2007. This forecastingoffers the potential for improved and consistentplanning of public health interventions.


Study area

Dhaka is the capital and principal city ofBangladesh located at 23° 42’ 0" N, 90° 22’

30" E, covering an area of 815.85 km2 (315 sqmiles). According to the World Gazetteer(2006), the population in the Dhaka region was11 million and the density was 14 608/km²(37 834.5/sq mile) making it the largest city inBangladesh and the eleventh most populouscity in the world[23]. The Dhaka region waschosen as the study area because of itsrelatively high incidence of DF between 2000and 2007 (average annual incidence: 2515.75cases).

Data collection

We obtained computerized data sets ofnotifications of monthly DF cases in the Dhakaregion for the period 1 January 2000 through31 October 2007 from the Directorate-Generalof Health Services, Mohakhali, Dhaka-1212[24].It may be noted that the Directorate-Generalof Health Services collects information ondengue cases separately as DF, DHF and DSSbut clubs this data as data for DF only.

Data analysis

A SARIMA (p,d,q)(P,D,Q)S model[25] was fitted,where p is the order of autoregression, d isthe order of integration, q is the order of movingaverage, P is the order of seasonalautoregression, D is the order of seasonalintegration, Q is the order of seasonal movingaverage and s is the length of seasonal period.The analyses were performed using SPSS 17software. The stationarity of the series wasmade by means of seasonal and non-seasonaldifferencing[25].Then the order of autoregressionand moving average were identified usingautocorrelation function (ACF) and partialautocorrelation function (PACF) of thedifferenced series. A model was fitted with atraining set of data from January 2000 toOctober 2007 and the fitted model was usedto predict values for a validation period (from



September 2006 to October 2007) to evaluatethe time series model. Most suitable modelswere selected on the basis of their ability ofreliable prediction. Two measures, namely,Normalised Bayesian Information Criteria(BIC)[26] and Mean Absolute Percent Error(MAPE)[25], were used. Lower values ofNormalised BIC and MAPE were preferable.Furthermore, Ljung-Box test (portmanteau test)was performed to test if the residual ACF atdifferent lag times was significantly differentfrom zero, where not being different from zerowas expected[27]. After the best model wasidentified, forecast for future values fromNovember 2007 to December 2008 was made.

Results and discussion

The observed series of DF (January 2000 toOctober 2007) shows that the series is non-stationary and there are seasonal fluctuationsin the dataset (Figure 1). ACF and PACF of oneseasonal differenced series (Figure 2a, Figure2b) as well as of one seasonal with one non-seasonal differenced series (Figure 2c, Figure

2d) instigated to explore a set of models basedon the training set of data (January 2000 toOctober 2007), which are listed in Table 1.Among these models, SARIMA(1,0,0)(1,1,1)12has both lowest normalised BIC (11.918) andMAPE (595.346) values and appeared to bethe best model. Moreover, the Ljung-Box testsuggested that the ACF of residuals for themodel at different lag times was not significantlydifferent from zero (Q18=15.266 and p>.10).All the coefficients of SARIMA (1,0,0)(1,1,1)12model were significant (Table 2). The modelhas been used to predict values from September2006 to October 2007 (Figures 3 and 4) forvalidation. It appeared that the predicted valuescould follow the upturn and downturn of theobserved series reasonably well. Finally, Figure5 represents the forecast values for the periodfrom November 2007 to December 2008,which indicates a seasonal pick in July 2008,with the estimated number of patients as 689and a sharp decrease from September 2008.The predicted values as well as the forecastvalues show some negative values, which is acommon case with a series with too many zerosas observed values in the series.

Figure 1: Observed dengue fever from January 2000 to October 2007

0

500

1000

1500

2000

2500

3000

3500

JAN

20

00

MA

Y2

00

0

SEP

20

00

JAN

20

01

MA

Y2

00

1

SEP

20

01

JAN

20

02

MA

Y2

00

2

SEP

20

02

JAN

20

03

MA

Y2

00

3

SEP

20

03

JAN

20

04

MA

Y2

00

4

SEP

20

04

JAN

20

05

MA

Y2

00

5

SEP

20

05

JAN

20

06

MA

Y2

00

6

SEP

20

06

JAN

20

07

MA

Y2

00

7

SEP

20

07



Table 1: MAPE and normalized BIC of Time Series Models

Models MAPE Normalized BICSARIMA(2,1,1)(1,1,0)12 1026.050 12.072SARIMA(2,1,0)(1,1,0)12 766.310 12.215SARIMA(1,1,1)(1,1,0)12 945.640 12.052SARIMA(0,1,0)(1,1,0)12 805.376 12.236SARIMA(1,1,0)(1,1,0)12 791.408 12.269SARIMA(1,1,1)(1,1,1)12 947.663 12.059SARIMA(1,1,1)(2,1,0)12 975.253 12.071SARIMA(1,0,1)(1,1,1)12 600.730 11.952SARIMA(1,0,1)(1,1,0)12 717.614 11.933SARIMA(1,0,0)(1,1,1)12 595.346 11.918

Table 2: Model parameters of SARIMA (1,0,0)(1,1,1)12

Variable βββββ SE p-valueAR (Lag 1) 0.385 0.106 0.000AR, Seasonal (Lag 1) -0.587 0.109 0.000Seasonal Difference 1.000MA, Seasonal (Lag 1) 0.405 0.151 0.009

Figure 2a: ACF of transformed series with seasonal difference (1, period 12)

CoefficientUpper Confidence LimitLower Confidence Limit

Lag Number

AC

F



Figure 2b: PACF of transformed series with seasonal difference (1, period 12)


Lag Number

Par

tial

AC

F

Figure 2c: ACF of transformed series with non-seasonal difference (1) andseasonal difference (1, period 12)


Lag Number

AC

F



Figure 2d: PACF of transformed series with non-seasonal difference (1) andseasonal difference (1, period 12)


Lag Number

Par

tial

AC

F

Figure 3: Dengue incidence: Observed values and predicted values of SARIMA (1,0,0)(1,1,1)12

-500

0

500

1000

1500

2000

2500

3000

3500

JAN

20

00

AP

R2

00

0

JUL

20

00

OC

T2

00

0

JAN

20

01

AP

R2

00

1

JUL

20

01

OC

T2

00

1

JAN

20

02

AP

R2

00

2

JUL

20

02

OC

T2

00

2

JAN

20

03

AP

R2

00

3

JUL

20

03

OC

T2

00

3

JAN

20

04

AP

R2

00

4

JUL

20

04

OC

T2

00

4

JAN

20

05

AP

R2

00

5

JUL

20

05

OC

T2

00

5

JAN

20

06

AP

R2

00

6

JUL

20

06

OC

T2

00

6

JAN

20

07

AP

R2

00

7

JUL

20

07

OC

T2

00

7

Observed Predicted



Figure 4: Dengue incidence – Observed values and predicted values of SARIMA(1,0,0)(1,1,1)12 for the period September 2006 to October 2007

-100

0

100

200

300

400

500

SEP

20

06

OC

T2

00

6

NO

V2

00

6

DEC

20

06

JAN

20

07

FEB

20

07

MA

R2

00

7

AP

R2

00

7

MA

Y2

00

7

JUN

20

07

JUL

20

07

AU

G2

00

7

SEP

20

07

OC

T2

00

7

Predicted

Observed

Figure 5: Forecast of dengue incidence from November 2007 to December 2008 bySARIMA (1,0,0)(1,1,1)12

-100

0

100

200

300

400

500

600

700

800

NO

V2

00

7

DEC

20

07

JAN

20

08

FEB

20

08

MA

R2

00

8

AP

R2

00

8

MA

Y2

00

8

JUN

20

08

JUL

20

08

AU

G2

00

8

SEP

20

08

OC

T2

008

NO

V2

00

8

DEC

20

08



Conclusion

The incidence of dengue fever every year inBangladesh, especially in Dhaka city, is aconstant threat to the population and a recurringproblem for the health authorities.Furthermore, all environmental conditions thatcan trigger an outbreak are more or less presentin the country. Forecasting a dengue outbreakcan help the authorities to take effectivemeasures to handle any unexpected situation.Such an effort is cost-effective considering thefinancial constraints of the health sector. Thepresent study is the first of its kind inBangladesh. The SARIMA results revealed thatthe number of dengue patients in 2008 willhave a seasonal pick with the highest value as689 in July, which is concordant with ourprevious experience. SARIMA models are well-practised tools in epidemiological researchwhich may offer further accuracy in predictionif some relevant variables[20,21, for example,temperature, humidity, rainfall, are consideredduring the modelling process. Efforts shouldbe made in the future to use such additional

information, which was not possible in thecurrent study due to lack of coordinationbetween different sources as well asdissimilarity of area of coverage by differentauthorities. Separate modelling approaches forDF, DHF and DSS would provide betterinformation to policy-makers and planners.Relevant data should be made available in atimely manner, possibly from one service point,with proper coordination between differentdata sources.

Acknowledgements

We thank the Disease Control Directorate,Directorate-General of Health Services, Dhaka,Bangladesh, for providing the data of notifieddengue cases between 2000 to 2007. We alsothank Professor Dr Md Alimul Islam for hissuggestions and inspiration. Finally, we wouldlike to thank the three anonymous reviewersfor their valuable comments on the previousversion of the manuscript.

References

[1] Rigau-Perez JG, Clark GC, Gubler DJ, Reiter P,Sanders EJ, Vorndam AV. Dengue and denguehaemorrhagic fever. Lancet. 1998 Sept 19;352(9132): 971-977.

[2] Nimmannitya S. Clinical manifestations ofdengue and dengue haemorrhagic fever:monograph on dengue and denguehaemorrhagic fever. WHO RegionalPublication SEARO No. 22. New Delhi: WorldHealth Organization, Regional Office forSouth-East Asia, 1983; pp. 48-53.

[3] Aziz MA, Graham RR, Gregg MB. “Daccafever”-an outbreak of dengue. Pak J Med Res.1967; 6: 83-92.

[4] Gaidamovich SY, Siddiqi SM, Haq F, KlisenkoGA, Melnikova EE, Obukhova VR. Serologicalevidence of dengue fever in the BangladeshRepublic. Acta Virol. 1980; 24: 153.

[5] Islam MN, Khan AQ, Khan AN. Viral disease inBangladesh: procedings of an InternationalSeminer on viral disease in South- East- Asianand Western Pasific, Feb. 8-12 Canbera,Australia, Washing DC, USA. Washington DC:Academic Press, 1982.

[6] Khan AM, Ahamed TU. Dengue status inBangladesh. Dengue News. 1986; 12-11.



[7] Amin MMM, Hussain AMZ, Murshed M,Chowdury IA, Banu D. Sero-diagnosis ofdengue infections by haemagglutinationinhibition (HI) in suspected cases in Chittagong,Bangladesh. Dengue Bull. 1999; 23: 34–38.

[8] Alam MN. Dengue and dengue haemorrhagicfever in Bangladesh. Orion Med J. 2000;6: 7-10.

[9] Hossain MA, Khatun M, Arjumand F, NisalukA, Breiman RF. Serologic evidence of dengueinfection before onset of epidemic,Bangladesh. Emerg Infect Dis. 2003 Nov; 9:1411-1414.

[10] Rahman M, Rahman K, Siddque A K, ShomaS, Kamal AHM, Ali KS, Nisaluk A, Breiman RF.First Outbreak of Dengue Hemorrhagic Fever,Bangladesh. Emerg Infect Dis. 2002 July; 8:738-740.

[11] Aziz MM, Hasan KH, Hasanat MA, SiddiquiMA, Salimullah M, Chowdhury AK, AhmedM, Alam MN, Hassan MS. Predominance ofthe DEN-3 genotype during the recentdengue outbreak in Bangladesh. SoutheastAsian J Trop Med Public Health. 2002; 33:42-48.

[12] Bowie C, Prothero D. Finding causes ofseasonal diseases using time series analysis. IntJ Epidemiol. 1981; 10: 87–92.

[13] Allard R. Use of time-series analysis ininfectious disease surveillance. Bull WorldHealth Organ. 1998; 76: 327–333.

[14] Helfenstein U. Box-Jenkins modelling of someviral infectious diseases. Stat Med. 1986; 5:37-47.

[15] Catalano R, Serxner S. Time series designs ofpotential interest to epidemiologists. Am JEpidemiol. 1987; 126: 724-731.

[16] Helfenstein U. Box-Jenkins modelling inmedical research. Stat Methods Med Res.1996; 5:3-22.

[17] Helfenstein U. The use of transfer functionmodels, intervention analysis and related timeseries methods in epidemiology. Int JEpidemiol. 1991; 20: 808–815.

[18] Abraham B, Ledolter J. Statistical Methods forForecasting. New York: Wiley, 1983. pp. 225-229, 336-355.

[19] Wongkoon S, Pollar M, Jaroensutasinee M,Jaroensutasinee K. Predicting DHF incidencein Northern Thailand using time series analysistechnique. Proceedings of World Academy ofScience, Engineering and Technology. 2007; 26:216-220.

[20] Naish S, Hu W, Nicholls N, Mackenzie JS,McMichael AJ, Dale P, Tong S. Weathervariability, tides, and barmah forest virusdisease in the Gladstone region, Australia.Environ Health Perspect. 2006; 114: 678-683.

[21] Hu W, Nicholls N, Lindsay M, Dale P,McMichael AJ, Mackenzie JS, Tong S.Development of a predictive model for RossRiver Virus Disease in Brisbane, Australia. Am JTrop Med Hyg. 2004; 71: 129-137.

[22] Promprou S, Jaroensutasinee M,Jaroensutasinee K. Forecasting DengueHaemorrhagic Fever Cases in SouthernThailand using ARIMA Models. Dengue Bull.2006; 30: 99-106.

[23] World Gazetteer. (Web site) http://world-gazetteer.com/ , 2006.

[24] Directorate General of Health Services.Dengue register, Disease Control Directorate.Dhaka: DGHS, 2000.

[25] Makridakis S, Wheelwright SC, Hyndman RJ.Forecasting: methods and applications. 3rd ed.New York: Wiley, 1998.

[26] Schwarz G. Estimating the dimension of amodel. Ann Statist. 1978; 6: 461-464.

[27] Brockwell P, Davis R. Introduction to time seriesand forecasting. New York: Springer, 2002.


Dengue vector surveillance in Hong Kong – 2007

M.W. Lee, M.Y. Fok

Food and Environmental Hygiene Department, Hong Kong Special Administrative Region, China

Abstract

A dengue vector surveillance programme was implemented in the city and port areas in Hong Kong. Asa result of the surveillance, only Aedes albopictus was detected to be present in various areas in summer.Aedes aegypti was, however, not detected in any area under surveillance. Although a rather high indexof 70.9% was recorded in July, the activity of Ae. albopictus was immediately brought down throughconcerted efforts of varies agencies and the public. The swift response of concerned agencies werefacilitated by the use of Geographic Information System (GIS) in the dissemination of surveillanceresults. Users were able to access the system at any time for the latest results of the surveillance for takingimmediate remedial measures. The public was also informed of the results regularly through theInternet and press releases to arouse awareness to prevent and control the local dengue vector.

Keywords: Dengue vector surveillance; Hong Kong SAR, Aedes albopictus; Community efforts; Vector control.

Introduction

Hong Kong is located on China’s south coast(22°20"N and 114°11"E); it is surrounded bythe South China Sea on the east, south andwest, and borders the city of Shenzhen inGuangdong Province to the north over the ShamChun river. The territory’s 1104 sq. km (426sq. miles) land area consists primarily of HongKong Island, Lantau Island, Kowloon Peninsulaand the New Territories as well as some 260other islands. Hong Kong has a hilly terrain withsteep slopes. Most of the urban developmentexists on the Kowloon peninsula, along thenorthern edge of Hong Kong Island and inscattered settlements throughout the NewTerritories. Hong Kong exhibits a monsoonalclimate, in which the south-west monsoonoccurs from May to September, characterizing

Hong Kong’s hot, wet summers; while thenorth-west monsoon occurs from Novemberto March, bringing to Hong Kong cold, drywinters. Because of the climatic influence, mostof the annual rainfall occurs in summer andthe mean air temperature ranges between 25–28 °C. Even during winter, the temperatureranges between 15–21 °C.

Hong Kong is one of the world’s leadingfinancial centres. It is an important centre forinternational finance and trade with the largestconcentration of corporate headquarters in theAsia-Pacific region, and is known as one of thefour “Asian Tigers” for its high growth rates andrapid industrialization between the 1960s and1990s. The territory’s population also increasedsharply throughout the 1990s, reaching 6.99million in 2006.



Dengue fever has been made statutorilynotifiable in Hong Kong since 1994[1]. All theinfections reported to the Department ofHealth of Hong Kong Special AdministrativeRegion, China, are investigated to establishtheir source. The Department of Health worksjointly with the Department of Food andEnvironmental Hygiene of Hong Kong SpecialAdministrative Region, China, which plays theleading role in the control of the diseasevector. Between 1994 and 2001, the annualnumber of notifications ranged from 3 to 17cases; all these cases acquired the infectionfrom outside of Hong Kong (i.e. importedcases), mostly from South-East Asiancountries. In 2002, there were 36 confirmedcases recorded, of which 20 cases were locallyinfected. There was another local caserecorded in 2003 but none since 2004. Thenumber of imported cases remained at 31from 2004 to 2006 while these increased to58 in 2007 (Table).

In Hong Kong, a total of 13 Aedes specieshave been recorded that include Ae. albopictusand Ae. aegypti[3,4]. Ae. albopictus is one of themost commonly found mosquitoes in HongKong. It has wide distribution both in urbanand rural areas. Ae. aegypti, on the other hand,probably has not been an indigenous speciesin Hong Kong. It was once discovered on boarda vessel from another country in mid-1950s.In 2000, a dengue vector surveillanceprogramme, using ovitraps at selected sites, tomonitor and evaluate the effectiveness ofdengue vector control work carried out byvarious agencies, and for making timelyadjustments to dengue vector control strategiesand measures, was put in place by the Foodand Environmental Hygiene Department. Theprogramme was expanded in 2003 with anincrease in the areas covered and the frequencyof surveillance. The surveillance programmewas further extended in 2004 to cover all majorport areas, including all seaports. A denguevector surveillance programme by using ovitrapshad already been in place for the Hong KongInternational Airport since 1998.

Methods and materials

The oviposition trap (ovitrap) was used in thissurveillance programme as a tool to detect theprevalence and distribution of aedinemosquitoes. The device was locallymanufactured as per specification. It comprisedof a simple plastic container of approximately200 ml capacity, painted black inside with astraight and slightly tapered sides. The openingmeasured 6.5 cm in diameter, the basediameter was 5.0 cm, and the container was10.0 cm in height. The ovitrap was covered bya black cap with four openings and a grey-colourumbrella-shape raised cover to protect thecontent inside the ovitrap from contaminationby unwanted materials. A brownish woodentongue depressor was placed diagonally insidethe container as an oviposition paddle.

Table : Number of imported andindigenous dengue fever cases from

1994 to 2007[2]

Year Imported Indigenous Totalcases cases

1994 3 0 31995 6 0 61996 5 0 51997 10 0 101998 15 0 151999 5 0 52000 11 0 112001 17 0 172002 24 20 442003 48 1 492004 31 0 312005 31 0 312006 31 0 312007 58 0 58



Thirty-eight areas with high humanconcentration were selected such as housingestates, schools and hospitals. All the 38 areaswere surveyed every month to closely monitorthe situation of each location and to obtain aterritory-wide picture of the vectorial situation.On an average 55 ovitraps were placed at eachselected site. The ovitraps were set at adistance of about 100 m from each other forone week and collected back to the laboratory.The percentage of positive ovitraps wasrecorded as the ‘Ovitrap Index’. To serve as aquick reference for taking prompt follow-upmosquito control actions, each of the ovitrapcollected was examined immediately for thepresence of mosquito larvae. The larvae foundwere identified under compound microscopeto species level and the Provisional OvitrapIndex (POI) was worked out. The ovitraps werethen incubated at room temperature for oneweek for the eggs in the ovitraps, if any, tohatch out. The number of ovitraps found withAe. albopictus or Ae. aegypti in the first andsecond examination were pooled together for

the calculation of the Area Ovitrap Index (AOI).Another index, Monthly Ovitrap Index (MOI),was then calculated by pooling the results ofall the ovitraps retrieved in the same monthfrom the 38 areas which reflected the overallvector situation of the month.

A total of 33 land ports, which arecategorized into seven groups according to thenature of the ports, were also surveyed. Twentyovitraps were used at land ports and 650ovitraps were used in the airport.

Results

Community surveillance

The MOIs of 2007 followed a similar trend asprevious years but were generally lower(Figure 1). The MOIs in the first quarter weremaintained at a rather low level of 0.2% to1.4%. However, the indices rose gradually inthe second quarter from 7.6% in April to 20.7%

Figure 1: Comparison of Monthly Ovitrap Index of 2007 with the average of previous years(from 2000 to 2006)

7.60.9

0.2

3.20.8

22.7

1.64.3

19.4

23.8

0.3

14.517.2 14.6

10.3

0.2 0.3

1.4

20.7

23.1

15.111.3 8.7 4.50.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

1 2 3 4 5 6 7 8 9 10 11 12Month

Average of Monthly Ovitrap Index from 2000 to 2006

Monthly Ovitrap Index for 2007

Ovi

trap

Ind

ex(%

)



in June and reached a peak of 23.1% in July.Although the MOI recorded in July 2007(23.1%) was the highest recorded in July since2003 and higher than the average MOI of Julyfrom 2000 to 2006 (19.4%), it was lower thanthe average MOI of June from 2000 to 2006(23.8%). A marked drop from 23.1% to 11.3%was observed in August and the MOIs declinedgradually thereafter and reached the lowest inDecember (0.2%).

In respect of individual survey areas, onlyone AOI exceeded 20% in April. The numberof AOIs greater than 20% increased sharply to12 in May and further to 18 in June. Threelocations were found to have AOIs greater than40% in June and increased further to 6 in July.A record high AOI of 70.9% was also recordedin July. After reaching the peak in July, theindices came down rapidly in August. All AOIsrecorded in August were lower than 40%. Thenumber of AOIs reaching 20% also decreasedfrom 16 in July to 8 in August and 3 inSeptember. The indices remained at a lowerlevel in the last quarter. Activity of Aedinemosquitoes was not detected in most of thesurvey areas after November.

Port surveillance

In 2007, the Port Monthly Ovitrap Index(PMOI) ranged from 0.0% in January throughFebruary to 3.2% in June. The variation inPMOIs showed a similar trend as in previousyears (Figure 2). The ovitrap indices of all portgroups were below 20.0%. The highest indexof 13.8% was recorded in the port group ofCross Boundary Check Points on Land in June.The ovitrap index at the Hong KongInternational Airport was also the highest inJune (2.6%). In the months of June, July andAugust, all port groups had records of positiveindices.

Discussion

The results of the urban and port areassurveillance indicated that Ae. albopictusexisted in various areas in summer. Thebreeding places of the vector include a varietyof small water bodies such as discardedbuckets, empty lunch boxes, sand pits, surfacedrainage channels, keyholes of manhole covers,bamboo stumps, and saucers underneath plant

Figure 2: Comparison of Port Area Ovitrap Index: 2004-2006 and 2007

0 0 0.30.9 1

3.2 2.9

1.3 1.6

0.3 0.1 0.1

0.1 0.1 0.00.6

3.0

4.13.4

2.83.3

1.20.4 0.1

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

Ovi

trap

Ind

ex(%

)

2007 Port Area Monthly Ovitrap Index

Average Figure of Port Area Monthly Ovitrap Index (2004-2006)



pots. High ovitrap indices were recordedrepeatedly in some of the areas covered bythe surveillance programme, indicating thepresence of persistent breeding grounds thatneeded particular attention. Ae. aegypti, theimportant vector for the transmission of denguefever and yellow fever, was however notdetected in all the areas covered by the urbanand port surveillance programmes.

It was well recognized that communityparticipation was the key to success incontrolling mosquitoes, particularly denguevectors, and an annual territory-wide anti-mosquito campaign was organized to promotecommunity participation and forge closepartnership of government departments andnongovernmental organizations in controllingthe mosquitoes. The dengue vector surveillanceprogramme served as a tool not only to monitorthe local dengue vector distribution but also toprovide objective information for takingappropriate actions by the community againstdengue vectors. The Area Ovitrap Index andMonthly Ovitrap Index numbers were releasedto the public through press releases and theInternet to arouse awareness in preventingmosquitoes. Government departments wereable to access detailed information of thesurveillance, including location of positiveovitraps through a Geographical InformationSystem which is accessible by registered usersthrough the government intranet. They areable to target mosquito control action at venuesthat fall within the 100 m radius of all positiveovitraps under their purviews. The people werealso advised to pay particular attention to anywater accumulation in and near theirresidences. A detailed and comprehensiveadvice on mosquito prevention and control wasissued together with the press release. Thepublic was also able to access the informationthrough the Internet.

For operational purposes, the ovitrapindices were classified into four different

categories – Level One: for indices less than5%; Level Two: for indices between 5% andless than 20%; Level Three: for indicesbetween 20% to less than 40% and Level Four:for indices at 40% or above. Different actionswere taken based on the levels reached. Atlower levels (levels 1–3), control measuresmainly relied on source reduction, e.g. properdisposal of disused articles, lunch boxes,containers, etc. Potential breeding sites suchas saucers underneath plant pots, surfacedrainage channels, roadside gully traps orkeyholes of manhole covers were inspectedweekly and accumulation of water wasremoved promptly. Larvicides were appliedwhenever immediate elimination of breedingsources was not feasible. When the OvitrapIndex reached Level Four, space spaying ofinsecticides was carried out at the resting placesof the adult mosquito to contain the mosquitoproblem.

On health education, health talks wereorganized for schoolchildren, managements ofestates, construction sites as well as localorganizations such as area committees todisseminate the message of mosquitoprevention and control. Training was alsoorganized for pest control personnel in thegovernment. Operatives of pest controlcontractors providing mosquito control servicesfunded by the government were also requiredto receive proper training on general pestcontrol, including mosquito control and denguefever.

Conclusions

According to the results of the dengue vectorsurveillance in 2007, Ae. aegypti was notdetected and the activity of Ae. albopictus was,in general, under control. The Monthly OvitrapIndices were mostly lower than the averagesof the past few years except in July where asurge in the ovitrap indices was observed.



However, with concerted efforts made andswift actions taken by relevant agencies andthe public, the indices were brought downquickly in the following month and maintainedat a lower level till the end of the year. Thisindicates that the vector problem had beenput under control in 2007.

Active participation of the government,local organizations and the public were the

key to success in controlling dengue vector.The results of dengue vector surveillancewere released to the public and other partiesconcerned through different channels tofacilitate prompt remedial actions. Timelytarget-specific control efforts were achievedthrough the coordination of district-basedanti-mosquito task force led by thegovernment.

References

[1] Dengue Fever in Hong Kong. Hong Kong MedJ. 2008; 14: 170-7.

[2] Department of Health, Government of HongKong Special Administrative Region, China.Web site: http://www.gov.hk/

[3] Department of Food and EnvironmentalHygiene, Government of Hong Kong Special

Administrative Region. China Mosquitoes ofHong Kong. Mosquitoes of Hong Kong, 2005.

[4] Chau G.W. An illustrated guide to theidentification of the mosquitoes of Hong Kong.Hong Kong: Urban Council Publication, HongKong Government, 1982.


Re-emergence of dengue in Argentina:Historical development and future challenges

Héctor Masuh#

Pest and Insecticide Research Centre (CIPEIN-CITEFA/CONICET/UNSAM),Juan Bautista de LaSalle 4397 – B1603ALO – Buenos Aires, Argentina

Abstract

After 82 years of the absence of dengue in Argentina, a dengue outbreak occurred in the northernprovinces of the country in 1998. Aedes aegypti, the vector mosquito, was eradicated in the 1960s,mainly due to the use of residual insecticides at an enormous cost of resources and through a verticalhealth programme. Since then, the country has gradually become reinfested due to the deterioration ofthe surveillance system and vector control programmes. At present, DENV-1 to 3 have been found incirculation and 3162 cases of dengue fever (DF) have been reported in the country. However, asautochthonous cases have been recorded during this epidemic only, the disease is still not consideredendemic in the country, although there is a regular occurrence of outbreaks in neighbouring countries.

The control strategies currently being used are the same ones as used in the past century althoughsocioeconomic and demographic conditions have greatly changed. Consequently, alternative methodsare proposed as potential tools to establish new ways of controlling the vector, which is the only way ofpreventing new outbreaks in the region.

Keywords: Dengue; Argentina; Control strategies; Aedes aegypti.


Introduction

Argentina is the southern-most country in LatinAmerica. With a surface area of 3 761 274km2, it has a wide diversity of geographical areassuch as the cold and dry steppes of Patagonia,the Pampa grasslands, the humid and dry Chacoregion and the jungle highlands or “yungas” inthe north[1]. The great climatic and topographicdiversity of this vast extension of landdetermines different forms of fauna and flora,as well as different types of human settlements

that develop different lifestyles andsocioeconomic activities that are directly relatedto their environment.

The growth of urban centres, viz. the cityof Buenos Aires, where nearly 40% of thecountry’s population is concentrated[2], inconjugation with movement of people fromand to the neighbouring countries, supportedby congenial environmental conditions in thenorth and centre, render this country prone toexplosive epidemic outbreaks. The prevailing

Re-emergence of dengue in Argentina


socioeconomic aspects of Latin America ingeneral and Argentina in particular, especiallythe extreme polarization of resources, areextremely relevant in the re-emergence ofdengue.

During the mid-20th century, the healthauthorities of American countries, together withthe Pan American Health Organization (PAHO),carried out important Ae. aegypti eradicationcampaigns, which were developed in Argentinain 1965[3]. However, by the end of the 1980s,the country was re-infested by the mosquito,a situation that currently prevails[4].

The present article describes some of thevariables that contributed to the re-emergenceof dengue in Argentina, placing particularemphasis on mosquito vector control, anddiscusses possible contributions to the currentvector control strategies.

History of dengue fever inArgentina

The first outbreak of dengue in Argentina wasrecorded by Nicolás Gaudino[5] in 1916. Thevirus entered the country via Paraguay andaffected the provinces of Corrientes and EntreRíos. Although no cases were reported in thecity of Buenos Aires, it affected 50% of themesopotamic population.

Since then, in Argentina, the disease wasnot recorded for 82 years, in spite of theoccurrence of severe outbreaks in theCaribbean and Central America in the 1960s,and the later appearance of denguehaemorrhagic fever (DHF) in the Cubanepidemic of 1981 which spread to all the otherAmerican countries except Canada andUruguay[6]. During those eight decades, denguewas considered a problem affecting south-eastAsia and other far-off regions. However, it has

slowly re-entered our continent via CentralAmerica. Today, almost all the Americancountries from Mexico to the southern tip ofthe continent are affected by this disease[7].

In 1998, there was an epidemic causedby DENV-2 restricted to the Chaco-Salta regionof Argentina, with its epicentre in the city ofTartagal. The epidemic reached its peak inMay[8], which caused several hundreds of casesof dengue fever (DF) (incidence rate: 45/10 000 inhabitants). All indications suggest thatthe virus was introduced from Bolivia[9].However, this was just the beginning. Sincethen, a series of outbreaks have occurred inArgentina – in 1998, 2000, 2002, 2003, 2004,2006, 2007 and 2008 (Figure 1). Fiveprovinces, namely Salta, Jujuy, Corrientes,Formosa and Misiones reported autochthonouscases. More than 70% of the cases werereported in the province of Salta[10]. Onlyimported cases were reported in 2005, amongpeople having travelled to Bolivia, Paraguay,Brazil, Puerto Rico and Nicaragua. Figure 2shows the provinces affected by the outbreaks,active serotypes and relationship with outbreaksin neighbouring countries. At present, theoutbreaks of dengue in Argentina have always

Figure 1: Dengue fever cases in Argentinasince the re-emergence of the disease

1600

1400

1200

1000

800

600

400

200

0

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Source: Sistema Nacional de Vigilancia de la Salud(National System of Surveillance of Health), NationalMinistry of Health of Argentina



had a direct relationship with neighbouringcountries, with the entry of viraemic subjectsto initiate transmission. As such, Argentina isstill considered a non-endemic country[11].Three serotypes have been detected inArgentina since the first emergency situation,and have only appeared simultaneously in 2003in the province of Salta.

In addition to DENV-2, serotypes DENV-3and DENV-4 started circulating in the north-eastern frontier with limited epidemic potential

until 2004, when there was an extendedoutbreak with thousands of DENV-3 cases inseveral cities of the Chaco-Salta region. Despitethe circulation of several serotypes in successiveyears and sequential infections, no clinical casesof DHF had been detected[12].

In 2006, the situation in the north-easternfrontier was aggravated by floods. Dengueoutbreaks were recorded in the area ofEmbarcación in Salta and Puerto Iguazú inMisiones due to DENV-1. Sixty-nine cases were

Figure 2: Outbreak localization and its relationships with outbreaks in border countries by year,province and circulating serotype in Argentina, 1998–2007

Source: Sistema Nacional de Vigilancia de la Salud (National System of Surveillance of Health),National Ministry of Health of Argentina



detected in Salta and 112 in Misiones, all ofwhich were confirmed by a laboratory orepidemiological nexus[13]. This was followed byyet another outbreak in north-eastern Argentinaand Iguazú (province of Misiones): where 55and 90 cases were reported in the Chaco-Salteño area and in Iguazú respectively. Thelatter cases were mostly imported through thesignificant flow of people in the “triple frontier”area around the falls[14].

Towards the end of 2006, the authoritiesof Paraguay reported cases of dengue in thecity of Asunción, which rapidly developed intoa great epidemic. Like in the beginning of2006, DENV-3 probably entered from Brazilvia the state of Mato Grosso. With the entryof new DENV-3 serotype, the population ofAsunción, which was previously exposed toDENV-1 in 1999-2000, presented DHF casesas expected due to sequential infections. Thisevent marked a turning point in the history ofdengue in the region as it was the first timethat this severe clinical form was recognizedin Paraguay[15]. Although by mid-February 2007there were under 20 cases of DHF, seriouscases of classical dengue were detectedwithout plasma extravasation, and thephysiopathological and clinical event definingDHF. Such DENV-3 cases had been previouslyobserved in Brazil. The affected individualspresented acute attacks in one or severalparenchyma: myocarditis, brain haemorrhage,or hepatocellular deficiency. Acute symptomsappeared 48–72 hours after the onset ofdengue, sometimes in the absence of anyapparent bleeding and without thehaematocrit modifications as normallyobserved in DHF. The term “visceral dengue”has recently been coined to name this clinicalvariant, which must be carefully consideredin the event of circulation of DENV-3. Due tothe dengue epidemic situation in Paraguay,Argentine provinces are now considered high-risk areas[16].

Historical evolution of Ae.aegypti in Argentina

At the beginning of the 20th century, Ae. aegyptiwas present in every American country exceptCanada, from the southern states of UnitedStates to Buenos Aires, Argentina. In Argentina,it was widely distributed, covering 14 provincesin the northern and central regions of thecountry[17]. In 1947, a continental programmecoordinated by PAHO was launched toeradicate yellow fever and its vector, Ae.aegypti[18]. It started out as a highly successfulcampaign and by 1954 and 1962 achieved itsgoal in 18 continental countries, includingArgentina. Since 1962, only three additionalcountries have managed to eradicate thisvector. During the 1970s, the support formosquito surveillance and control programmegot slackened, with the result that Ae. aegyptire-infested. By 1995, Ae. aegypti had adistribution similar to that in the 1940s beforethe eradication effort was initiated. OnlyBermuda and Chile remained free of thisinfestation[19].

Presence of Ae. albopictus

In August 1998, the presence of Ae. albopictuswas reported in the locality of San Antonio,province of Misiones; in February 2004 it wasalso found in Eldorado, another locality inMisiones. These are the first reports of thisspecies in our country[20,21,22]. In the surveillancestudies performed during February and Marchof 2007 in the open spaces and suburbs of thecity of Puerto Iguazú, 24 foci of Ae. albopictuswere detected, 18 of which were shared withAe. aegypti[23]. The presence of Ae. albopictusconveys a potential risk in the epidemiologicalcontext of the region regarding the circulationand transmission of dengue, yellow fever andother related arboviruses[24].



Current situation

The reinfestation of this region with Ae. aegyptiforced the authorities to re-launch monitoringand control activities based on the new criteriaof health service decentralization establishedby PAHO. According to these norms, theNational Government transferred theresponsibility of monitoring and controlactivities to the local municipalities, contributingto them with supplies and staff training. Thisnew modality made it necessary to modify oldcriteria used by the centralized system,generating local difficulties in the provincialfacilities regarding their resources and stafftraining. Furthermore, it is still difficult tocombine criteria regarding the monitoringmethod, rational purchase of supplies(equipment, insecticides, security equipment,etc.) and development of control activities[25].

The high-risk situation of viral transmissionstill prevails in several localities in northArgentina despite the intervention of national,provincial and local governments, as well as ofNGOs, that have been working on vectorcontrol for several years. The Ae. aegypti indicesare still high enough to produce autochthonousoutbreaks. In most municipalities (Figure 3),the House Index (HI) (Ae. aegypti breedingsites/houses inspected) remains over 10%[26].Therefore, the entire northern region of thecountry must be considered a high-risk area.The current floods in Santa Cruz de la Sierra,Bolivia, put the provinces of Salta and Jujuy inan outbreak-prone area with the additional riskof yellow fever transmission as the flooded ruralareas being evacuated lie in the jungle yellowfever-endemic zone.

In 2008, 2 996 183 tourists arrived inArgentina from dengue-endemic neighbouringcountries, 285 073 of which entered fromParaguay. Approximately 46% of these touristsarrived by plane. In 2008 >2 400 000Argentines left the country via Buenos Aires

to travel to dengue-endemic countries. Thelevel of migration in border areas, especiallyin the tropical regions of northern Argentina,is under-reported[27]. The number of importeddengue cases in Buenos Aires and other citiesof Argentina detected during the currentperiod is substantially higher than the numberdetected in previous years.

During 2008, the National Ministry ofHealth reported only 28 cases of dengue inthe country, 9 of which were imported.National government workers together with thelocal provincial staff of Salta are currentlycarrying out intense house-by-house controlactivities against mosquito breeding sites, withthe collaboration of the community and usinginsecticide space spraying. These activities have

Figure 3: Ae. aegypti infestation in Argentinaby province. Cumulative values of 2008.Numbers indicate municipalities with the

presence of the vector

Source: Sistema Nacional de Vigilancia de la Salud(National System of Surveillance of Health),National Ministry of Health of Argentina



extended to the border town of Yacuiba incoordination with the health workers of thisBolivian district. No dengue deaths have beenreported as yet in Argentina[28].

Current control strategies

Ever since the outbreak in Tartagal in 1998, allthe routine and emergency activitiesrecommended for the control of dengue wereimplemented in the country by the NationalMinistry of Health. The monitoring, control andevaluation methods implemented were theclassical methods used for many years in similarsituations[29,30]. The necessary supplies andequipment were purchased and field workerswere trained on their correct usage. Anemergency control strategy included theapplication of ultra low volume (ULV) thermalfog spray treatments, portable mist blowers,and house-by-house focal treatment.Simultaneously, diffusion activities were carriedout to alert the population of the currentsituation. Adulticide treatments were onlyperformed during epidemics and not as ameans of prevention.

The active substances used wereTemephos® sand granules as a larvicide, andthe organophosphate Sumithion® and thepyrethroid Deltamethrin® as adulticides inspatial sprays in an oily base using gas oil assolvent. These are obviously not the best toolsfor implementing control activities in urbanareas where the inhabitants suffer a high degreeof exposure to the insecticides used.

Innovation in controlstrategies

Although certain epidemic outbreaks werecontrolled in some areas of northern Argentina,the inadequacy of implementing actionsextrapolated from similar situations in other

countries or regions with differentsocioeconomic conditions was soon obvious.We needed to modernize, improve, changeand/or adapt future vector control strategies tomeet our national and local requirements.

Some social events have triggered thesechanges. For example, focal treatments inArgentina were possible due to theimplementation of social plans during the 2001recession for the unemployed, who wereobliged to contribute four working hours forvector control activities. However, since theeconomic recovery of the country, these planswere de-activated and now it is impossible tocarry out these activities. Major constraintsincluded security risk, refusals, locked houses,etc. These obstacles and inconveniencesrequired the development of alternativestrategies.

The CIPEIN, Pest and Insecticide ResearchCentre, in Buenos Aires, Argentina, is a WorldHealth Organization Collaborating Centre forthe evaluation of Chagas disease and denguevector resistance. Among other tasks, it carriesout basic and operational research studies withthe object of optimizing control activities forinsect vectors of human disease. Among theCentre’s many contributions to mosquitocontrol, we can mention the development ofnew active substances (permethrin cis-isomer,permethrin trans-isomer)[31,32,33], isolation ofnatural products with insecticide properties[34,35]

and new insecticide formulations as fumigantsin cans or tablets[36,37], Insect Growth Regulators(IGR) formulations in sand[38], and ULVformulations for spatial treatments[39].

Another proposal is the use of adulticidesin complete cycles throughout the city, inaddition to the control of immature forms, asa control strategy in case of an imminentoutbreak of dengue. In countries like Argentina,this type of methodology would be particularlyimportant due to the short periodicity of riskof transmission, which is generally from January



to April, coinciding with the period of highertemperatures and greater rainfall. Furthermore,the outbreaks of dengue in Argentina are closelyassociated with the epidemiological situationin neighbouring countries, evidenced by thecoincidence in time and circulating serotypesin each affected area. Therefore, it has beensuggested that the control of adult mosquitoesin border areas with epidemiological risk is astrategy that might avoid autochthonousoutbreaks and, at the same time, is cost-beneficial at an incidence of more than 29 casesfor every 1000 inhabitants[40].

However, developing such a tool is onlypart of the vector control challenge. Tosupplement focal house-to-house treatment,and in the frame of an integral mosquito vectorcontrol, a combination of treatments has beenproposed that involves spraying a larvicidal-adulticidal mixed formulation[41] using units setup on vehicles in addition to intra-domiciliaryactions performed by the dwellers themselves.This proposal is currently under evaluation andcould constitute an efficient alternative forcontrolling this disease.

In spite of the lack of an extended successof campaigns based only on the use ofinsecticide tools, other strategies of vectorcontrol without chemical treatment involvingthe community have not been organized either.A good review of the achievements of thecommunity-based dengue control programmeswas done by Heintze et al.[42].

The PLICOV (Latin American Programmefor Innovation in Vector Control) initiative wasconceived due to the need of regional countriesto develop novel strategies, which can beadapted to the particular situation of eachcountry[43]. A group of six countries, comprisingof Argentina, Bolivia, Peru, Panama, Cuba andColombia, are jointly developing evaluation andcontrol activities of new tools to verify theirpotential use for vector control in our continent.

This objective has been supported not only byfield studies, but also by laboratory researchcarried out in Latin American countries.

Resistance to insecticides

As recommended by the World HealthOrganization[44], the main preventive activitiesinclude monitoring of Ae. aegypti ovipositionand larviciding sites. Since 1998, extensivechemical control operations were performedin the northern part of Argentina. A massivecontrol programme began in 2002 in Clorinda(Formosa)[45] and in 2003 in Iguazú (Misiones),carried out by the Mundo Sano Foundation incollaboration with the National Ministry ofHealth, the local municipal government, andCIPEIN. The insecticides generally used in theevent of an outbreak were temephos forlarvicidal treatment in water containers (focaltreatment) and cis-permethrin as an adulticidalULV formulation (spatial treatment). For thecontrol strategies to succeed it is important toknow the level of susceptibility to theinsecticides used, because the developmentof resistance could lead to control failures[46].Therefore, our Centre implemented the firstmonitoring programme in Argentina in the citiesof Clorinda and Iguazú, based on a protocolestablished during a meeting of the LatinAmerican Network for Vector Control held inIguazú (Misiones) in December 2004[47], andcompared the susceptibility data obtained tothe mosquito reference strain at CIPEIN. Theresults indicated an incipient resistance totemephos in these mosquito populations,posing an alert for this region. The BrazilianMinistry of Health considers that ResistanceRatio (RR) values of 3 are a reason to alternatetemephos with another insecticide such asBacillus thuringiensis var. israelensis ormethoprene[48]. No control failures have beenobserved yet, but if these values rise to 10,the current control strategies would need tobe completely revised[49].



Conclusions

The vertical plans of the mid-20th century basedon the mobilization of huge resources andDDT, and centred on the eradication of Ae.aegypti, the vector, provided extraordinaryresults. However, their application in thecurrent situation is highly impracticable and,as demonstrated by Brazil, not only a budgetaryissue.

Judging by the progression of the diseasein our continent, and in the world in general,the problem is far from being solved. Thecomplex situation that Argentina and the restof the South American countries face not onlydepends on the development of new activesubstances or more efficient formulations, butalso on adopting an integral approach to the

problem that includes active participation ofall parties, reasonable allocation of resources,cost-benefit analyses, insecticide-resistancemonitoring , establishing adequateentomological and epidemiological surveillanceand, most importantly, the political will.

Acknowledgments

The author thanks Dr Mario Zaidenberg andLic. Pa blo Orellano, from the National Ministryof Health, Dr Alfredo Seijo, from theGovernment of the Autonomous City of BuenosAires, and Dr Rolando Boffi for theircontributions to this manuscript. The authoralso wishes to thank Dr Paola González Audinofor the revision of this work.

References

[1] Secretaría de Turismo de la Nación. El turismoen cifras: 1990-1997. Buenos Aires: theSecretaría. 1998:1-9.

[2] Instituto Nacional de Estadísticas y Censos deArgentina. Resultados Provinciales del Censo2001. (http://www.indec.mecon.gov.ar /webcenso/provincias_2/provincias.asp)accessed 29 November 2008).

[3] Ousset JH, De Ustarán KJ, Lombardo B.Erradicacion del Aedes aegypti en laRepública Argentina (Infestacion Inicial -Adaptacion de procedimientos). In: SegundasJornadas Entomoepidemiologicas Argentinas.1965; II:81-88.

[4] Boffi R. Dengue en la República Argentina.2do Congreso Argentino de Zoonosis, 1ro

Argentino y Latinoamericano deEnfermedades Emergentes. In: AsociaciónArgentina de Zoonosis (eds), 1998:133.

[5] Gaudino NM. Dengue. Revista de SanidadMilitar Argentina. 1916; 15: 617-627.

[6] Gubler DJ, Clark GG. Dengue/denguehemorrhagic fever. The emergence of a globalhealth problem. Emerging Infect Dis. 1995; 1:55-57.

[7] Organización Panamericana de la Salud.Resurgimiento del dengue en las Américas.Boletin epidemiológico. 1997; 18(2): 1-16.

[8] Seijo A. El dengue como problema de saludpública. Arch.argent.pediatr, 2001; 99:/510-521.

[9] Zaidenberg M. Emergencia de dengue en laArgentina: Epidemia de dengue en Salta.Epidemiología y Vacunas. 1999; 3: 1-4.

[10] Avilés G, Rangeon G, Baroni P, Paz V, MonterosM, Sartini JL, Enría D. Epidemia por virusDengue-2 en Salta, Argentina, 1998. Medicina(Buenos Aires). 2000; 60: 875-879.

[11] Boletín Epidemiológico Nacional. Zoonosis yenfermedades transmitidas por vectores:Dengue. Ministerio de Salud de la Nación.2006: 40-41.



[12] Enría D, Morales MA, Fabbri C. Dengue. In:Cecchini E and González Ayala SE. eds. Librode Infectología y Enfermedades infecciosas.1° Edición, Ediciones Journal. Buenos Aires,Argentina. 2008, (91): 638-642.

[13] Pan American Health Organization. Numberof reported cases of dengue and denguehemorrhagic fever, by country andsubregion of the Americas, PAHO,Washington. (http://www.paho.org/Spanish/AD/DPC/CD/dengue.htm. - accessed 26November 2008).

[14] Ministerio de Salud de la Nación de laRepública Argentina. Actualización de lasituación de dengue. Argentina, al 1 de juniode 2007. (http://www.msal.gov.ar/htm/Site/actualizacion_limitrofes_dengue_y_argentina_01_de_junio_de_2007.xls.) accessed 23November 2008.

[15] Seijo A, Cernigoi B, Deodato B. Dengueimportado del Paraguay a Buenos Aires.Estudio clínico y epidemiológico de 38 casos.Medicina (Buenos Aires), 2001; 61:137-141.

[16] Seijo A. Situación del Dengue en Argentina.(http://www.aam.org.ar/Dengue.pdf.) accessed29 November 2008.

[17] Del Ponte EF. Enfoque Sanitario de la FiebreAmarilla para la Republica Argentina. PrimerasJornadas Entomoepidemiologicas Argentinas1959; 1:217-274.

[18] Organización Panamericana de la Salud –Organización Mundial de la Salud. PlanContinental de Ampliacion e Intensificaciondel Combate al Aedes aegypti. Informe de ungrupo de trabajo, Caracas, Venezuela. OPS-HCP-HCT, 1997; 90-97.

[19] Gubler DJ. Dengue and dengue haemorrhagicfever: its history and resurgence as a globalhealth problem. In: Gubler DJ and Kuno G(Eds.) Dengue and dengue haemorrhagic fever.New York: CAB International; 1997.

[20] Rossi GC, Pascual NT, Krsticevic FJ. First recordof Aedes albopictus (Skuse) from Argentina. JAm Mosq Control Assoc. 1999; 15: 422.

[21] Schweigmann N, Boffi R. Aedes aegypti yAedes albopictus: Situación entomológica enla región. 2do Congreso Argentino de Zoonosisy 1er Congreso Argentino y Lationoamericanode Enfermedades Emergentes. 1998; 259-263. Asociación Argentina de Zoonosis.Buenos Aires, Argentina.

[22] Schweigmann N, Vezzani D, Orellano P, KurucJ, Boffi R. Aedes albopictus in an area ofMisiones, Argentina. Revista de Saúde Pública.2004; 38:136-138.

[23] Espinosa M, Weinberg D, Coto H. Presenciade Aedes albopictus en la localidad de PuertoIguazú, provincia de Misiones, Argentina. XSimposio internacional de controlepidemiologico de enfermedades transmitidaspor vectores, Agosto 2007, Buenos Aires,Argentina.

[24] Vezzani D, Carbajo AE. Aedes aegypti, Aedesalbopictus, and dengue in Argentina: currentknowledge and future directions. Mem InstOswaldo Cruz. 2008; 103: 66-74.

[25] Organización Panamericana de la Salud.Resolución XV. Desarrollo y fortalecimientode los sistemas locales de salud en latransformacion de los sistemas nacionales desalud. Comité Regional de la OMS para lasAméricas, Documento Oficial 225 1988; 1-60.Washington DC.

[26] Masuh H, Seccacini E, Zerba E, Licastro S.Aedes aegypti (Diptera: Culicidae): monitoringof populations to improve control strategies inArgentina. Parasitol Res. 2008; 103: 167-170.

[27] Secretaría de Turismo de la RepúblicaArgentina, Dirección Nacional de DesarrolloTurístico. SECTUR. Centro deDocumentación. Boletín técnico. Direcciónde estadísticas, December 2008.

[28] Situación de Fiebre Amarilla en Argentina(Yellow fever situation in Argentina). Ministeriode Salud de la Nación. República Argentina.Consejo Federal de Salud (COFESA). BuenosAires. Mayo 2008. (http://www.msal.gov.ar/htm/Site/pdf/anexo-viii-acta-02-08.pdf.)accessed 29 November 2008.



[29] World Health Organization, Regional Officefor the Western Pacific Guidelines for denguesurveillance and mosquito control. WHOWestern Pacific Education in Action Series 8.Manila: WHO WPRO, 2003..

[30] Pan American Health Organization. Dengueand dengue hemorrhagic fever in the Americas:guidelines for prevention and control. Scientificpublication No. 548. Washington, D.C.:PAHO, 1994.

[31] Masuh H, Seccacini E, Licastro S, Zerba E. Anew aqueous suspension formulation of cis-permethrin and its insecticidal activity. PestManagement Sci 2000; 56:1073-1076.

[32] Masuh H, Licastro S, Zerba E. Método deaislamiento del insecticida Permetrinaenriquecido en el isómero cis. ArgentinePatent. Nº P 96 01 05372. 1996. In Spanish.

[33] Masuh H, Licastro S, Zerba E. Síntesisestereoselectiva de isómeros trans deesteres de ácidos 3-vinil-2,2-dimetilciclopropanocarboxilicos. ArgentinePatent. Nº P9701-04482. 2003. In spanish.

[34] Lucia A, Gonzalez Audino P, Seccacini E,Licastro S, Zerba E, Masuh H. Larvicidal effectof Eucalyptus grandis essential oil andturpentine and their major components onAedes aegypti (L.) larvae (Diptera: Culicidae). JAm Mosq Control Assoc. 2007; 23:299-303.

[35] Lucia A, Licastro S, Zerba E, Masuh H. Yield,chemical composition, and bioactivity ofessential oils from 12 species of Eucalyptus onAedes aegypti larvae. Entomol Experim etApplicata. 2008; 129: 107-114.

[36] Masuh H, Lopez PA, Vega C, Licastro S, ZerbaE. Field evaluation of a smoke generatingformulation containing b-cypermethrin againstthe dengue vector in Argentina. J Am MosqControl Assoc. 2003; 19: 53-57.

[37] Gonzalez Audino PG, Masuh H, Zerba E.Thermal behaviour, biological activity andconformational study of a-Methoprene – b-cyclodextrin complex in a smoke generatingformulation. Molecules, a Journal of Synthetic

Chemistry and Natural Product Chemistry.2005; 10: 534-544.

[38] Seccacini E, Lucia A, Harburguer L, Zerba E,Licastro S, Masuh H. Effectiveness ofpyriproxyfen and diflubenzuron formulationsas larvicides against Aedes aegypti (Diptera:Culicidae). J Am Mosq Control Assoc. 2008 Sep;24(3): 398-403.

[39] Seccacini E, Masuh H, Licastro S, Zerba E.Laboratory and scaled up evaluation of cis-permethrin applied as a new ultra low volumeformulation against Aedes aegypti (Diptera:Culicidae). Acta Trop. 2006; 97: 1-4.

[40] Orellano P, Pedroni E. Análisis costo-beneficiodel control de vectores en la transmisiónpotencial de dengue. Rev Panam Salud Publica.2008, 24: 113-119. (http://www.scielosp.org/scielo.php?script=sci_arttext&pid=S1020-49892008000800005&lng=en&nrm=iso.)accessed 29 November 2008.

[41] Lucia A, Harburguer L, Licastro S, Zerba E,Masuh H. Efficacy of a new combinedlarvicidal-adulticidal ultralow volumeformulation against Aedes aegypti (Diptera:Culicidae), vector of dengue. Parasitol Res.2009 Apr; 104(5): 1101-7.

[42] Heintze C, Velasco Garrido M, Kroeger A. Whatdo community-based dengue controlprogrammes achieve? A systematic review ofpublished evaluations. Transactions of the RoyalSociety of Tropical Medicine and Hygiene.2007; 101: 317-325.

[43] RELCOV: Red Latinoamericana de Control devectores. The PLICOV Innovative. (http://relcov.org/cgi-bin/textonoticias.asp?ID_NOTICIA=43.) accessed 29 November 2008.

[44] World Health Organization. Vector control formalaria and other mosquito borne diseases.WHO Technical Report Series 857. Geneva:Switzerland, 1995.

[45] Masuh H, Coto H, Licastro S, Zerba E. Controlde Aedes aegypti (L.) en Clorinda: un modelopara areas urbanas. Entomología y Vectores.2003; 10: 485-494.



[46] World Health Organization, Regional Officefor South-East Asia. Striving for better health inSouth-East Asia: selected speeches by Dr UtonMuchtar Rafei Regional Director, WHO South-East Asia Region. Volume II: 1997-2000. NewDelhi: WHO SEARO, 2001. IntercountryWorkshop on Insecticide Resistance ofMosquito Vectors, Salatiga, Indonesia, August1997. (http://www.searo.who.int/en/Section980/Section1162/Section1167/Section1171_4749.htm.) accessed 29November 2008.

[47] Bisset J, Blanco S, Braga I, Coto H, Masuh H,Moncayo A, Nathan M, Orellano P, VazquezCangas J, Zerba E. 2005. Protocolo paradeterminar la susceptibilidad o resistencia a

insecticidas de mosquitos de la especie Aedesaegypti [Protocol to evaluate the susceptibilityor resistance to insecticides by Aedes aegyptimosquitoes]. (http://www.mundosano.org/publicaciones/publicaciones3.php.) accessed29 November 2008.

[48] Braga IA, Valle D. Aedes aegypti: surveillance,resistance monitoring and control alternativesin Brazil. Epidemiol Serv Saude Brasilia. 2007;16:295-302.

[49] Seccacini E, Lucía A, Zerba E, Licastro S,Masuh H. Aedes aegypti resistance totemephos in Argentina. J Am Mosq ControlAsso.c 2008; Dec;24(4):608-9.


Duration of short-lived cross-protective immunityagainst a clinical attack of dengue:

A preliminary estimate

Hiroshi Nishiura#

Theoretical Epidemiology, University of Utrecht, Yalelaan 7, 3584 CL, Utrecht, The Netherlands

Abstract

It is believed that primary infection with a single serotype of dengue virus elicits short-lived cross-protective immunity against other heterologous serotypes; however, the duration of cross-protectionhas not been explicitly estimated using epidemiological data. To offer an empirical estimate of theduration, the present study re-analysed historical cohort data of multiple clinical attacks of dengueamong American soldiers in the Philippines from 1922–24. In the original study, the historical cohortof 299 cases with a first clinical attack of dengue were closely surveyed; 99 (33.1%) experienced asecond attack, while the remaining 200 returned to the United States without further attacks. The timeintervals from first to second attack among the 99 cases, and from first attack to departure to the UnitedStates among the 200 soldiers, were used for estimating the duration of cross-protective immunitybased on a simple mathematical model. Employing an exponential distribution or Kronecker’s deltafunction as the loss function of cross-protection against a second clinical attack, the mean duration ofcross-protective immunity since the first clinical attack was estimated as 6.90 (4.87, 11.83) days and7.52 (4.88, 16.38) days, respectively. The force of infection, which was jointly estimated with theduration of cross-protection, reasonably explained the other observed epidemiological information inthe data, supporting the finding of a short cross-protection period. Even though the estimates suggestedthat the first clinical attack most likely elicited cross-protective immunity, the length of cross-protectionlasted only 1–2 weeks, far shorter than previously believed.

Keywords: Dengue; Epidemiology; Immunity; Serotype; Statistical model.


Introduction

Dengue fever (DF) is a vector-borne diseasecaused by four closely related dengue viruses(DENV-1-4)[1-2]. It is distributed in most tropicaland subtropical areas where Aedes aegypti and/or Aedes albopictus are abundant[3]. Infection

with DENV can also cause denguehaemorrhagic fever (DHF), which is a clinicalsyndrome characterized by increased vascularpermeability, plasma leakage, hypovolemia andshock[4,5]. Although the pathogenesis of DHFhas yet to be fully clarified, several risks havebeen reported; these include secondaryinfection with heterologous serotypes[6,7],


Duration of cross-protection against second attack of dengue

primary infection in infants born to dengue-immune mothers[8], differing virulence of astrain[9] and differing human susceptibilityaccording to race or genetic factors[10,11].

Although the epidemiological risks of DHFhave been explored for more than 30 years,many aspects of the transmission dynamics ofdengue remain to be clarified[12]. Thetransmission dynamics, especially of theinteractions between two or more serotypes,have been explored using mathematicalmodels[13-22]. Despite a recent increase in thenumber of relevant studies, only those basedon epidemiological observations in the fieldhave provided detailed insights into thepathogenesis of DHF or interactions betweendengue transmission and disease[13-16]. There isa general lack of field data complete withserotype, time, age and space measurementsthat would allow scientists to investigate andmodel dengue at a population level. Questionsthat could be clarified with modeling exercisesinclude: (i) more specific information on themechanisms of innate dengue viral virulence(if any); (ii) the result of infections with anytwo specific heterologous serotypes; and (iii)the mechanisms and duration of protectiveimmunity.

Among these unknowns, the present studyfocuses on cross-protective immunity amongthose who have experienced primary infectionagainst further infection caused by aheterologous serotype. The duration ofacquired cross-protective immunity has neverbeen explicitly estimated and variousepidemiological models have employed anumber of different and unsupportedassumptions. For instance, Ferguson et al.[14]

assumed the absence of cross-protection,although a historical study conducted bySabin[23] suggests that the presence of cross-protective immunity for a short time-period isplausible following primary infection. Thepresence of transient cross-protective immunity

was once supported by explicit data analysesby Adams et al.[15], but the data were fromDHF cases in Thailand in a time-series (withserotype-specificity) that required a number ofother epidemiological assumptions. Otherstudies have assumed differing mechanisms ofcross-protective immunity following anexposure to a heterologous serotype shortlyafter primary infection (e.g. exposure to theheterologous serotype results either in infectionor sero-conversion[20] and/or permits developingimmunity against the heterologousserotype[16,19]). Even though the presence ofcross-protective immunity seems likely[15,23], andalthough the majority of previous studies haveacknowledged the critical importance of short-lived cross-protective immunity in describingthe oscillatory transmission dynamics of dengue,the actual duration remains unknown. Animplicit suggestion has been that the durationis 2–9 months[23].

Accordingly, it would be important to offeran empirical estimate of the duration usingexisting informative data. The present studyaims to estimate the duration of cross-protective immunity against a second clinicalattack of dengue as a function of the time sincethe first attack. For the estimation, historicalcohort data from American soldiers in thePhilippines from 1922–24 are re-analysed.

Methods

The historical data of DENV infection originatedfrom a well-known and rigorous study byJoseph Franklin Siler, Milton Weston Hall andArthur Parker Hitchens that took place in 1924–25[24]. The study was originally published in thePhilippine Journal of Science[24] and wasreprinted with appendices by the Bureau ofPrinting, Manila[25]. Further details of thepublication were revisited by Nishiura andHalstead in a recent study[26]. The experimentaltransmission of DENV-4 in human volunteers



recruited from US Army personnel is widelyknown[26,27]. Siler et al. also conducted anepidemiological study of the natural infectionof dengue among American soldiers in theThirty-first Infantry from 1922–24[24,25]. Thisinvestigation revisits that study.

As well as the results from thetransmission experiment in human volunteers,Siler et al.[24,25] hoped that the study wouldachieve further insights into the frequency ofinfection, acquired immunity and recurrenceof dengue in the average American soldier inManila under natural conditions. Accordingly,an epidemiological survey of the clinical attackof dengue was attempted among Americansoldiers; multiple clinical episodes of denguein each individual were recorded. Afterobtaining the preliminary results of theepidemiological observations, the authors notedserious technical problems in interpreting thedata and precisely estimating the frequency ofinfection at a population level. The issuesincluded: (i) varying time-intervals between thefirst attack of dengue and the end of militaryduty (i.e. some cases experienced the firstattack at a time close to the end of duty, andthus were unlikely to experience a secondattack); (ii) the duration of military service wasvariable and some soldiers left for home duringthe period of observation while others remainedin the Philippines; and (iii) mild attacks wereless likely to be recorded compared with severecases. To resolve these technical problems, Sileret al. conducted further epidemiologicalobservations in the Thirty-first Infantry; subjectswere limited to those who had their initialattack of dengue between 1 July 1922 and 30June 1923. All of the enrolled soldiers startedtheir duty on or after 1 January 1922 and leftthe Philippines no earlier than 31 December1923. The usual length of duty in the PhilippineIslands was 2 years. The period of observationended 31 December 1924, a time by whichthe soldiers had been closely monitored forany possible signs or symptoms of dengue.

Individuals with irregular military assignmentsor transfer were excluded because of theabove-mentioned epidemiological problems.Except for a few individuals, all of the includedsubjects were men from the United States whocould be assumed to be fully susceptible atthe beginning of their tour of duty in thePhilippines, or at least were stated as “couldnot have been exposed to dengue for morethan six months before.” Unfortunately, theseverity of the cases was not well detailed,and it is unknown if there was any indicationof DHF among those with clinical attacks.

The Thirty-first Infantry numbered 1086personnel, among which there were 562potential episodes of dengue. Because ofmissing observations of 28 potential cases, theauthors proportionally decreased the totalsample (n = 1032). The first attack of denguewas clinically assessed and detailed clinicalrecords were obtained for 421 cases.Furthermore, strictly applying the exclusioncriteria to satisfy the authors’ concern regardingepidemiological problems (especially, to meetthe condition of the time of assignment beingon or after 1 January 1922), only 299 cases(71.0% of those with clinical records) wereselected for further analyses. Again,proportionally decreasing the total sample size(n = 733), the authors concluded that a clinicalattack of dengue was observed at least onceamong 40.8% of the soldiers. Of the 299 casesthat had a first attack, 99 (33.1%) experienceda second attack while the remaining 200 leftthe Philippines without any further clinicalattacks. For all of the included subjects, thetime interval between their arrival in thePhilippines and the first attack was recorded.Moreover, the time from the first attack todeparture to the United States was recordedamong the 200 cases without a second attack.The time interval from the first to the secondattack as well as the time interval from thesecond attack to departure back to the UnitedStates was recorded among the 99 cases with



a second clinical attack. In the originalpublication, the data for those with only a firstclinical attack were reported as a group (i.e.given as just summary tables) by discrete timeintervals and there was no individualinformation (such as time from arrival-to-attackand attack-to-departure), but the data for thosewith a second attack (n = 99) were recordedfor each individual. Although third and fourthclinical attacks were observed among 14(14.1%) and 1 (1.0%) cases among the 99experiencing a second attack, the informationwas discarded in the present study for simplicity.

Using the historical cohort data of thosewho experienced at least a first clinical attackof dengue, the present study estimates theduration of cross-protective immunity againsta second clinical attack as a function of timesince the first attack. Cases, both with andwithout a second clinical attack (n = 99 and200), are analysed. First, the descriptivestatistics of the time intervals were examined.The time intervals between arrival and the firstattack and between the first attack anddeparture to the United States were comparedbetween those who did and those who didnot experience a second clinical attack. Forthese comparisons, a t-test and the Welchanalysis of variance (ANOVA) were employed

following the use of a F-test[28]. Subsequently,a mathematical model was developed andapplied to estimate the duration of cross-protective immunity against a second clinicalattack of dengue.

Figure 1 shows a schematic diagram of amathematical model that describes the cohortepisode of first and second clinical attacks ofdengue. Let t denote the time since the firstclinical attack. Moreover, I1(t), S(t) and I2(t)denote, respectively, the fractions of those whoexperienced a first attack and are still immuneto another clinical attack, who are susceptibleto another clinical attack caused by aheterologous serotype, and who experienceda second clinical attack, at time t since the firstattack of dengue. Supposing that the rate toloose cross-protective immunity and the forceof infection (i.e. the rate at which susceptibleindividuals experience infection) are δ and λ(per day), respectively, the model for theobserved intervals is given by

(1)

Figure 1: Compartmental model to describe the time interval between first and second clinicalattacks of dengue in the Thirty-first Infantry in the Philippines from 1922-24

[Although infected soldiers are assumed as transiently immune against other serotypesimmediately after the first clinical attack, they loose the cross-protective immunity at rate δ and

become susceptible to other heterologous serotypes. The susceptible individuals experienceinfection at rate λ and experience a second clinical attack.]

Loss of immunity secondary infection

First attack Susceptiblae Second attack

l t1( ) S t( ) l t2( )



It should be noted that a constant forceof infection λ assumes an endemic equilibriumin the Philippines (see discussion onseasonality). Since I1(0) = 1 and S(0) = I2(0) =0, the probability density and the cumulativedistribution of the second clinical attack at timet since the first attack, f(t) and F(t), respectively,are

which was also applied to the data usingthe likelihood function (3).

It should be noted that the followingassumptions were made for inference: (i) allincluded subjects were fully susceptible todengue at the beginning of their military duty inthe Philippines; (ii) the first clinical attack elicitedlife-long immunity against the causativehomologous serotype; (iii) the force of infectionwas independent of time, and seasonality wasignored because of the absence of adequatedata; (iv) multiple serotypes were co-circulatingduring the period of observation in thePhilippines with identical transmission potential(though the exact number of co-circulatingserotypes does not have to be known); and (v)the second clinical attack does not have to reflectsecondary infection, and the assumed loss ofimmunity reflected the waning of cross-protection against a second ‘clinical’ attack.Although the results in the present study aredeemed preliminary because of these simplisticassumptions, it is critically important to validatethe realism of these assumptions (especially, iii,iv and v) to appropriately interpret the results.Thus, these points are later discussed in moredetail (see Discussion).

Results

Figure 2A shows the distribution of time fromarrival in the Philippines to the first attack for299 cases. The mean (and standard deviation(SD)) and median (and lower-upper quartiles)were 153.9 (115.1) and 144 (48-213) days,respectively. The mean (SD) intervals fromarrival to the first attack among those who didor did not experience a second clinical attackwere 124.7 (103.6) and 168.4 (118.0) days;significantly different by a t-test (t ratio = -3.27,p < 0.01). Figure 2B shows the distribution oftime from the first to the second attack ofdengue among 99 cases. The mean (SD) andmedian (lower-upper quartiles) were 216.9(142.4) and 213 (92-279) days, respectively.

(2)

Let the time interval from the first to thesecond attack of case i be ti (where i belongsto the 99 cases with a second attack), and letthe time interval from the first attack to thereturn to the United States of case j be tj (wherej belongs to the 200 cases without a secondattack). The observed tj among the 200 withouta second clinical attack are dealt with ascensored data. That is, the likelihood of notobserving a second clinical attack for tj days isgiven by 1-F(tj). Accordingly, the total likelihoodis given by

The parameters, δ and λ, were estimatedby minimizing the negative logarithm ofequation (3). Profile likelihood confidenceintervals were computed. In addition to theexponentially distributed immune-loss functionin model (1), Kronecker’s delta function wasalso employed as the loss function of cross-protective immunity. Delta function assumesthat the duration of cross-protection does notdiffer between individuals (i.e. is constant), andyields an estimate of the duration that can beregarded as maximum. Under the alternativeassumption, f(t) and F(t) are replaced by

(4)

(3)



Figure 2: Frequency distributions of the time from arrival to the first attack and the time fromthe first to the second attack in the Thirty-first Infantry in the Philippines from 1922-24

[A. The time since arrival in the Philippines to the first clinical attack of dengue (n = 299).B. The time interval between the first and second attacks among 99 American soldiers.

The other 200 soldiers did not experience a second clinical attack.]

0 100 200 300 400 500 600 700Time from arrival to first attack (days)

0 100 200 300 400 500 600 700Time from first to second attack (days)

Freq

uen

cy

Freq

uen

cy

0.15

0.10

0.05

0.13

0.10

0.08

0.05

0.03

A B

Figure 3: Comparison of the time from first attack to departure between those with and thosewithout a second attack of dengue in the Thirty-first Infantry in the Philippines from 1922-24

[Departure denotes the end of military service in the Philippines (soldiers then returned to theUnited States)].

With second attack= 99n

Without second attack= 200n

Tim

efr

omfir

stat

tack

tode

part

ure

(day

s)

900

800

700

600

500

400

300

200

100

0



Figure 3 compares the time from the firstattack to departure to the United Statesbetween those with and those without a secondclinical attack. The mean (SD) and median(lower-upper quartiles) lengths for the entiresamples (n = 299) were 574.2 (141.0) and575 (475-675) days, respectively. The mean(SD) intervals from the first attack to departurefor those who did and those did not experiencea second attack were 600.8 (107.7) and 561.0(153.4) days, respectively. Since the F-testrevealed a significant difference in variance (F-ratio = 2.03, p < 0.01), a Welch ANOVA wassubsequently employed for the comparison.This showed that the time from the first attackto departure among those with a second attackwas significantly longer than those without (F-

ratio = -6.75, p = 0.01). All of the above-mentioned time intervals among those with asecond attack were given as individual data,permitting an estimation of the total length ofstay in the Philippines. The mean (SD) lengthof stay was 725.5 (100.2) days, roughlycorresponding to 2 years as described in theoriginal study[24,25].

Assuming an exponentially distributedimmune-loss function, the maximum likelihoodestimates (and the corresponding lower andupper 95% CI) of δ and λ were 0.14 (0.08,0.21) and 7.52×10-4 (6.13×10-4, 9.12×10-4)per day, respectively. The mean length of cross-protective immunity against a second clinicalattack is given by 1/δ, i.e., 6.90 (4.87, 11.83)

Figure 4: Estimated duration of cross-protective immunity against a secondclinical attack of dengue

[The estimated fraction of those still protected against a second clinical attack, from aheterologous serotype, is shown as a function of time since the first attack. Two different

models, exponential distribution (thick line with dotted-and-dashed 95% confidence intervals)and Kronecker’s delta function (thin line with dotted 95% confidence intervals), were assumed

as the survival function of cross-protective immunity. The 95% confidence intervals werederived from profile likelihood.]

Frac

tion

ofim

mun

ein

divi

dual

sag

ains

t sec

onda

ryin

fect

ion 1.0

0.8

0.6

0.4

0.2

0.0

Time since first attack (days)0 5 10 15 20



days. Similarly, 1/λ details the mean length oftime to experience a second attack after acomplete loss of cross-protective immunity,which was estimated as 3.64 (3.00, 4.47)years. Similarly, assuming Kronecker’s deltafunction for the loss of cross-protectiveimmunity, the maximum likelihood estimates(and 95% CI) of δ and λ were 0.13 (0.06, 0.20)and 7.53×10-4 (6.89×10-4, 7.95×10-4) per day,respectively, indicating that the mean lengthof cross-protective immunity was 7.52 (4.88,16.38) days and the mean length of time fromcomplete loss of cross-protection to a secondattack was 3.64 (3.44, 3.98) years. Figure 4shows the estimated survivorship functions ofthose who would still have possessed cross-protective immunity as a function of time sincethe first attack. Both distributional assumptionsyielded similar mean lengths of the cross-protective immunity.

Discussion

Despite the critical importance of cross-protective immunity for understanding theepidemiological dynamics of dengue, there hasbeen no previous determination of the durationof cross-protection against a heterologousserotype; thus, the present study re-analysedhistorical case cohort data among US Armypersonnel in the Philippines from 1922-24. Asit was observed from the experimentaltransmission of dengue in humanvolunteers[26,27], another original data set of Sileret al.[25] also yielded critically importantinformation on the time intervals of exposureand transmission events, permitting empiricalassessment of the length of acquired cross-protective immunity against a second clinicalattack. The most important conclusion drawnfrom the simple exercise undertaken in thepresent study is that the duration of transientcross-protective immunity was estimated asshort as 1 or 2 weeks, which is far shorter than

has been implicitly suggested (i.e. 2-9months)[23]. The finding of short-lived cross-protective immunity can explain Sabin’s notein which mild systemic inflammation wasobserved by inoculating subjects who had beenthought to be cross-protected[23].

Although it was not possible to quantifythe degree (or strength) of cross-protection, tothe best of the author’s knowledge, the presentstudy is the first to explicitly estimate theduration based on empirical epidemiologicaldata. The role of cross-protective immunity hasbeen recognized as critical for describing thetransmission dynamics of dengue, andespecially, oscillatory epidemiologicalpatterns[15,20]. Despite the successfulquantification of the short length of cross-protection, it should be noted that theestimated duration reflected cross-protectionagainst a second clinical attack. Considering thatDENV infection involves a substantial fractionof sub-clinical infections[28,29], cases with asecond attack could have been infected beforetheir second clinical attacks were observed.Considering that experimental infection ofanimals with silent sero-conversion has beenobserved[30], the sub-clinical secondary infectionis indeed plausible. Nevertheless, the possiblepresence of sub-clinical secondary infectionindicates that the duration of cross-protectiveimmunity against infection (rather than clinicalattack) is shorter than the estimated durationin the present study. This supports the mainconclusion of the present study, i.e. that theduration of cross-protection is extremely short.Considering the fact that infection with asecond heterologous serotype tends toenhance the severity of a secondaryinfection[6,7], the duration of cross-protectiveimmunity against infection with a secondheterologous serotype may be reasonably closeto the estimate in the present study, and maybe slightly shorter than the estimated durationagainst a second clinical attack.



Since the present study was intended topresent preliminary results of estimates basedon a simple model structure, the mathematicalmodel employed a number of unrealisticassumptions, among which (v), the interpretationof a second ‘clinical’ attack, was discussedabove. The remaining two important issues, (iii)the constant force of infection, and (iv) equalfrequency of co-circulation among heterologousserotypes, are discussed here. Although the forceof infection could vary as a function of time inreality (e.g. reflecting seasonal ecologicaldynamics of the vector population), during thestudy period from July 1922 to December 1924,there was no month without a clinical attack ofdengue[25]. Moreover, the largest difference inincidence (i.e. highest minus lowest incidence)was as small as 1560 per 1000 per annum amongthe entire population of American soldiers[25],indicating that the seasonal forcing was not criticalquantitatively. That is, even though assumption(iii) could have influenced the precision of theestimate, the main conclusion of an extremelyshort duration of cross-protection is still deemedvalid.

The contention of short-lasting cross-protection (and validity of the model) is alsosupported by two other calculations, one ofwhich is also relevant to the interpretation ofthe above-mentioned point (iv). First, using theestimated force of infection λ = 7.5×10-4 perday, the fraction of those who experienced aclinical attack of dengue during 2 years ofmilitary service would be given by 1-exp(-2×365×λ) = 0.422, which gives an estimatevery close to the observed cumulativeincidence of first clinical attack by the end ofthe study period (i.e. 40.8%). Accounting forthe possible presence of sub-clinical infection,λ could have been greater than the estimate,but, in fact, a greater λ supports the finding ofa short duration of cross-protection (i.e. if λ isgreater than 7.5×10-4 per day, the estimatedmean duration of cross-protection, 1/δ, will beshorter than that in the present study). Second,

if there were 2, 3 or 4 co-circulating serotypes,the assumption of identical transmissibilitywould yield the force of infections for all 2, 3or 4 circulating serotypes as 15.0, 11.3 and10.0×10-4, respectively, per day. The mediantime from arrival to first attack is then given by–ln(0.5)/λ = 462, 616 and 693 days. Althougha direct comparison between these estimatesand the observed data cannot be made becauseof the absence of detailed data about thosewho avoided a clinical attack of dengue, theestimated median times from arrival to firstattack are shorter than the observed medianamong those experiencing a first attack (144days in Figure 2A), indicating that the force ofinfection could have been larger than thatestimated, which again supports the finding ofshort-lived cross-protective immunity. Also, ifthe force of infection of a specific serotypewas much greater than those of otherserotypes, this could reasonably explain thediscrepancy between the expected mediantime from arrival to first attack and the medianestimate in Figure 2A. Assuming that theaverage life expectancy of the host was L =50 years and exponentially distributed, andadopting an approximate estimator of the basicreproduction number (i.e. the average numberof secondary cases generated by a singleprimary case in a fully susceptible population),R0 = πL (where π is the serotype-specific forceof infection), equal frequency of transmissibilityamong co-circulating 2, 3 or 4 serotypes yieldsR0 of 9.13, 6.84 and 6.08[31,32]. Obtaining a moreprecise estimate utilizing an extendedmodelling approach is a subject for future study.

In summary, the present study re-analysedthe distribution of time intervals between firstand second clinical attacks of dengue, jointlyestimating the duration of cross-protectiveimmunity against a second clinical attack andthe force of infection among US Armypersonnel who experienced at least a first attackin the Philippines. Although transient cross-protective immunity most likely exists, the



length is estimated to be just 1-2 weeks, farshorter than previously suggested. Consideringthat DENV infection involves a substantialnumber of sub-clinical infections, and becausethe present study reported preliminary resultsbased on simplistic model assumptions, futureimprovements that address the presence ofsub-clinical infection, adjust seasonalcharacteristics of infection, and obtain more

precise estimate of the duration of cross-protection, are considered crucial.

Acknowledgment

The work of HN was supported in part by theNetherlands Organization for ScientificResearch (NWO).

References

[1] Gubler DJ. The global emergence/resurgenceof arboviral diseases as public health problems.Archives of Medical Research. 2002; 33:330-342.

[2] Westaway EG, Blok J. Taxonomy andevolutionary relationships of flaviviruses. In:Gubler DJ, Kuno G. eds. Dengue and denguehemorrhagic fever. New York, NY: CABInternational, 1997. pp. 147-73.

[3] Guzman MG, Kouri G. Dengue: an update.Lancet Infectious Diseases. 2002; 2: 33-42.

[4] World Health Organization (WHO). Denguehaemorrhagic fever: diagnosis, treatment,prevention and control. 2nd ed. Geneva:WHO, 1997.

[5] Gubler DJ. Epidemic dengue/denguehemorrhagic fever as a public health, socialand economic problem in the 21st century.Trends in Microbiology. 2002; 10: 100-103.

[6] Halstead SB. Pathogenesis of dengue:challenges to molecular biology. Science.1988; 239: 476-481.

[7] Sangkawibha N, Rojanasuphot S, Ahandrik S,Viriyapongse S, Jatanasen S, Salitul V,Phanthumachinda B, Halstead SB. Risk factorsin dengue shock syndrome: a prospectiveepidemiologic study in Rayong, Thailand. I.The 1980 outbreak. American Journal ofEpidemiology. 1984; 120: 653-69.

[8] Kliks SR, Nimmanitya S, Nisalak A, Burke D.Evidence that maternal dengue antibodies areimportant in the development of denguehemorrhagic fever in infants. American Journalof Tropical Medicine and Hygiene. 1988; 38:411-419.

[9] Cologna R, Armstrong PM, Rico-Hesse R.Selection for virulent dengue viruses occurs inhumans and mosquitoes. Journal of Virology.2005; 79: 853-859.

[10] Stephens HA, Klaythong R, Sirikong M, VaughnDW, Green S, Kalayanarooj S, Endy TP, LibratyDH, Nisalak A, Innis BL, Rothman AL, EnnisFA, Chandanayingyong D. HLA-A and -Ballele associations with secondary dengue virusinfections correlate with disease severity andthe infecting viral serotype in ethnic Thais.Tissue Antigens. 2002; 60: 309-18.

[11] Guzman MG, Kouri GP, Bravo J, Soler M,Vazquez S, Morier L. Dengue hemorrhagicfever in Cuba, 1981: a retrospectiveseroepidemiologic study. American Journal ofTropical Medicine and Hygiene. 1990; 42:179-84.

[12] Nishiura H. Mathematical and statisticalanalyses of the spread of dengue. DengueBulletin. 2006; 30: 51-67.



[13] Fischer DB, Halstead SB. Observations relatedto pathogenesis of dengue hemorrhagic fever.V. Examination of age-specific sequentialinfection rates using a mathematical model.Yale Journal of Biology and Medicine. 1970;42: 329-349.

[14] Ferguson NM, Donnelly CA, Anderson RM.Transmission dynamics and epidemiology ofdengue: insights from age-stratified sero-prevalence surveys. Philosophical Transactionsof the Royal Society of London Series B, BiologicalSciences. 1999; 354: 757-768.

[15] Adams B, Holmes EC, Zhang C, MammenMP, Nimmannitya S, Kalayanarooj S, BootsM. Cross-protective immunity can accountfor the alternating epidemic pattern ofdengue virus serotypes circulating in Bangkok.Proceedings of the National Academy ofSciences of the United States of America.2006; 103: 14234-14239.

[16] Nagao Y, Koelle K. Decreases in denguetransmission may act to increase the incidenceof dengue hemorrhagic fever. Proceedings ofthe National Academy of Science, the UnitedStates of America. 2008; 105: 2238-2243.

[17] Esteva L, Vargas C. Coexistence of differentserotypes of dengue virus. Journal ofMathematical Biology. 2003; 46: 31-47.

[18] Feng Z, Velasco-Hernandez JX. Competitiveexclusion in a vector-host model for thedengue fever. Journal of Mathematical Biology.1997; 35: 523-544.

[19] Kawaguchi I, Sasaki A, Boots M. Why aredengue virus serotypes so distantly related?Enhancement and limiting serotype similaritybetween dengue virus strains. Proceedings ofthe Royal Society of London, Series B. 2003;270: 2241-2247.

[20] Wearing HJ, Rohani P. Ecological andimmunological determinants of dengueepidemics. Proceedings of the NationalAcademy of Sciences of the United States ofAmerica. 2006; 103: 11802-11807.

[21] Cummings DA, Schwartz IB, Billings L, ShawLB, Burke DS. Dynamic effects of antibody-dependent enhancement on the fitness ofviruses. Proceedings of the National Academyof Sciences of the United States of America.2005; 102: 15259-15264.

[22] Chowell G, Diaz-Duenas P, Miller JC, Alcazar-Velazco A, Hyman JM, Fenimore PW, Castillo-Chavez C. Estimation of the reproductionnumber of dengue fever from spatial epidemicdata. Mathematical Biosciences. 2007; 208:571-589.

[23] Sabin AB. Research on dengue during WorldWar II. American Journal of Tropical Medicineand Hygiene. 1952; 1: 30-50.

[24] Siler JF, Hall MW, Hitchens AP. Dengue: Itshistory, epidemiology, mechanism oftransmission, etiology, clinical manifestations,immunity, and prevention. Philippine Journalof Science. 1926; 29: 1-302.

[25] Siler JF, Hall MW, Hitchens AP. Dengue: itshistory, epidemiology, mechanism oftransmission, etiology, clinical manifestations,immunity, and prevention. Monograph of thePhilippine Bureau of Science, No. 20. Manila,Bureau of printing, 1926.

[26] Nishiura H, Halstead SB. Natural history ofdengue virus (DENV)-1 and DENV-4 infections:reanalysis of classic studies. Journal of InfectiousDiseases. 2007; 195: 1007-1013.

[27] Halstead SB. Etiologies of the experimentaldengues of Siler and Simmons. AmericanJournal of Tropical Medicine and Hygiene. 1974;23: 974-982.

[28] Guzman MG, Kouri G, Valdes L, Bravo J,Alvarez M, Vazques S, Delgado I, Halstead SB.Epidemiologic studies on Dengue in Santiagode Cuba, 1997. American Journal ofEpidemiology. 2000; 152: 793-799.



[29] Yamashiro T, Disla M, Petit A, Taveras D, Castro-Bello M, Lora-Orste M, Vardez S, Cesin AJ,Garcia B, Nishizono A. Seroprevalence of IgGspecific for dengue virus among adults andchildren in Santo Domingo, DominicanRepublic. American Journal of TropicalMedicine and Hygiene. 2004; 71: 138-143.

[30] Kochel TJ, Watts DM, Gozalo AS, Ewing DF,Porter KR, Russell KL. Cross-serotypeneutralization of dengue virus in Aotusnancymae monkeys. Journal of InfectiousDiseases. 2005; 191: 1000-1004.

[31] Dietz K. Transmission and control of arbovirusdiseases. In: Ludwig D, Cooke KL ed.Epidemiology. Philadelphia: SIAM, 1975. pp.104-121.

[32] Dietz K. The estimation of the basicreproduction number for infectious diseases.Statistical Methods in Medical Research. 1993;2: 23-41.


Discrimination between primary and secondarydengue virus infection by using an

immunoglobulin G avidity test

A. Chakravarti#, M. Matlani, A. Kumar

Maulana Azad Medical College and Associated Lok Nayak Hospitals, New Delhi, India

Abstract

Discrimination between primary and secondary dengue infections is important, as the possibility ofDHF is more in secondary infection. Therefore, there is need to develop a test that can distinguishbetween primary and secondary serological responses. The traditionally-used haemagglutinationinhibition (HI) test, which is recommended by the World Health Organization, is complicated toperform. We standardized an enzyme-linked immunosorbent assay kit with some modifications todiscriminate between primary and secondary dengue infections. Sera from 72 patients with acutedengue infection were tested. Seventy-one of the 72 patients were correctly classified (18 of 18 patientswith primary dengue and 53 of 54 patients with secondary dengue). We conclude that this rapid andsimple test is an excellent alternative to the HI test for discriminating between primary and secondarydengue virus infections during the acute phase of dengue.

Keywords: Dengue; Discrimination; Primary and secondary infections; Immunoglobulin G avidity test.


Introduction

Dengue infection (DI) is among the mostimportant arboviral diseases in India in termsof both morbidity and mortality[1]. Dengue virusis a member of the flaviviridae family, with fourserologically related but antigenically distinctiveserotypes (DENV-1, DENV-2, DENV-3 andDENV-4). Acute infection due to dengue virusis generally asymptomatic and may presentwith classical dengue fever (DF), a mild illness,or its severe form, dengue haemorrhagic fever(DHF) or dengue shock syndrome (DSS)[1]. DHF,

which is life-threatening, has been postulatedto result from immune enhancement after asecond (heterotypic) infection by a differentserotype. The hypothesis on antibody-dependent enhancement can be used for theestablishment of an early diagnostic test todistinguish the primary from the secondaryinfection and to know the immunological statusof the patients infected with dengue virus.Keeping in view the increased possibility ofDHF in secondary infections, it is important todiscriminate between primary and secondaryinfections[2,3].


Immunoglobulin G avidity test to discriminate between primary and secondary dengue virus infections

Traditionally, the haemagglutinationinhibition (HI) test has been used to detectand differentiate between primary andsecondary dengue virus infections[4]. Patients areclassified as having secondary dengue virusinfections when the HI test titre in their sera isgreater than or equal to 1:2560, and areclassified as having primary dengue virusinfection if the HI test titre is less than 1:2560[4].However, when the interval between theacute- and the convalescent-phase samples isless than 7 days, or the convalescent phasespecimens are not available, haemagglutinationinhibition test is difficult to interpret[4,5].Moreover, the requirements of serum pre-treatment with acetone or kaolin to removenon-specific inhibitors makes HI test a tediousone. Furthermore, this test cannot give an earlydiagnosis[4,5].

Innis et al.[6] first proposed the classificationof primary and secondary dengue infectionsby determining the ratio of dengue virus IgMantibodies to the dengue virus IgG antibodies.The acute-phase sera of patients with primarydengue virus infections show higher IgM/IgGratios, as compared to the patients withsecondary infections who show lower IgM/IgGratios. The ratio of IgM/IgG higher than 1.78was considered as a marker of primary infectionand less than that was considered as a markerof secondary infection[6]. The IgG antibodyavidity test is a very useful tool fordifferentiating between primary and secondaryimmune responses[7]. The avidity assay is basedon the fact that the first antibodies synthesizedafter an antigenic challenge or primary infectionshave a lower affinity for the antigen than thoseproduced later on. In the secondary infection,the rapid antibody response is characterizedby the production of high-avidity antibodies[2].The avidity levels are reported as the avidityindex, expressing the percentage of IgG boundto the antigen following treatment withdenaturing agents. Recently, a few studies have

standardized the avidity test and discriminatedbetween primary and secondary dengueinfections. A study carried out by de Souza etal.[2] reported for the first time that by using acommercial kit for serological dengue diagnosis,it was possible to discriminate between aprimary dengue infection and a case ofsecondary dengue infection by detecting avidIgG[2]. Matheus et al.[3] developed an in-houseELISA to standardize avidity test and concludedthat the avidity test was more useful than theHI test for the discrimination of primary fromsecondary dengue virus infection, whatever thetype of dengue antigen used[3]. Since thisaspect has not been worked in our settings,there is a dearth of data from India. Thus, theaim of the present study was to standardize anIgG avidity test using a commercially availablekit for differentiating between primary andsecondary dengue infections in our settings.


The present study was performed in theDepartment of Microbiology, Maulana AzadMedical College and Associated Lok NayakHospitals, New Delhi, from September 2005to December 2006. The study group included150 patients clinically suspected of havingdengue infection, attending the outpatientdepartment and admitted in medical wards ofLok Nayak Hospital, New Delhi. The WHOcriteria were followed for inclusion or exclusionof a case of dengue infection[4]. Acute-phaseblood samples were collected within 4–8 daysof infection, and a convalescent-phase samplewas obtained after 8–15 days of onset of fever.Confirmation of acute dengue was obtainedby the detection of IgM antibodies anddemonstration of a ≥ 4-fold change in reciprocalIgM antibody titres in paired serum samples.IgG and IgM antibodies were detected by usingPanBio IgG and IgM capture ELISA kit.



Primary and secondary dengue infectionswere defined by using the following diagnosticcriteria. Primary dengue virus infection wascharacterized by the presence of IgMantibodies during the acute-phase andseroconversion (appearance of the IgGantibodies along with IgM antibodies) in theconvalescent phase[2], or the acute phase serumsample was positive for both IgM and IgG andthe ratio of IgM/IgG was greater than 1.78[6]. Asecondary dengue infection was characterizedby the presence of both IgM and IgG antibodiesin the acute-phase serum sample and also theIgM/IgG ratio was lesser then 1.78[3].

Following the above criteria, 18 patientswere classified as primary dengue infection casesand 54 patients as secondary infection cases.Samples whose absorbances were above thelimit of the ELISA reader were retested usinghigher dilution, i.e. 1:1000. The IgG avidity testwas standardized by using dengue indirect IgGELISA kit (PanBio), the procedure of which wasmodified by introducing an urea incubation step.The test was performed in the same manneras reported in a previous study onstandardization of dengue IgG avidity test[2],excepting that the commercial kit used by uswas PanBio dengue IgG indirect ELISA kit.Serum samples were diluted 1:100. Thesamples were then dispensed in duplicate intodengue antigen-coated wells. The sampleswere then incubated for half-an-hour at roomtemperature. After the incubation period firstdifferential washing with PBS was done. Afterfirst washing half of the wells were washedwith phosphate-buffered saline (pH 7.2) whichcontained urea, and the other half were rinsedwith phosphate-buffered saline without urea.After five washing cycles, the test wasperformed as per the manufacturer’sinstructions. The avidity index (AI), expressedas a percentage, was calculated as the ratio ofthe optical density with urea to the opticaldensity without urea multiplied by 100.

The test was performed several times usingdifferent concentrations of urea, 6M urea for10 min, 7M urea for 10 min, and 8M urea for5 min. Since variable results have beenobtained by several investigators when ELISAwas used to test avidity using differentcommercial plates, different sources of antigenand[8] different urea concentrations[9], we testeddifferent formats such as concentration of ureaand time for urea incubation step to standardizethe procedure and followed the same methodas de Souza et al.[2].

Statistical analysis

We used SPSS version 12 statistical softwarefor the statistical analysis. Mann Whitney’s testwas used to check whether the avidity levelwas significantly different between primary andsecondary dengue infections. A receiveroperating characteristic (ROC) curve analysiswas employed using Analyze-it software, toevaluate the accuracy of the test.

Results

The study used various incubation schedulesof 6M for 10 min, 7M for 10 min and 8M for 5min. The mean avidity index for primaryinfection was 38.24, 22 and 15.11 for urea at6M for 10 min, 7M for10 min and 8M for 5min respectively, while the mean avidity indicesfor secondary infection were 78.94, 72 and43.7 for urea at 6M for 10 min, 7M for 10 minand for urea at 8M for 5 min, respectively(Figure). The use of 6M urea for 10 min coulddifferentiate only 13 out of 18 primary dengueinfections and 44 out of 54 secondary dengueinfections. Whereas 8M urea for 5 min coulddifferentiate 15 out of 18 primary dengueinfections and 49 out of 54 secondary dengueinfections. The use of 7M urea for 10 mindifferentiated best between the primary and



the secondary infections. This washing schedulecould correctly classify all 18 primary and 53secondary dengue infections, and was chosento evaluate IgG AI. Avidity indices ranged from14–32 for primary infections and 27–106 forsecondary infections with 7M for 10 min. The

mean AI for primary infection was 22±5.4 andfor secondary infection was 72±12.2. Themean AI of the 18 primary dengue infectionswas significantly lower than the 54 secondarydengue infections (P < 0.001) (Table). The cut-off point of ≥ 27.4% IgG AI was chosen for the

Figure: IgG antibody avidity indices in sera from patients with primary andsecondary dengue infections with washing schedules of 7M urea for 10 min

and 8M urea for 5 min and 6M for 10 min

100.00

80.00

60.00

40.00

20.00

0.00

Primaryinfections at 7M

Secondaryinfections at 7M





8

Table: Performance of IgG avidity test

≥ 27.4 0 53 72.87(0.0) (98.14%) (14.2)

< 27.4 18 22 1 26(100.0) (5.4) (1.85%)

Total 18 22 54 72(100.0) (5.4) (100.0) (12.2)

Data for patients withPrimary infection Secondary infection

No. (%) of patients Mean AI (SD) No. (%) of patients Mean AI (SD)

Avidity Index(AI) range



classification of primary and secondaryinfections. At this cutoff point, the IgG aviditytest provided correct classifications of 71 of 72patients [18 of 18 patients with primary dengue(100%) and 53 of 54 patients with secondarydengue (98.14%)].

The avidity test showed 98.67% sensitivity,100% specificity.

Discussion

Dengue hemorrhagic fever and dengue shocksyndrome have been observed to occurfrequently with secondary dengue infection.Therefore, it is important to discriminatebetween primary and secondary infections andto asses the immunological status of patients toknow the progression of the disease[4]. Thehaemagglutination inhibition test isconventionally used as a standard test todifferentiate between primary and secondarydengue virus infections. The main disadvantagesof the HI test are the requirement of pairedsamples, serum pre-treatment with acetone orkaolin and goose red blood cells[5,10]. Keeping inmind the disadvantages of HI, alternative assaysare needed for differentiating between primaryand secondary dengue infections. The utility of

the assay in diagnosing a primary infection hasbeen reported for a variety of parasites andviruses like leishmania[11], respiratory syncytialvirus (RSV)[12], and rubella[13]. Recently, a fewstudies have standardized the avidity test anddiscriminated between primary and secondarydengue infection. To the best of our knowledge,no study is reported from India which has usedavidity test for differentiating between primaryand secondary dengue infections. The studiescarried out by de Souza et al.[2] have shown forthe first time that, by using a commercial kit forthe diagnosis of dengue, it was possible todiscriminate between a case of primary dengueinfection and a case of secondary dengueinfection by detecting avid IgG antibodies. Ourresults, obtained by using a commercial IgGindirect ELISA kit, confirmed the results obtainedby de Souza et al.[2] We observed a mean AI of22% during a primary infection and 72% duringa secondary infection. The sensitivity andspecificity of the test were 98.67% and 100%,respectively, with a single sample. These findingswere in tune with the previous studies, whichshowed that the avidity test was an excellentalternative to HI assay for differentiatingbetween primary and secondary dengueinfections. Thus, the avidity test standardizedby this study is a simple test which candifferentiate between primary and secondarydengue infections.

References

[1] Chakravarti A, Kumaria R. Eco-epidemiological analysis of dengue infectionduring an outbreak of dengue fever. India VirolJ. 2005; 2: 32.

[2] De Souza VA, Fernandes S, Araujo ES, TatenoAF, Olivera OM. Use of an immunoglobulin Gavidity test to discriminate between primaryand secondary dengue virus infections. J ClinMicrobiol. 2004 Apr ; 42: 1782-1784.

[3] Matheus S, Deparis X, Labeau B, Lelarge J,Movran J, Dussart P. Discrimination of primaryand secondary dengue virus infection by animmunoglobulin G avidity test using a single

acute phase serum sample. J Clin Microbiol.2005; 43: 2793-2797.

[4] World Health Organization. Denguehaemorrhagic fever: diagnosis, treatment,prevention and control. Geneva: World HealthOrganization, 1997.

[5] Gubler DJ. Dengue and dengue hemorrhagicfever. Clin Microbiol Rev. 1998; 11: 480-496.

[6] Innis BL, Nisalak A, Nimmannitya S,Kusalerdchariya S, Chongswasdi V,Suntayakorn S, Puttisri P, Hoke CH. Anenzyme-linked immunosorbent assay to



characterize dengue infections where dengueand Japanese encephalitis co-circulate. Am JTrop Med Hyg. 1989; 40(4): 418-27.

[7] Hedman KM, Lappalainen M, Söderlund M,Hedman L. Avidity of IgG in serodiagnosis ofinfectious diseases. Rev Med Microbiol. 1993;4: 123-129.

[8] Inouye S, Hasegawa A, Matsuno S, Katow S.Changes in antibody avidity after virusinfections: detection by an immunosorbentassay in which a mild protein-denaturing agentis employed. J Clin Microbiol. 1984; 20(3):525-9.

[9] Blackburn NK, Besselaar TG, Schoub BD,O’Connell KF. Differentiation of primarycytomegalovirus infection from reactivationusing the urea denaturation test formeasuring antibody avidity. J Med Virol.1991; 33(1): 6-9.

[10] Chakravarti A, Parthsarathy P, Chakravarti A.Evaluation of a rapidimmunochromatographic test in the diagnosisof dengue fever. Indian J Pathol Microbiol.2003; 46(1): 127-8.

[11] Redhu NS, Dey A, Balooni V, Singh S. Use ofimmunoglobulin g avidity to determine thecourse of disease in visceral and post-kala-azar dermal leishmaniasis patients. Clin VaccineImmunol. 2006; 13(8): 969-71.

[12] Meurman O, Waris M, Hedman K.Immunoglobulin G antibody avidity inpatients with respiratory syncytial virusinfection. J Clin Microbiol. 1992; 30(6):1479-84.

[13] Hedman K, Rousseau SA. Measurement ofavidity of specific IgG for verification of recentprimary rubella. J Med Virol. 1989; 27(4):288-92.


DENV-3 genotype III circulating in São Paulo, Brazil,from 2003 to 2008 is not associated with dengue

haemorrhagic fever/dengue shock syndrome

Luiza Antunes de Castro-Jorgea, Daniel Macedo de Melo Jorgeb,Benedito Antônio Lopes da Fonsecaa#

aLaboratory of Molecular Virology, Virology Research Centre, School of Medicine of Ribeirão Preto,University of São Paulo, Avenida dos Bandeirantes, 3900, 14049 -900 Ribeirão Preto, São Paulo, Brazil

bBioinformatics Laboratory, Department of Genetics, School of Medicine of Ribeirao Preto,University of São Paulo, Avenida dos Bandeirantes, 3900, 14049 - 900 Ribeirão Preto, São Paulo, Brazil

Abstract

Dengue viruses (DENV) are the most important arboviruses of public health significance, and compriseof four distinct antigenic serotypes (DENV-1 to 4) that show substantial genetic diversity. These virusesusually cause dengue fever (DF) but some patients progress to a more severe form of the illness, i.e.dengue haemorrhagic fever/dengue shock syndrome (DHF/DSS). The first reports of DENV-3 cases inBrazil occurred in the year 2000 with co-circulation of DENV-1 and 2. Thereafter, DENV-3 spreadthroughout the country. DENV-3 phylogenetic analysis has revealed the existence of four to five DENV-3 genotypes. Genotype III of DENV-3 has been the main genotype circulating in Brazil, but recent studieshave indicated that DENV-3 genotype I and genotype V are also circulating in some states of Brazil. Inorder to evaluate DENV-3 genotypes circulating in São Paulo state from 2003 through 2008 we analyzedthe NS1 region of DENV-3 isolated from patients residing in Ribeirão Preto and presenting with differentclinical manifestations of dengue disease. Nucleotide sequences from 31 viruses were obtained andcompared to 105 DENV-3 corresponding sequences retrieved from GenBank. Phylogenetic analysisshowed that São Paulo DENV-3 sequences belong to genotype III and that Puerto Rico strains are closelyrelated to South American strains. There was no association between DENV-3 genotype and DHF/DSS.

Keywords: Dengue genotyping; Phylogenetic analysis; São Paulo; Brazil.

#E-mail: [email protected]; Fax: +55-16-3633-0036

Introduction

Dengue is an acute febrile disease caused bya flavivirus with four antigenically distinctserotypes (DENV-1, -2, -3, and -4) and is mainly

transmitted to humans by Aedes aegyptimosquitoes. Dengue virus (DENV) infectionsare currently the most important humanarboviral disease in terms of morbidity andmortality in several countries of America, Asiaand Africa[1]. DENV contains a single-stranded


DENV-3 genotype III circulating in São Paulo, Brazil

positive-sense RNA genome of about 10.7 kbthat encodes three structural proteins (envelopeglycoprotein, E; membrane, M; and capsid, C)and seven non-structural proteins (NS1, NS2a,NS2b, NS3, NS4a, NS4b and NS5)[2]. Theclinical manifestations of dengue range frominapparent or mild disease (dengue fever) tosevere forms known as dengue haemorrhagicfever and dengue shock syndrome (DHF/DSS).It currently affects around 100 million peopleevery year worldwide, and around 500 000people have DHF/DSS with 2.5% deaths[3]. Themortality rates vary according to the affectedregion, being close to 11% in Brazil which hasaccounted for nearly 65% of the reported casesof dengue fever in the American regions inthe last 10 years[4].

The sequential introduction of differentdengue serotypes in Brazil has contributed tothe high incidence of the disease. The firstepidemic, which occurred in the north-westAmazon region (Roraima state) in 1982, wasassociated with DENV-1 and DENV-4[5]. Aftera 4-year interval without any confirmed denguecases, an epidemic due to DENV-1 occurredin Rio de Janeiro state and was followed byseveral epidemics in highly populated cities inthe south-east and north-east regions of Brazil[6].In 1990–1991 an outbreak of DHF wasrecorded in Rio de Janeiro, and it was associatedwith DENV-2[7]. The first reports on DENV-3cases in Brazil occurred in 2000[8] and a periodof co-circulation of DENV-1, -2, and -3 wasobserved. Serotyping analysis of dengue virusstrains isolated after 2002 showed that DENV-3 has spread to new areas of the country andreplaced the other dengue virus serotypes[9],confirming a high infection capacity of this virusin both humans and vectors. In 2008, DENV-2was the predominant serotype isolated in Riode Janeiro and Ceará states, which increasedthe disease severity after almost seven yearsof DENV-3 circulation. However, DENV-3 isstill the predominant serotype detected inBrazil, and in São Paulo state. The city of

Ribeirão Preto (estimated population 558 137),located in the north-eastern region of São Paulostate, is among the cities with highest incidenceof dengue in the state. Since 1990, yearlyepidemics have occurred in the city of RibeirãoPreto, and, in 2006, the dengue incidencereached its peak with 1153 cases/100 000inhabitants, but, fortunately, this incidence ratehas been reduced in recent years[10].

Molecular analyses showed the existenceof different variants among DENV serotypeswhich led to the recognition of differentgenotypes within each serotype. Geneticdiversity of DENV-1 and DENV-2 identified fiveviral genotypes[11,12]. DENV-4 viruses were firstseparated into two distinct genotypes[13], butmore recently, a third genotype has beenidentified[14]. DENV-3 was initially classified intofour geographically distinct genotypes[15];nevertheless, recent studies have suggested theexistence of an additional group withingenotype I that was named genotype V[16].DENV-3 strains circulating in Brazil since 2000belong to genotype III[17,18], and are closelyrelated to strains from Sri Lanka and India,which are associated with DHF/DSS cases inthose countries[15]. However, circulation ofDENV-3, genotype I, has been reported inMinas Gerais state, a state located in the south-eastern region of Brazil, in 2003–2004[19], andthe co-circulation of genotypes III and V wasreported in patients living in the northern regionof Brazil during the 2002–2004 epidemics[20].

Molecular characterization of DENV withthe identification of circulating genotypes is animportant task for laboratories performingvirological surveillance, as it has beendemonstrated that intratypic variations amongdifferent genotypes could be associated withthe disease severity[21,22,23]. In order to evaluateDENV-3 circulating genotypes in São Paulo statefrom 2003 to 2008, we analysed the NS1 regionof DENV-3 isolated in the state from patientswith different clinical presentations of denguedisease.




Clinical samples

This study was conducted between 2003 and2008 at the School of Medicine of Ribeirão Preto– University of São Paulo (FMRP-USP), and the

samples were collected at the Clinical Hospitalof FMRP-USP and at the Centro de Saúde Escola(CSE), a community health centre supervisedby the School of Medicine of Ribeirão Preto –University of São Paulo. Ethical clearance wasgranted by the Research Ethical Committee ofthe Clinical Hospital of FMRP-USP and by the

Table: Brazilian strains of DENV-3 sequenced in the present study

Strains GenBank accession # Year isolated Clinical manifestationD3/BR/RPR17/03 FJ608533 2003 DHFD3/BR/RPM8/04 FJ608504 2004 DFD3/BR/RPM1/04 FJ608505 2004 DFD3/BR/RPM3/04 FJ608508 2004 DFD3/BR /RPM4/04 FJ608506 2004 DFD3/BR/RPM7/04 FJ608507 2004 DFD3/BR/RPc7/06 FJ608529 2006 DFD3/BR/RPc9/06 FJ608530 2006 DFD3/BR/RPc12/06 FJ608531 2006 DFD3/BR/RPc18/06 FJ608532 2006 DFD3/BR/RP5/07 FJ608509 2007 DFD3/BR/RP9/07 FJ608510 2007 DFD3/BR/RP12/07 FJ608511 2007 DFD3/BR/RP13/07 FJ608512 2007 DFD3/BR/RP15/07 FJ608534 2007 DHFD3/BR/RP22/07 FJ608513 2007 DFD3/BR/RP23/07 FJ608514 2007 DFD3/BR/RP24/07 FJ608515 2007 DFD3/BR/RP25/07 FJ608516 2007 DFD3/BR/RP27/07 FJ608517 2007 DFD3/BR/RP32/07 FJ608518 2007 DFD3/BR/RP33/07 FJ608519 2007 DFD3/BR/RP35/07 FJ608520 2007 DFD3/BR/RP36/07 FJ608521 2007 DFD3/BR/RP37/07 FJ608522 2007 DFD3/BR/RP47/07 FJ608523 2007 DFD3/BR/RP52/07 FJ608524 2007 DFD3/BR/RP61/08 FJ608525 2008 DFD3/BR/RP68/08 FJ608526 2008 DFD3/BR/RP72/08 FJ608527 2008 DFD3/BR/RP74/08 FJ608528 2008 DF



Academic Board of Teaching and Research atCSE. Patients were recruited if they gaveinformed written consent and if the responsiblephysician suspected of dengue virus infectionon clinical grounds based on the World HealthOrganization (WHO) guidelines[24]. Thirty-oneisolates obtained from acute-phase serumsamples from patients with DF or DHF(previously identified by RT-PCR as DENV-3)were selected for this study (Table 1). All serumsamples were from patients living in the city ofRibeirão Preto or neighbouring cities.

RNA extraction and RT-PCR

Viral RNA was isolated from 140 µL of eachserum sample using QIAamp® Viral RNA MiniKit (QIAGEN, USA) according to themanufacturer’s directions. The RT-PCR wascarried out using QIAGEN® OneStep RT-PCRkit in a 25 µL final volume containing 5 µL 5XOne-Step RT-PCR buffer, 5 µL of dNTPs (10mM), 0.5 pmol/µL of each primer, 1 µL ofenzyme mix (Omniscript and Sensiscript) and 5µL of RNA. The reaction was performed usingserotype-specific primers described previouslyby Lanciotti[25] and also using primers describedby Henchal[26]. The amplifications wereperformed under the following parameters:50 °C for 30 min and 95 °C for 15 min for reversetranscription, followed by 35 cycles of 94 °C for1 min, 55 °C for 1 min, 72 °C for 1 min andfinal cycle of 72 °C for 10 min. The ampliconswere detected on a 2% agarose gelelectrophoresis stained with 1 µg/mL of ethidiumbromide, and then visualized with the KodakElectrophoresis Documentation and AnalysisSystem 120 (Kodak, USA).

Sequencing, multiple sequencealignment and phylogenetic analysis

In order to obtain the nucleotide sequencesof NS1 region from 31 virus samples, PCRamplicons were purified with Wizard® SV Gel

and PCR Clean-Up System (Promega, USA) anddirectly sequenced twice in both orientationsusing the ABI Prism Big Dye Terminator CycleSequencing Ready Kit (Applied Biosystems,UK).

Phylogenetic analysis of the NS1 partialcoding sequence (cds) also included 105sequences of DENV-3 retrieved from GenBank.The new sequences obtained in this work weresubmitted to the GenBank (Table). Prior tophylogenetic analysis, DENV-3 sequences werealigned by using the multiple sequencealignment program CLUSTAL W vs.1.8[27], andedited using the BioEdit software v7.0.0[28] andMEGA 4.1[29]. Phylogenetic trees wereconstructed according to the best-fit model ofnucleotide substitution implemented inModelTest (TrN+G)[30] and were implementedwith model Maximum Composite Likelihood(ML). The phylogenetic relationships amongstrains were reconstructed by the neighbour-joining (NJ) and maximum parsimony (MP)methods using Mega 4.1. Branch topology wasverified by generating 1000 bootstraps for NJand 10 for MP and a representative sequenceof DENV-4 was used to root the trees.Phylogenetic approaches yielded identical ornearly identical topologies, but only NJ tree isshown.


In order to characterize the DENV-3 strainscirculating in Ribeirão Preto, São Paulo state,and to determine their relationships with otherDENV-3 strains, clinical samples from differentepidemics were analysed. This study isimportant due to the fact that despite havingone of the highest incidence rates of denguein the country, and having experiencedepidemics caused by DENV-1 and DENV-2, thecity of Ribeirão Preto has a very low mortalityrate associated with dengue illness. Thirty-onesamples collected from 2003 to 2008 in



Ribeirão Preto were evaluated in this study.The majority of the patients had clinicalsymptoms of DF and only two patientspresented with symptoms of DHF. The samplesbelonged mainly to adults ranging from 19 to50 years and the male-to-female ratio was 1:1.The nucleotide sequences were obtained fromviral RNA extracted directly from the patients’serums, and then submitted to GenBank(Table). Sequence analysis revealed that allviruses sequenced in this study belonged toDENV-3 genotype III (Figure 1).

NS1 was chosen for DENV-3 genotypingbecause it has been shown to be important indisease outcome[31], for having a number of Band T cell epitopes[32,33], and being increasinglyimportant in the diagnosis of dengue acutedisease[26,34,35]. Thus, even though NS1 has notbeen well evaluated in genotyping studies, inour view, it represented a good target forgenotyping DENV-3 isolates from Brazil[36]. Allthe NS1 DENV-3 sequences segregated intofour distinct genotypes as established on thebasis of C, prM and E genes in earlierstudies[15,37]. Genotype IV is not shown due tounavailability of sequences. Genotype Icomprised viruses isolated in Taiwan (China),Indonesia, Philippines, Timor-Leste and FrenchPolynesia from 1978 to 2005. Genotype IIconsisted of viruses from Thailand, Taiwan(China), Bangladesh and Viet Nam isolated from1987 to 2007. Genotype III was further dividedin three clades: the Indian clade consisting ofIndian isolates of 2006; the American cladeconsisting of isolates from Martinique,Argentina, Venezuela, Puerto Rico and Brazilisolated from 1999 to 2008; and the Asian cladecomprising of Singapore, Sri Lanka and Taiwan(China) isolates. The existence of intragenotypicgroups have also been observed by otherauthors, e.g. Messer[37] analysing the C, prMand a portion of E genes showed that genotypeIII viruses includes four groups: Latin America,East Africa and groups A and B from Sri Lanka.Kochel et al.[38] have also observed four main

groups within genotype III, a South Americangroup, a Central American group and alsogroups A and B from Sri Lanka. Our analysesshowed a similar distribution of genotype IIIviruses in the three analysed trees; however,by analysing the available sequences for theNS1 region we have found that genotype IIIforms three main clades as already describedabove.

Within the American clade two main clusterswithin genotype III were observed (Figure 1).One cluster groups Puerto Rico DENV-3 strainsisolated in 1998 to 2007 and Venezuelan strainsisolated in 2001. The second cluster containsfour groups of isolates. One group is composedby isolates from Puerto Rico and one Venezuelanstrain of 2005. There is another minor groupwith Argentinean strains isolated in 2007, a biggroup with two Argentinean isolates, theMartinique isolate and some Brazilian isolatesfrom São Paulo, Rio de Janeiro and Acre states,and another group with two isolates from PuertoRico from 2003–2004 and the Brazilian isolatesfrom Acre and São Paulo states.

Several studies have applied phylogeneticmethods to analyse the epidemiology of dengueviruses and to understand the geneticrelationships between them[37-40]. These studieshave shown that dengue viruses could travelshort distances between neighbouringcountries[38,40] as well as long distances betweencontinents[39]. In this study, it was possible todetermine that Brazilian DENV-3 isolatesgrouped into separate clusters: the majority ofsamples isolated in Ribeirão Preto were groupedaltogether and were related to another Brazilianstrain from Acre state, whereas another isolatesgrouped with two Puerto Rico strains from 2003and 2004, and the rest of sequences fromBrazilian isolates were grouped with Argentineanand Martinique viruses (Figure 1). These differentgroups show that the viruses are constantlymoving within the country, as described byAquino et al.[18]. This tree also suggests that some



Figure 1: Neighbour-joining phylogenetic tree of DENV-3 using a 419 bp fragmentof the NS1 gene

[This phylogenetic tree is showing the presence of genotype III and V in Brazil. The Tamura-Neinucleotide substitution model was used to estimate distance matrix. Sequences obtained in the

present study are marked with . It is possible to observe two main clusters within the Americanclade in genotype III. Cluster A is composed of strains from different Latin American countriesand cluster B is composed only with strains from Puerto Rico and Venezuela. Bootstrap valuesgreater than 80% were maintained in the tree. Horizontal branch lengths are drawn to scale.]

D3/BR/RPM8/04D3/BR/RPc12/06D3/BR/RPM1/04D3/BR/RPM7/04D3/BR/RPM3/04D3/BR/RPc9/06D3/BR/RP24/07D3/BR/RP13/07D3/BR/RP12/07

D3/BR/RP23/07D3/BR/RP22/07

D3/BR/RP68/08D3/BR/RP72/08D3/BR/RP74/08D3/BR/RP61/08

D3/BR/RP52/07D3/BR/RP5/07

D3/BR/RP32/07D3BR/RP1/2003

D3/BR/RPM4/04BR DEN3 97 04

D3/BR/RP47/07BID V1730BID V1608

ARG6768 07ARG6733 07

D3/H/IMTSSA MART/1999/1243D3/BR/RP37/07

BR DEN3 290 02D3/BR/RP36/07

BR DEN3 95 04BR74886/02BR DEN3 98 04

D3/BR/RP27/07D3/BR/RP33/07D3/BR/RPc18/06

D3/BR/RPR17/03D3/BR/RPc7/06D3/BR/RP25/07D3/BR/RP9/07D3/BR/RP35/07D3/BR/RP15/07

ARG6645 07ARG6541 07

ARG11586 07ARG6475 07

ARG6694 07ARG11595 07

BID V1623BID V1078

BID V1043BID V1593

A

BID V1451BID V1617

BID V1450BID V1453BID V1088BID V1737BID V1476

BID V2105BID V1621

BID V1607BID V1490

BID V1417DENV 3/US/BID V2123/2002

BID V916BID V1116

BID V913

B

American Clade

Singapure99TW628

D3/H/IMTSSA SRI/2000/1266Asian Clade

D3/05NS1/del200629D3/06

D3/01NS1/del2006Indian Clade

Genotype III

BR DEN3 RO3 02 2BR DEN3 RO2 02BR DEN3 RO1 0280 2H87

Genotype V

95TW466PhMH J1 97

Taiwan 739079AU2082

PF92/2956PF94/136116

PF89/320219PF90/6056

Sleman/7898902890 DF DV 3

den3 8898901590

D3/Hu/TL129NIID/2005D3/Hu/TL109NIID/2005

D3/Hu/TL029NIID/2005D3/Hu/TL018NIID/2005ET00 209

KJ71KJ46

PI64PH86FW06

FW01den3 98

BA5198901403 DSS DV 3

98901640TB55iTB16

Genotype I

ThD3 0010 87BDH02 1

BDH02 5Singapore 8120/9598TW182

ThD3 0055 93C0331/94

98TW41498TW50398TW38898TW390

C0360/94KPS 4 0657/207

BID V1013ThD3 1283 98

BID V101007CHLS001

BID V1009BID V1015BID V1327BID V1016

BID V1012BID V1329BID V1326BID V1331BID V1882

BID V1008BID V1014

Genotype II

AY618992 DV4 Th01

99

97

89

85

81

93

90

86

86

80

0.05



Brazilian strains are similar to the Puerto Ricostrains. In fact, Puerto Rico strains are related toSouth American strains included in this analysis,which could indicate a strong relationshipbetween these strains, reinforcing the theoryof a single introduction of DENV-3 genotype IIIin Latin America[18,41]. Even though it is not clearwhether the American genotype III lineagecame from Africa or Asia, a migration analysisstudy done by Araujo et al.[41] has suggested thatthe genotype III was first introduced into the

Americas through Mexico, and from there theseviruses spread to other countries in the regionusing independent migration routes to reachCentral America, the Caribbean and SouthAmerican countries.

Another interesting point to make is thatthese closely related strains are responsible forvery distinct disease manifestations in eachcountry, inducing severe disease in Puerto Ricoand mild disease in Ribeirão Preto city.

Figure 2: Distribution of DENV-3 genotypes in Brazil

[A red represents the presence of genotype I; a green represents the presence ofgenotype V; and a yellow represents the circulation of genotype III. The co-circulation of two

genotypes is registered in two states of Brazil (Minas Gerais and Rondonia).]



We have shown that DENV-3 genotypeIII is the circulating genotype in São Paulo stateand is still the most prevalent genotype in Braziland the Americas. However, the recent reportsof DENV-3 genotype I and V co-circulation inBrazil associated with cases of DF and DHF[19,20]

may change this situation. These genotypeswere isolated in different states of Brazil (Figure2), and given that Brazil is a tropical country,under optimal conditions, this DENV-3genotype may spread to other areas of thecountry and cause a more severe disease.

Studies such as this one are importantsurveillance strategies that should be taken tofollow the DENV path across the country, andto investigate the association of a specificgenotype with distinct disease manifestations.Also, the emergence of DENV-3 genotype I andV in the Americas supports future research tofollow up the movement of these genotypes inBrazil, and also to identify the possible

introduction and emergence of these genotypesinto other countries in South America. Moreover,given the limited options available for denguecontrol, active surveillance programmes withcontinuous monitoring of dengue infection incommunities is still the best strategy availableto detect the introduction of new serotypes/genotypes, and, consequently, to prevent theoccurrence of epidemics. Genetic studiesinvestigating substitutions across different geneswithin the DENV-3 viruses are also necessary toknow, with more certainty, the evolutionarydirections of DENV-3 in South America.

Acknowledgments

This study was supported by a grant from theFundação de Amparo à Pesquisa do Estado deSão Paulo (FAPESP) (Grant # 2007/04326-7).LAC-J was recipient of scholarship from São PauloResearch Foundation (Grant # 06/58792-6).

References

[1] Fonseca BAL, Figueiredo LTM. Dengue. In:Veronesi R, Focaccia R (3th ed.), Tratado deInfectologia. São Paulo: Livraria Atheneu,2005. pp. 343-356.

[2] Chambers TJ, Hahn CS, Galler R, Rice CM.Flavivirus genome organization, expression,and replication. Annu Rev Microbiol. 1990;44: 649-688.

[3] World Health Organization. Dengue, denguehaemorrhagic fever and dengue shock syndromein the context of the integrated management ofchildhood illness. Geneva, World HealthOrganization, 2005. Document WHO/FCH/CAH/05.13.

[4] Pan American Health Organization. Denguesurveillance–featured surveillance items (http:// w w w. p a h o . o r g / e n g l i s h / a d / d p c / c d /dengue.htm, accessed 06 January 2009).

[5] Osanai CH. A epidemia de dengue em BoaVista, Território Federal de Roraima, 1981-1982. Dissertação de mestrado, EscolaNacional de Saúde Pública, Rio de Janeiro,RJ, 1984.

[6] Schatzmayr HG, Nogueira RMR, Rosa APAT.An outbreak of dengue virus at Rio de Janeiro,Memórias do Instituto Oswaldo Cruz. 1986;81: 245-246.

[7] Nogueira RMR, Zagner SMO, Martins ISM,Lampe E, Miagostovich MP, Schatzmayr HG.Dengue Heamorrhagic Fever/Dengue ShockSyndrome (DHF/DSS) caused by serotype 2in Brazil. Memórias do Instituto Oswaldo Cruz.1991; 86: 269.

[8] Nogueira RMR, Miagostovich MP, Filipis AMP,Pereira MAS, Schatzmayr HG. Dengue virustype 3 in Rio de Janeiro, Brazil. Memórias doInstituto Oswaldo Cruz. 2001; 96: 925-926.



[9] Nogueira RMR, Araújo JMG, Schatzmayr HG.Aspects of dengue virus infections in Brazil1986-2007. Virus Reviews & Research. 2007;12: 1-17.

[10] Prefeitura Municipal de Ribeirão Preto.Secretaria Municipal de Saúde. (http://www.r ibei raopreto.sp.gov.br / s saude/i16principal.asp?pagina=/ssaude/vigilancia/vigep/dengue/i16casos.htm, accessed 05January 2009).

[11] Rico-Hesse R. Molecular evolution anddistribution of dengue viruses type 1 and 2 innature. Virology. 1990; 174(2): 479-493.

[12] Lewis JA, Chang GJ, Lanciotti RS, Kinney RM,Mayer LW, Trent DT. Phylogenetic relationshipof dengue - 2 viruses. Virology. 1993; 197:216-224.

[13] Lanciotti R, Gubler DJ, Trent DW. Molecularevolution and phylogeny of dengue-4 viruses.J Gen Virol. 1997; 78: 2279-2284.

[14] Wang E, Ni H, Xu R, Barrett A, Watowich S,Gubler D, Weaver S. Evolutionaryrelationships of endemic/epidemic andsylvatic dengue viruses. Journal of Virology.2000; 74(7): 3227-3234.

[15] Lanciotti RS, Lewis JG, Gubler DJ, Trent DW.Molecular evolution and epidemiology ofdengue-3 viruses. J Gen Virol. 1994; 75: 65-75.

[16] Wittke V, Robb TE, Thu HM, Nisalak A,Nimmannitya S, Kalayanrooj S, Vaughn DW,Endy TP, Holmes EC, Aaskov JG. Extinctionand rapid emergence of strains of dengue 3virus during an interepidemic period. Virology.2002; 301: 148-156.

[17] Miagostovich MP, Santos FB, Simone TS, CostaEV, Filippis AMB, Schatzmayr HG, NogueiraRMR. Genetic characterization of dengue virustype 3 isolates in the State of Rio de Janeiro,2001. Braz J Med Biol Res. 2002; 35: 869-72.

[18] Aquino VH, Anatriello E, Gonçalves PF, DaSilva EV, Vasconcelos PFC, Vieira DS, BatistaWC, Bobadilla CV, Moran M, Figueiredo LTM.Molecular epidemiology of dengue type virus3 in Brazil and Paraguay, 2002-2004. Am JTrop Med Hyg. 2006; 75: 710-715.

[19] Figueiredo L, Cecillio AB, Ferreira GP,Drumond BP, Oliveira JG, Bonjardim CA,Ferreira PCP, Kroon EG. Dengue virus 3genotype 1 associated with dengue fever anddengue hemorrhagic fever, Brazil. Emerg InfectDis. 2008; 14: 314-316.

[20] Nogueira MB, Stella V, Bordignon J, Batista WC,Borba L, Silva LHP, Hoffmann FG, Probst CM,Santos CND. Evidence for the co-circulationof dengue virus type 3 genotypes III and V inthe Northern region of Brazil during the 2002-2004 epidemics. Mem Inst Oswaldo Cruz.2008; 103(5): 483-488.

[21] Leitmeyer KC, Vaughn DW, Watts DM, SalasR, Chacon IV, Ramos C, Rico-Hesse R. Denguevirus structural differences that correlate withpathogenesis. J Virol. 1999; 73: 4738-4747.

[22] Pandey BD, Igarashi A. Severity-relatedmolecular differences among nineteen strainsof dengue type 2 viruses. Microbiol Immunol.2000; 44: 179-188.

[23] Rico-Hesse R. Microevolution and virulenceof dengue viruses. Adv Virus Res. 2003; 59:315-341.

[24] World Health Organization. Denguehaemorrhagic fever: diagnosis, treatment andcontrol. Geneva: WHO, 1997.

[25] Lanciotti RS, Calisher CH, Gubler DJ, ChangGJ, Vorndam AV. Rapid detection and typingof dengue viruses from clinical samples byusing reverse transcriptase-polymerase chainreaction. J. Clin. Microbiol. 1992; 30(3):545-551.

[26] Henchal EA, Polo SI, Vorndam V, Yaemsiri C,Innis BI, Hoke CH. Sensitivity and specificityof a universal primer set for the rapid diagnosisof dengue virus infections by polymerasechain reaction and nucleic acid hybridization.Am. J. Trop. Med. Hyg. 1991; 45(4): 418-428.

[27] Thompson JD, Gibson TJ, Plewniak F,Jeanmougin F, Higgins DG. The Clustal Xwindows interface: Flexible strategies formultiple sequence alignment aided by qualityanalysis tools. Nucl Ac Res. 1997; 24:4876-4882.



[28] Hall TA. BioEdit: a user-friendly biologicalsequence alignment editor and analysisprogram for Windows 95/98/NT. Nucl AcidsSymp Ser. 1999; 41: 95-98.

[29] Tamura K, Dudley J. Nei M, Kumar S. MEGA4.Molecular Evolutionary Genetic analysis(MEGA) software version 4.0. Molecular Biologyand Evolution. 2007; 24: 1596-1599.

[30] Posada D, Crandall KA. Modeltest: testing themodel of DNA substitution. Bioinformatics.1998; 14(9): 817-818.

[31] Libraty DH, Young PR, Pickering D, Endy TP,Kalayanarooj S, Green S, et al. Highcirculating levels of the dengue virusnonstructural protein NS1 early in dengueillness correlate with the development ofdengue hemorrhagic fever. J Infect Dis. 2002;186(8): 1165-1168.

[32] Falconar AKI, Young PR, Miles MA. Preciselocation of sequential dengue virussubcomplex and complex B cell epitopes onthe nonstructural-1 glycoprotein. Arch Virol.1994; 137: 315-326.

[33] Garcia G, Vaughn DW, del Angel RM:Recognition of synthetic oligopeptides fromnonstructural proteins NS1 and NS3 ofdengue-4 virus by sera from dengue virus-infected children. Am J Trop Med Hyg. 1997;56: 466-470.

[34] Dussart P, Labeau B, Lagathu G, Louis P, NunesMR, Rodrigues SG, et al. Evaluation of anenzyme immunoassay for detection of denguevirus NS1 antigen in human serum. Clin VaccineImmunol. 2006; 13: 1185-1189.

[35] Kumarasamy V, Wahab AHA, Chuab SK,Hassan ZA, Chema YK, Mohamada M, et al.Evaluation of a commercial dengue NS1antigen-capture ELISA for laboratory diagnosisof acute dengue virus infection. J Virol Methods.2007; 140: 75-79.

[36] King CC, Chao DY, Chien LJ, Chang GJJ, LinTH, Wu YC, Huang JH. Comparative analysisof full genomic sequences among differentgenotypes of dengue virus type 3. VirologyJournal. 2008; 5: 63.

[37] Messer WB, Gubler DJ, Harris E, SivananthanK, Silva AM. Emergence and global spread ofa dengue serotype 3, subtype III virus. EmergingInfectious Diseases. 2003; 9(7): 800-809.

[38] Kochel T, Aguilar P, Felices V, Comach G, CruzC, Alava A, Vargas J, Olson J, Blair P. Molecularepidemiology of dengue virus type 3 inNorthern South America: 2000-2005.Infection, Genetics and Evolution. 2008; 8:682-688.

[39] Rico-Hesse R, Harrison LM, Salas RA, Tovar D,Nisalak A, Ramos C, Boshell J, Mesa MTR,Nogueira RMR, Rosa AT. Origins of denguetype 2 viruses associated with increasedpathogenicity in the Americas. Virology. 1997;(230): 244-251.

[40] Kobayashi N, Thayan R, Sugimoto C, Oda K,Saat Z, Vijayamalar B, et al. Type-3 dengueviruses responsible for the dengue epidemicin Malaysia during 1993-1994. Am J Trop MedHyg. 1999; 60: 904-909.

[41] Araujo JMG, Nogueira RMR, Schatzmayr HG,Zanotto PMA, Bello G. Phylogeography andevolutionary history of dengue virus type 3.Infection, Genetics and Evolution. 2009; 9(4):716-725.


Application of monoclonal antibody DSSC7 fordetecting dengue infection in Aedes aegypti based onimmunocytochemical streptavidin biotin peroxidase

complex assay (ISBPC)

Umniyati SRa#, Sutaryoa, Wahyono Db, Artama WTc, Mardihusodo SJa, Soeyokoa,Mulyaningsih Ba, Utoro Ta

aFaculty of Medicine, Gadjah Mada University, Jalan Farmako, Sekip Utara, Yogyakarta 55281, Indonesia

bFaculty of Pharmacy, Gadjah Mada University, Yogyakarta 55281, Indonesia

cFaculty of Veterinary Medicine, Gadjah Mada University, Yogyakarta 55281, Indonesia

Abstract

Aedes aegypti mosquito is the important vector of dengue fever and chikungunya fever. Therefore, forvirus detection in the mosquito, the possibility of cross-reactivity with chikungunya virus must beconsidered. The laboratory studies were aimed at characterizing the monoclonal antibody DSSC7,and its application for detecting dengue (DENV) antigen on the various organs of orally-infected Ae.aegypti in the paraffin-embedded tissue sections, viz. head squash, abdomen squash, based onimmunocytochemical streptavidin-biotin-peroxidase complex (ISBPC) assay. Determination of theantibody class and subclass was based on antigen-mediated ELISA (enzyme-linked immunosorbentassay). The specificity of monoclonal antibody DSSC7 was determined by Western blotting method,using DENV-1, DENV-2, DENV-3, DENV-4, and chikungunya antigen. The presence of DENV antigen inthe various organs of the orally-infected Ae. aegypti were microscopically optimized in the paraffin-embedded tissue section using ISBPC assay and monoclonal antibody DSSC7 (diluted 1:10, 1:50,1:100 in phosphate buffer saline) as a primary antibody. The specificity of the immunocytochemicalprocedure is validated by negative controls and by positive controls that show that the antibody isbinding to an appropriate structure. The monoclonal antibody DSSC7 recognize DENV complexspecific and does not recognize chikungunya antigen. The monoclonal antibody belongs to IgG class,IgG1 subclass, and it is used as primary antibody to detect DENV infection in Ae. aegypti on tissuesection in experimental infection based on ISBPC method. The infection rates of abdomen squash andhead squash after incubation period of five days were 75% and 33.33% respectively.

Keywords: Aedes aegypti; Antigen; DENV; Chikungunya; Monoclonal antibody DSSC7.

#E-mail: [email protected]; Fax: 062-274-581876


Application of monoclonal antibody DSSC7 for detecting dengue infection in Aedes aegypti

Introduction

The incidence of dengue fever (DF) occursevery year in Indonesia, but the number ofcases were unusually high in at least 12 of the32 provinces of the country during 2004. From1 January to 4 April 2004 a total of 52 013cases, mainly hospitalized cases of DENV and603 deaths, have been registered with theIndonesian Ministry of Health. It was doublethe figure when compared with the same periodin the previous year. Provinces in Java, includingWest Java, Central Java and East Java, wereparticularly severely affected, with more than35% of the cases reported from DKI-Jakarta[1].Aedes aegypti is the important vector of denguehaemorrhagic fever (DHF) as well aschikungunya fever[2]; therefore, detection ofdengue virus in the mosquito is a matter ofgreat concern in view of its cross-reactivity withchikungunya virus.

There are several methods of virusdetection in the mosquito, such as the directfluorescent-antibody (DFA) test on mosquitotissues, usually brain or salivary glands or headsquashes, and reverse transcriptase polymerasechain reaction (RT-PCR). However, the DFAmethod is labour-intensive and requiresfluorescent microscope and cryofreezer. RT-PCR provides a rapid serotype-specific diagnosisfor RNA viruses. The method is rapid, sensitive,simple, and reproducible if properly controlledand can be used to detect viral RNA in humanclinical samples, autopsy tissues, ormosquitoes[3,4,5]. Although immunofluorescencetests were used in the past, newer methodsinvolving enzyme conjugates such as peroxidaseand phosphatase in conjunction with eitherpolyclonal or monoclonal antibodies havegreatly improved[6].

Monoclonal antibodies (mAb) are usedextensively in basic biomedical research, in the

diagnosis of disease, and in the treatment ofillnesses, such as infections and cancer.Antibodies are important tools used and recentresearch works have led to many medicaladvances. Producing mAb requires immunizingan animal, usually a mouse; obtaining immunecells from its spleen; and fusing the cells witha cancer cell (such as cells from a myeloma) tomake them immortal, which means that theywill grow and divide indefinitely. A tumor ofthe fused cells is called a hybridoma, and thesecells secrete mAb. A major advantage of usingmAb rather than polyclonal antiserum is thepotential availability of almost infinite quantitiesof a specific monoclonal antibody directedtoward a single epitope (the part of an antigenmolecule that is responsible for specific antigen-antibody interaction). To produce the desiredmAb, the cells must be grown in either of twoways: by injection into the peritoneal cavity ofa suitably prepared mouse (the in vivo, ormouse ascites, method) or by in vitro tissueculture. The mouse ascites method is generallyfamiliar, well understood, and widely availablein many laboratories; but the mice requirecareful handling to minimize pain or distressinduced by excessive accumulation of fluid inthe abdomen or by invasion of the viscera.When injected into a mouse, the hybridomacells multiply and produce fluid (ascites) in itsabdomen; this fluid contains a highconcentration of antibody[7,8,9].

Monoclonal antibody against DENV-3 wasproduced by the Dengue Team of GadjahMada University through three-time fusion from1993–1995. The first fusion generated 4hybridomas (clones) producer, whereas thesecond fusion generated 13 clones producer;meanwhile, the third fusion generated 22clones producer. Hybridoma cell linesproducing the DENV antibodies were storedin the liquid nitrogen tank in the LaboratoriumHayati, Gadjah Mada University[10].



Among the hybridoma cell lines generatedfrom the third fusion, DSSC7 and DSSF1 cloneswere still growing very well in the tissue culture,after storing it in the liquid nitrogen for severalyears[11]. Once a cellular source of monoclonalantibody has been established in culture, it isusual to obtain a small quantity of ascitic fluidfor further characterization before preparing alarger stock of antibody. A number of tests needto be carried out in order to relate the outcometo the final application and required specificityof the antibody such as isotyping, cross-reactivity and epitope analysis[8,12].

The objective of this study was aimed atproducing, characterizing, optimizing andapplying the monoclonal antibody DSSC7 forthe detection of DENV antigen on the variousorgans of the orally-infected Ae. aegypti in theparaffin-embedded tissue sections, viz. headsquash, abdomen squash, based onimmunocytochemical streptavidin-biotin-peroxidase complex (ISBPC) assay.


Large-scale production ofmonoclonal antibody

Hybrid cells producing antibodies were grownin flask containing complete RPMI medium.The cells were collected by centrifugation,twice washed by phosphate buffered saline(PBS), resuspended in normal saline (106–107

hybrid cells in 0.5 ml normal saline) andinoculated i.p. into prisrane-reared BALB/cmice. After 2–3 weeks, the ascitic fluidproduced by each mouse was collected with asyringe (19 g) or by puncturing the abdomen,then stored for further steps, i.e. antibodypurification, while the hybrids were stored inliquid nitrogen[7,8,12].

Characterization of monoclonalantibody

Determination of isotype andcross-reactivity

Determination of the antibody class and subclasswas carried out based on antigen-mediated ELISA(enzyme linked immunosorbent assay). Thespecificity of monoclonal antibody DSSC7 wasdetermined by Western blotting method, usingDENV-1, DENV-2, and DENV-3, DENV-4, andchikungunya antigen[13]. In a previous study,analysis of the DENV novel anti-denguemonoclonal antibodies (DSSC7 and DSSF1) withdifferent binding specificities for DENV-1, DENV-2, DENV-3, DENV-4 and other flavivirus(Japenese B encephalitis virus) and chikungunyaantigens were carried out based on indirect andinhibition ELISA. The result showed that the mAbDSSC7 showed high immunoreactivity towardDENV-1, DENV-2, DENV-3, DENV-4 and nocross-reactivity toward Japanese encephalitis andchikungunya antigen based on indirect ELISA[10].The viruses were obtained from Naval MedicalResearch Unit 2 (NAMRU-2), Jakarta. Antigenswere prepared as follows: monolayers of C6/36cells were grown to 90% confluence in 75 cm2

flasks, then inoculated with dengue virus, andincubated for 1 hour at 28 °C in an atmosphereof 5% CO2. Flasks were supplemented with 15mL of maintenance medium (minimal essentialmedium, 2% fetal bovine serum [FBS], 1x non-essential amino acids, 100 U/mL of penicillin, and100 µg/mL of streptomycin) and maintained at28 °C in an atmosphere of 5% CO2. Infectionwas monitored daily by an inverted microscopeand cell supernatants were harvested at sevenor eight days post-infection. Maintenance mediumwas changed after 2 to 4 days (depending on thevirus) and the culture supernatants and infectedcells were harvested when cytopathic effect wasapparent throughout the monolayer. The culturesupernatants were clarified by centrifugation for10 minutes at 1000 rpm at 4 °C, and stored in



aliquots at –80 °C until use. The infectedmonolayers were washed with PBS and lysed in2 ml of a hypotonic buffer containing 1% TX-100. Intact nuclei were removed by briefcentrifugation at 14 000 rpm in a micro centrifugeand the lysate supernatants (referred to as“lysates”) were aliquoted and stored at –80 °Cuntil use. Virus stocks were stored as individual 1mL aliquots in 20% FBS at –70 °C.

Immunoblotting

DENV-1, DENV-2, DENV-3, DENV-4 andchikungunya antigens were prepared from lysatesupernatants in eppendorf tubes by mixing 25 µlantigen with gliserol 3 µl, SDS (10%) 3 µl,(bromophenol blue) and xylene cyanol 0.1% 2 µland boiling it for a minute. Antigen were separatedby sodium dodecyl sulfate-polyacrylamide gelelectrophoresis (SDS-PAGE) and transferred tonitrocellulose membranes. Blots were probedwith DSSC7 antibodies from ascites at a 1:100dilution. Bound antibodies were detected withan alkaline phosphatase-conjugated goat anti-mouse IgG antibody, followed by the addition ofsubstrate containing 0.15 M sodium barbital 9ml, BCIP 1mg in DMF 200 µl, NBT 1mg and 40µl M MgCl2. Protein loading and transferefficiency were monitored by Coomassie blueand silver staining and the use of pre-stainedmolecular weight markers, respectively.

Detection of DENV infection ofAe. aegypti based on ISBPC assay

Positive control tissue specimens

Five-day-old male Ae. aegypti mosquitoes wereinjected intrathoracally with DENV-3 strain H87,and female Ae. aegypti were orally infected withthe blood sample of patient. DENV-3 strain H87virus was obtained from NAMRU-2 Jakarta.Blood samples from patients clinically diagnosedas DHF were also used as source of infectionbecause they were positive dengue IgM and

IgG based on the immunochromatography test.Head squash preparation made at incubationperiod of 8 weeks was used as positive controls.The slides were air-dried, wrapped in aluminumfoil, and stored at –80 °C. Before use, the headsquash preparation was fixed in acetone for 20min at 4 °C.

Negative control tissue specimens

Negative controls comprised (1) uninfected Ae.aegypti mosquitoes from non-endemic area ofDHF and Anopheles mosquitoes from Bantuldistrict, Yogyakarta province. The Anophelesmosquitoes were used as negative control tissuebecause they are not the vector of dengue virus.

The infected mosquitoes were held in smallcylindrical cages covered with mosquitonetting, and they were incubated at 27±1 °Cand a relative humidity of 88±6%. Thepresence of dengue antigen on head squashof intrathoracally-infected male Ae. aegyptiwere detected based on ISBPC assay using mAb2D3B10 and commercially mAb as positivecontrols. The mAb 2D3B10 is specific forDENV-3 and it reacts to viral E protein at themolecular weight of 57.9 kDa, whereas thecommercially mAb reacts to DENV-1, DENV-2, DENV-3, and DENV-4. Negative controltissue specimens without primary antibodywere used as negative controls. Positive resultwas detected as discrete brownish granulardeposits throughout most fields having braintissue. Negative result were detected as bluecolour throughout most field having brain tissueand no brownish colour other than the chitinousmosquito tissues and non-specific backgrounddistinctly different from specific positive result[6].

Infection of mosquitoes

Four- to-five-day-old Ae. aegypti females wereorally infected with DENV-3 strain H87 virus.The infected mosquitoes were held in small cages



covered with mosquito netting and incubated at27±1 °C and a relative humidity of 88±6%. Oneday after inoculation, the presence of dengueantigen on the various organs of orally-infectedAe. aegypti mosquitoes were optimized in theparaffin-embedded tissue section based onISBPC assay using monoclonal antibody DSSC7(diluted 1:10, 1:50, 1:100 in phosphate buffersaline) as a primary antibody. The presence ofdengue antigen on the various organs of orally-infected Ae. aegypti mosquitoes were alsoobserved 2 days, 3 days, 4 days and 5 days afterinoculation in the paraffin-embedded tissuesections, head squashes and abdomen squashespreparations based on ISBPC assay usingmonoclonal antibody DSSC7 (1:50 in phosphatebuffer saline) as a primary antibody.

Result

Production of monoclonal antibody

It was possible to produce 50 cc of antibodyagainst DENV secreted by DSSC7 hybridoma inascites of two BALB/c mice. Monoclonalantibodies in ascitic fluid were also secreted byDSSF1. Typical antibody concentrations in ascitesof hybridoma-bearing mice ranged from 2–20mg/l, and they represent a significant fraction ofall protein present. In contrast, the antibody levelsin culture supernatants of hybridomas were ofthe order of 5–50 µg/ml. It is, therefore, obviousthat purification of antibodies from serum orascites will be much easier than from culturesupernatants. If antibodies must be purified fromculture supernatants, affinity chromatography isusually the method of choice[7]. In this study,monoclonal antibody secreted by DSSC7 clonewere not purified. According to Sutaryo et al.[10],monoclonal antibodies secreted by 3E9E12,2D3B10 and 1D10C5 clones were purified byaffinity chromatography on protein A, and themAb concentrations were 2.955 mg/ml, 2.645mg/ml and 2.485 mg/ml respectively.

Characterization of monoclonalantibody

Determination of isotype

A commercially available testing kit isotypic-specific reagent (Sigma ISO-2) based on antigen-mediated ELISA was used to perform isotypeanalysis during this study. In this study,monoclonal antibody secreted by DSSC7 clonebelonged to IgG class, IgG1 subclass.

Protein dengue

Dengue virus is a single-stranded, positive-senseRNA virus with 11 kb unfragmented genomesurrounded by a lipid bilayer envelope. RNA codesfor three structural proteins and seven non-structural proteins. Three structural proteins areC (nucleocapsid), M (membrane-associatedprotein) and E (envelope protein), and the sevennon-structural proteins have been named asNS1,NS2, NS3, NS4, NS5, NS6 and NS7[14]. Theantigen used in this study comprised NS-3 protein(68.9 kDa), E protein (57.9 kDa), and NS-1protein (48.0 kDa) (Figure 1).

Figure 1: DENV-3 viral protein (Yogyakartaisolate) gel (SDS-PAGE) that has been stained

with silver dye showing strong bands atmolecular weight of 68.9 kDa (NS-3

protein), 57.9 kDa (E protein), and a weakband at 48.0 kDa (NS-1 protein)

KDa

205

116

97

84

66

45

36

29

24

68.9 kDa (kilo Dalton)

57.9 kDa (kiloDalton)

48.0 kDa (kilo Dalton)

1. Standard protein

2. Dengue antigen

1 2



Monoclonal antibody secreted by a singlehybridoma (DSSC7) was generated from thethird fusion-recognized dengue complex-specific epitope (DENV-1, DENV-2, DENV-3,DENV-4) at molecular weight of about 48.000Da (48 kDa) based on Western blotting analysis(Figure 2).

Figure 2: Monoclonal antibody DSSC7(1:100) recognized dengue antigen (DENV-1,

DENV-2, DENV-3, DENV-4) epitope atmolecular weight of 48 kDa showing no

cross-reactivity toward chikungunya antigenbased on Western blotting method

116

97

66484529

24

KDa Standard DENV-1 DENV-2 DENV-3 DENV-4 CHIKProtein

According to Henchal and Putnak[14],nonstructural (NS-1) DENV protein hasmolecular weight of 48.000 Dalton. The NS-1protein could be found in the cell, at plasmamembrane, or it is secreted out of the cellduring the infection. The figure also showedthat there was no cross-reactivity between themAb DSSC7 toward chikungunya antigen. Thisfinding supported the previous study that themAb DSSC7 showed high immunoreactivitytoward DENV-1, DENV-2, DENV-3, DENV-4and no cross-reactivity toward Japaneseencephalitis and chikungunya antigen based onindirect ELISA[10]. Meanwhile, the previousstudy showed that mAb secreted by a singlehybridoma (2D3B10), which generated fromthe second fusion, recognized a DENV-3 virus

type-specific determinant (epitope), based onimmune dot blot assay by using DENV-1, DENV-2, DENV-3, and DENV-4 antigen (Yogyakartaisolate)[15].

Detection of DENV infection ofAe. aegypti based on ISBPC assay

Positive control

Dengue antigen was detected as brownishcolour in the cytoplasm of infected cellthroughout most fields having brain tissue ofinfected Ae. aegypti mosquito with DENV-3strain H87 at incubation period of 11 daysunder light microscope based on ISBPC assayusing mAb 2D3B10 (1:50) as primary antibody.The previous study showed that the mAb2D3B10 recognized DENV-3 viral E protein.Antigen was also detected as discrete brownishgranular deposits throughout most fields havingbrain tissue of orally-infected Ae. aegyptimosquito with DENV-3 strain H87 at incubationperiod of 8 days under light microscope basedon ISBPC assay using commercial mAb asprimary antibody. According to themanufacturer (Chemicon Laboratory), the mAbreacts to DENV-1, DENV-2, DENV-3, andDENV-4. Both brownish colour in thecytoplasm of infected cells and discretebrownish granular deposits throughout mostfields having brain tissue of orally-infected Ae.aegypti mosquito with DENV-3 strain H87 atincubation period of 11 days were shown underlight microscope based on ISBPC assay usingmAb DSSC7 (1:50) as primary antibody. Thenegative result was shown on head squashesof uninfected Ae. aegypti from non-endemicarea of DHF and head squashes of Anophelesmosquitoes as blue colour throughout mostfields having brain tissue and no brownish colourother than the chitinous mosquito tissues andnon-specific background distinctly differentfrom specific positive result (Figure 3).



Figure 3: Head squashes immunocytochemical preparation of orally-infected Ae. aegypti withDENV-3 strain H87 at incubation period of 11 days showing positive reaction as brownishcoloration in the cytoplasm of infected cells with mAb 2D3B10 (A), and showing positive

reaction as discrete brownish granular deposits throughout most fields having brain tissue withcommercial mAb as positive controls (B). Positive reaction was also shown as brownishcoloration in the cytoplasm and discrete brownish granular deposits with mAb DSSC7 asprimary antibody (C) and negative reaction was shown on head squashes of uninfected

Ae. aegypti preparation as blue coloration (D).

A B

D

250 µM

C

100 µM

The result indicated that DENV antigen wasdetected on head squash preparation of Ae.aegypti infected with DENV-3 at incubationperiod of 11 days based on ISBPC assay usingmAb DSSC7 at concentration of (1:50) asprimary antibody.

The result indicated that DENV antigen wasalso detected on paraffin-embedded sectionof Ae. aegypti orally infected with DENV-3 atincubation period of 1 day based on ISBPCusing mAb DSSC7 at concentration of (1:10)and (1:50) and mAb 2D3B10 (1:50) as primaryantibodies (Table).



Table: Result of microscopic examination of dengue antigen on paraffin-embedded section ofAe. aegypti orally infected with DENV-3 at incubation period of 1 day based on ISBPC assay

using mAb DSSC7 1:10, DSSC7 1:50, DSSC7 1:100 from ascitic fluid and purifiedmAb 2D3B10 1:50

DSSC7 1:10 12 0 12DSSC7 1:50 12 0 12DSSC7 1:100 0 8 8Negative control 1 14 15Positive control 12 0 12Total 37 32 69

Result of immunocytochemistryPositive Negative Total

Monoclonal antibody

The table indicates that there was nosignificant difference between the positive rateof mosquito preparation using mAb DSSC7 1:10,DSSC7 1:50, and mAb 2D3B10 1:50 as primaryantibody based on Fisher’s Exact Test (P=1;>0.05). The result also revealed that there wasno significant difference between the negativecontrol and mosquito preparation using mAbDSSC7 (1:100) as primary antibody (P=0.62;>0. 05).

The monoclonal antibody DSSC7 (diluted1:50 in PBS) was optimum to be used asprimary antibody to detect DENV antigen invarious organs of orally-infected Ae. aegypti inthe tissue section preparation based on ISBPCassay under light microscope (Figures 4–7).

The DENV viral antigens were detectedin the mid gut in the cytoplasm cells atincubation period of 1–3 days, in thehaemocytes at incubation period of 1–2 days,in the fat (1–2 days) (Figure 4), in the brain atincubation period of 2 days (Figure 5), in thesalivary glands at incubation period of 2 days(Figure 6), in the ovaries at incubation periodof 1–2 days (Figure 7).

The DENV viral antigen was also detectedon head squash and abdomen squashpreparation (Figure 8).

In the head squash preparation, positiveresult was detected as discrete brownishgranular deposits throughout most fields havingbrain tissue. DENV viral antigen wereimmunolocalized to the cytoplasm of >100cells per field at 100x magnification in highinfection, 10–100 cells per 100x field inmoderate infection. Low infection cannot beseen at 100x magnification, and the infectioncan be seen at 400x magnification. Meanwhile,negative result was detected as blue colourthroughout most fields having brain tissue andno brownish color other than the chitinousmosquito tissues and non-specific backgrounddistinctly different from specific positive result.The result also exhibited that DENV viralantigens were detected in oocytes at incubationperiod of 4 days and 5 days and it also indicatedthat the infection rate of abdomen squash andhead squash at incubation period of five dayswere 75% and 33.33% respectively.



Figure 4: Tissue section preparation of orally-infected Ae. aegypti with DENV-3 at differenceincubation period showing dengue viral antigen (brownish colour) in various organs of orally-infected Ae. aegypti in the mid gut in the cytoplasm cells at incubation period of 1 day (B),

3 day (C), haemocytes at incubation period of 1 day (E), and fat at incubation period of1 day (H) 2 day (I) based on ISBPC assay using monoclonal antibody DSSC7 (1:50).

Blue colour is shown in the cytoplasm cells of mid gut (A), hemocyte (B), and fat (G)as negative control without primary antibody

A B C

D E F

G H I

250 µM

250 µM

250 µM

100 µM

100 µM

100 µM



Figure 5: Tissue section preparation of orally-infected Ae. aegypti with DENV-3 at differentincubation periods showing dengue viral antigen in the brain –deutocerebrum (A),

protocerebrum, tritocerebrum ang facet eye (C) based on ISBPC assay usingmonoclonal antibody DSSC7 (1:50)

Facet eye

Protocerebrum

Protocerebrum

Facet eye

Tritocerebrum

Tritocerebrum

A B

C D

Deutocerebrum Deutocerebrum

100 µm100 µm

250 µm

250 µm



Figure 6: Tissue section preparation of orally-infected Ae. aegypti with DENV-3 at incubationperiod of 2 days showing dengue viral antigen in the cytoplasm of salivary gland based on

ISBPC assay using monoclonal antibody DSSC7 (1:50)

250 µm

250 µm

Figure 7: Tissue section preparation of orally-infected Ae. aegypti with DENV-3 showing dengueviral antigen in the ovary at incubation period of 1 day (B) and at incubation period of 2 days

(C, D) based on ISBPC assay using monoclonal antibody DSSC7 (1:50)

A B

C D

100 µm



Figure 8: Head squash and abdomen squash of orally-infected Ae. aegypti with DENV-3at incubation period of 5 days showing dengue viral antigen in the brain (A), and eggs in the

ovary (D,F) based on ISBPC assay using monoclonal antibody DSSC7 (1:50)

A B C

D E F

Discussion

Hybrid cells (DSSC7) producing antibodieswere grown in flask containing complete RPMImedium. The hybrid cells were inoculated intraperitoneal into pristane-treated BALB/c mice.After 2–3 weeks, the ascitic fluid produced byeach mouse was collected with a syringe or bypuncturing the abdomen, and then stored forfurther steps, while the hybrids were stored inliquid nitrogen. It has produced 50 cc ofantibody against DENV secreted by DSSC7hybridoma in ascites of two BALB/c mice.Monoclonal antibody secreted by DSSC7 clonebelongs to IgG class, IgG1 subclass. Animportant early characterization test of anypanel of antibodies is the analysis of whetherthey react with the same, close or totally

different epitopes. Monoclonal antibodysecreted by a single hybridoma (DSSC7) whichgenerated from the third fusion recognized asDENV complex specific epitope and showingno cross-reactivity toward CHIK antigen basedon Western blotting analyses. The mAb DSSC7reacts to non-structural protein (NS1).

The viral NS1 protein circulates in the seraof infected patients throughout the clinical phaseof the disease. Novel diagnostic tests based onNS1 detection have been recently developedand marketed. During in vitro infection, theflavivirus NS1 protein is expressed as anintracellular membrane-associated form essentialfor viral replication[16,17]. In solution, secreted NS1protein behaves as a hexamer; it circulates andaccumulates in the sera of dengue virus-infectedpatients throughout the clinical phase of the



disease[18,19,20]. A recent study demonstrated thatsoluble NS1 protein binds to endothelial cellsand, following recognition by anti-NS1antibodies, could contribute to plasma leakageduring severe dengue virus infection[21]. Thedetection of secreted NS1 protein represents anew approach to the diagnosis of acute dengueinfection. A recently developed, commerciallyavailable diagnostic test based on dengue NS1antigen-capture ELISA (Platelia Dengue NS1 Agtest, Bio-Rad Laboratories, Marnes la Coquette,France), was investigated in two studies (one inSouth America and the other in South-East Asia);the test had an overall sensitivity of 88.7% and93.4% in the two studies, with 100%specificity[22,23].

The monoclonal antibody DSSC7 was ableto be used as primer antibody to detect DENVinfection in Ae. aegypti on tissue section, headsquash, abdomen squash preparation of orally-infected Ae. aegypti with DENV-3 at differentincubation periods based on ISBPC assay.Therefore, it will be suitable for detecting theDENV infection in Aedes.

Immunocytochemistry is a powerfulmethod for the identification of proteins orantigen in cells and tissues. However, thismethod is dependent on the specificity of theantibody binding to the epitope of the proteinused as an immunogen. The specificity of theresult depends on two independent criteria:the specificity of the antibody and of themethod used. The antibody specificity is bestdetermined by immunoblot and/orimmunoprecipitation. The specificity of themethod is best determined by both a negativecontrol, replacing the primary antibody withnon-immune serum, and a positive control,using the antibody with cell, known to containthe protein or antigen[24].

This study followed the principle of ISBPCtechniques to demonstrate antigen in cells ortissue as follows (Figure 9).

Figure 9: The principle ofimmunohistochemistry (immunoperoxidaseSBC) techniques to demonstrate antigen in

cells or tissue

(1) The endogenous peroxidase activitymay be destroyed by treating aspecimen with hydrogen peroxidesolution.

(2) The non-specific background iseliminated by incubating the specimenwith non-immune serum.

(3) The primary antibody to specificantigen is incubated to target antigens.This is followed by addition ofbiotinilated second antibody whichserves as the linker between theprimary antibody and peroxidase-streptavidin conjugate.

(4) Streptavidin-peroxidase is then addedto bind to the biotin residues on thelinking antibody.

The presence of enzyme can be revealedby addition of a mixture of substrate-chromogen solution. The enzyme peroxidasewill catalyse the substrate, hydrogen peroxide,and convert the chromogen to a browncoloured deposit demonstrating the locationof the antigen.

B A B P

B

A

B

P

B

P B B P

Biotinylatedsecondary

antibody (30 mins)Primary antibody

(1 hour)

Antigen

StreptavidinHRP Complex

(30 mins)

ColourPrecipitate

DAB

H O22

+

B

B

Conventional ABC Technique



This finding indicated that the transovarialinfection of DENV in Ae. aegypti could beobserved both in the tissue section andabdomen squash preparation of the mosquitobased on immunocytochemical assay usingmAb DSSC7 as primary antibody. Other study(Umniyati, unpublished data) indicated thatnatural transovarial infection of DENV virus inAe. aegypti were identified in Gondokusumansub-district, Yogyakarta Municipality, postoutbreak of DHF in 2004, based on theidentification of DENV antigen in the brain onhead squash of mosquito reared from larvaeand pupae collected from domestic householdwells and water containers for bathing (bakmandi) without blood feeding, using mAbDSSC7. Based on these reports, naturaltransovarial transmission in the domestichousehold wells and bak mandi have greatepidemiological significance and may play animportant role in the maintenance of virus innature, and may act as reservoirs of theseviruses. Therefore, vector surveillance andcontrol activities in domestic household wellsas part of an active community DENV controlstrategy should be performed.

Several studies suggest the existence oftransovarial DENV transmission in Aedesinfected female mosquitoes, allowingpropagation of virus to their progeny. Such aprocess would allow it to act as a reservoir forvirus maintenance during interepidemicperiods. The transovarial transmission rate ofDENV was found to be seven times higher inthe high susceptible isofemale lines than in thelow susceptible lines. The rate of transovarialtransmission initially increased in the initial twogenerations (F1-F2), but in further generations

it was steady. It was also reported that a highertransovarial transmission rate in the progenywas obtained from the longer desiccatedeggs[25].

Conclusion

Monoclonal antibody DSSC7 recognized DENVcomplex specific epitope at molecular weightof 48 kDa (NS1) protein and showed no cross-reactivity toward chikungunya antigen. Themonoclonal antibody belongs to IgG class, IgG1subclass, and it is able to be used as primaryantibody to detect DENV infection in Ae. aegyption tissue section, head squash and abdomensquash preparation in experimental infectionbased on ISBPC method.

Suggestion

Monoclonal antibody DSSC7 could be appliedas primary antibody to investigate the naturaltransovarial infection of DENV in Ae. aegypti,for developing vector surveillance and earlywarning system to anticipate a DHF epidemic.

Acknowledgement

The authors gratefully thank the Dean of Facultyof Medicine, and the Head of Department ofParasitology, the Head of Department ofPathology – Anatomy, Gadjah Mada University,for their support in the conduct of thesestudies. Thanks are also due to Suprihatin,Purwono, T.M. Joko, Agustin and Yunadir fortheir valuable assistance in the laboratory.

References[1] World Health Organization. Dengue fever in

Indonesia – update. Communicable DiseaseSurveillance & Response. Geneva: WHO,2004. (http://www.who.int/csr/don/2004_05_11a/en/- accessed 16 June 2009).

[2] Kusriastuti R. Epidemiologi Penyakit DemamBerdarah Dengue dan KebijaksanaanPenanggulanganmnya di Indonesia. TheProceedings of the Symposium on DengueControl up date, Yogyakarta. 2005.



[3] Deubel V.The contribution of moleculartechniques to the diagnosis of dengue infection.In: Gubler DJ, Kuno G (ed.), Dengue anddengue hemorrhagic fever. London: CABInternational, 997. pp. 335-366.

[4] Guzman MG and Kouri G. Advances indengue diagnosis. CDLI. 1996; 3: 621-627.

[5] Lanciotti RS, Calisher CH, Gubler DJ, ChangGJ and Vorndam AV. Rapid detection andtyping of dengue viruses from clinical samplesusing reverse transcriptase chain reaction. J CliMicrobiol. 1992; 30:545-551.

[6] Anonymous. Histology andImmunocytochemistry. Available at website.(URL www.hmds.org.uk/histology.html). 2005.

[7] Goding JW. Monoclonal antibodies principleand practice. London: Academic press, Inc.,1983.

[8] Artama WT. Pedoman kuliah antibodimonoklonal, teori, produksi, karakterisasi danpenerapan. Yogyakarta: PAU-BioteknologiUGM, 1992.

[9] Lidder and Cryer A. Practical guide tomonoclonal antibodies. Singapore: John WillySons, 1991.

[10] Sutaryo, Umniyati SR, Wahyono D. Produksiantibodi monoklonal terhadap virus dengue-3untuk deteksi penderita DBD dan vektornya.Laporan Penelitian RUT-3 Tahun I, FK UGM,Yogyakarta. 2005.

[11] Umniyati SR, Soeyoko, Mulyaningsih, B.Pengembangan antibodi monoklonal antidengue-3 produksi lokal Universitas GadjahMada untuk deteksi infeksi virus dengue padanyamuk Aedes spp. Laporan Penelitian HibahBersaing X/1, Universitas Gadjah Mada,Yogyakarta. 2002.

[12] Wahyono D, Pichazyk M, Mourton C, MarieBastide M, Bernard PAU. Novel anti-digoxinmonoclonal antibodies with different bindingspecificities for digoxin metabolites and otherglycosides. Hybrid. 1990; 9: 619-30.

[13] Deubel V, Despress P. Molecular Biology ofDengue Virus. The Proceeding of theWorkshop on Molecular Biology of DengueVirus, Center for Trop. Med., Gadjah MadaUniversity, Yogyakarta. 1997.

[14] Henchal EA and Putnak JR. The dengueviruses. Clin Microbiol Rev. 1990; 3(4):376–396

[15] Sutaryo, Artama, Umniyati SR, Wahyono D.Produksi dan karakterisasi antigen dengueisolat Yogyakarta dengan antibodi monoklonalanti dengue produksi UGM. Laporan, BlockGrant, Pusat Kedokteran Tropis UGM,Yogyakarta. 2002.

[16] Lindenbach BD, Rice CM. Trans-complementation of yellow fever virus NS1reveals a role in early RNA replication. J Virol.71. 1997; 9608–9617

[17] Mackenzie JM, Jones MK, Young PR.Immunolocalization of the dengue virusnonstructural glycoprotein NS1 suggests a rolein viral RNA replication. Virology. 1996; 220:232–240.

[18] Young PR, Hilditch PA, Bletchly C, HalloranW. An antigen capture enzyme-linkedimmunosorbent assay reveals high levels ofdengue virus protein NS1 in the sera of infectedpatients. J Clin Microbiol. 2000; 38:1053–1057.

[19] Libraty DH, Young PR, Pickering D, Endy TP,Kalayanarooj S. .High circulating levels of thedengue virus nonstructural protein NS1 earlyin dengue illness correlate with thedevelopment of dengue hemorrhagic fever. JInfect Dis. 2002. 186: 1165–1168.

[20] Alcon S, Talarmin A, Debruyne M, FalconarA, Deubel V. Enzyme-linked immunosorbentassay specific to Dengue virus type 1nonstructural protein NS1 reveals circulationof the antigen in the blood during the acutephase of disease in patients experiencingprimary or secondary infections. J ClinMicrobiol . 2002; 40: 376–381



[21] Kumarasamy V, Wahab AH, Chua SK, HassanZ, Chem YK. Evaluation of a commercialdengue NS1 antigen-capture ELISA forlaboratory diagnosis of acute dengue virusinfection. J Virol Methods. 2007;140: 75–79.

[22] Avirutnan P, Zhang L, Punyadee N,Manuyakorn A, Puttikhunt C. Secreted NS1of dengue virus attaches to the surface of cellsvia interactions with heparan sulfate andchondroitin sulfate. E. PLoS Pathog. 2007 Nov;3(11): e183.

[23] Dussart P, Labeau B, Lagathu G, Louis P, NunesMR. Evaluation of an enzyme immunoassay

for detection of dengue virus NS1 antigen inhuman serum. Clin Vaccine Immunol. 2006;13: 1185–1189.

[24] Burry RW. Specificity Controls forImmunocytochemical Methods. Journal ofHistochem and Cytochem. 2000; 48:163-166.

[25] Joshi V, Mourya DT, Sharma RC. Persistence ofdengue-3 virus through transovarialtransmission passage in successive generationsof Aedes aegypti mosquitoes. Am J Trop MedHyg. 2002; 67: 158–161.


Enhancement of MHC class I binding and immunogenicproperties of the CTL epitope peptides derived from

dengue virus NS3 protein byanchor residue replacement

Hideyuki Masakia#, Yoshiki Fujiic,d, Kiyohiro Irimajirib, Takanori T. Tomuraa,Ichiro Kuranec,d

aDepartment of Biochemistry, Kinki University School of Medicine, Osaka-Sayama, Osaka 589-8511, Japan

bKinki University Life Science Research Institute, Osaka-Sayama, Osaka, Japan

cDepartment of Virology 1, National Institute of Infectious Diseases, Shinjuku, Tokyo, Japan

dDepartment of Infectious Biology, Institute of Basic Medical Sciences, University of Tsukuba, Tsukuba,Ibaraki, Japan

Abstract

The immunogenecity of the defined H-2Kd-restricted, murine cytotoxic T lymphocyte (CTL) epitopesof dengue viruses were examined for CTL induction in epitope peptide / H-2Kd tetramer assays. Thepeptides used in the study included those corresponding to amino acid (a.a.) residues 298-306(GYISTRVEM) of NS3 of dengue virus types 2 and 4 (named DENV-2/4), and to a.a. residues 299-307(GYISTRVGM) of NS3 of dengue virus types 1 and 3 (named DENV-1/3), and their respective modifiedepitope peptides, DENV-2/4-9L (GYISTRVEL) and DENV-1/3-9L (GYISTRVGL), in which the C-terminalresidue M of the original epitope peptide was replaced by L, in order to provide the complete H-2Kd-binding motif. Immunization of BALB/c mice with the original epitope peptide, DENV-2/4 or DENV-1/3, did not induce specific CTLs, while that with the modified epitope peptide, DENV-2/4-9L or DENV-1/3-9L, induced epitope peptide/H-2Kd tetramer-binding CD8+ cells indicating specific CTLs.Competition-based binding assay with biotinylated epitope-related reference peptides (DENV-2/4-9L-Biotin and DENV-1/3-9L-Biotin) demonstrated that the modified epitope peptide, DENV-2/4-9L andDENV-1/3-9L, had higher avidity to H-2Kd than the respective original epitope peptides. These resultsindicate that modification of dengue virus-derived CTL epitope peptide by replacing a.a. residue at theposition of anchor residue increases the binding avidity to MHC class I, resulting in the induction ofspecific CTLs. The strategy to enhance the immunogenicity of CTL epitope peptide may contribute toinvestigation of CTL biology in dengue virus infection.

Keywords: Dengue virus; CTL epitope; Binding motif; Anchor residue; MHC class I; Affinity; Immunogenicity.

#E-mail: [email protected]; Tel: +81-72-366-0221 Ext.3152; Fax: +81-72-366-0206


Avidity and immunogenecity of dengue virus CTL epitope

Introduction

MHC class I – restricted, CD8+ cytotoxic Tlymphocytes (CTLs) play an important role inthe elimination of virus-infected and tumor cellsby antigen-specific lysis[1,2,3]. They recognizespecific structures on the surface of target cellsas their antigens, which are composed of selfMHC class I molecules and the peptides of 8to 11 a.a. in length. The peptides are derivedfrom the endogenous protein, fitting to thegroove of the MHC class I molecule[4,5,6].Recently, there has been a great deal of interestin the CTL epitope-based immunomodulationtherapy, including peptide vaccines, mainly forthe treatment of malignancies[7,8,9,10,11,12].

Dengue viruses, of which there are fourserotypes, cause dengue fever and denguehaemorrhagic fever/dengue shock syndrome(DHF/DSS), the severe manifestation ofinfection, which is often fatal[13]. They are of agreat global health importance, particularly inthe tropical regions, causing up to 100 millioninfections including a couple of thousands ofdeaths each year[14]. Protective immunityagainst the same serotype virus is life-long,while re-infection with a different serotype canoccur. The secondary infections are oftencomplicated with DHF/DSS, suggesting thatpre-existing immunity to a different serotypevirus may contribute to the pathogenesis ofDHF/DSS[15]. One of the strategies for theprevention of dengue virus infection isvaccination. However, a vaccine has not beendeveloped yet. Immunization with denguevaccine may have the potential risk of inducingDHF/DSS manifestation. In this context,peptide vaccine based on CTL epitopes thatbinds to MHC class I molecule is thought tobe a candidate, because it is anticipated thatthis vaccine induces the least cross-reaction dueto its minimal component. Furthermore,immune response elicited by immunizationwith a single epitope peptide is thought to be

much simpler than those elicited by virusinfection, which evokes multiple immuneresponses against various epitopes on viruses.Establishment of a strategy by immunizationwith a single dengue virus-derived epitopepeptide, thus, is anticipated to facilitatedissection of immunobiology of dengue virusinfection. This strategy is expected tocontribute to investigation of theimmunopathogenesis (DHF/DSS may beinvolved).

Rothman et al. [16] elucidated CTLresponses to an immunodominant epitope onthe dengue virus NS3 protein in BALB/c miceafter primary infection. They mapped theminimal CTL epitopes consisting of nine aminoacids. By using CTL clones, they defined theH-2Kd-restricted CTL epitopes, whichcorresponded to the amino acid (a.a.) residues298-306 (GYISTRVEM) of NS3 of dengue virustypes 2 and 4, or a.a. residues 299-307(GYISTRVGM) of NS3 of dengue virus types 1and 3[16,17]. Immunodominant epitopes forhuman CD8+ CTLs have been also defined ondengue virus NS3 protein[18,19,20,21] suggestingthat NS3 is the main target for the CTL responsein humans as well.

Previously, by using cytotoxicity assay, wehave demonstrated that immunization with themodified epitope peptide, DENV-1/3-9L(GYISTRVGL) or DENV-2/4-9L (GYISTRVEL), inwhich the original epitope peptide C-terminalresidue M was replaced by residue L to providea complete H-2Kd- binding motif[22,23], inducedspecific CTLs with little affection to antigenspecificity, while immunization with originalone, DENV-1/3 or DENV-2/4, whichcorresponded to the defined CTL epitopespanning a.a. residues 299-307 (GYISTRVGM)of NS3 of dengue virus types 1 and 3 or whichcorresponded to that spanning a.a. residues298-306 (GYISTRVEM) of NS3 of dengue virustypes 2 and 4, respectively, did not[24].



In the present study, we analysedimmunogenic properties of the modified epitopepeptides for CTL induction more in detail. Wefirst examined whether immunization with themodified epitope peptide induces epitopepeptide/H-2Kd tetramer-binding CD8+ cells, thespecific CTLs more significantly than that withthe original peptide, as was observed incytotoxicity assays. We also analysed the avidityof the original epitope peptides and modifiedpeptides to H-2Kd molecule by competition-based binding assay, using the biotinylatedepitope peptide-related reference peptides. Wedemonstrated that the modification of theoriginal epitope peptides by substitution of theC-terminal a.a. residue increased the bindingavidity to H-2Kd, and that this modificationenhanced the immunogenicity of the epitopepeptides in CTL induction.


Mice: Female BALB/cAJcl mice were purchasedfrom Clea Japan (Tokyo, Japan), and maintainedin the Animal Facility of Kinki University School

of Medicine under the conventional condition.Mice were used at the ages of 6 to 12 weeks.

Cells: Murine mastcytoma line P815 (H-2d), fibroblast cell line L929 (H-2k), and cellline L-Kd-172 (kindly provided by Dr Jack R.Bennink, NIAID, NIH), which is H-2Kd- genetransfectant cell line derived from L929, weremaintained in RPMI 1640 medium (Sigma, St.Louis, MO) with 5x10-5 M 2-mercaptoethanol(2-ME), 100 U penicillin, 100 mg/mlstreptomycin, 10 mM HEPES, and 10% heat-inactivated fetal calf serum at 37 °C in 5% CO2.

Peptides: The sequences and derivationof peptides DENV-2/4 (GYISTRVEM), DENV-2/4-9L (GYISTRVEL), DENV-2/4-9L-Biotin(GYISTRVELGEAC-Biotin), DENV-1/3(GYISTRVGM), DENV-1/3-9L (GYISTRVGL),and DENV-1/3-9L-Biotin (GYISTRVGLGEAC-Biotin) are shown in the Table. They weresynthesized with 9-fluorenylmethoxycarbonylchemistry by Sigma Genosis Japan (Ishikari,Hokkaido, Japan). Peptides were purified byreverse phase HPLC in the conditions of 5%to 80% gradient elution with acetonitrile in 0.1%

Table: Synthetic peptides used in the study

Peptide Sequence Virus derivationDENV-2/4 GYISTRVEM Dengue virus types2/4 NS3 298-306DENV-2/4-9L GYISTRVEL *1DENV-2/4-9L-Biotin GYISTRVELGEAC-Biotin *2DENV-1/3 GYISTRVGM Dengue virus types1/3 NS3 299-307DENV-1/3-9L GYISTRVGL *3DENV-1/3-9L-Biotin GYISTRVGLGEAC-Biotin

Peptide sequence is expressed by Dayhoff’s one-letter code of amino acid except “-Biotin”.

Note that the difference between DENV-2/4 and DENV-1/3 is only the residue at position 8, that is “E”in DENV-2/4 or “G” in DENV-1/3.*1 The sequence corresponds to the residues NS3 298-306 of dengue virus types 2 and 4 except the residue ofC-terminus substituted for “L”, and to the residues NS3 299-307 of kunjin virus.*2 The sequence corresponds to the residues NS3 298-310 of dengue virus type 4 except the substitutedresidue “L” and the biotinylated residue “C”.*3 The sequence corresponds to the residues NS3 299-307 of dengue virus types 1 and 3 except the residue ofC-terminus substituted for “L”.



trifluoroacetic acid using TSKgel ODS-80Ts QAcolumn (Toso, Tokyo, Japan). The purity ofpeptides was determined to be more than95.0%, and mass-spectrometry (AppliedBiosystems Voyager System 1162) analysisproved every peptide molecular weight to bethe anticipated one.

Immunization and CTL induction:Immunization and CTL induction were carriedout as described before. Briefly, mice wereimmunized by subcutaneous injection withpeptide DENV-2/4-9L or peptide DENV-1/3-9Lwith complete Freund adjuvant (CFA). Then,more than three weeks later, lymph node cellsprepared from the draining lymph nodes wereco-cultured with irradiated syngeneic spleencells pulsed with the same peptide in EHAAmedium (Sigma), supplemented with 100 µg/ml nucleic acid precursors, 2mM L-glutamine,5x10-5M 2-ME, 100U penicillin, 100 µg/mlstreptomycin, 10mM HEPES, and 10% fetal calfserum (FCS) at 37 °C in 5% CO2. On day 4,half volume of the medium was replaced witha fresh one, and 10 U recombinant mouse IL-2 was added. On day 7, viable cells wereharvested and used as CTL effector cells.

Detection of MHC tetramer-bindingcells: The phycoerythrin (PE)-labelled MHCtetramer composed of four monomericcomplexes which consisted of the modifiedH-2Kd heavy chain, ß2-microglobulin, andpeptide DENV-2/4-9L (named DENV-2/4-9L-tetramer) or peptide DENV-1/3-9L (namedDENV-1/3-9L-tetramer) were prepared by MBLCo., Ltd. (Nagoya, Aichi, Japan). The CTLeffector cells (1x106) suspended in 100 µl ofphosphate buffered saline (PBS) containing0.02% NaN3 (PBS/NaN3) were incubated with5 µl of FITC-conjugated anti-mouse CD8antibody (Immunotech, Marseille, France) and5 µl of DENV-2/4-9L-tetramer or DENV-1/3-9L-tetramer at room temperature for 30 minutes.Cells were washed three times with PBS/NaN3

at 4 °C, then fixed with 1 ml of PBS containing1% paraformaldehyde, and analysed by a FACSCalibur (Becton Dickinson, San Jose, CA) andCELL QuestTM version 3.3 software.

Analysis of the avidity between MHCmolecule and peptide: Peptide bindingcompetition assay was carried out to determinethe relative avidity of the original epitopepeptides (DENV-2/4 and DENV-1/3) and thatof the modified epitope peptides (DENV-2/4-9L and DENV-1/3-9L) to H-2Kd molecule, usingH-2Kd-gene transfectant L-Kd-172 cells and thebiotinylated, epitope-related reference peptide.Briefly, 1x106 of L-Kd-172 cells suspended incomplete medium were incubated with variousconcentration of non-biotinylated, original ormodified epitope peptide in the presence of 1µM of biotynilated, epitope-related referencepeptide (DENV-2/4-9L-Biotin or DENV-1/3-9L-Biotin, respectively) at 37 °C for two hours.Cells were washed three times with PBS/NaN3at 4 °C, and then incubated with 0.5 µg ofstreptoavidin-conjugated Cy-ChromeTM (SA-CyC) (BD PharMingen) at 4 °C for 30 minutes.Cells were washed three times, fixed with 1ml of PBS containing 1% paraformaldehyde,and subjected to FACS analysis.

Geometric mean fluorescence intensity(MFI) was measured, and we defined delta MFI(ΔMFI) by subtracting the background MFI (i.e.MFI in the case stained with SA-CyC only). Percent fluorescence intensities were calculatedby the formula: % fluorescence intensity = 100x (ΔMFI with the non-biotinylated epitopepeptide/ΔMFI without the non-biotinylatedepitope peptide), and plotted against non-biotinylated epitope peptide concentrations inlogarithmic scale. Concentration of the non-biotinylated epitope peptide that correspondto the per cent fluorescence intensity fifty (IC50:50% inhibitory concentration) was obtained bythe chart, and used as an index of the relativeavidity of each peptide to H-2 Kd molecule.




Because of incomplete set of anchor residues(only one anchor residue Y at position 2) inthe original epitope peptide for preparation ofpeptide/H-2Kd tetramers, and because of lessaffection to the specificity recognized by CTLswith replacement of C-terminal residue M byL[24], we prepared phycoerythrin-labelled DENV-1/3-9L/H-2Kd and DENV-2/4-9L/H-2Kd

tetramers. We then examined whether themodified epitope peptides, DENV-2/4-9L andDENV-1/3-9L, induce DENV-2/4-9L/H-2Kd

tetramer-binding cells and DENV-1/3-9L/H-2Kd

tetramer-binding cells, respectively, moreefficiently than the original peptides, DENV-2/4 and DENV-1/3, as was observed incytotoxicity assays.

Immunization with the modified peptideDENV-2/4-9L followed by in vitro stimulationwith same peptide-pulsed APC induced 5.59%of CD8-positive DENV-2/4-9L/H-2Kd tetramer-binding cells, while that with the originalpeptide DENV-2/4 did 1.25% (Fig . 1A).Similarly, immunization with the modifiedpeptide DENV-1/3-9L followed by in vitrostimulation with same peptide-pulsed APCinduced 5.81% of CD8-positive DENV-1/3-9L/H-2Kd tetramer-binding cells, while that withthe original peptide DENV-1/3 did 0.34% (Fig.1B). In addition, immunization with PBSemulsified with CFA followed by in vitrostimulation with non-pulsed spleen cells (i.e.,pulsed with PBS only) induced 0.15% of CD8-positive DENV-1/3-9L/H-2Kd tetramer-bindingcells (data not shown). The results indicate thatimmunization and in vitro stimulation with themodified epitope peptides, which possesseda complete binding motif to H-2Kd (i.e. Y atposition 2 and hydrophobic L or I at C-terminusof 9-mer peptide), efficiently induced CD8-positive, epitope peptide/H-2Kd tetramer-binding cells implying CTLs, as was observedin cytotoxicity assays.

It has been reported that the affinity of apeptide for MHC binding is an importantparameter determining the immunogenicity ofan MHC-presented epitope peptide[25,26].Indeed, only peptides derived from tumour-associated antigens, hepatitis B virus or influenzaA virus with a high binding affinity for MHC classI molecules, have been demonstrated to beimmunogenic enough for inducing CTLresponse[7,9,11,26,27,28,29,30]. However, so far, therehad been no report regarding the correlation ofbinding affinity to MHC class I molecules andimmunogenicity of dengue virus-derived CTLepitope peptides for CTL induction. Thus, weassessed the avidity of the modified epitopepeptides, DENV-2/4-9L and DENV-1/3-9L, to H-2Kd molecules in comparison with the originalepitope peptides, DENV-2/4 and DENV-1/3,respectively. Rothman et al.[16] demonstrated thatelongation of the epitope peptide at the C-terminus did not affect the specific lysis of thetarget cells[17]. We, thus, prepared thebiotinylated epitope peptides of 13-merelongated at the C-terminal side with 4 a.a.residues spanning NS3 307-310 of dengue virustypes 2 and 4, except for C-terminal residue Areplaced with C to conjugate biotin, DENV-2/4-9L-Biotin (GYISTRVELGEAC-Biotin) and DENV-1/3-9L-Biotin (GYISTRVGLGEAC-Biotin) (Table),and examined whether they bound to H-2Kd

molecules. L-Kd-172 cells, which are H-2Kd-genetransfectant cells derived from cell line L929,gained increased fluorescence intensity in dose-dependent manner after incubation with variousconcentration of the biotinylated-modifiedepitope peptide (Fig. 2). In contrast, no specificstaining with SA-CyC was observed in L929 cellsincubated with the biotinylated epitope peptide.The results indicate that the biotinylated epitopepeptides, DENV-2/4-9L-Biotin (GYISTRVELGEAC-Biotin) and DENV-1/3-9L-Biotin(GYISTRVGLGEAC-Biotin), bound to H-2Kd

molecule specifically, suggesting that they couldbe used as epitope-related reference peptidesfor binding competition.



Figure 1: Induction of the tetramer-binding cells by immunization with the modified epitopepeptides

[Source of the CTLs was pooled lymph node cells (5 mice each).(A) CD8-positive, DENV-2/4-9L/H-2Kd-tetramer-binding cells (right upper quadrant) accounted

for 5.59% of the analysed cells after immunization and in vitro stimulation with modifiedepitope peptide DENV-2/4-9L, while 1.25% after those with the original epitope peptide

DENV-2/4.(B) CD8-positive, DENV-1/3-9L/H-2Kd-tetramer-binding cells (right upper quadrant) accountedfor 5.81% of the analysed cells after immunization and in vitro stimulation with the modifiedpeptide DENV-1/3-9L, while 0.34% after those with the original epitope peptide DENV-1/3.]

0.34% 5.81%

(B) DENV-1/3-immunized DENV-1/3-9L-immunized

DEN

V-2

/4-9

L-

Tet

ram

er

CD8

1.25% 5.59%

DENV-2/4-immunized(A)

DEN

V-2

/4-9

L-

Tet

ram

er

DENV-2/4-9L-immunized

CD8



Based on the findings mentioned above,we carried out competitive binding inhibitionassays with non-biotinylated peptide (i.e. theoriginal epitope peptide or the modified epitopepeptide), and compared the relative avidity toH-2Kd molecule. L-Kd-172 cells were incubatedwith various concentrations of the non-biotinylated peptides in the presence of 1 mMbiotinylated reference peptide, and stained with

SA-CyC. The geometric mean fluorescenceintensity was measured by FACS analysis. Percent fluorescence intensities was plotted againstnon-biotinylated peptide concentrations, and50% inhibitory concentrations (IC50) wereestimated to evaluate the avidities, meaning thatthe lower the IC50 value is, the higher the avidityof the peptide to H-2Kd is. Per cent fluorescenceintensity decreased in a dose-dependent

Figure 2: Specific binding of the biotinylated peptide to H-2Kd molecule[L929 cells (H-2k) and L-Kd-172 cells (H-2k+H-2Kd) were incubated with the biotinylated

peptide at concentrations of 10 µM, 1 µM, 0.1 µM and 0.01 µM. After incubation, cells werestained with SA-CyC, and analysed by flowcytometry. The cells without incubation with

biotinylated peptide, but stained with SA-CyC were used as controls. Fluorescence intensityincreased with DENV-2/4-9L-Biotin (A) and DENV-1/3-9L-Biotin (B) on L-Kd-172 cells in a

dose-dependent manner. No specific staining was observed on L929 cells withthe biotinylated peptide.]

control

DENV -1/3-9L-Biotin 0.01µM

DENV-1/3-9L-Biotin 0.1µM

DENV -1/3-9L-Biotin 1µM

DENV-1/3-9L- Biotin 10µM

(B)

Fluorescence intensity

cell

nu

mb

er

L929 L-K -172d

(A)

control



DENV-2/4-9L-Biotin 1µM

DENV-2/4-9L-Biotin 10µM

Fluorescence intensity

cell

nu

mb

er

L929 L-K -172d



manner as concentration of the non-biotinylatedpeptide increased (Fig. 3). The IC50 of DENV-2/4-9L (4.5 µM) was 10.7 times lower than thatof DENV-2/4 (48.0 µM). Similarly, the IC50 ofDENV-1/3-9L (7.3 µM) was 2.4 times lower thanthat of DENV-1/3 (17.8 µM). These resultsindicate that the modified epitope peptides(DENV-2/4-9L and DENV-1/3-9L) demonstratedhigher avidity to H-2Kd molecule than theoriginal epitope peptides (DENV-2/4 and DENV-1/3), respectively. These findings, taken together,indicate that modification of dengue virus-

derived CTL epitope peptide by replacing a.a.residue at the position of anchor residue toprovide a complete binding motif for MHC classI increases the binding avidity to MHC class I,resulting in immunogenicity augmentation forCTL induction by its immunization as has beenreported about the CTL epitope peptides derivedfrom tumour-associated antigens or otherviruses, and that this strategy may be applicablefor induction of dengue virus-specific CTLs byimmunization with other CTL epitope peptidesof relatively poor immunogenicity.

Figure 3: Increase in the avidity of the peptides to H-2Kd molecule by substitution of aminoacid residue to provide a complete H-2Kd-binding motif

[L-Kd-172 cells were incubated with various concentrations of non-biotinylated peptides in thepresence of 1 µM of the biotinylated peptide. Percent fluorescence intensity (See Materials andmethods.) was plotted against logarithmic scale of non-biotinylated peptide concentration, andlines were drawn by using Microsoft Excel 2003® software (DENV-2/4-9L : y = – 44.237x +

78.87 , DENV-2/4 : y = – 94.446x + 208.76 , DENV-1/3-9L : y = – 57.164x + 99.239 ,DENV-1/3 : y = – 47.947x + 109.9). The similar experiments were repeated more than threetimes, and the representative data are shown. The fluorescence intensity of the biotinylated

peptide decreased in a dose-dependent manner as concentration of the non-biotinylatedpeptides increased.

(A) *A : The 50 % inhibitory concentration (IC50) of peptide DENV-2/4-9L was 4.5 µM.*B : The IC50 of peptide DENV-2/4 was 48.0 µM.

(B) *C : The IC50 of peptide DENV-1/3-9L was 7.3 µM.*D : The IC50 of peptide DENV-1/3 was 17.8 µM.

Note that X-axis (peptide concentration) is expressed as logarithmic scale. The modifiedepitope peptides (DENV-2/4-9L and DENV-1/3-9L) with the substitution of C-terminal residue Lfor M, that possessed a complete H-2Kd-binding motif, demonstrated higher avidity to H-2Kd

molecule than the original epitope peptides (DENV-2/4 and DENV-1/3).]

(A)

50

*A *B

peptide concentration (log M)µ

%Fl

uo

resc

ence

inte

nsi

ty

100

80

60

40

20

00.0 0.5 1.0 1.5 2.0 2.5

DENV-2 4/

DENV-2 4-/ 9L

50

*C *D

DENV-1/3

DENV- -1/3 9L

100

80

60

40

20

0

%Fl

uo

resc

ence

inte

nsi

ty

0.0 0.5 1.0 1.5 2.0 2.5

peptide concentration (log M)µ

(B)



There was no significant difference in theavidity to H-2Kd between DENV-2/4-9L (IC50:4.5 µM) and DENV-1/3-9L (IC50: 7.3 µM).However, there was an apparent differencebetween DENV-2/4 (IC50: 48.0 µM) and DENV-1/3 (IC50: 17.8 µM). It was reported that theimmunogenicity of MHC class I-restrictedpeptide is determined not only by bindingaffinity to MHC molecule but also by T cellrepertoire[31]. Because immunization with themodified epitope peptide, DENV-1/3-9L,induced the CTLs that lysed the target cellspulsed with the original epitope peptide,DENV-1/3, and that these CTLs are cross-reactive to the other original epitope peptide,DENV-2/4 (data not shown), it is not plausiblethat low immunogenicity of DENV-1/3 isattributed to poor T cell repertoire. We, thus,think of other two possibilities. One possibilityis that the avidity to H-2Kd needs to bebetween 7.3 µM and 17.8 µM of IC50 to induceCTLs. The other possibility is that not onlyanchor residues but also the non-anchorresidues at secondary position contribute toincreased MHC class I avidity and peptide-MHCcomplex stability[32,33]. In the present study, we

only evaluated the binding avidity. Thus, adetailed study including analysis of the stabilityof H-2Kd/DENV-1/3 complex and H-2Kd/DENV-2/4 complex will be a future subject.

In this paper, we present the first reportthat modification of dengue virus-derived CTLepitope peptide increasing the binding avidityto MHC class I augmented immunogenicity forCTL induction. The strategy to augment theimmunogenicity of dengue virus-derived CTLepitope peptide, established here, is expectedto contribute to the immunobiology analysisof dengue virus infection, such as investigationof the immunopathogenesis, protection againstinfection, and vaccine development.

Acknowledgments

This study was supported by the grant forResearch on Emerging and Re-emergingInfectious Diseases from the Ministry of Health,Labour and Welfare, Japan H15-shinkou-17 andH18-shinkou-ippan-009.

References

[1] Gotch F, Gallimore A, McMichael A. CytotoxicT cells - protection from disease progression -protection from infection. Immunol Lett. 1996;51: 125-8.

[2] Kast WM, Melief CJ. In vivo efficacy of virus-derived peptides and virus-specific cytotoxicT lymphocytes. Immunol Lett. 1991; 30(2):229-32.

[3] Melief CJ, Kast WN. T-cell immunotherapy ofcancer. Res Immunol. 1991; 142(5-6): 425-9.

[4] Germain RN, Margulies DH. The biochemistryand cell biology of antigen processing andpresentation. Annu Rev Immunol. 1993; 11:403-50.

[5] Matsumura M, Fremont DH, Peterson PA,Wilson IA. Emerging principles for therecognition of peptide antigens by MHCclass I molecules. Science. 1992; 257:927-34.

[6] Townsend A, Bodmer H. Antigen recognitionby class I - restricted T lymphocytes. Annu RevImmunol. 1989; 7: 601-24.

[7] Bakker ABH, Van der Burg SH, Huijbens RJF etal. Analogues of CTL epitopes with improvedMHC class-I binding capacity elicit anti-melanoma CTL recognizing the wild-typeepitope. Int J Cancer. 1997; 70: 302-9.



[8] Baratin M, Kayibanda M, Ziol M et al. Aminoacid modifications in the wild type sequencep53 232-240 overcome the poorimmunogenicity of this self tumour epitope. JPept Sci. 2002; 8(7): 327-34.

[9] Bristol JA, Schlom J, Abrams S. Developmentof a murine mutant ras CD8+ CTL peptideepitope variant that possesses enhancede MHCclass I binding and immunogenic properties. JImmunol. 1998; 160: 2433-41.

[10] Parkhurst MR, Salgaller ML, Southwood S etal. Improved induction of melanoma-reactiveCTL with peptides from the melanoma antigengp100 modified at HLA-A0201 - bindingresidues. J Immunol. 1996; 157(6): 2539-48.

[11] Petersen TR, Buus S, Brunak S, Nissen MH,Sherman LA, Claesson MH. Identification anddesign of p53-derived HLA-A2 - bindingpeptides with increased CTL immunogenicity.Scand J Immunol. 2001; 53: 357-64.

[12] Thompson LW, Garbee CF, Hibbitts S et al.Competition among peptides in melanomavaccines for binding to MHC molecules. JImmunother. 2004; 27(6): 425-31.

[13] Halstead SB. Epidemiology of dengue anddengue hemorrhagic fever. In: Gubler DJ,Kuno G, editors. Dengue and denguehemorrhagic fever. Oxon: CAB International,1997. pp. 23-44.

[14] Gubler DJ. Dengue and dengue hemorrhagicfever. Clin Microbiol Rev. 1998; 11(3): 480-96.

[15] Thein S, Aung MM, Shwe TN et al. Risk factorsin dengue shock syndrome. Am J Trop MedHyg. 1997; 56(5): 566-72.

[16] Rothman AL, Kurane I, Ennis FA. Multiplespecificities in the murine CD4+ and CD8+ T-cell response to dengue virus. J Virol. 1996;70(10): 6540-6.

[17] Spaulding AC, Kurane I, Ennis FA, RothmanAL. Analysis of murine CD8+ T-cell clonesspecific for the dengue virus NS3 protein:flavivirus cross-reactivity and influence ofinfecting serotype. J Virol. 1999; 73(1):398-403.

[18] Livingston PG, Kurane I, Dai CJ et al. Denguevirus-specific, HLA-B35 - restricted, humanCD8+ cytotoxic T lymphocyte (CTL) clones.Recognition of NS3 amino acids 500 to 508by CTL clones of two different serotypespecificities. J Immunol. 1995; 154(3):1287-95.

[19] Mathew A, Kurane I, Green S et al.Predominance of HLA-restricted cytotoxic T-lymphocyte responses to serotype - cross-reactive epitopes on nonstructural proteinsfollowing natural secondary dengue virusinfection. J Virol. 1998; 72(5): 3999-4004.

[20] Mongkolsapaya J, Dejnirattisai W, Xu X-n etal. Original antigenic sin and apoptosis in thepathogenesis of dengue hemorrhagic fever.Nat Med. 2003; 9(7): 921-7.

[21] Simmons C, Dong T, Chau NV et al. Early T-cell responses to dengue virus epitopes inVietnamese adults with secondary dengue virusinfections. J Virol. 2005; 79(9): 5665-75.

[22] Rammensee H-G, Friede T, Stevanovic S. MHCligands and motif: first listing. Immunogenetics.1995; 41(4): 178-228.

[23] Romero P, Corradin G, Luescher IF, MaryanskiJL. H-2Kd - restricted antigenic peptides sharea simple binding motif. J Exp Med. 1991; 174:603-12.

[24] Masaki H, Fujii Y, Irimajiri K, Munakata H,Tomura T, Kurane I. Modification of theanchor residue to MHC class I augments CTL-inducing ability of epitope peptides derivedfrom dengue virus NS3 protein. Denguebulletin. 2006; 30: 171-6.

[25] Kozhich AT, Caspi RR, Berzofsky JA, Gery I.Immunogenicity and immunopathogenicity ofan autoimmune epitope are potentiated byincreasing MHC binding through residuesubstitution. J Immunol. 1997; 158: 4145-51.

[26] Sette A, Vitiello A, Reherman B et al. Therelationship between class I binding affinityand immunogenicity of potential cytotoxicT cell epitopes. J Immunol. 1994; 153:5586-92.



[27] Lipford GB, Bauer S, Wagner H, Heeg K.Peptide engineering allows cytotoxic T-cellvaccination against human papilloma virustumour antigen, E6. Immunology. 1995; 84(2):298-303.

[28] Tirosh B, el-Shami K, Vaisman N et al.Immunogenicity of H-2Kb - low affinity, highaffinity, and covalently-bound peptides in anti-tumor vaccination. Immunol Lett. 1999; 70(1):21-8.

[29] Tourdot S, Oukka M, Manuguerra JC et al.Chimeric peptides: a new approach toenhancing the immunogenicity of peptideswith low MHC class I affinity: application inantiviral vaccination. J Immunol. 1997;159(5): 2391-8.

[30] Vierboom MP, Feltkamp MC, Neisig A et al.Peptide vaccination with an anchor-replacedCTL epitope protects against humanpapillomavirus type 16- induced tumors

expressing the wild-type epitope. J Immunother.1998; 21(6): 399-408.

[31] Oldstone MBA, Lewicki H, Borrow P, HudrisierD, Gairin JE. Discriminated selection amongviral peptides with the appropriate anchorresidues: implications for the size of thecytotoxic T-lymphocyte repertoire and controlof viral infection. J Virol. 1995; 69(12):7423-9.

[32] Ruppert J, Sidney J, Celis E, Kubo RT, GreyHM, Sette A. Predominant role of secondaryanchor residues in peptide binding to HLA-A2.1 molecules. Saibo. 1993; 74: 929-37.

[33] Sarobe P, Pendleton CD, Akatsuka T et al.Enhanced in vitro potency and in vivoimmunogenicity of a CTL epitope fromhepatitis C virus core protein following aminoacid replacement at secondary HLA-A2.1binding positions. J Clin Invest. 1998; 102(6):1239-48.


Liver function tests in patients with dengueviral infection

Rajoo Singh Chhinaa#, Omesh Goyala, Deepinder Kaur Chhinab, Prerna Goyala,Raj Kumarb, Sandeep Puric

aDepartment of Gastroenterology, Dayanand Medical College and Hospital, Ludhiana, Punjab,India - 141001

bDepartment of Microbiology, Dayanand Medical College and Hospital, Ludhiana, Punjab,India - 141001

cDepartment of Medicine, Dayanand Medical College and Hospital, Ludhiana, Punjab,India - 141001

Abstract

To assess the frequency and degree of hepatic dysfunction in patients with dengue infection, records of214 serologically confirmed cases of dengue infection with available biochemical liver tests, admittedto our tertiary-care institute, were analysed. Patients were classified as classical dengue fever (DF) –81.3%, dengue haemorrhagic fever (DHF) – 13.6% and dengue shock syndrome (DSS) – 5.1%. Themean age was 31.6 years (male:female = 3.3:1). Deranged total bilirubin, aspartate aminotransferase(AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), albumin and prothrombin timeindex (PTI) [international normalized ratio (INR)] was present in 19.5% (29/143), 97.7% (209/214),93.9% (199/214), 32.6% (47/144), 29.1% (44/151) and 15.5% (22/156) patients respectively. Themean (± SE) total bilirubin, AST, ALT, ALP, albumin and INR values were 0.93 ± 0.09 mg/dl, 353.7 ±49.6 U/L, 218.6 ± 27.2 U/L, 135.2 ± 6.5 U/L, 3.2 ± 0.04 g/dl and 1.2 ± 0.03 respectively. The meanvalue of AST was significantly higher than ALT. The degree of rise of AST and ALP was significantly morein DHF and DSS, as compared to DF; but the frequency of rise was similar in all groups. Mean serumbilirubin, ALT and ALP values were significantly higher in patients with haemorrhage as compared tothose without haemorrhage, in patients with secondary dengue infection as compared to primaryinfection, and in non-survivors. Hepatic dysfunction was very common in all forms of dengue infection,with AST rising significantly more than ALT. Serum bilirubin, ALT and ALP were significantly higher inpatients with DSS, haemorrhage, sequential infection and non-survivors. While preferentially high ASTmay serve as an early indicator of dengue infection, high bilirubin, ALT and ALP may act as poorprognostic markers.

Keywords: Dengue infection; Hepatic dysfunction; Aspartate aminotransferase (AST); Alanine aminotransferase(ALT); Alkaline phosphatase (ALP).

#E-mail: [email protected]; Fax: 0161-2302620

Liver function tests in patients with dengue viral infection


Introduction

Dengue infection, an arthropod-borne viralhaemorrhagic fever, continues to be a majorchallenge to public health, especially in South-East Asia[1]. It has a wide geographical distributionand can present with a diverse clinicalspectrum[2]. Although dengue virus is a non-hepatotropic virus, liver injury due to dengueinfection is not uncommon and has beendescribed since the 1960s[3]. The degree of liverdysfunction in dengue infection varies from mildinjury with elevation of aminotransferases aloneto severe injury with jaundice and evenfulminant hepatic failure[2,4]. The liver dysfunctioncould be a direct viral effect or an adverseconsequence of dysregulated host immuneresponse against the virus[2]. Several outbreaksof dengue infection have been reported fromIndia[5,6,7,8]. However, large clinical studiesdocumenting hepatic involvement in dengueinfection, especially in adults, are scarce.

The aim of this study was to assess thefrequency and degree of hepatic dysfunction inpatients with dengue infection presenting to atertiary-care medical facility in Punjab. Punjab isa state located in north India with an area of50 362 sq km, extending from latitudes 29° 30’to 32° 32’ North and longitudes 73° 55’ to 76°50’ East.


An outbreak of dengue infection was noted inthe state of Punjab during the last quarter of2006. During this period, 2205 patientspresented to Dayanand Medical College andHospital, Ludhiana, Punjab, India, with acutefebrile illness. The provisional clinical diagnosisof dengue infection was made on the basis of ahistory of fever of short duration (<15 days)and constitutional symptoms, with or withouthaemorrhagic manifestations. These patientswere screened for dengue-specific IgM and IgGantibodies by IgM and IgG capture ELISA

respectively (Panbio Diagnostics, Brisbane,Australia). IgM and IgG antibody positivity wasfound in 366 and 76 patients respectively. Themortality rate was 1.9% (7/366). Medical recordsof all adult patients with available liver functiontests (LFT) (n=214) were analysed for theirclinical presentation, degree of hepaticinvolvement, hospital course and outcome.Hepatitis markers (HBsAg, IgM antibody to HAVand HEV) were done in the clinically suspectedcases (53 patients).

Patients were divided into three categories:(a) classical dengue fever (DF); (b) denguehaemorrhagic fever (DHF); and (c) dengueshock syndrome (DSS), according to the WHOcriteria[9]. Non-survival was taken as pooroutcome and survival as good outcome in ourstudy.

The statistical analysis was done using theFischer’s exact test and Student’s unpaired t-test for significance of difference in proportionsand means between two groups respectively.

Results

Of the 214 patients reactive for dengue virus-specific IgM antibody, dengue virus-specific IgGantibody was also positive in 43 (20.1%)patients. One hundred and seventy four (81.3%)patients were classified as dengue fever, 29(13.6%) as dengue haemorrhagic fever, and 11(5.1%) as dengue shock syndrome. Further,17.8% (31/174) patients with DF, 13.8% (4/29)patients with DHF, and 72.7% (8/11) patientswith DSS had positive IgG antibody, indicatingsequential infection. Markers for hepatitis A, B,C and E viruses were done in 53 patients andfound to be negative in all.

The mean age of patients in our study was31.6 years with a range of 15 to 80 years. Themaximum number of patients (n=71; 33.2%)belonged to the age group of 21–30 years.There were 164 (76.6%) males and 50 (23.4%)females (male:female ratio = 3.3:1).



The main presenting symptoms were fever(100%, 214/214), myalgias (43%, 92/214),haemorrhagic manifestations (40.6%, 87/214),vomiting (37.4%, 80/214) and abdominal pain(20%, 43/214). Encephalopathy was observedin 3 (1.4%) patients; one each belonged toDF, DHF and DSS groups. The gastrointestinaltract was the most common site of haemorrhage(n=42/214, 19.6%), followed by skin rash/petechiea (n=22/214, 10.3%). The clinicalexamination revealed hepatomegaly in 26(12.1%) patients, splenomegaly in 4 (1.9%) andascites in 4 (1.9%) patients. The mean (±SE)haemoglobin, haematocrit, total leukocytecount and platelet count at admission were13.8 ± 0.17 gm/dl, 40.6 ± 0.5%, 6123 ±339 cells/mm3 and 48.5 ± 2.6 x 1000/mm3

respectively.

Hepatic dysfunction, in the form ofderanged total bilirubin, AST, ALT, ALP, albuminand PTI (INR) was present in 19.5% (29/143),97.7% (209/214), 93.9% (199/214), 32.6% (47/144), 29.1% (44/151) and 15.5% (22/156)patients respectively. The mean (±SE) totalbilirubin, AST, ALT, ALP, albumin and INR valueswere 0.93 ± 0.09 mg/dl, 353.7 ± 49.6 U/L,218.6 ± 27.2 U/L, 135.2 ± 6.5 U/L, 3.2 ±0.04 g/dl and 1.2 ± 0.03 respectively. Themean value of AST was significantly higher thanthe mean ALT value (p=0.017). A comparisonbetween the degree of rise of ALT and AST is

Figure 1: Comparison between ALT and ASTelevation in patients with dengue infection

(n=214)

Number of times increased

Per

cen

tage

of

pat

ien

ts

AST

ALT

>3-10x >10x>1-3x≤1x

60%

50%

40%

30%

20%

10%

0%

Per

cen

tage

of

pat

ien

tsDengue fever groups

DF DHF DSS

60%

50%

40%

30%

20%

10%

0%

>1-3 timesthan ULN

>3-10 timesthan ULN

>10 timesthan ULN

Figure 2(a): Comparison of the pattern ofrise of AST in patients with DF, DHF and DSS

Figure 2 (b): Comparison of the pattern ofrise of ALT in patients with DF, DHF and DSS

DF DHF DSS

Dengue fever groups

Per

cen

tage

of

pat

ien

ts

50%

40%

30%

20%

10%

0%

>1-3 timesthan ULN

>3-10 timesthan ULN

>10 timesthan ULN

shown in Fig. 1. Significantly more percentageof patients had AST values >10 times elevatedas compared with ALT (p < 0.0001).

Figure 2 (c): Comparison of the pattern ofrise of ALP in patients with DF, DHF and DSS

30%

25%

20%

15%

10%

5%

0%

DF DHF DSS

Dengue fever groups

Per

cen

tage

of

pat

ien

ts

>1-2timesthan ULN

>2-5 xtimesthan ULN

DF= Classical dengue fever, DHF = Denguehemorrhagic fever, DSS = Dengue shock syndrome,ULN = Upper limit of normal



A comparison between the mean valuesof various liver biochemical tests in differentgroups of dengue infection, and the numberof patients with abnormal tests is shown in

Table 1 and Figures 2(a), 2(b) and 2(c). All theliver biochemical tests were significantly morederanged in the DSS group as compared tothe DF group. Also, the percentage of patients

Table 1: Comparison of biochemical liver test derangements in patients with DF, DHF and DSS

Liverbiochemical

test

DF(n= 174)

DHF(n=29)

pvalue*

DSS(n=11)

pvalue**

Mean ± SE 0.79 ± 0.080.89 ±

0.130.623

2.7 ±0.97

0.0001T. bilirubin(mg/dl)

(n=143)† No (%) of patientswith > ULN

16/118(13.5 %)

4/14(28.6%)

0.2269/11

(81.8%)0.0001

Mean ± SE277.18 ±

20.5478.4 ±

163.50.016

1234.7 ±787.6

0.0001AST (IU/L)(n=214)†

No (%) of patientswith > ULN

169/174(97.1%)

29(100%)

1.0011

(100%)1.000

Mean ± SE174.6 ±

11.05254.5 ±

800.059

819 ±431.5

0.0001ALT (IU/L)(n=214)†


161/174(92.5%)

27/29(93.1%)

1.0011

(100%)1.000

Mean ± SE122.44 ±

4.7166.7 ±

28.90.007

241.5 ±60

0.0001ALP (IU/L)(n=144)†


36/119(30.3%)

6/15(40 %)

0.0565/10(50%)

0.29

Mean ± SE 3.3 ± 0.04 3.0 ± 0.17 0.0132.7 ±0.18

0.0004Albumin(g/dl)

(n=151)† No (%) of patientswith < LLN

32/121(26.4%)

8/20(40 %)

0.2834/10

(40 %)0.46

Mean ± SE 1.1 ± 0.02 1.2 ± 0.06 0.0691.3 ±0.11

0.018PTI

(n=158)†No (%) of patients

with > ULN12/127(9.4%)

4/20(29%)

0.2366/11

(54.5%)0.0007

All data expressed as mean ± standard error (SE).† = Number of patients in which the particular test was available.

*p value = p value between DF and DHF groups.

**p value = p value between DF and DSS groups.

DF = Classical dengue fever, DHF = Dengue hemorrhagi c fever, DSS = Dengue shock syndrome, ALP = Alkalinephosphatase, ALT = Alanine transaminase, AST = Aspartate transaminase, PTI = Prothrombin time index, LLN =Lower limit of normal, ULN = Upper limit of normal.



Table 2: Comparison of biochemical liver test derangements between various patient groups

CharacteristicBilirubin(mg/dl)

AST(IU/L)

ALT(IU/L)

ALP(IU/L)

1. Sex

Male (164) 0.99 ± 0.1 380 ± 63.4 241 ± 34.9 129 ± 6.7

Female (50) 0.69 ± 0.08 266.4 ± 40.3 144.3 ± 14.9 159 ± 17.4

p value 0.109 0.333 0.131 0.053

2. Haemorrhage

With haemorrhage(n=87) 1.13 ± 0.18 464.5 ± 116.3 290.4 ± 64.0 154.6 ± 13.6

Without haemorrhage(n=127) 0.79 ± 0.08 277.8 ± 24.4 169.3 ± 11.8 121.8 ± 5.5

p value 0.05 0.064 0.028 0.012

3. Gastrointestinal haemorrhage

GI haemorrhage (n=42) 1.4 ± 0.32 591.4 ± 212.9 326.9 ± 107.7 158.9 ± 22.3

Without haemorrhage(n=127) 0.79 ± 0.08 277.8 ± 24.4 169.3 ± 11.8 121.8 ± 5.5

p value 0.008 0.016 0.016 0.021

4. Primary/secondary dengue infection

IgM reactive (n=171) 0.68 ± 0.05 312 ± 33.1 190 ± 16.7 123.5 ± 5.8

IgM and IgG reactive(n=43) 1.8 ± 0.36 519 ± 56.8 332 ± 117.2 174.5 ± 20.9

p value 0.0001 0.096 0.036 0.0012

5. Survival

Survivors (n=207) 0.84 ± 0.07 349.9 ± 50.6 207.5 ± 25.1 130.1 ± 5.9

Non-survivors (n=7) 3.15 ± 1.6 463.3 ± 250.5 544.7 ± 372.1 241.7 ± 72.4

p value 0.0001 0.684 0.026 0.001

All data expressed in mean ± standard error.

ALP = Alkaline phosphatase, ALT = Alanine transaminase, AST = Aspartate transaminase



having hyperbilirubinemia and deranged PTIwas significantly more in the DSS group ascompared to the DF group. On a comparisonof the DF and DHF groups, it was observedthat the mean values of AST, ALP and albuminwere significantly different in the two groups.

A comparison of the biochemical livertests in various patient subgroups (Table 2)showed that serum bilirubin, ALT and ALPvalues were significantly higher in patientswith haemorrhage as compared to thosewithout haemorrhage, and were even higherin the patients with GI haemorrhage. It wasalso noted that bilirubin, ALT and ALP weresignificantly higher in patients with secondaryinfection as compared to primary infectionand in non-survivors as compared to thosewho survived. The AST value was significantlymore deranged only in patients with GIhaemorrhage as compared to those withouthaemorrhage. There was no significantdifference in the LFTs between male andfemale patients and in the patients with orwithout encephalopathy.

Discussion

The clinical and biochemical impact of denguevirus on liver function was studied on 214serologically confirmed cases of dengueinfection during an outbreak in north India in2006. In this study, DHF and DSS were presentin 13.6% (29/214) and 5.1% (11/214) patientsrespectively. This is in accordance with theresults of a recent study from Delhi[8] (DHFand DSS in 9.3% and 2.2% respectively).However a few other studies had reported ahigher percentage of DHF[6,7].

Impaired consciousness was seen in onlythree patients in our study. LFT abnormalitiesin these patients were not significantly differentfrom those without encephalopathy, indicatingthat liver failure was not the cause of altered

sensorium in these patients. Other possiblereasons for the neurological symptoms indengue infection are metabolic acidosis, severedisseminated intravascular coagulation, grosshaemorrhage or edema in the brain, orhyponatremia due to excessive fluidadministration.

Hepatomegaly was observed in 12.1%patients in this study, compared to 17.6%–20.4% in other Indian studies[5,6]. The relativehigher incidence of hepatomegaly reported bySharma et al.[5] could be attributed to the factthat all their patients belonged to the DHFgroup. Although liver size does not correlatewith disease severity, an enlarged liver isobserved more frequently in shock than in non-shock cases[9]. In our study, too, hepatomegalywas more frequent in the DSS group ascompared to DF group (45.5%; 5/11 v/s 10.9%;19/174; p < 0.05).

Biochemical liver dysfunction, in the formof increased transaminases, was found in mostof the patients in our study (93.9%–97.7%),similar to the results of other studies[5,6,7,10].However, in a study by Souza et al.[11], ASTand ALT were deranged only in 63.4% and 45%patients respectively. In our study, increasedlevels of ALP and serum bilirubin were notedin a smaller proportion of patients, inaccordance with the results of Itha et al.[7]

The aspartate aminotransferase (AST) levelsin dengue infection tend to be greater thanalanine aminotransferase (ALT) levels[12,13]. Thisdiffers from the pattern in viral hepatitis but issimilar to that seen in alcoholic hepatitis. Theexact cause of this is uncertain, but it has beensuggested that it may be due to excess releaseof AST from damaged monocytes duringdengue infection[10]. We also noted apreferential elevation of liver enzymes, withAST being significantly higher than ALT. Thisabnormality may act as an early indicator ofdengue infection.



Comparing the three subgroups of dengueinfection (DF, DHF and DSS), we observed thatthe frequency of liver dysfunction (raised AST,ALT and ALP) was equally common in all thegroups (Table 1, Figures 2(a), 2(b) and 2(c)).Similar results were noted in another Indianstudy[7]. However, Wahid et al.[14] found liverdysfunction to be more common in DHF thanin DF patients.

The severity of hepatic dysfunction indengue infection has been associated withdisease severity. Indeed, liver injury has beenproposed to be a good positive predictive factorfor the development of DHF[13]. We noted agreater degree of hepatic injury in the DHFgroup (significantly higher values of AST andALP) and DSS group (significantly higher valuesof all biochemical liver tests) as compared tothe DF group, suggesting that the degree ofliver injury may be related to the severity ofdengue infection. Similar data have beensuggested by Seneviratne et al.[2] and Souza etal.[11]. However, in two other studies, thedegree of elevation of liver enzymes in the DFand DHF groups was not significantlydifferent[7,14].

In our study, the mean bilirubin, ALT andALP values were significantly higher in patientswith haemorrhage as compared to thosewithout haemorrhage, and were even higherin those with GI haemorrhage. Wahid et al.[14]

also observed that the ALT and ALP levels weresignificantly higher in DHF patients withspontaneous bleeding than those withoutbleeding (p < 0.05), while Nguyen et al.[15]

noted significantly higher elevation of AST andALT in DHF patients with gastrointestinalhaemorrhage. A possible reason for this couldbe an ischaemic injury to the liver due to acuteblood loss.

In the present study, the mean bilirubin,ALT and ALP values were significantly higherin patients with secondary dengue infection as

compared to those with primary dengueinfection, while the mean AST value in thetwo groups was similar. Nguyen et al.[15]

observed that the results of transaminases didnot differ significantly between the two groups,while Souza et al.[11] noted that transaminaseswere significantly higher in cases with secondaryinfection.

Jaundice in dengue infection has beenassociated with fulminant liver failure and byitself is a poor prognostic factor[15]. We foundhyperbilirubinemia to be significantly morecommon in patients with DSS, in patients withhaemorrhage and in non-survivors. Thus, ourobservations support the fact that high bilirubinmay act as a bad prognostic marker in patientswith dengue infection.

The percentage of patients with derangedPTI was significantly more in the DSS group ascompared to the DF group. There was nosignificant difference of biochemical liver testsbetween males and females in our study.However, in another study[11], transaminaseswere significantly higher among females.

Liver dysfunction was found to besignificantly more severe in non-survivors ascompared with survivors (Table 2). The causesof mortality (n=7) were adult respiratorydistress syndrome in three patients, underlyingcardiac dysfunction causing arrhythmias in onepatient, underlying decompensated cirrhosis inone patient, sepsis in one patient and refractoryshock in one patient.

We thus report the liver function testabnormalities and their clinical implications ina large group of patients with dengue infection.Non-availability of the baseline LFT values inthe same patient group is a possible limitingfactor. Future studies with assessment of viraltiters, and their correlation with LFTs, may helpto define the cause of hepatic injury in dengueinfection.



In summary, liver involvement is verycommon in all forms of dengue infection, withAST rising significantly more than ALT. Serumbilirubin, ALT and ALP are significantly higherin patients with DSS, haemorrhage and

sequential infection and in non-survivors.Therefore, while preferentially high AST mayserve as an early indicator of dengue infection,high bilirubin, ALT and ALP may act as poorprognostic markers.

References

[1] Halstead SB. Dengue. Curr Opin Infect Dis.2002; 15(5): 471-476.

[2] Seneviratne SL, Malavige GN, deSilva HJ.Pathogenesis of liver involvement duringdengue viral infections. Trans R Soc Trop MedHyg. 2006; 100 (8): 608-614.

[3] Burke T. Dengue haemorrhagic fever: apathological study. Trans R Soc Trop Med Hyg.1968; 62(5): 682-692.

[4] Lum LC, Lam SK, George R, Devi S. Fulminanthepatitis in dengue infection. Southeast AsianJ Trop Med Public Health. 1993; 24(3):467-471.

[5] Sharma S, Sharma SK. Clinical profile of DHFin adults during 1996 outbreak in Delhi, India.Dengue Bulletin. 1998; 22: 20-27.

[6] Daniel R, Rajamohanan, Philip AZ. A study ofclinical profile of dengue fever in Kollam,Kerala, India. Dengue Bulletin. 2005; 29:197-202.

[7] Itha S, Kashyap R, Krishnani N, Saraswat VA,Choudhuri G, Aggarwal R. Profile of liverinvolvement in dengue virus infection. NatlMed J India. 2005; 18(3): 127-130.

[8] Makroo RN, Raina V, Kumar P, Kanth RK. Roleof platelet transfusion in the management ofdengue patients in a tertiary care hospital. AsianJ Transfus Sci. 2007; 1(1): 4-7.

[9] World Health Organization. Denguehaemorrhagic fever: diagnosis, treatment,prevention and control. 2nd ed. Geneva: WorldHealth Organization, 1997.

[10] Kuo CH, Tai DI, Chang-Chein CS, Lan CK,Chiou SS, Liaw YF. Liver biochemical tests anddengue fever. Am J Trop Med Hyg. 1992; 47(3):265-270.

[11] Souza LJ, Alves JG, Nogueira RM, GicovateNeto C, Bastos DA, Siqueira EW, Souto FilhoJT, Cezário Tde A, Soares CE, Carneiro RdaC. Aminotransferase changes and acutehepatitis in patients with dengue fever:analysis of 1,585 cases. Braz J Infect Dis. 2004;8(2): 156-163.

[12] Souza LJ, Gonçalves Carnerio H, Souto FilhoJT, Souza TF, Cortes VA, Neto CG, Bastos DA,Siqueira EWS. Hepatitis in dengue shocksyndrome. Braz J Infect Dis. 2002; 6(6):322-327.

[13] Kalayanarooj S, Vaughn DW, NimmannityaS, Green S, Suntayaorn S, Kunentrasai N,Viramitrachai W, Ratanachu-eke S, KiatpolpojS, Innis BL, Rothman AL, Nisalak A, Ennis FA.Early clinical and laboratory indicators ofacute dengue illness. J Infect Dis. 1997; 176(2):313-321.

[14] Wahid SF, Sansui S, Zawawi MM, Ali RA. Acomparison of the pattern of liver involvementin dengue hemorrhagic fever with classicaldengue fever. Southeast Asian J Trop Med PublicHealth. 2000; 31(2): 259-263.

[15] Nguyen TL, Nguyen NT, Tieu NT. The impactof dengue fever on liver function. Res Virol,1997; 148(4): 273-277.


Changing clinical manifestations of dengue infection innorth India

Chandrakanta, Rashmi Kumar#, Garima, Jyotsana Agarwal,Amita Jain, Rachna Nagar

Departments of Paediatrics and Microbiology, Chhatrapati Shahuji Maharaj Medical University (CSMMU),Lucknow 226003, India

Abstract

Dengue infection is endemic in many parts of India, including the state of Uttar Pradesh. This studydescribes the changing clinical picture of dengue viral infections observed by us in children admittedto a teaching hospital in Lucknow, India.

A total of 139 children with suspected dengue were admitted during this period, of which 124 couldbe tested by dengue IgM capture ELISA and 102 were positive. However, only 80 of these 102 patientscould be followed up. Average age was 5.9 (±3.1) years and 87.5% of them were from rural areas. Themale:female ratio was 1.6:1. Seizures were observed in 45% cases, altered sensorium in 53.7%, vomitingin 41.2%, haemorrhage in 38.8%, skin rash in 37.5%, abdominal pain in 25%, headache in 18.8% andjaundice in 2% cases. Gastrointestinal tract was the commonest site of bleeding. On examination,edema was present in 47.5% cases, hepatomegaly in 62.5%, splenomegaly in 60.0% and hypotensionin 10.0% cases. The investigations revealed a low platelet count of less than 100 000/mm3 in 60.3%cases. Mean liver enzyme levels were mildly raised. Definitions of WHO criteria for DHF were presentin only 18 (22.5%) cases. Mean total duration of fever in survivors was 14.9±7.3 days. The overallfatality rate in hospital was 5.0%.

The results indicated a significant proportion of children presented with little-described features ofencephalopathy, edema, splenomegaly and prolonged fever rather than the typical dengue presentation.These features were not noted during the past epidemics and in previous years.

Keywords: Dengue viral infection; Dengue fever; Dengue encephalopathy; Dengue haemorrhagic fever.


Introduction

Dengue infection is the most importantarbovirus infection of humans and the mostimportant tropical infectious disease after

malaria. Although dengue fever is a very olddisease, more complicated forms of theinfection – dengue haemorrhagic fever (DHF)and dengue shock syndrome (DSS) have beenrecognized in the last century[1]. In India, the

Changing clinical manifestations of dengue infection in north India


first virologically confirmed epidemic occurredin Calcutta (now known as Kolkata) and theeastern coast of India in 1963–1964[2]. All fourserotypes of the virus are circulating now[3]. Amajor widespread epidemic of DHF occurredin 1996 involving areas around Delhi, and,since then, there has been a remarkableresurgence of the infection in north Indianplains that include the state of Uttar Pradesh.Once considered an urban problem, it has nowpenetrated into rural areas also due to highpopulation density and other factors[4].

As observed in this part of the country,dengue infection is showing an increasingtrend. The illness occurs throughout the yearwith a peak during monsoon and post-monsoonseason due to high vector density. Majoroutbreaks have occurred in this region in 2003and 2006. Besides the increasing frequency ofthe infection, even the manifestations observedhave been varied. In 2008, we observedmanifestations of dengue which were notcommonly observed in the previous years. We,therefore, undertook to prospectively study anddescribe the varied manifestations of dengueviral infection as seen in hospitalized childrenin northern India.


This study was conducted in the Departmentof Paediatrics of Chhatrapati Shahuji MaharajMedical University (CSMMU) Hospital,Lucknow – a tertiary-care teaching hospital.

Over a period of five months from Augustto December 2008, we carefully screenedadmissions for suspected diagnosis of dengue,as made by the admitting physician, usuallyon the basis of febrile illness with rash, orbleeding with or without alteration ofconsciousness. A detailed clinical history wastaken, physical examination was performedand baseline investigations were noted using

a structured proforma. Laboratoryinvestigations and treatment of the patientswere decided by the treating physician. Teststhat were usually done included haemoglobin(Hb), total and differential leukocyte count(TLC and DLC), platelet count (PLT count),haematocrit (HCT), and liver function tests(LFT) including prothrombin time (PT), serumproteins and albumin. Lumbar puncture andcerebrospinal fluid (CSF) examination wasusually done in patients who presented witha history of altered sensorium and/or seizures.Serology for dengue infection was also doneas part of routine clinical work. Blood sampleswere collected and transported to theDepartment of Microbiology, CSMMU. Theywere tested for dengue IgM by antibodycapture ELISA (Mac ELISA) test usingcommercial kits marketed by IVD ResearchInc., USA. Thus, the study was purelyobservational.

Diagnosis of dengue infection, DF andDHF was made according to WHO criteria[5].If altered sensorium was present, the child wasclassified as dengue encephalopathy (DE) withor without DHF.

Statistical analysis

Data were entered into a Microsoft Excelsheet. Frequencies, mean and standarddeviation was calculated by using Epi-infosoftware for statistical analysis.

Results

During the period of the study, a total of 139suspected dengue patients were admitted tothe hospital, of which 124 could be tested fordengue IgM, and 102 were positive. Of these102 patients, 80 could be followed up anddocumented. The clinical features of these 80patients are given in Table 1.



In the 80 serologically-confirmed cases, 18(22.5%) satisfied WHO criteria for DHF, while43 (53.7%) had encephalopathy. Seven patientswith DHF had encephalopathy also. One casewith clinical presentation of Guillian Barrésyndrome was also dengue IgM-positive. Meanduration of fever at presentation was 10.7±6.2days. After follow-up, mean total duration offever in survivors was found to be 14.9±7.3days. Other main complaints besides feverwere: swelling over body, rash, altered

sensorium, seizures, vomiting , bleeding,abdominal pain and headache.

On examination, a discrete maculopapularerythematous rash was present in 37.5% cases.Edema was present in 47.5% children. It wasgeneralized in 36.3%, over extremities in 6.2%and facial in 5% children. Haemorrhage wasfound in 31 (38.8%) children. Gastrointestinaltract was the most common site for bleeding(23.7%) followed by the skin (16.2%).

Table 1: Clinical features of dengue IgM-positive cases

S. Clinical features Dengue IgM +ve casesNo. (n=80) No (%)

1. Mean age in years ± SD 5.9±3.1

2. Male:Female ratio 1.6:1

3. Residence in rural area 70 (87.5)

4. Fever 80 (100)

5. Average duration of fever at admission in days ± SD 10.7±6.2

6. Altered sensorium 43 (53.7)

7. Seizures 36 (45)

8. Abdominal pain 20 (25)

9. Haemorrhage 31 (38.8)

10. Diarrhoea 5 (6.2)

11. Vomiting 33 (41.2)

12. Headache 15 (18.8)

13. Rash 30 (37.5)

14. Edema 38 (47.5)

15. Hepatomegaly 50 (62.5)

16. Splenomegaly 48 (60.0)

17. Hypotension 8 (10.0)

18. Meningeal signs 7 (8.7)

19. Raised intracranial tension (ICT) 5 (6.2)

20. Jaundice 2 (2.5)

21. Total duration of fever in days ± SD 14.9±7.3

22. DHF 18 (22.5)DE 43 (53.7)DE+DHF 7 (8.7)

23. Duration of hospital stay in days ± SD 7.6±4.7

24. Mortality in hospital 4 (5.0)



Conjunctival haemorrhage and epistaxis werenoted in 2 patients (4.1%) and 1 patient (2.0%)respectively. Intracranial haemorrhage wassuspected in one child, but cranial imagingcould not be done in this case. One childdeveloped haemorrhagic pleural effusion andanother had pulmonary haemorrhage. Gumbleeding was not present in any child.Hepatomegaly and splenomegaly were presentin 62.5% and 60% cases respectively. Meanliver size was 4.1±1.1 cm below costal marginand mean spleen size was 2.8±1.4 cm belowcostal margin at the time of admission.

Seizures occurred in 36 (45%) of cases.Altered sensorium was present on admissionin 41 (51.2%). In children with alteredsensorium, rash was seen in 18 (43.9%),bleeding manifestations were seen in 15(38.5%) and swelling in 13 (31.7%). Alterationof sensorium developed later in another 2patients. The average duration of alteredconsciousness on admission was 2.8±5.2 daysand for seizures 2.8±5.3 days. Generalisedhypertonia was found in 21 (48.8%) subjectswith encephalopathy, meningeal signs in 7(16.8%) and focal neurological deficit in 2

Table 2: Laboratory investigations in dengue IgM-positive cases

S. Investigation Dengue IgM +ve casesNo. (n=80) No (%)

1. Mean Hb (gm%) ± SD 9.8±2.0

2. Mean total leukocyte count (per mm3) ± SD 9848±3908

3. Mean % polymorphs in blood 61.5±11.9

4. Platelet count (per mm3) in blood<20 000 11 (13.7)21 000–40 000 12 (16.2)41 000–60 000 11 (13.7)61 000–80 000 9 (11.2)81 000–100 000 5 (6.2)>100 000 32 (40.0)

5. Mean packed cell volume (PCV) (%) ± SD 26.8±5.7

6. Mean serum bilirubin (in mg%) ± SD 1.0±0.7

7. Mean sGOT (IU) ± SD sGOT(IU) >40 IU 98.4±69.820/26 (76.9)

8. Mean SGPT (IU) ± SD sGPT(IU)> 40 IU 78.1±66.422/33 (66.7)

9. Mean International Normalized Ratio (INR) ± SD 1.8±2.0

10. Mean serum sodium (in mEq/l) ± SD 132.5±5.3

11. Mean urea (in mg%) ± SD 35.0±18.2

12. Mean serum protein (in gm%) ± SD <6.1 gm% 5.9±0.814/26 (53.8)

13. Mean serum albumin (in gm%) ± SD 3.1±0.45

14. CSF findings (38 patients)CSF pleocytosis (>10 cells/mm3) 17/38 (44.7) 30.4±80.6Mean cell count ± SD (per mm3) 11.9±24.2Mean polymorphs% ± SD Mean CSF protein (in mg%) ± SD 84.4±61.1CSF sugar – normal (>=2/3rd of blood sugar) 29 (80.6)CSF sugar – decreased (<2/3rd of blood sugar) 7 (19.4)



(4.6%) children. Five (6.2%) patients developedfeatures of raised intracranial tension (ICT) suchas hypertension, bradycardia andhyperventilation.

The laboratory findings are given in Table2. The platelet count was below 100 000/mm3

in 48 (60%) cases. In 13.7% cases the plateletcount was below 20 000 mm3. Liver enzymessGOT and sGPT were raised above the normallimit in 76.9% and 66.7% cases respectively.Packed cell volume was greater than 36 in 2patients only.

Examination of the cerebrospinal fluid wasdone in 38 patients, of which 17 (44.7%)showed pleocytosis with mean cell count of30.4±80.6/mm3. Mean CSF protein was84.4±61.1 mg%.

Discussion

Dengue is a major public health problem inLucknow and surrounding districts in the stateof Uttar Pradesh in north India. Over the last7–8 years we have been observing variedclinical manifestations of dengue, which arerather different from the past reports from thisregion as well as from other parts of thecountry. An ‘encephalopathic’ presentation wasnoted by us from 2003 itself, which led us totest consecutive children hospitalized withacute febrile encephalopathy (AFE) for dengueIgM and genome in CSF and serum. Of a totalof 265 patients of AFE tested, 39 (14.7%) wereconclusively proven to have dengue viralinfection[6]. We also observed dengue viralinfection presenting as acute hepatic failure. Atotal of 27 children admitted with acute hepaticfailure were tested for dengue IgM of which13 were unequivocally positive, and 7 of thesewere tested for dengue genome by RT-PCR,of which 4 were positive[7]. In 298 patients ofacute undifferentiated febrile illness, dengue

IgM was positive in 56 (18.8%) and denguegenome was detected in 15 of 44 IgM-positivecases[8]. It was observed that in addition to thewell-known WHO criteria for case definitionof DF, altered liver function with moderateelevation of transaminases is a differentiatingfeature of dengue. Over the last two seasonswe have observed a further shift in the clinicalmanifestations, which we think is worthy ofdissemination through this communication.

Our patients did not include all dengueIgM-positive cases over the study period.However, they were unselected cases andtherefore unlikely to represent a biased group.Most of the dengue cases belonged to ruralareas. This may only reflect the predominantlyrural population admitted to this hospital. Intwo studies conducted between 2003 and 2006we found no significant difference in theincidence between rural and urban areas[4].

The major difference from the previousreports is the frequent occurrence ofencephalopathy, swelling, splenomegaly andprolonged fever. Encephalopathy, an importantmanifestation of dengue infection seen here,has been reported by us previously. It wasobserved in 53.7% patients in this series, butwas not reported by earlier workers fromLucknow[9,10] and was seen in only 4% cases inthe Delhi epidemic of 1996[11]. Rash, swellingand/or bleeding manifestations in these patientsare suggestive of dengue and prompts testingfor dengue IgM. Encephalopathy has beenreported in several studies from Thailand[12-16].Encephalopathy in dengue was believed to bedue to cerebral edema, hyponatremia,hypoperfusion or intracranial bleed, but, morerecently, the actual dengue viral invasion ofthe brain is recognized[17,18].

Another manifestation observed by usfrequently over the last few years is thepresence of swelling which was found in almost



half of our patients (Figure). This is a peculiargeneralized non-pitting edema which may beexplained by plasma leak in DHF. However,no earlier workers from India or abroad havementioned this finding. Only 15 of the patientswith swelling had received intravenous fluidsprior to presentation here. Swelling was notseen in patients with other diagnoses seenhere even if fluids had been administeredoutside.

Although hepatomegaly is among theWHO clinical criteria for DF, splenomegalyis not generally held to be a feature ofdengue infection. Earlier reports fromLucknow[9,10] and other parts of India[11,19-21]

do not describe a high frequency ofsplenomegaly. In our earlier studies we toodid not usually find splenomegaly in DF, DHFor DE[6-8]. However, in this season (2008), itwas observed commonly in almost 3/5th ofthe cases. Peripheral smears for malaria werenegative in all cases. A recent study fromDelhi has reported a somewhat highpercentage (32.4%) of splenomegaly inchildren with dengue[22].

Dengue fever is generally described as ashort febrile illness. The WHO criteria mentionan illness of 2–7 days’ duration[5]. Over the

last two years we have observed a longerduration of fever than in previous years. Themean duration of fever in survivors in this studywas almost 15 days.

On an analysis of the laboratory findings,it was observed that platelets were below100 000/mm3 in a majority of the cases, withroughly 1/6th having counts below 20 000/mm3. Liver transaminases showed a mild-to-moderate elevation in around 3/4th patients.Alterations in liver functions are well-knownin dengue infection[23-26], but are not listed inthe WHO criteria for case definition[5]. Packedcell volumes (PCV) were almost always lowin our patients, presumably due to highprevalence of anaemia. The diagnosis of DHFrests heavily on finding a high PCV, but in ourpatients, we have to rely on other evidencesof capillary leak like low serum proteins. Eventhis may be misleading because serumproteins may be low due to malnutrition also.Demonstration of pleural fluid or ascites isdifficult because these findings may not befound in all stages of the illness and involvestransportation of a sick child. Therefore, onlyabout a fourth of our hospitalized patients haddefinite WHO features for case definition ofDHF. We strongly feel that these casedefinitions need revision as these cannot beapplied in settings such as ours where dengueregularly occurs.

Our study patients had a severepresentation and rather high mortality. This isbecause all were hospitalized patients. If alldengue cases occurring in the community wereto be described then we expect the severemanifestations and mortality to be certainlylower. However, such severe manifestationsand encephalopathy are not described fromother parts of the country even in hospitalizedpatients, which makes us think that thedifferences are at least in part due to greatervirulence and neurotropism of the serotypescirculating in this region.

Figure: Characteristic swelling on the face ofa child with dengue encephalopathy



Only serological diagnosis by IgM ELISA waspossible in our patients. IgM, however, has itslimitations in the diagnosis of dengue infection.Studies have shown 80% positivity in the first 5days of illness, 93% positivity between 6 and10 days of illness onset, and 99% positivity afterthe 10th day[1]. A negative IgM in a patientsampled early in the illness therefore does nottotally rule out dengue infection. On the otherhand, a positive IgM does not always mean thatthe current illness is dengue, but that dengueinfection has occurred in the recent past, i.e.60–90 days. IgM positivity therefore may merelymean that dengue transmission is going on.Further, although IgM-type antibodies are heldto be specific between flaviviruses, some cross-reactivity was still possible, especially withJapanese encephalitis, which is endemic here.Due to these limitations, WHO has put thediagnosis of dengue infection on the basis ofpositive acute phase IgM ELISA test as ‘probable’

only. However, all our patients were clinicallysuspected as dengue, and, therefore, are verylikely to have had dengue viral infection.

In conclusion, clinical manifestations ofdengue as seen by us in Lucknow over the lastfew years appear to be different from thoseseen in other parts of the country, or even inthe same region in earlier epidemics. Themanifestations also seem to be changing overthis period. DF and DHF/DSS are not the onlyclinical presentations. Encephalopathy is animportant presentation in hospitalized children.The spectrum of findings may be explained bythe presence of different circulating serotypesin this region. It would be interesting tocorrelate serotype with clinical features in thisinfection. Our report, however, is based ononly a small study and should be corroboratedby a larger, detailed study.

References

[1] Gubler DJ. Dengue and dengue hemorrhagicfever. Division of Vector Borne InfectiousDiseases, National Center for InfectiousDiseases, Centers for Disease Control andPrevention, 1999.

[2] Ramakrishanan SP, Gelfand HM, Bose PN,Sehgal PN, Mukharjee RN. The epidemic ofacute haemorrhagic fever, Calcutta, 1963:epidemiological Inquiry. Indian J Med Res.1964; 52: 633-50.

[3] Bharaj P, Chaher HS, Pandey A, Diddi K, LalitD, Guleria R, Kabra SK, Broor S. Concurrentinfections by all four dengue virus serotypesduring an outbreak of dengue in 2006 inDelhi. India. Virol. J. 2008; 5: 1-4.

[4] Tripathi P, Kumar R, Tripathi S, Tambe JJ,Venktesh V, 2008. Descriptive epidemiologyof dengue transmission in Uttar Pradesh. IndianPaed. 2008, Apr, 45, pp. 315-8.

[5] World Health Organization. Denguehaemorrhagic fever: diagnosis, treatment,prevention and control, 2nd ed. Geneva, 1997.

[6] Kumar R, Tripathi S, Tambe JJ, Arora V,Srivastava A, Nag VL. Dengue encephalopathyin children in northern India: Clinical featuresand comparision with nondengue. J. Neurol.Sci. 2008, Jun, 269(1-2): 41-8.

[7] Kumar R, Tripathi P, Tripathi S, Kanodia A,Venkatesh V. Prevalence of dengue infectionin north Indian children with acute hepaticfailure. Ann Hepatol. 2008; 7: 58-61.

[8] Kumar R, Tripathi P, Tripathi S, Kanodia A, PantS, Venkatesh V. Prevalence and clinicaldifferentiation of dengue fever in children innorthern India. Infection. 2008; 16: 444-9.

[9] Agarwal R, Kapoor S, Nagar R, Misra A. Aclinical study of the patients with dengue



haemorrhagic fever during epidemic of 1996at Lucknow, India. Southeast Asian J Trop. Med.Pub. Hlth. 1999; 30: 735-740.

[10] Kishore J, Singh J, Dhole TN, Ayyagari A.Clinical and serological study of first largeepidemic of dengue in and around Lucknow,India. Dengue Bulletin. 2006; 30: 72-79.

[11] Aggarwal A, Chandra J, Aneja S, Patwari AK,Dutta AK. An epidemic of dengue hemorrhagicfever and dengue shock syndrome in childrenin Delhi. Indian Pediatr. 1998; 35: 727-32.

[12] Hendarto SK, Hadinegoro SR. Dengueencephalopathy. Acta Paediatrica Jpn. 1992;34(3): 350-7.

[13] Cam BV, Fonsmark L, Hue NB, Phuong NT,Poulsen A, Heegaard ED. Prospective case –control study of encephalopathy in childrenwith dengue hemorrhagic fever. Am J Trop MedHyg. 2001; 65: 848-51.

[14] Kankirawatana P, Chokephaibolkit K, YoksanS, Pathavathana P. Dengue infection presentingwith central nervous system manifestations. JChild Neurol. 2000; 15: 544-7.

[15] Thisyakorn U, Limpitikul W. Dengue infectionwith central nervous system manifestations.Southeast Asian J Trop Med Public Health.1999; 30(3): 504-6.

[16] Panchareon C, Thisyakorn U. Neurologicalmanifestations in dengue patients. SoutheastAsian J Trop Med Public Health. 2001; 32(2):341-5.

[17] Lum LC, Lam SK, Choy YS, George R, Harun F.Dengue encephalitis: a true entity? Am J TropMed Hyg. 1996; 54(3): 256-9.

[18] Solomon T, Dung MN, Vaughn DW, Kneen R,Thao LTT, Boonyos R, Loan HT, Day NPJ,Farrar J, Myint KSA, Warrell M, James WS,Nisalak A, White NJ. Neurological

manifestations of dengue infection. Lancet.2000; 355: 1053-9.

[19] Rategeri VH, Shepur TA, Wari PK, Chavan SC,Mujahid IB, Yergolkar PN. Clinical profile andoutcome of dengue fever cases. Indian J Pediatr.2005; 72: 705-6.

[20] Narayanan M, Aravind MA, Thilothammal N,Prema R, Sargunam CS, Ramamurty N.Dengue fever epidemic in Chennai – a studyof clinical profile and outcome. Indian Pediatr.2002; 39: 1027- 33.

[21] Singh NP, Jhamb R, Agarwal SK, Guha M,Dewan R, Daga MK, Chakravarti A, Kumar S.The 2002 outbreak of dengue fever in Delhi,India. Southeast Asian J Trop Med Public Health.2005; 36: 15-21.

[22] Faridi MM, Aggarwal A, Kumar M, SarafrazulA. Clinical and biological profile of denguehemorrhagic fever in children in Delhi. TropDoct. 2008; 38: 2 8-30.

[23] Larreal Y, Valero N, Estevez J, Reyes I,Maldonado M, Espina LM, Arias J, Melean E,Anez G, Atencio R. Hepatic alterations inpatients with dengue. Invest Clin .2005; 46(2):169-78.

[24] Kou CH, Tai DI, Chang-Chien CS, Lan CK,Chiou SS, Liaw YF. Liver biochemical tests anddengue fever. Am J Trop Med Hyg. 1992; 47(3):265-70.

[25] Nguyen TL, Nguyen TH, Tieu NT. The impactof dengue hemorrhagic fever on liver function.Res Virol. 1997; 48: 273.

[26] Souza LJ, Alvez JG, Nogueira RM, GicovateNC, Bastos Da, Siqueira EW, Souto Filho JT,Cezario T, Soares CE, Carneiro RC.Aminotransferase changes and acute hepatitisin patients with dengue fever: analysis of 1,585cases. Braz J Infect Dis. 2004; 8(2): 156-63.


Dengue virus serotype 3 (genotype III) from Colombia:A perspective of its pathogenic potential

Sergio Yebrail Gómez Rangela, Christian Julián Villabona-Arenasa,b,Flor Angela Torres Pimientoa, Daniel Rafael Miranda-Esquivelb,

Raquel Elvira Ocazionez Jimeneza#

aLaboratorio de Arbovirus, Centro de Investigaciones en Enfermedades Tropicales,Universidad Industrial de Santander, Sede Guatiguará Km 2 Autopista Piedecuesta, Colombia

bLaboratorio de Sistemática & Biogeografía, Escuela de Biología,Universidad Industrial de Santander, A.A. 678 Bucaramanga, Colombia

Abstract

The introduction of DENV-3 genotype III in Latin American countries has been associated with dengueoutbreaks, and the role of the virus with respect to the occurrence of dengue haemorrhagic fever (DHF)has been different depending on the country. We have conducted research on the relative abundanceof DENV-3 in relation to the incidence of DHF in a Colombian endemic area. Additionally, it wasexplored using phylogenetic analyses whether or not viruses are genetically distinct in relation to theseverity of dengue. Viral isolation was made from serum samples collected during the period fromJanuary 2007 to October 2008. Sequences from the envelope gene of viruses from Colombia andLatin American countries isolated from DF and DHF patients and submitted to GenBank were compared.We found that in 2007–2008 the predominance of DENV-3 declined as compared to 2002–2004(28.3% versus 87.8%), whereas the DENV-1 and DENV-2 predominance increased (54.7% versus2.7% and 16.9% versus 5.4%, respectively). This relative abundance of serotypes coincided with anincrease of DHF compared with the period of the highest DENV-3 dominance (25.9% versus 4.6%).Phylogenetic analyses showed that: (i) there is no relationship between DENV-3 clades and the severityof the disease; and (ii) Colombian viruses clustered apart from those coming from countries whereDENV-3 has caused severe dengue. The results suggest that DENV-3 could not play any important rolein the occurrence of DHF in Colombia, and that local viruses are genetically distinct from LatinAmerican viruses associated with epidemics of DHF.

Keywords: Dengue serotypes; Dengue haemorrhagic fever; DENV-3 genotype III; Colombia.

#E-mail: [email protected]; Fax: (+57)-7-6455693, Voice: (+57)-7-6455693

Introduction

Dengue virus exists as four antigenically distinctviruses designated as serotypes (DENV-1, -2,

-3, and -4), belonging to genus Flavivirus offamily Flaviviridae[1]. Infection with any one ofthese serotypes generally leads to a mild, self-limiting febrile illness called dengue fever (DF).

Dengue virus serotype 3 (genotype III) in Colombia


Nonetheless, in a few cases, the viral infectionleads to severe, sometimes fatal, denguehaemorrhagic fever (DHF) and dengue shocksyndrome (DSS)[2]. Epidemiological studies haveidentified sequential infection with differentserotypes as a risk factor for DHF/DSS[3,4].Despite the much higher frequency ofsecondary infections in areas where two ormore DENV serotypes are present, only a smallpercentage of patients develop DHF[5,6]. Theinfecting viral strain is hypothesized to influencethe severity of dengue. It has beendemonstrated that dengue virus serotypes andstrains within a serotype may vary in their abilityto cause DHF[7-9].

DENV-3 viruses are phylogeneticallygrouped into four genotypes by Lanciotti etal.[10] or five genotypes by Wittke et al.[11]. InLatin America, this serotype was presentbetween 1963 and 1977, and reappeared inNicaragua and Panama in 1994[12]. It thendispersed to Central and Caribbean Americancountries[13-15]. In South America, DENV-3appeared first in Brazil[16] and Venezuela[17] in2000, and then dispersed to neighbouringcountries in the following years[18-21]. Virusesisolated before 1994 were DENV-3 genotypeIV, and those isolated after 1994 were DENV-3 genotype III[10,17,18-20]. Recently, Brazilianisolates in 2002 and 2004 were grouped intogenotype I[22], but the precise classification hasbeen controversial, considering that thisgenotype was classified as genotype V byNogueira et al[23].

In Colombia, the presence of DENV-3genotype III was detected for the first time in2001 in the Departamento de Santander, theregion where the present study was conducted.The reappearance of the virus coincided withan extended epidemic but increase in thenumber of DHF cases was not observed[21,24].The same was seen in Mexico[13], PuertoRico[14], Venezuela[17] and Peru[20], where a greatnumber of dengue cases occurred after the

re-introduction of the virus, but DHF cases wererare. In contrast, in Brazil[25], Paraguay[26] andCuba[15], DHF/DSS in DENV-3-infected patientswas frequent and some of them died. Likewise,the predominance of this dengue serotype inIndia and Sri Lanka has been associated withan increased incidence of DHF/DSS[27,28].

It is difficult to determine the causes fordifferent clinical outcomes in dengue patientsinfected with DENV-3 of genotype III. This isdue to the limitations of our knowledge aboutthe role of host factors and the virus-specificdeterminants of virulence. In this study weinvestigated the predominance of the viruscirculating during 2007 and 2008, five yearsafter its occurrence in the Departamento deSantander, with relation to the occurrence ofDHF. We have also studied the geneticrelationships of Colombian DENV-3 isolates todetermine if viruses in DF patients diverge orhave distinct geographical origin from those inDHF patients.

Methods

Study area

Santander is one of the 32 states (departments)of Colombia with a total number of 87municipalities. Located in the north-central partof the country close to Venezuelan border, itcovers an area of 30 537 km². Its Capital isBucaramanga, which together with threenearby municipalities, constitutes the seventhlargest metropolitan area of Colombia with onemillion inhabitants (2005). It has a populationdensity of 1012/km2, and an annual averagetemperature of 22 °C. At least 94% of denguecases occurring in Santander originated inBucaramanga and the metropolitan area. Thenumbers of DHF cases out of total denguecases in Santander during the period 1998–2008, as reported by the state Secretariat ofHealth[29,30], are shown in Table 1.



Serosurveillance

A total of 680 serum samples from clinicallysuspected dengue patients, reported to thedengue surveillance programme set up by thestate Health Secretariat, were included in thestudy. Between 18 January 2007 and 11October 2008, acute serum samples for virusisolation were selected every week from totalsamples sent to the Public Health Laboratoryin Bucaramanga. Only sera from patients withfever of unknown origin (not from respiratory,diarrhoea or other apparent causes) wereincluded, and a reporting form was completedwith clinical and laboratory data collected fromeach patient.

Virus isolation

The isolation of viruses from the acute phasesamples was attempted in C6/36 cells aspreviously described[21]. Briefly, 100 µl of serum

was added onto cell monolayers and, aftercentrifugation, 1 ml of culture medium wasadded. Cells were incubated at 32 °C, andanalysed for the presence of virus on the 12th

post-infection day by using a polyclonal anti-dengue antibody (Instituto Evandro Chagas,Brazil) in a direct immunofluorescence assay.

Typing of viruses

Serotype identification of the virus isolates wascarried out by a seminested reversetranscription-PCR (RT-PCR) protocol on the basisof that described by Lanciotti et al[31]. Briefly,viral RNA was extracted from 140 µl of cell-infected culture supernatant by using Trizol®(GIBCO BRL, Grand Island, NY), followed byreverse transcription with forward primer D2.cDNA was subjected to PCR amplification withD1 and D2 primers for 42 cycles, and a secondround of amplification was conducted with amixture of type-specific reverse primers (TS1-TS4). PCR reaction product was electrophoresedthrough a 2.5% agarose gel, stained withethidium bromide, and photographed.

Anti-dengue IgG antibodies

IgG antibodies were screened in the seracollected 0–4 days after the onset of symptomsfrom dengue virus isolation-positive cases byusing the PanBio IgG ELISA kit (PanBio Inc.,Brisbane, Australia). Primary or secondaryinfection status was determined by the absenceor presence of IgG antibody in an acute-phasesample.

Sequence analysis

The envelope gene sequences of ColombianDENV-3 genotype III isolates were used alongwith some representative global isolates.Sequences of the remainder genotypes wereincluded as outgroup. All sequences used in the

Table 1: Annual DHF cases reported by thestate Secretariat of Health, Departamento de

Santander, Colombia, 1998–2008

Year Dengue DHFcases Total %

1998 23 826 881 3.71999 4956 25 0.52000 1525 130 8.52001 10 530 779 7.42002 10 356 523 5.02003 6638 288 4.32004 1669 50 2.92005 1586 419 26.42006 2341 496 21.22007 5167 1 331 25.72008* 2579 676 26.2

Source[29,30]: *From January to October, informationsupplied by Luis Gualdrón, División de Epidemiología,Secretaría de Salud, Santander, Colombia.



Table 2: Isolates of DENV-3 analysed

Location* Strain Year Clinical Accessionstatus no.

Brazil 68784 2000 DF AY038605Brazil BR74886/02 2002 DSS AY679147Brazil/Porto Velho D3BR/PV5/02 2002 DF DQ118875Brazil/Porto Velho D3BR/PV2/03 2003 DF DQ118872Brazil/Porto Velho D3BR/PV4/03 2003 DF DQ118874Brazil/Riberão Petro D3BR/RP1/03 2003 DF DQ118877Brazil/Riberão Petro D3BR/RP2/03 2003 DF DQ118879Colombia/Norte de Santander COD3_ OC092 2005 DF FJ189462Colombia/Santander COD3_ 01072 2001 DF FJ189450Colombia/Santander COD3_ 02200 2002 DF FJ204475Colombia/Santander COD3_ LV073 2003 DF FJ189458Colombia/Santander COD3_ LV016 2003 DHF FJ189454Colombia/Santander COD3_ LV038 2003 DHF FJ189455Colombia/Santander COD3_ LV058 2003 DHF FJ189457Colombia/Santander COD3_ LV057 2004 DF FJ189456Colombia/Santander COD3_ LV433 2004 DHF FJ189461Colombia/Santander COD3_ LV428 2004 DHF FJ189460Cuba Cuba116/00 2000 DF AY702032Cuba Cuba580/01 2001 DSS AY702030Cuba Cuba21/02 2002 DF AY702031Mexico MEX6097 1995 DF AY146763Mexico/Oaxaca OAXACA-MX/00 2000 DF DQ341207Mexico/Quintana Roo 6889/QUINTANA ROO-MX/97 1997 DF DQ341205Mexico/Quintana Roo 6896/QUINTANA ROO-MX/97 1997 DF DQ341206Mexico/Yucatan 4841/YUCATAN-MX/95 1995 DF DQ341202Mexico/Yucatan 6584/YUCATAN-MX/96 1996 DF DQ341203Mexico/Yucatan 6883/YUCATAN-MX/97 1997 DF DQ341204Nicaragua Nicaragua24/94 1994 DHF AY702033Paraguay/Asunción D3PY/AS10/03 2003 DF DQ118883Paraguay/Asunción D3PY/AS9/03 2003 DF DQ118885Paraguay/Fernando de la Mora D3PY/FM11/03 2003 DF DQ118886Paraguay/Pedro Juan Caballero D3PY/PJ4/03 2003 DF DQ118887Venezuela/Aragua LARD6007 2000 DF AY146765Venezuela/Aragua LARD6315 2000 DF AY146767Venezuela/Aragua LARD6318 2000 DF AY146768Venezuela/Aragua LARD6397 2000 DF AY146769Venezuela/Aragua LARD6411 2000 DF AY146770



study were deposited in GenBank (Table 2).Colombian viruses were isolated in ourlaboratory from patients suffering from eitherDF or DHF in previous studies[21,24]. Thesequences were aligned in the Muscle softwarev. 3.7[32] using the default parameters and themodel of nucleotide substitution that best fitsthe data set was determined using a hierarchicallikelihood ratio test[33] using the Modeltestsoftware[34]. A maximum likelihood phylogenetictree was reconstructed in the phyML softwarev. 3.0[35] where the starting tree was found usingthe neighbour-joining method. A Bootstrap

analysis with 10 000 pseudo-replicates wasconducted to place confidence values ongrouping within the tree.

Results

Dengue serotypes and infectionpattern

A total of 53 dengue viruses were isolated from426 and 254 serum samples collected fromfebrile cases enrolled in 2007 and 2008

Venezuela/Aragua LARD5990 2000 DF AY146771Venezuela/Aragua LARD6218 2000 DHF AY146766Venezuela/Aragua LARD7110 2001 DHF AY146776Venezuela/Aragua LARD7812 2001 DHF AY146777Venezuela/Aragua LARD7984 2001 DHF AY146778

Isolates used as outgroup in this studyLocation* Strain Year Accession no.Malaysia LN6083 1994 AF147460China 80-2 1980 AF317645Fiji 29472 1992 L11422India 1416 1984 L11424Indonesia 1280 1978 L11426Malaysia 1981 L11427Philippines H87 1956 L11423Puerto Rico PR6 1963 L11433Puerto Rico 1339 1977 AY146761Samoa 1696 1986 L11435Sri Lanka 1326 1981 L11431Sri Lanka 1594 1985 L11436Sri Lanka 260698 1989 L11437Sri Lanka 2783 1991 L11438Thailand D86-007 1986 L11441Thailand D88-303 1988 AY145714Thailand D95-0014 1995 AY145724Thailand D97-0106 1997 AY145728

Location* Strain Year Clinical Accessionstatus no.

* Country and/or state or city



respectively. The serotypes detected wereDENV-1 (n=29), DENV-2 (n=9) and DENV-3(n=15), and no isolates of DENV-4 wasobtained. Primary infection was more frequentin DENV-1 (86.9%) and DENV-3 (73.3%) thanDENV-2 (55.6%)-infected patients.

DENV-3 predominance and DHF

In 2007, DENV-3 (42.2%) was the mostprevalent serotype followed by DENV-1(36.3%) and DENV-2 (21.2%). In 2008, incontrast, DENV-1 (85%) became the dominantserotype followed by DENV-2 (10%), whileDENV-3 (5%) was detected to a much lesserextent. This temporal relative abundance ofdengue serotypes coincided with an increasein the frequency of DHF cases with respect tothe period of the highest DENV-3 dominance.This is, from 4.6% (861/18 663; DENV-3=87.8%) between 2002 and 2004 to 25.9%(2922/11 673; DENV-3=28.3%) between2007 and 2008 (Tables 1 and 3).

DENV-3 phylogenetic diversity

The aligned final data set comprised 60sequences. The ingroup included 42 sequences

of the entire E gene (1479 bp in length) frompatients suffering from either DF or DHF/DSS(Figure). The Tamura and Nei plus Gamma (TrN+ Γ) model was the best fit to the data withan α value (shape parameter) of 0.24. Thesingle phylogenetic tree obtained revealed fivedifferent groups of DENV-3 viruses that couldbe assigned to genotypes. The analysis clearlydistinguished the two different genotypes (IIIand IV) detected in Latin America. All the2001–2004 Colombian DENV-3 isolatesgrouped into genotype III, along with virusesfrom Latin American countries that wereisolated after 1994. Viruses isolated from DHFpatients did not cluster apart with isolates fromDF patients. Consequently, there were nophylogenetically distinct groups related withdisease severity when using the envelopegene. Nonetheless, Latin American virusescould be grouped in two clades. One cladegrouped strains from Mexico, Venezuela,Colombia and Nicaragua, where the diseaseoutcome has been benign in the majority ofDENV-3-infected patients[13,17,24,36]. The secondclade grouped the isolates from Cuba, Braziland Paraguay, where infections resulted ineither fatalities or serious visceral and nervoussystem involvement[15,25,26].

Table 3: Annual temporal predominance of dengue virus serotype, Departamento deSantander, Colombia

YearSerotype: count (%)

SourceDENV-1 DENV-2 DENV-3 DENV-4

1998 8 (57.1) 6 (42.9) 0 0 *1999 2 (50) 2 (50) 0 0 *2000 1 (10) 7 (70) 0 2 (20) *2001 1 (4) 10 (40) 9 (36) 5 (20) *2002 1 (6.7) 1 (6.7) 13 (86.6) 0 *2003 0 1 (2.7) 37 (97.3) 0 *2004 1 (4.8) 2 (9.5) 15 (71.4) 3 (14.2) *2007 12 (36.3) 7 (21.2) 14 (42.2) 0 §2008 17 (85) 2 (10) 1(5) 0 §

*: Ocazionez et al.[21]. §: this study.



Figure: Maximum likelihood phylogenetic tree of 60 DENV-3 envelope gene sequences.Viruses are listed by abbreviation for country, year and strain (Table 2). Bootstrap values above

50 are shown above the branches.



Discussion

Our results show that at least three dengueserotypes have been simultaneously present inthe Departamento de Santander between 2007and 2008, and that the relative abundance ofserotypes had a distinct pattern each year. Toour knowledge, dengue virological surveillancein 2005 and 2006 was not carried out inSantander. Under these circumstances, we useddata from previous studies in the same regionof the country for the period 1998–2006 toidentify the relationship between the occurrenceof DHF and the abundance of DENV-3.

DENV-3 was the predominant serotype inSantander in the period 2002–2004, whileDENV-1, DENV-2 and DENV-4 were found inconsiderably lower frequency. This temporalserotypes distribution coincided with a decreasein the frequency of DHF with respect to theprevious year[21,24]. Between 2007 and 2008, incontrast, DENV-3 declined and DENV-1 andDENV-2 increased, and the frequency of DHFwas six-fold higher with respect to 2002–2004(Tables 1 and 3). Although the severity ofdengue in accordance with WHO parameters[37]

in patients enrolled in the present study wasnot determined, however, in previousstudies[21,24] conducted in Santander, DHF wasless frequent in DENV-3 patients compared withDENV-2-infected patients (10.9% versus 27.5%),and that there was no DSS or fatal case causedby DENV-3. Moreover, the period of the highestpredominance of DENV-3 coincided with adecrease of DHF, compared with the period ofDENV-2 dominance.

Phylogenetic analyses of isolates of DENVfrom severe cases or DHF epidemics suggestthat viral factors can have an influence on theoutcome of viral infection[8]. In this study,segregation of DF- versus DHF-associatedviruses on the basis of E gene sequences wasnot observed. This finding is in agreement witha study of DENV-3 viruses from Venezuela[17].

Likewise, Miagostovich et al.[38] did not find anydifferences among the untranslated region(UTR) sequences of viruses isolated from fatalor DF patients in Brazil. Additional studiesinvestigating other genes within the DENV-3viruses are necessary to infer with morecertainty the genetic basis of virulence.

It seems that the strains of DENV-3genotype III circulating in the Americas exhibitdifferent pathogenic potential. During the 1998epidemic in Nicaragua, 11.8% out of DENV-3-infected patients developed DHF and fatalitieswere not registered[36]. In Venezuela, eventhough the virus caused the largest epidemicseen after 1989, during the period of its highestpredominance only 8% of dengue cases weresevere and death in DENV-3-infected patientswas not reported (Dirección de Epidemiologíay Analisis Estratégico, 2001). Although, inMexico, the severity of dengue increased inthe mid-1990s after the introduction of DENV-3, the continuous presence of Asian genotypeof DENV-2 seems to have played a moreimportant role in DHF outbreaks[13]. On thecontrary, the introduction of DENV-3 into Brazil,Paraguay and Cuba was associated with severedisease. The virus caused an epidemic in thestate of Rio de Janeiro with 1831 DHF cases,and a total of 91 deaths in DENV-3-infectedpatients[16]. In Paraguay, the predominance ofthe virus resulted in 28 129 dengue cases, outof which 55 were haemorrhagic and 14 endedfatally. The fatality rate was 25.4%[26]. In Cuba,12 886 dengue cases occurred during the 2001DENV-3 epidemic, of which 70 suffered fromDHF and 3 of them ended fatally[15].

Our phylogenetic analyses clearly revealedthat isolates of DENV-3 from Colombia belongto the same genotype as isolates from Brazil,Paraguay and Cuba. Why did the predominanceof the virus result in epidemics with a distinctseverity of dengue? One explanation involvesgenetic diversity within the isolates of the virus.It has been suggested that DENV-3 viruses



might have gone through a period of in situevolution within Latin American countries afterits introduction, diversifying into distinctphylogenetic groups[17]. A genetic shift in DENV-3 of genotype III has been suggested as thecause for the emergence of an invasive strainresponsible for the increased frequency of DHFin Sri Lanka[9,28]. We found that isolates comingfrom countries where DENV-3 has not beenassociated with severe disease (Colombia,Venezuela and Mexico) grouped apart fromisolates coming from countries where the virushas caused deaths (Brazil, Paraguay andCuba)[19,39,40]. Additional studies are necessaryto evaluate whether or not a genetic shift inviruses from Latin America is occurring.

Other aspects might influence the severityof dengue during the DENV-3 outbreaks. Virus-specified determinants of virulence at the levelof susceptibility to cross-neutralization, moreprone to enhancement by dengue antibodies,and variation in the ability to infect and betransmitted by their mosquito vector, have beenproposed[27]. On the other hand, human geneticresistance to infection caused by viral strains morevirulent, and herd immunity of the respectivepopulations, should also be considered.

An observation that caught our attentionin the present study was the increasedincidence of DHF during the period 2005–2008(25%) compared with the period 1998–2004(4.5%). We did not investigate the temporaldistribution of dengue serotypes in Santanderin 2005–2006. Nonetheless, we conducted astudy in the municipality of Ocaña, located 299km from Bucaramanga in the north-eastern partof Santander (data not published). In Ocaña,in 2005, DENV-2 (77.5%) was the dominantserotype followed by DENV-3 (12.5%), andDENV-1 (5%) was isolated to a much lesserextent. In 2006, DENV-1 (57.5%) becamedominant, followed by DENV-3 (27.5%) andDENV-2 (22.5%). We can speculate that inBucaramanga the predominance of DENV-2

and DENV-1 could have also increased in 2005and 2006, respectively, despite DENV-3continuing as the prevalent serotype; and, inthis context, the frequency of DHF cases couldhave increased. One explanation could be thatthe co-dominance of DENV-2 and DENV-3 in2001 coincided with increased DHF cases inBucaramanga during the period 1998–2004(Tables 1 and 3), and that DENV-1 and DENV-2-infections were more associated with DHFthan DENV-3-infections[21]. In Nicaragua (1999–2003), the predominance of DENV-2 wasassociated with infections with shock to agreater extent, and the predominance ofDENV-1 with an increased number of infectionswith severe manifestations[41]. In addition, themarked increase of DHF in Latin America hasbeen largely attributed to the increasedfrequency of DENV-2 infections[4,42].

We cannot exclude that the increase ofDHF cases in Santander between 2005 and2008 could be due to the failure of physiciansin clinics to collect sufficient data to fulfil therequirements for the WHO case definition. Inthe case of some patients, serial haematocrittests required to estimate the degree ofhaemoconcentration were not available; andas such, patients with platelet counts below100 000/mm3 and/or with haemorrhages mighthave been classified as a DHF case.

Taken together, our results in this studyand previous studies suggest that the presenceof DENV-3 in the north-central part ofColombia since 2001 has had a minor role inthe occurrence of DHF. More studies need tobe conducted to clarify the pathogenic potentialof the virus in Colombia. The phylogeneticanalysis suggests that DENV-3 Colombianviruses could be genetically distinct from theviruses of the same serotype coming fromneighbour countries with potential to causeDHF. Thus, continuous virological surveillanceshould be a priority in the Colombian dengueendemic areas.



Acknowledgments

We thank Luis Gualdrón from the SecretariaDepartamental de Salud, Bucaramanga. Thisstudy was funded by the Gobernación de

Santander, Secretaria Departamental de Salud,Bucaramanga (Grant # 2007-068000-0070) andthe Universidad Industrial de Santander (Grant# 5636) from Colombia.

References

[1] Lindenbach BD, Rice CM. Flaviviridae: Theviruses and their replication. In: Knipe DM,Howley PM. Ed. Fields virology. 4th ed.Philadelphia: Lippincott Williams and Wilkins,2001. pp. 991-1041.

[2] Gubler DJ. Dengue and dengue haemorrhagicfever. Clin Microbiol Rev. 1998; 11: 480-496.

[3] Burke DS, Nisalak A, Johnson DE, Scott RM. Aprospective study of dengue infections inBangkok. Am J Trop Med Hyg. 1988; 38:172-180.

[4] Kouri GP, Guzmán MG, Bravo JR, Triana C.Dengue haemorrhagic fever/dengue shocksyndrome: lessons from the Cuban epidemic,1981. Bull WHO. 1989; 67: 375-380.

[5] Gubler DJ, Reed D, Rosen L, Hitchcock JR.Epidemiologic, clinical, and virologicobservations on dengue in the kingdom ofTonga. Am J Trop Med Hyg. 1978; 27: 581-589.

[6] Watts DM, Porter KR, Putvatana P, Vasquez B,Calampa C, Hayes CG, Halstead SB. Failureof secondary infection with Americangenotype dengue 2 to cause denguehaemorrhagic fever. Lancet. 1999; 354:1431-1434.

[7] Rosen L. The Emperor’s New Clothes revisited,or reflections on the pathogenesis of denguehemorrhagic fever. Am J Trop Med Hyg. 1977;26: 337-343.

[8] Rico-Hesse R, Harrison L, Salas R, Tovar D,Nisalak A, Ramos C, Boshell J, de Mesa M,Nogueira R, Travassos da Rosa A. Origins ofdengue type 2 viruses associated withincreased pathogenicity in the Americas.Virology. 1997; 230: 244-251.

[9] Messer WB, Gubler DJ, Harris E, SivananthanK, de Silva AM. Emergence and global spreadof a dengue serotype 3, subtype III virus. EmergInfect Dis. 2003; 9: 800-809.

[10] Lanciotti RS, Lewis JL, Gubler DJ, Trent DW.Molecular evolution and epidemiology ofdengue-3 viruses. J Gen Virology. 1994; 75:65-75.

[11] Wittke VR, Thu TE, Nisalak HM, NimmannityaA, Kalayanrooj S, Vaughn DW, Endy TP, HolmesEC, Aaskov JG. Extinction and rapidemergence of strains of dengue virus duringan interepidemic period. Virology. 2002; 301:148-156.

[12] CDC – Centers for Disease Control. Dengue type3 infection-Nicaragua and Panama, October-November, 1994. MMWR. 1995; 44: 21-24.

[13] Briceño-García B, Gómez-Dantés H, Argot-Ramirez E, Montesano R, Vázquez-MartínezAL, Ibañez-Bernal S, Madrigal-Ayal G, Ruiz-Matus C, Flisser A, Tapia-Coyner R. Potentialrisk for dengue hemorrhagic fever: the Isolationof serotype Dengue-3 in Mexico. Emerg InfectDis. 1996; 2: 133-135.

[14] Rigau J, Ayala A, García E, Hudson S, VordamV, Reiter P, Cano M, Clark G. The reappearanceof dengue-3 and a subsequent dengue-4 anddengue-1 epidemic in Puerto Rico. Am J TropMed Hyg. 2002; 67: 355-362.

[15] Peláez O, Guzmán MG, Kouri G, Pérez R, SanMartín JL, Vásquez S, Rosario D, Mora R,Quintana I, Bisset J, Cancio R, Mesa AM, CastroO, González D, Avila LC, Rodriguez R, AlvarezM, Pelegrino JL, Bernardo L, Prado I. Dengue3 Epidemic, Havana, 2001. Emerg Infect Dis.2004; 10: 719-722.



[16] Nogueira RM, Schatzmayr HG, Filippis AMB,Santos FV, Cunha RV, Coehlo JO, Souza LJ,Guimarães FR, Araújo ESM, De Simone TS,Baran M, Texeira G, Miagostovich MP. Denguevirus type 3, Brazil, 2002. Emerg Infect Dis.2005; 11: 1376-1381.

[17] Uzcategui N, Comach G, Camacho D,Salcedo M, Cabello de Quintana M, JimenezM, Sierra G, Cuello de Uzcategui R, James WS,Turner S, Holmes E, Gould E. Molecularepidemiology of dengue virus type 3 inVenezuela. J Gen Virology. 2003; 84:1569-1575.

[18] Aquino VH, Anatriello E, Gonçalves PF, DaSilva EV, Vasconcelos PF, Vieira D, Batista W,Bobadilla M, Vazquez C, Moran M,Figuereido LT. Molecular epidemiology ofdengue type 3 in Brazil and Paraguay, 2002-2004. Am J Trop Med Hyg. 2006; 75:710-715.

[19] Barrero PR, Mistchenko AS. Genetic analysisof dengue virus type 3 isolated in Buenos Aires,Argentina. Virus Res. 2008; 135: 83-88.

[20] Montoya Y, Holechek S, Caceres O, PalaciosA, Burans J, Guevara C, Quintana F, Herrera V,Pozo E, Anaya E, Mamani E, Guitierrez V,Ladron de Guevara A, Fernandez E, Asmat P,Alva-Dávalos V, Holguin C, Laguna VA, MoralesAM, Minaya P, Kochel T. Circulation of dengueviruses in north-western Peru, 2000-2001.Dengue Bulletin. 2003; 27: 52-62.

[21] Ocazionez RE, Cortés FM, Villar LA, GómezSY. Temporal distribution of dengue virusserotypes in Colombian endemic area anddengue incidence. Re-introduction of dengue-3 associated to mild febrile illness and primaryinfection. Mem Inst Oswaldo Cruz. 2006; 101:725-731.

[22] Barcelos FL, Batista CA, Portela FG, Paiva DB,Germano de Oliveira J, Bonjardim CA,Peregrino PC, Geessien E. Dengue virus 3genotype 1 associated with dengue fever anddengue hemorrhagic fever, Brazil. Emerg InfectDis. 2008; 14: 314-316.

[23] Nogueira MB, Stella V, Bordignon J, Batista WC,De Borda L, Pereira LH, Hoffmann FG, ProbstCM, Duarte CN. Evidence for the co-circulation of dengue virus type 3 genotypesIII and V in the Northern region of Brazil duringthe 2002-2004 epidemics. Mem Inst OswaldoCruz. 2008; 103: 483-388.

[24] Ocazionez RE, Gómez SY, Cortés FM. Denguehemorrhagic fever serotype and infectionpattern in a Colombian endemic area. RevSalud Publica (Bogota). 2007; 9: 262-274.

[25] Passos M, Santos L, Pereira M, Casali CG, FortesB, Ortiz Valencia L, Alexandre J, Medronho R.Clinical differences observed in patients withdengue caused by different serotypes in theepidemic of 2001/2002, occurred in Rio deJaneiro. Rev Soc Bras Med Trop. 2004; 37:293-295.

[26] Pan American Health Organization, RegionalProgram on Dengue through the PAHO/WHORepresentation in Paraguay (PAHO-Paraguay).Ministry of Public Health and SocialWelfare(Ministerio de Salud Publica y Bienestar Social/MSPBS) of Paraguay in EID Updates, Regionof the Americas, 2007. Dengue OutbreakContinues to Subside in Paraguay. Emergingand Reemerging Infectious Diseases 4.

[27] Dash PK, Parida MM, Saxena P, Abhyankar A,Singh CP, Tewari KN, Jana AM, Sekhar K, RaoPV. Reemergence of dengue virus type-3(subtype-III) in India: Implications forincreased incidence of DHF & DSS. Virol J.2006; 6; 3:55.

[28] Messer WB, Virarana UT, Sivananthan K,Elvtigala J, Preethimala LD, Ramesh R, WithanaN, Gubler DJ, de Silva AM. Epidemiology ofdengue in Sri Lanka before and after theemergence of epidemic dengue hemorrhagicfever. Am J Trop Med Hyg. 2002; 66: 765-773.

[29] García CA, Gutiérrez M, Ramirez AN, NiñoME. Estrategia de Gestión Integrada (EGI) parala prevención y control del dengue enSantander. Revista del Observatorio de SaludPública de Santander. 2007; 2: 31-39.



[30] Álvarez A, Martínez H, Millán S, López L.Diagnóstico de la Situación de Salud enSantander. 2nd ed. Santander : Secretaría deSalud de Santander, 2004. pp. 56.

[31] Lanciotti RS, Calisher CH, Gubler DJ, ChangGJ, Vorndam AV. Rapid detection and typingof dengue viruses from clinical samples byusing reverse transcriptase-polymerase chainreaction. J Clin Microbiol. 1992; 30:545-551.

[32] Edgar RC. MUSCLE: multiple sequencealignment with high accuracy and highthroughput. Nucleic Acids Res. 2004; 32:1792-1797.

[33] Posada D, Crandall KA. Selecting the Best-FitModel of Nucleotide Substitution. Syst Biol.2001; 50: 580–601.

[34] Posada D, Crandall KA. Modeltest: testing themodel of DNA substitution. Bioinformatics.1998; 14: 817-818.

[35] Guindon S, Gascuel O. A simple, fast, andaccurate algorithm to estimate largephylogenies by maximum likelihood. Syst Biol.2003; 52: 696-704.

[36] Harris E, Videa E, Perez L, Sandoval E, Tellez Y,Perez ML, Cuadra R, Rocha J, Idiaquez W,Alonso RE, Delgado MA, Campo LA, AcevedoF, Gonzalez A, Amador JJ, Balmaceda A.Clinical, epidemiologic, and virologic featuresof dengue in the 1998 epidemic in Nicaragua.Am J Trop Med Hyg. 2000; 63: 5-11.

[37] World Health Organization. Guide fordiagnosis, treatment and control of denguehemorrhagic fever. 2nd edn. Technical AdvisoryCommittee on Dengue hemorrhagic fever forthe South East Asian and Western Pacificregions. Geneva: WHO, 1980.

[38] Miagostovich MP, dos Santos FB, Fumian TM,Guimarães FR, da Costa EV, Tavares FN,Coelho JO, Nogueira RM. Complete geneticcharacterization of a Brazilian dengue virustype 3 strain isolated from a fatal outcome.Mem Inst Oswaldo Cruz. 2006; 101:307-313.

[39] Kochel T, Aguilar P, Felices V, Comach G, CruzC, Alava A, Vargas J, Olson J, Blair P. Molecularepidemiology of dengue virus type 3 inNorthern South America: 2000-2005. InfectGenet Evol. 2008; 8: 682-688.

[40] Diaz FJ, Black WC, Farfán-Ale JA, Loroño-PinoMA, Kenneth E. Olson KE; Beaty BJ. Denguevirus circulation and evolution in Mexico: aphylogenetic perspective. Arch Med Res. 2006;37: 760-773.

[41] Balmaseda A, Hammond SN, Pérez L, TellezY, Saborío SI, Mercado JC, Cuadra R, Rocha J,Pérez MA, Silva S, Rocha C, Harris E. Serotype-specific differences in clinical manifestationsof dengue. Am J Trop Med Hyg. 2006; 74:449-456.

[42] Guzman MG, Kouri G. Dengue and denguehemorrhagic fever in the Americas: lessons andchallenges. J Clin Virol. 2003; 27: 1-13.


Applied informatics manipulation for fightagainst dengue

Viroj Wiwanitkit#

Wiwanitkit House, Bangkhae, Bangkok, Thailand 10160

Abstract

Dengue is an important tropical infectious disease. This infection has been widely studied and there aremany reports on its pre-clinical and clinical aspects. In the era of information technology, scientists cansuccessfully manipulate the large amount of information available on dengue for use in the diagnosis,treatment and control of this disease. In this article, the author briefly summarizes and comments onapplied informatics manipulation for the prevention and control of dengue.

Keywords: Dengue; Informatics; Manipulation.


Introduction

Dengue is an important tropical infectiousdisease that has become a focused disease andits effective control for sustainability requiresall-round efforts. There are many researchstudies on different aspects of this effort:epidemiology, clinical, diagnostics, treatment,and prevention and control of dengue. Varioustypes of dengue researches can be summarizedby their characteristics into pre-clinical andclinical studies. For pre-clinical study, scientistsusually study the pathophysiology of dengueby standard methods of scientific research. Forclinical study, physicians are concerned withdiagnosis and therapeutic research on dengue.

Basically, these dengue research areas canbe divided into four main groups. The first grouprelates to descriptive study focusing mainly on

the natural history of dengue. The goodexamples are studies on dengue prevalenceand clinical cases summarization. The secondgroup pertains to analytical study. This kind ofwork focuses mainly on the cause and resultto derive odd ratio and risk estimation. Thethird group of studies focuses on thedevelopment of sensitive and dengue-specificdiagnostic tests. The last group of work is theexperimental group. This mainly focuses onclinical studies for dengue as previouslymentioned. This kind of work is considereduseful for establishing of real-time clinicalpractices. However, in addition to theseclassical approaches, use of database andinformation technology is important. In the eraof information technology, scientists cansuccessfully scan the heap of informationavailable on dengue for usage in diagnosis,treatment and control of the disease. This

Applied informatics manipulation for fight against dengue


article tries to briefly summarize the applicationof informatics technology for the developmentof new technologies/strategies in all aspects ofcontrol and management of DF/DHF.

Database searching fornatural history study ofdengue

As previously mentioned, case summarizationis a basic descriptive study for dengue research.In the past, the summarization could beundertaken when there was enough data foranalysis. The analysis can be done only if thesample size of the data is sufficient for statisticalanalysis. However, with advances in databasetechnology, there are numerous databases inmedicine, which can serve as primary datasource. Of several databases, there is aninteresting specific database, DengueInfo,which can be a good gateway to dengueinformation resources[1]. Many useful data canbe freely accessed using this database.

There are many new research studieswhere scientists have made use of databasesearching and utilized the same in metanalysistechnique to derive the natural history ofdengue. The best example is the series ofCochrane Reviews. The paper by Panpanich etal. is within this group[2]. Panpanich et al. carriedout metanalysis on the use of corticosteroidsfor treating dengue shock syndrome (DSS)[2].In this review, Panpanich et al. searched theCochrane Infectious Disease Group SpecializedRegister (January 2006), CENTRAL (TheCochrane Library 2005, Issue 4), MEDLINE(1966 to January 2006), EMBASE (1974 toJanuary 2006), LILACS (1982 to January 2006),and concluded that there was insufficientevidence to confirm the efficacy of the use ofcorticosteroids in managing dengue shocksyndrome[2]. Wiwanitkit recently reported onliver dysfunction due to dengue infection by

an analysis of the previously published Thaicases[3]. In this work, Wiwanitkit used Pubmedand Thai Index Medicus database search forderiving primary data to conclude his researchresults[3]. According to this work, Wiwanitkitnoted the importance of detection of abnormalhigh transaminase enzyme among patients withdengue infection, which consequentlydeveloped into hepatic encephalopathy[3].Wiwanitkit also used a similar technique tosummarize the magnitude and pattern ofneurological pathology in fatal denguehaemorrhagic fever and found that neurologicalpathology was also common[4].

Prediction for pathobiologymechanism in dengue

With the advent of computational biology,some new research studies have been carriedout on predictive pathobiology of dengue bymeans of “omics” science. Indeed, there aremany newly launched bioinformatics tools thatcan be applied for dengue research. The basictools on genomics and proteomics can providemany new data to the scientific community.For example, Wiwanitkit recently explained thepathobiology of DSS by means of predictivebioinformatics. Firstly, Wiwanitkit usedphylogenomics analysis to predict thephylogenetic interrelationship and reported thatplatelet CD61 might have an important role incausing haemorrhagic complication in dengueinfection[5]. However, Wiwanitkit furtherclarified by means of gene ontology that therewas no existence of functional similaritybetween DNS1 and CD61[6]. A study onfunctional similarity between dengue non-structural protein 1 and platelet integrin/adhesinprotein, CD61, showed no confirmativeresult[6]. Secondly, Wiwanitkit also focused thework on the immune complex generation indengue. Wiwanitkit successfully demonstratedweak binding affinity of immunoglobin G,which could be an explanation for the immune



mimicking theory in pathophysiological findingsin the recovery phase of dengue[7]. Themolecular docking technique was mainly usedfor this predictive-specific study[7]. Wiwanitkitalso used the docking study to estimate thesize of immune complex and reported itspossible relationship to renal pathology indengue[8]. The author concluded that becauseentrapment of the immune complex isbelieved to occur when a previous glomerularlesion causes narrowing of the glomerulus’sdiameter, the immune complex should not havea significant role in the pathogenesis of renalfailure in dengue infection[8]. Hibberd et al.used a genomics approach to understand thehost response during dengue infection andfound many new host pathways involved inviral replication in vitro, and also host immuneresponses that were influenced by viralsequence[9]. For the nature of outbreak, Halideand Ridd recently used a predictive model fordescribing dengue haemorrhagic feverepidemics[10]. Halide and Ridd also found thatthe most important determinant in thepredictive model was the present number ofcases followed by the relative humidity threeto four months previously[11].

In silico mapping of denguevirus epitopes

Epitope finding is the basic principle in appliedimmunology. This activity is useful forunderstanding the immunopathology ofinfectious diseases as well as to help searchfor vaccine candidate. In silico mapping ofdengue virus epitopes is the current usefulapplication of bioinformatics technology indengue research. Kutubuddin et al. used basicbioinformatics to describe recognition ofCD4(+) T cell epitopes in envelope (E)glycoprotein of Japanese encephalitis, WestNile and dengue viruses[12]. In this work,analysing the occurrence of amphipathic

segments, Rothbard-Taylor tetra/pentamermotifs and presence of alpha helix-preferringamino acids were used for epitopesprediction[12]. Wen et al. recently usedcomputational prediction to identify denguevirus-specific CD4(+) T-cell epitopes[13].According to this work, As a result, C(45-57)(KLVMAFIAFLRFL), E(396-408) (SSIGKMFEATARG),NS3(23-35) (YRILQRGLLGRSQ), and NS3(141-155) (NREGKIVGLYGNGVV) were theidentified epitopes[13]. A similar work was alsorecently published by Leclerc et al.[14].

Vaccine search by means ofimmunomics

Immunomics is the new specific “omics”science for the study of epitope for productionof new vaccines. Immunomics is presentlyfocused on vaccine research. For dengue, thedisease without an effective vaccine,immunomics can be useful. In 2007, Khan etal. introduced a systematic bioinformaticsapproach for the selection of an epitope-basedvaccine, targeted to assess its efficacy againstdengue[15]. In this study the number of uniqueprotein sequences required to representcomplete antigenic diversity of short peptidesin dengue virus is significantly smaller than thatrequired to represent complete proteinsequence diversity[15]. Recently, in 2008, Khanet al. also published another work on theidentification of conservation and variability ofdengue virus proteins by mean ofbioinformatics[16]. In 2007, Mazumder et al.reported on computational analysis andidentification of amino acid sites in dengue Eproteins relevant to development of diagnosticsand vaccines[17]. They found that six singularsites [N(37), Q(211), D(215), P(217), H(244),K(246)] in dengue E protein that wereconserved, were part of the predictedconsensus T(h)-cell epitopes and were exposedin the dimer or trimer[17]. They also proposed



these sites and corresponding epitopic regionsas potential candidates for prioritization byexperimental biologists for development ofdengue vaccines[17].

Computer-aided drug design

Computer-aided drug design is another usefulbioinformatics application in dengue research.Similar to vaccine search, it is possible to searchfor a new antiviral compound for dengue basedon bioinformatics technology. For example, Zhouet al. recently used virtual screening of small-molecule libraries against dengue virus E proteinto identify new antiviral compound for dengue[18].According to the study of Zhou et al., P02, anew compound with a change for thedevelopment of an effective treatment againstdengue virus and related flaviviruses could beidentified[18]. Luzhkov et al. used a similartechnique, virtual screening and bioassay study,to find new inhibitors for dengue virus mRNAcap (nucleoside-2’O)-methyltransferase[19].According to this study, a novel inhibitor, with apreviously unknown scaffold that has an IC(50)value of 60 microM, could be identified[19]. Yanget al. also used combinatorial computationalapproaches to identify tetracycline derivatives asflavivirus inhibitors[20]. Yang et al. described thatrolitetracycline and doxycycline were the twocompounds that have their inhibitory effect ondengue virus propagation with IC(50)s estimatedto be 67.1 microM and 55.6 microM,respectively[20].

Plan for dengue controlby GIS

Finally, it should also be noted that not onlythe information in textual format but also infigure format can be manipulated. TheGeographic Information System (GIS) is a good

example of figure format data manipulation.There are some recent interesting reports onGIS and dengue. Wiwanitkit reported anobservation on the correlation between rainfalland the prevalence of clinical cases of denguein Thailand[21]. In this work, the collectedprimary data in general report was used forfurther manipulation and a predictive map wasalso generated. A similar work was alsoreported by Bonet et al. in the Latin America-Caribbean region[22]. In addition, GIS can alsobe applied for the actual field study data. Thisis usually used for the vector or mosquitosurvey. A good example is the paper byChansand and Kittayapong[23]. According to thiswork, the immature survey data and the GPScoordinates of house location were combinedinto GIS maps showing the distribution ofimmature density and clustering of immaturestages and positive containers in the studyarea[23]. The authors concluded that thisapproach could be used to improve theefficiency and accuracy of dengue vectorsurveillance for targeting vector control[23].Vanwambeke et al. concluded that the greatvariation of determinants for recent dengueinfection in space and time should be takeninto account when designing local denguecontrol programmes with the help of GIS[24].

Diagnosis and prediction ofdisease outbreaks

As mentioned before, the bioinformaticstechnology is useful for the assessment ofpathobiology of dengue. The diagnosis andprediction of disease outbreaks can successfullybe performed by the bioinformatics technique.Shaw et al. recently used dimension reductionto improve outbreak predictability of multistraindiseases including dengue[25]. Shaw et al.showed that this technique allowed the centreto use manifold equations, which are mainlyused for prediction and are applicable even to



noisy systems[25]. Jackson et al. also proposeda new technique, mass cataloguing, based onmass spectrometry and genotyping, to identifyoutbreaks of flaviviruses including dengue[26].For dengue, the method can help distinguishmajor subgroupings within each serotype[26].

Informatics analysis ofdengue vector genomes

Not only the applications for the dengue virusbut also for the dengue vector can beperformed based on advanced informaticstechnology. Informatics analysis of denguevector genomes is presently performed inentomology. For example, Lobo et al.performed analysis of 14 BAC sequences from

the Aedes aegypti genome[27]. The data fromthis study is useful for genome annotation andassembly[27]. Takahashi et al. used mathematicalmodels for the Ae. aegypti dispersaldynamics[28]. Travelling waves by wing and windcould be identified in this study[28].

Conclusion

This paper has tried to review the applicationof informatics for the diagnosis, treatment andcontrol of dengue virus. As the field of denguevirus research has seen an increased applicationof informatics over the years, presentation ofa review summarizing informatics applications,their impacts and future potentials seems tobe in order.

References

[1] Schreiber MJ, Ong SH, Holland RC, HibberdML, Vasudevan SG, Mitchell WP, Holmes EC.Infect Genet Evol. 2007 Jul; 7(4): 540-1.

[2] Panpanich R, Sornchai P, KanjanaratanakornK. Corticosteroids for treating dengue shocksyndrome. Cochrane Database Syst Rev. 2006Jul 19; 3: CD003488.

[3] Wiwanitkti V. Liver dysfunction in Dengueinfection: an analysis of the previouslypublished Thai cases. J Ayub Med CollAbbottabad. 2007 Jan-Mar; 19(1):10-2.

[4] Wiwanitkti V. Magnitude and pattern ofneurological pathology in fatal denguehemorrhagic fever: a summary of Thai cases.Neuropathology. 2005 Dec; 25(4): 398.

[5] Wiwanitkit V. Platelet CD61 might have animportant role in causing hemorrhagiccomplication in dengue infection. Clin ApplThromb Hemost. 2005 Jan;11(1): 112.

[6] Wiwantikti V. A study on functional similaritybetween dengue non structural protein 1 andplatelet integrin/adhesin protein, CD61. J AyubMed Coll Abbottabad. 2006 Jul-Sep;18(3):13-6.

[7] Wiwanitkit V. Weak binding affinity ofimmunoglobin G, an explanation for theimmune mimicking theory inpathophysiologic findings in the recoveryphase of dengue. Nanomedicine. 2005 Sep;1(3): 239-40.

[8] Wiwanitkti V. Immune complex: does it havea role in pathogenesis of renal failure indengue infection? Ren Fail. 2005; 27(6):803-4.

[9] Hibberd ML, Ling L, Tolfvenstam T, MitchellW, Wong C, Kuznetsov VA, George J, Ong SH,Ruan Y, Wei CL, Gu F, Fink J, Yip A, Liu W,Schreiber M, Vasudevan SG. A genomicsapproach to understanding host response



during dengue infection. Novartis Found Symp.2006; 277:206-14, 214-7, 251-3.

[10] Halide H, Ridd P. A predictive model fordengue hemorrhagic fever epidemics. Int JEnviron Health Res. 2008 Aug; 18(4): 253-65.

[11] Khan AM, Miotto O, Heiny AT, Salmon J,Srinivasan KN, Nascimento EJ, Marques ET Jr,Brusic V, Tan TW, August JT. A systematicbioinformatics approach for selection ofepitope-based vaccine targets. Cell Immunol.2006 Dec; 244(2): 141-7.

[12] Kutubuddin M, Kolaskar AS , Galande S, GoreMM, Ghosh SN, Banerjee K. Recognition ofhelper T cell epitopes in envelope (E)glycoprotein of Japanese encephalitis, westNile and Dengue viruses. Mol Immunol. 1991Jan-Feb; 28(1-2): 149-54.

[13] Wen JS, Jiang LF, Zhou JM, Yan HJ, Fang DY.Computational prediction and identificationof dengue virus-specific CD4(+) T-cellepitopes. Virus Res. 2008 Mar; 132(1-2): 42-8.

[14] Leclerc C, Dériaud E, Megret F, Briand JP, VanRegenmortel MH, Deubel V. Identification ofhelper T cell epitopes of dengue virus E-protein. Mol Immunol. 1993 May; 30(7):613-25.

[15] Khan AM, Heiny AT, Lee KX, Srinivasan KN,Tan TW, August JT, Brusic V. Large-scale analysisof antigenic diversity of T-cell epitopes indengue virus. BMC Bioinformatics. 2006 Dec18; 7 Suppl 5: S4.

[16] Khan AM, Miotto O, Nascimento EJ, SrinivasanKN, Heiny AT, Zhang GL, Marques ET, TanTW, Brusic V, Salmon J, August JT. PLoS NeglTrop Dis. 2008 Aug 13; 2(8): e272.

[17] Mazumder R, Hu ZZ, Vinayaka CR, SagripantiJL, Frost SD, Kosakovsky Pond SL, Wu CH.Computational analysis and identification ofamino acid sites in dengue E proteins relevantto development of diagnostics and vaccines.Virus Genes. 2007 Oct; 35(2): 175-86.

[18] Zhou Z, Khaliq M, Suk JE, Patkar C, Li L, KuhnRJ, Post CB. Antiviral compounds discoveredby virtual screening of small-molecule librariesagainst dengue virus E protein. ACS Chem Biol.2008 Dec 19; 3(12): 765-75.

[19] Luzhkov VB, Selisko B, Nordqvist A, PeyraneF, Decroly E, Alvarez K, Karlen A, Canard B,Qvist J. Virtual screening and bioassay studyof novel inhibitors for dengue virus mRNAcap (nucleoside-2’O)-methyltransferase.Bioorg Med Chem. 2007 Dec 15; 15(24):7795-802.

[20] Yang JM, Chen YF, Tu YY, Yen KR, Yang YL.Combinatorial computational approaches toidentify tetracycline derivatives as flavivirusinhibitors. PLoS ONE. 2007 May 9; 2(5): e428.

[21] Wiwanitkit V. An observation on correlationbetween rainfall and the prevalence of clinicalcases of dengue in Thailand. J Vector BorneDis. 2006 Jun; 43(2): 73-6.

[22] Bonet M, Spiegel JM, Ibarra AM, Kouri G,Pintre A, Yassi A. An integrated ecosystemapproach for sustainable prevention andcontrol of dengue in Central Havana. Int JOccup Environ Health. 2007 Apr-Jun; 13(2):188-94.

[23] Chansang C, Kittayapong P. Application ofmosquito sampling count and geospatialmethods to improve dengue vectorsurveillance. Am J Trop Med Hyg. 2007 Nov;77(5): 897-902.

[24] Vanwambeke SO, van Benthem BH, KhantikulN, Burghoorn-Maas C, Panart K, Oskam L,Lambin EF, Somboon P. Multi-level analysesof spatial and temporal determinants fordengue infection. Int J Health Geogr. 2006 Jan18; 5: 5.

[25] Shaw LB, Billings L, Schwartz IB. Usingdimension reduction to improve outbreakpredictability of multistrain diseases. J MathBiol. 2007 Jul; 55(1): 1-19.



[26] Jackson GW, McNichols RJ, Fox GE, WillsonRC. Toward universal flavivirus identificationby mass cataloging. J Mol Diagn. 2008 Mar;10(2):135-41.

[27] Lobo NF, Campbell KS, Thaner D, DebruynB, Koo H, Gelbart WM, Loftus BJ, SeversonDW, Collins FH. Analysis of 14 BAC sequencesfrom the Aedes aegypti genome: a benchmarkfor genome annotation and assembly.Genome Biol. 2007; 8(5): R88.

[28] Takahashi LT, Maidana NA, Ferreira WC Jr,Pulino P, Yang HM. Mathematical models forthe Aedes aegypti dispersal dynamics:travelling waves by wing and wind. Bull MathBiol. 2005 May; 67(3): 509-28.


Community participation and social engagement inthe prevention and control of dengue fever in

rural Cambodia

Sokrin Khuna,b#, Lenore Mandersonb

aNational Center for Health Promotion, Ministry of Health, Phnom Penh, Cambodia

bSchool of Psychology, Psychiatry and Psychological Medicine, Faculty of Medicine, Nursing and HealthSciences, Monash University, Victoria 3800, Australia

Abstract

The prevention and control strategies for dengue fever require community involvement to succeed.Drawing on data collected in 2003–2004 as part of an ethnographic study in eastern Cambodia, weexplore the role of community participation and the factors that influence its success in the preventionand control of dengue fever in Cambodia. Community participation has the potential for effective andefficient control of the disease, but this is subject to how communities are engaged in specific activities.Historical, political, social and economic factors have undermined the social institutions and conventionsin the study villages that could facilitate community involvement. In particular, poverty and differencesin local interests influence the capacity for people to be involved. Villagers regarded the maintenanceof the domestic environment as a personal responsibility and were reluctant to extend their action to awider domain. Comprehensive programmes, which draw on local institutions and understandings ofcommunity and enable community members to participate in the planning and management ofprevention and control activities, are essential to ensure programme sustainability and effectiveness.

Keywords: Cambodia; Community participation; Dengue; Social engagement.

#E-mail: [email protected]; Tel.: + 61 3 03 4506; Fax: +613 9903 4508

Introduction

Community participation is a process ofengaging various stakeholders and members ofcommunities, however defined, to participatein the development and management ofparticular programmes or projects. It isconventionally represented as the lynchpin forthe success of targeted health interventions and

sustainable programmes. Various pilot projectshave been conducted, well described in theliterature in relation to the strategies andprocesses of participation[1-5]. However, littleresearch has been conducted on howcommunity participation, once under way, isperceived by community members, healthworkers or other stakeholders. Similarly, thetranslation of community participation from


Community participation in the prevention of dengue in Cambodia

policy to practice in health, development anddisease control programmes has received littleattention, and the effectiveness of communityparticipation both as a process and an outcomein disease prevention and control, including fordengue fever, remains unclear[2,3,6,7].

Cambodia adopted communityparticipation as one of the principles of itsprimary health care policy and health systemin 1999. Yet, dengue fever (DF), denguehaemorrhagic fever (DHF) and dengue shocksyndrome (DSS) remain critical public healthproblems. Because of the domestic habitat andbehaviour of Aedes mosquitoes, communityparticipation and effective health education arecentral to sustainable dengue prevention andcontrol. Most interventions in which communitymembers are urged to be involved arerelatively straightforward – maintaining safewater storage to prevent breeding, ensuringthat there are no pools of stagnant water,carefully disposing of hard waste and using localbiocontrol agents (e.g. copepods). Yet, therehas been uneven success in the sustainedcommitment of communities in DF/DHF-endemic areas to environmental management,in part because of the lack of attention to socialfactors influencing dengue transmission, andin part because of the superficial nature ofcommunity involvement.

In this article, we explore how Cambodianvillagers perceive community participation inthe prevention and control of dengue fever.We focus on women, as the central players indengue-related interventions, for it is they whoare assumed to be responsible for permittingcontrol programme staff to distribute temephos(as 1% sand granules with a dosage of 1 ppmper litre of water) and to fog and spray outsidehouses, and for maintaining the domesticenvironment, disposing of waste, ensuringlarvae-free water supplies, protecting childrenfrom bites, and taking appropriate action if andwhen their children are sick[8,9]. Women are

also more likely than men to face challengesin seeking health care for themselves and theirchildren, due to multiple demands on their timeand their lesser role in determining howhousehold cash resources are used[10]. Thisstudy tries to highlight how the socialinstitutions within the study villages, that havesupported community cooperation andreciprocity in the past, have been eroded inrecent decades for political and economicreasons.


Data were collected from March 2003 toFebruary 2004 in the province of KampongCham (KPC), eastern Cambodia, wheredengue has especially high prevalencecompared to other regions of the country.During 2002–2003, there were 3713 cases ofDF, DHF and DSS and 49 deaths in KPC. Inthe first eight months of 2007, there were5105 cases of DF/DHF/DSS and 65 deaths inKPC[11]. Dengue continues to be endemicdespite the National Dengue ControlProgramme (NDCP) that conducts activities inthe area.

Study area

An ethnographic study was conducted in twovillages with the highest reported incidence ofthe disease in the province. The villages, Khunand Nekry (pseudonyms), are locatedapproximately 30 kilometres in oppositedirections from the provincial town centre (alsoknown as Kampong Cham) and around 100kilometres from the Capital, Phnom Penh. Mostvillagers are poor farmers, growing rice forsubsistence and sale, supplementing this byselling other produce and re-selling non-foodgoods in small quantities. Environmentalconditions in the villages are conducive todengue transmission. Broken coconut shells,



plastic bags, used packages and otherdisposable items are indiscriminately scatteredin house yards and lanes; these provide idealconditions for the breeding of the vector, Aedesmosquito. Wooden houses are built on stiltsusing bamboo and thatch. Water jars, storedunder houses, are rarely covered and breedlarvae all the year round.

Data collection

Data collection methods included keyinformant interviews, focus group discussions,in-depth interviews and ongoing participantobservation, as well as structured observationsand entomological surveys[12]. Key informantinterviews were conducted with all villagehealth volunteers about villagers’ awarenessand perceptions about DF, and theirparticipation in prevention, control anddevelopment activities. Four focus groupdiscussions were conducted with mothers orother family caretakers of children who hadbeen infected with dengue, to gain insight intotheir understanding of DF and theirinvolvement in prevention and controlmeasures. In-depth interviews were thenconducted with 29 women whose children hadbeen infected in the past year or during theresearch period, including about theirparticipation in dengue prevention and controlactivities and their views about controlprogramme activities. Women were also askedabout changes in their village, and so couldspeak of issues about which they may havebeen reticent publicly. These data weresupplemented by questionnaires with 38 otherwomen, representing 15% of householdswhere children had no history of dengue, ontheir participation in prevention and controlactivities. Sixteen interviews were conductedwith health staff at health centres, provincialand national levels on community participation.All data were entered into computer, codeswere developed, and the data were analysedthematically.

Ethics approval to conduct this researchstudy was granted by the Human ResearchEthics Committee of The University ofMelbourne (Australia), the Ministry of Health(Cambodia), and WHO/TDR. All potentialparticipants were provided with plain languageParticipation Information Sheets in Khmer(Cambodia’s national language), and the projectwas explained to them verbally. Consent toparticipate was verbal, since the collection ofsignatures held negative connotations for mostpeople.

Results

Social engagement in villagesin Cambodia

The idea of community, if taken to refer not toco-location but also to a spirit of commonpurpose, shared identity and trust, has declinedin Cambodia as a result of its recent bitterhistory of autocracy, violence, genocide, andpoverty. Interviews with participants aged 60and older, born in both villages, illustrated theextent to which villages were believed to havechanged since the 1940s. Prior to politicalturbulence and civil war in Cambodia in 1970,development activities in the villages, primarilyroad maintenance and building wells, wasconducted under the leadership of Buddhistabbots and monks, senior citizens and localleaders. Under Pol Pot’s regime, in contrast,the population was forced to provide labourfor developmental purposes in the villages.

Today, villagers passively participate invillage development. Most roads and waterwells are constructed with substantial financialand logistic support and management from thegovernment, international organizations andnational NGOs. This partly reflects changes inideas of the role of government, and in somecases, the implementation style ofdevelopment agencies and NGOs. Villagers



have contributed some labour, such as diggingwater drains along the road in front of theirown houses. While this involvement issometimes relatively spontaneous, it oftenfollows the explicit requests of outsideorganizations, which undertake developmentactivities on the provision that householderscontribute in cash or kind. The rationale is thatby contributing, people will have greaterownership of the projects, and so willparticipate in their maintenance.

But various factors discourage communityengagement and involvement. Villagers believethat the level of trust among them decreasedsignificantly during the three decades ofdomestic political unrest since 1970, and thishas continued. During the period of KhmerRouge rule under Pol Pot (1975–1979), peoplewere divided into “locals” (nak mul tharn) and“evacuees” (nak chum leas). Most evacueeswere urban dwellers, sent to rural areas by forceto contribute to economic development, inwhich context “locals” exerted extremedominance and authority. Those who wereforcibly resettled were subject to random terror,murder, forced labour, starvation and absenceof basic services. This significantly underminedtrust and ties among villagers. As one womanexplained, “From the time of the Khmer Rouge,there were more and more unreliable peopleand there were fewer good people, most ofthem were cheats and they even cheated thegovernment” (In-depth interview, 11).

Elderly interviewees believed that the levelof cooperation among villagers – referred to inKhmer as provass, has gone down. They claimedthat in the past, fellow villagers wereenthusiastic about helping each other, and wereinvolved on a voluntary basis in a range ofactivities, from digging water wells to helpingto build entire houses; now almost all workinvolves payment. Political party representativesdonate gifts in cash or kind to villagers to gainpopularity, but there is little evidence of long-

term political commitment to development.Others make public donations to those whoare especially disadvantaged or vulnerable, ordonate to temples or schools when high profilepoliticians visit. But the motivation of theseactivities is questionable, and is seen byvillagers themselves as having less to do withaspirations for community development andmore with power and popularity.

The standard of living of villagers has alsodeclined. Most villagers reported that, overtime, they have had poorer crop yields as aresult of lack of irrigation and because ofdrought and floods in dry and rainy seasons.People reported that they can no longer makea profit from farming because of the expensesof petrol and hiring water pumps and oxen. Tomeet various recurrent and emergency cashneeds, including for medical care, villagers sellor pawn their land and other resources such ascows or pigs and take out loans with exorbitantinterest rates. Most villagers generate too littlecash to repay loans or retrieve property, andso lose their property and continue to payinterest, so spiralling into extreme poverty[13].

Villagers have to meet direct and indirectcosts for medical care to treat sick children.Delays occur because of the difficulty inlocating resources, and the poorest householdsat times have no choice but to rely on homecare only[13]. Villagers felt that the level of careand support from others, and from social welfareand public health services, had declined. Oneelderly woman reported that her neighboursused to accompany her at the Kampong ChamProvincial Hospital whenever she had a medicalproblem, because she had no relatives. Thesurgery was free of charge if a person was toopoor to pay the fee. Now, this is rare:

“I had a surgery in 1962 or 1963. Thesurgery was very clean and I was dischargedfrom the hospital after staying only one week.At first I was afraid that I had no money and



no relatives to look after me. My neighborssaid, you tell the health staff that you have nomoney, then the health staff will exempt thefee. And the doctor said, don’t worry, we won’tcharge you. I didn’t pay even one cent. I boughtsome bananas for the head of the ward. Butshe refused to take them and said to me, pleasekeep them for yourself, we don’t need anythingfrom you (In-depth interview, 36).”

Competing interests

People do not necessarily share the samehealth-related problems and socioeconomic orpolitical status that could result in collectiveaction[3,14]. This is true for dengue and for theprevention and control activities in the studyvillages. Key informants in Khun village arguedthat women whose children had been affectedby dengue were far more interested indiscussions and health education messagesthan those whose children had never haddengue. Similarly, mothers of small childrenwere more influenced than mothers of olderchildren because of the pervasive belief thatdengue was unlikely to affect older children.Other participants had other health priorities.But, in addition, participants mentioned lackof water for their rice fields and lack of moneyto buy petrol for water pumps to irrigatedrought-affected rice fields, buy piglets or calvesto raise them for sale, or buy oxen for farmwork. Others reported difficulties in findingemployment to earn money to meet their basicexpenses or to repay interest and loans to fellowvillagers. Dengue prevention was a minorconcern against these major problems.

Political divisions at the national levelstrongly influence village politics. The threemain parties in the country have their ownparty activists and supporters in villages, andeverywhere, billboards promote their particularinterests. In 2001, Commune Councils wereestablished, to be responsible for the

management and development of communes.Although their members were chosen throughgeneral election, candidates were electedunder the name of their political party. As aresult, each elected council member wasstrongly allied to his or her political party, usingvillage issues for political advantage and workingwithin the community to build networks fortheir own party and gather support for theirplatforms. The politicization of villagegovernment and of social relations withinvillages also appears to hamper the spirit oftogetherness in village development amongpeople with different political ideologies.

Village cooperation andparticipation

Development committees were established in1998 in Nekry and Khun villages, as in mostvillages nationwide; these included a villagedevelopment committee (VDC), schoolcommittee, temple committee, canalcommittee, water-well committee, women’sassociation, and village health volunteers (VHV)and others belonging to the village healthsupporting group (VHSG)[15]. The tasks of VHVsand VDC included providing health education,assisting with health outreach activities, givingfirst aid to villagers, and referring patients tohealth centres. Few of the committeesreceived technical and logistic support however,so they quickly dissolved. In the study villages,the VDC and VHSG are still prominent asentities but are not functional; many othercommittees no longer exist even in name. Asone member of a VHSG explained, “I had notime. I had nothing for my children to eat. Itwasn’t just me, everyone quitted” (Keyinformant interview, 38).

Even so, the cooperation of villagers iscritical for village development. Mostcooperation occurs through social and culturalactivities, but also through the exchange of



labour or provass. Provass is a Khmer term,connotative of reciprocity: according to Nekryand Khun villagers, the term means “workingtogether to share outcomes”. Three types ofprovass occur in the villages. At the time ofthe study, the first and dominant type of provasswas shared animal husbandry. Many householdsraised young pigs or cows which belonged toanother household, and would share with theowners any piglets or calves according to prioragreement. A second type of provass was theexchange of labour, when householders farmedfor a short period for another householderwhen extra labour was needed. The latterhouseholder would later work for the formerto return the labour. Provass also referred toborrowing oxen for farming, with the loanrepaid with rice or labour. Provass providesmutual benefits and only occurs betweenvillagers of relatively equal socioeconomicstatus. In Khun and Nekry, this meant that poorvillagers were generally excluded. But thisreciprocal assistance has also diluted as a casheconomy has taken hold in the villages andother types of help no longer exist as familieslack the means to reimburse in cash or kind:“Nobody helped. The word ‘help’ doesn’texist. I had no oxen to farm so I had to use ahoe to dig the soil. If I wanted oxen, I wouldhave to hire and pay for them. Once I’d paidfor the oxen, I wouldn’t have any rice left”(In-depth interview, 5).

Participation in prevention andcontrol of dengue

Aedes mosquito breeding sites areubiquitous[12,16], explaining the continuedendemicity of DF. Women who participated ininterviews were asked what villagers shoulddo to prevent dengue. Many suggested thatvillagers should clean up their own house yards,collecting or burning rubbish such as tyres andcoconut shells, to get rid of mosquito breedingsites. The majority believed that people knew

of the danger of such breeding sites, and thatthey should work together to get rid of themto prevent disease. Many said that they wouldclean up their own yards if someone requestedthem to do so but they were reluctant to askothers. “I was afraid that they’d be angry. Theywould say they didn’t need to be told to cleantheir house. But if someone told me to do so,I’d be happy to follow their advice” (In-depthinterview, 14). However, most womencomplained that while they cleaned up theirown houses and yards, others did not, resultingin indiscriminate garbage throughout in thevillages; they claimed too that neighboursignored them when they asked them to cleantheir yards. Many villagers also felt that allvillagers should use temephos to preventdengue but this did not happen.

Mothers of children not infected withdengue reported individual and collectiveactivities to prevent and control the disease(Table). All but five of the women knew aboutthe disease and appropriate prevention andcontrol activities. Most perceived denguecontrol to be a personal responsibility, withabout half (19/33) reporting that they regularlycleaned their water jars and a third (10/33)stating that they kept their houses clean todiscourage mosquitoes. Almost one of fourwomen (9/33) claimed that they usedtemephos in water jars. A few women alsoused other (ineffective) activities to preventand control dengue, such as using mosquitonets at night, removing sewage, clearing bushesand using mosquito coils. The majorityoverlooked discarded containers, the mostcommon source of larval breeding in the rainyseason[12], and made no effort to get rid ofthem. Only three women reported telling theirneighbours to turn coconut shells upside down,or to remove discarded cans and plasticpacking bags. Hence despite claims of highknowledge, few women undertook allnecessary tasks on a regular basis.



Health workers at the village health centresand at the NDCP faced significant challengesin encouraging villagers’ participation in diseaseprevention. They complained that villagersprioritized income-generation activities overvector source reduction because of the dictatesof their economic status, their reliance on theNDCP to undertake such activities, and thelow effectiveness of the dengue healtheducation campaign[12,16]:

“In my opinion, people do not haveenough time to clean up their yard becausetheir standard of living is so low; they even donot have enough food to eat… from season toseason. When they return home from work inthe late afternoon, they cook for their children,

then find another job… they have no time toclean up the house. They are too busy to cleanthe jars even once a month or wash theirclothes. Some people never think about theirhouse or hygiene, they only think about foodto eat; that was why our health education wasnot successful.” (Health worker, FGD, 30).

Discussion

Community participation as an approach seemsto have been most successful in countries withstrong political authority, as in Cuba[1,17]. Incontrast, in fragmented societies wherecommunity members have different interestsand problems, and lack trust and confidence

Table: Villagers’ activities to prevent and control DF, mothers of children not infected with DF(n=33, survey data)

Reported Activities Numbers PercentageReported individual household’s dengue fever prevention and control activitiesClean my water jars 19/33 58Clean up the house 10/33 30Use Abate in my water jars 9/33 27Clean up the village to rid of stagnant water 1/33 3Cover my water jars 1/33 3Clear the bush 1/33 3Mosquito coil 1/33 3Burn the rubbish 1/33 3Reported collective dengue fever prevention and control activitiesNeighbours to clean their water jars 3/33 9Neighbours to cover jars 1/33 3Neighbours to collect coconut shells 3/33 9Neighbours to collect discarded cans 3/33 9Neighbours to collect plastic packing bags 1/33 3Neighbours to do the clean up 1/33 3Neighbours know everything 1/33 3Neighbours to use the net at night 3/33 9Neighbours to remove sewage 1/33 3Don’t know 5/33 15



in political leadership[7,10], communityparticipation has faced significant challenges.Local social, political and economic factors andassociated structural barriers and inequalitiescompound to affect the ability of members ofcommunities to sustain the activities requiredof them for disease control[18]. These variousfactors have influenced the introduction andsustainability of community participation inCambodia.

The literal translation of community inKhmer is sahakum, indicating a group of villagesor regions whose residents share the samejobs[19]. While sahakum is used in reference tocommunity participation in official contexts,Cambodians prefer to speak of “villagers” (nakphum), “residents” (nak strok), “provincialdwellers” (nak khet) and “city-dwellers” (nakkrong) to indicate geographical identity, invokingthe administrative structures of village,commune, district, province and city. The termsahakum only became popular when healthand development-related programmes wereintroduced in the 1990s. In the study on whichwe report, villagers rarely understood questionswhen the word sahakum was used in relationto participation in preventing or controllingdengue. Women stated bluntly that they didnot understand the term and could not explainwhat it meant. However, all villagers clearlyunderstood the terminology and ideasassociated with provass, with tveu kar cheamuyknea (working together) and nak phum tveukarcheamuy knea (villagers working together) toprevent or control dengue.

A number of scholars have argued thatcommunity-based programmes are moresustainable than vertical ones[4,20,21]. Studiesshow that top-down dengue prevention andcontrol activities have a temporary effect butdo not lead to the behavioural changes neededto reduce larval indices from the local domesticenvironment to ensure prevention and

control[22]. However, a recent review ofcommunity-based dengue control studies[7]

indicated that the implementation ofcommunity-based interventions has beenvariable, and noted the lack of involvement byvillagers, and specifically village committees,in planning and implementation, so threateningsustainability. The study recommendedintersectoral cooperation and stressed theimportance of involving local health services,civil authorities and key community membersto encourage individuals to take part in andsustain dengue prevention and controlstrategies.

Village participation in development inCambodia, as has occurred with roadconstruction, and the tradition of reciprocallabour exchanges, points to the potential forparticipatory development. However, the lowlevel of village cooperation, the lack of a spiritof collaboration, and economic pressurescombine to create significant challenges.Cooperation occurs in villages only if there ismaterial or financial involvement[13,23]. As notedabove, willingness and ability to work togetheris relatively low even when the rewards aretangible (as occurs with labour exchange). Theidea of working for the public good is far lessfamiliar. Villagers work individually, includingundertaking dengue control activities onrequest, but they are reluctant to encourageeach other to do so. Effective dengue healtheducation is needed to encourage people toundertake such activities on a continuingbasis[16]. There is an urgent need to restore trust,confidence and cooperation between villagersand in the society as a whole. This requiresthe political commitment of the government.

Whiteford[7,10] has suggested that people’svision of their future plays a major role in theirparticipation in government programmes. Shehas argued that Cubans consider health to bea collective achievement, and this, coupled



with confidence in the government and feelingsof hope for the future, have supportedgovernment-community partnerships fordisease control, including for dengue. Incontrast, in the Dominican Republic, unfulfilledpolitical promises, lack of political will and thelack of belief in community-based social actionhave resulted in the failure of communityparticipation despite good communityunderstandings of dengue control[10]. Thesituation in Cambodian villages echoes that ofthe Dominican Republic. Suspicion, distrust,increasing poverty, food insecurity,unemployment, landlessness andindebtedness, against the backdrop of violenthistory, contribute to a lack of confidence ingovernment capacity, its long-termcommitment to village development or toimproved health, and consequently, peopleshow little interest in community participation.

Conclusions

Community-based programmes involving localresponsibility and for the participation in theelimination of breeding sites are the only cost-effective and sustainable ways to ensure controlin any dengue-affected country, in particular, inpoorly resourced countries. However, inCambodia, community participation has beenimplemented primarily by internationalorganizations, NGOs, government departmentsand vertical disease control programmes. Localknowledge and local institutions, including thosethat would serve to achieve the same goals,have largely been overlooked. In this study,community members claimed that community-based dengue control occurred, but, in practice,people had limited opportunity to participate inplanning and managing dengue prevention andcontrol in their own villages, and so had littleinterest in or awareness of the need to ensurethat basic control activities were sustained.

In poor communities and poor countriessuch as Cambodia, disease control programmesneed to take into account factors affectingcommunity welfare, engagement andparticipation. In Cambodia, despite differencesin wealth, different health problems and often,different political views and affiliations,everyday life is dominated by the struggle tosurvive. Given this scenario, health programmessuch as DF programmes, using communityparticipation, need to deal not only with sourcereduction to control the larvae and themosquitoes but to address larger questions ofpoverty and income. Until such fundamentalissues are addressed, people’s engagement inprogramme planning and management and inthe work of disease prevention will remainpartial and episodic.

Acknowledgments

Sokrin Khun was supported financially by theUNICEF-UNDP-World Bank-WHO SpecialProgramme for Research and Research Trainingin Tropical Diseases (WHO/TDR) to undertakethe degree of Doctor of Philosophy in medicalanthropology at the University of Melbourne,Australia, and was supervised by LenoreManderson. We acknowledge with muchgratitude the support of WHO/TDR and of theSchool of Public Health, The University ofMelbourne, throughout his candidature, as wellas the support of Monash University when thisarticle was written. The authors thank sincerelythe Cambodian National Dengue ControlProgram, the Ministry of Health, the healthworkers at the village health centers andreferral hospitals, and school teachers, for theirinterest and cooperation. We are especiallygrateful to the women and children in the studyareas for their generous participation.



References

[1] Baly A, Toledo ME, Boelaert M, Reyes A,Vanlerberghe V, Ceballos E, Carvajal M, MasoR, La Rosa M, Denis O and Van der Stuyft P.Cost effectiveness of Aedes aegypti controlprogrammes: participatory versus vertical.Transactions of the Royal Society of TropicalMedicine and Hygiene. 2007; 101(6):578-86.

[2] Butterfoss FD. Process evaluation forcommunity participation. Annual Reviews ofPublic Health. 2006; 27: 323–40.

[3] Espino F, Koops V and Manderson L.Community participation and tropicaldisease control in resource-poor settings.Geneva: UNICEF/UNDP/World Bank/WHOSpecial Programme for Research andTraining in Tropical Diseases (WHO/TDR),2004.

[4] Gubler DJ and Clark GG. Community-basedintegrated control of Aedes aegypti: a briefoverview of current programmes. AmericanJournal of Tropical Medicine and Hygiene. 1994;50 (6): 50-60.

[5] Winch PJ, Kendall C and Gubler DJ.Effectiveness of community participation invector-borne disease control. Health Policy andPlanning. 1992; 7(4): 342-51.

[6] Heintze C, Velasco GM and Kroeger A. Whatdo community-based dengue controlprogrammes achieve? A systematic review ofpublished evaluations. Transactions of the RoyalSociety of Tropical Medicine and Hygiene.2007; 101: 317-25.

[7] Whiteford L. Local identity, globalization andhealth in Cuba and the Dominican Republic.In: Whiteford L and Manderson L, eds. Globalhealth policy, local realities: the fallacy of thelevel playing field. Boulder, Lynne RiennerPublishers, 2000. pp. 57-101.

[8] Allotey P and Gyapong M. The gender agendain the control of tropical diseases: A review ofcurrent evidence. Geneva: UNICEF/UNDP/World Bank/WHO Special Programme for

Research and Training in Tropical Diseases(WHO/TDR), 2005.

[9] Winch P.J., Lloyd L., Hoemeke L. and LeontsiniE. Vector control at household level – ananalysis of its impact on women. Acta Tropica.1994; 56(4): 327-39.

[10] Whiteford L. The ethnoecology of denguefever. Medical Anthropology Quarterly. 1997;11(2): 203-23.

[11] Ngan C. Dengue cases in Cambodia in 2007.National Dengue Control Program of theNational Malaria Control Program, PhnomPenh. 2008; 1-8

[12] Khun S and Manderson L. Abate distributionand dengue control in rural Cambodia. ActaTropica. 2007a; 101(2): 139–46.

[13] Khun S and Manderson L. Poverty, user feesand ability to pay for health care for childrenwith suspected dengue in rural Cambodia.BMC International Journal for Equity in Health.2008; 7: 10.

[14] Khun S. Community participation in theprevention and control of dengue fever: a casestudy in Cambodia. Global forum update onresearch for health, combating infectious andchronic disease. Pro-Brook Publishing. 2006;3: 21-26.

[15] CIMC-PHC. Implementation guidelines for thenational policy on primary health care.Interministerial Committee on Primary HealthCare, Phnom Penh. 2002; 1-20.

[16] Khun S and Manderson L. Community andschool-based health education for denguecontrol in rural Cambodia: a processevaluation. Plos Neglected Tropical Disease.2007b; 1 (3), e143: 1-10.

[17] Toledo ME, Vanlerberghe V, Perez D, LefevreP, Ceballos E, Bandera D, Gil AB and Van derStuyft P. Achieving sustainability of community-based dengue control in Santiag de Cuba.Social Sciences & Medicine. 2007; 64(4):976-88.



[18] Manderson L. Community participation andmalaria control in Southeast Asia: defining theprinciples of involvement. Southeast AsianJournal of Tropical Medicine and Public Health.1992; 23(1): 9-17.

[19] Chuon N. Khmer dictionary. Phnom Penh: TheBuddhism Institute, 1967. p 820.

[20] Gubler DJ and Clark GG. Communityinvolvement in the control of Aedes aegypti.Acta Tropica.1996; 61(2): 169-79.

[21] Gomez F, Suarez C and Cardenas R.Educational campaign versus malathionspraying for the control of Aedes aegypti inColima, Mexico. Journal of Epidemiology andCommunity Health. 2002; 56: 148-52.

[22] Toledo M.E., Baly A., Vanlerberghe V., MaritzaR., Benitez J.R., Duvergel J. and Van der StuyftP. The unbearable lightness of technocraticefforts at dengue control. Tropical Medicineand International Health. 2008; 13 (5):728–36.

[23] Khun S and Manderson L. Health seeking andaccess to care for children with suspecteddengue in Cambodia. An ethnographic study.BMC Public Health. 2007c; 262 (7): 1-10.


Dengue in the National Capital Territory (NCT) of Delhi(India): Epidemiological and entomological profile for

the period 2003 to 2008

J. Nandia#, R.S. Sharmaa, P.K. Duttab, G.P.S. Dhillona

aDirectorate of the National Vector Borne Disease Control Programme, 22 Shamnath Marg,Delhi – 110 054, India

bEx Associate Professor, Armed Forces Medical College, Pune, India

Abstract

Dengue is endemic in the National Capital Territory (NCT) of Delhi. During the period 2003 to 2008,9737 confirmed cases of dengue fever (DF)/dengue haemorrhagic fever (DHF) and 115 deaths wererecorded compared with 1341 cases and 6 deaths that happened during 1997 to 2002, representingan increase of 626%. During this period two outbreak peaks were also recorded. In addition, thesatellite town of Gurgaon (Haryana) bordering Delhi also suffered a severe outbreak of DF/DHF during2008.

Aedes aegypti, the responsible vector, is fully entrenched in both urban and rural areas. DF/DHFtransmission in years of extended winter rains occurs both during the summer and rainy seasons.Evaporation coolers during summer maintain low temperature and high humidity to ensure denguetransmission in some highly congested localities.

Keywords: National Capital Territory (NCT); Delhi; Endemic; Room water coolers; DF/DHF transmission.


Introduction

The National Capital Territory (NCT) of Delhiis endemic for dengue and has experiencedseveral outbreaks since 1967[1]. In the recentpast, Delhi recorded an outbreak in 1988,which resulted in 33% mortality amongchildren admitted in hospitals[2]. This wasfollowed by yet another severe outbreak in1996 throughout the NCT region of Delhi when

a total of 10 252 cases and 423 deaths wererecorded[3]. All the serotypes (DENV 1-4) havebeen detected circulating in the NCT Region[1].

In view of the severity of the outbreak in1996, the Directorate of the National Vector-Borne Disease Control Programme (NVBDCP)initiated several measures. These includedregular epidemiological and entomologicalsurveillance for timely prediction of an

Dengue in the National Capital Territory (NCT) of Delhi (India)


impending outbreak. The presentcommunication incorporates theepidemiological and entomological databasesfor the years 2003 to 2008.

Study area

The National Capital Territory (NCT) of Delhi,with an area of 1485 sq km, is located between28°75’ north latitude and 76°22’ eastlongitude[4]. The population of NCT increasedfrom 9.43 million in 1991 to 13.7 million in2001, recording a decadal growth rate of51.3% (Census report, Government of India,2001). Since 2001, the population has further

increased to 15.46 million with a growth rateof 12.2% (communication from MunicipalCorporation of Delhi).

Local civic bodies: The NCT of Delhi hasmultiple agencies responsible for the controlof vector-borne diseases. These include: twocivic bodies, the cantonment authority and theRailways. The Municipal Corporation of Delhi(MCD) has twelve zones and 272 wardscovering about 1399 sq km area. The NewDelhi Municipal Council (NDMC) has one zoneand eight wards covering 42.74 sq km area.The cantonment covers 42.89 sq km area[4]

(Figure 1). The railways look after its residentialcolonies and railway stations.

Figure 1: Map showing the zones of Municipal Corporation of Delhi, New Delhi MunicipalCouncil and Cantonment




Epidemiological surveillance: The NationalVector-Borne Disease Control Programme(NVBDCP) is a nodal agency for the monitoringof DF and DHF throughout the country.Hospitals follow the under-mentioned casedefinitions for the purpose of reporting.

• Patients with clinical symptoms likesudden onset of high fever, severebody pain and headache, myalgia,nausea, vomiting and rash, withpositive dengue-specific IgM in a singleserum specimen, to be considered asa dengue case.

• Clinical symptoms with lowthrombocytopenia and leucopenia arealso taken as cases of dengue fever.The presence of both these twocriteria with haemorrhagicmanifestation and death are taken asdeath due to dengue fever.

Entomological monitoring: Monitoring ofAedes aegypti throughout the year has beencarried out by the Central Cross CheckingOrganization (CCCO) under NVBDCP to workout the House Index (HI), Container Index (CI)and Breteau Index (BI) as per WHOguidelines[5].

Results

Epidemiological profile of dengue fever anddengue deaths in NCT Delhi: The DFincidence in NCT Delhi has been showing arising trend over the last six years (2003 to2008), when 9737 confirmed DF cases and115 deaths were recorded compared with 1341DF cases and 6 deaths that were reportedduring 1997–2002 (Source: NVBDCP). Thisrepresented an increase of 626% in theincidence, with an average mortality rate of1.2%. The DF attack rate recorded was 20.9,

4.2, 6.9, 22.3, 3.5 and 7.9 per 100 000population respectively during 2003 to 2008.During this period, two outbreak peaks wererecorded, first in 2003 and then second in 2006(Figure 2). The satellite town of Gurgaon(Haryana) bordering Delhi recorded a severeoutbreak of DF/DHF during 2008 (Source:NVBDCP).

Figure 2: Status of dengue fever and deathsin NCT Delhi (2003–2008)

2882

6061023

548

1312

336665

93 1 2

35

0

500

1000

1500

2000

2500

3000

3500

4000

2003 2004 2005 2006 2007 2008

Cas

es

0

10

20

30

40

50

60

70

Deat

hs

Cases Deaths

Seasonal distribution: Although DF caseswere reported throughout the year but amajority, i.e. 96%, were recorded duringSeptember to November, with a peak in themonth of October (Figure 3). Dengue deathsalso coincided with this period.

Spatial distribution of dengue cases inNational Capital Territory of Delhi: Thedengue incidence pertaining to NCT areas andother agencies is given in Table 1. The

Figure 3: Seasonal incidence of dengue feverand deaths (2003–2008)

2 3 4 1 3 2 25 204

1615

5504

2196

178

0 0 0 0 0 0 27

10

71

25

00

1000

2000

3000

4000

5000

6000


Cas

es

0

10

20

30

40

50

60

70

80

Dea

ths

Cases Deaths



Municipal Corporation of Delhi comprises 12zones (Figure 1). Six zones border the statesof Haryana and Uttar Pradesh and theremaining six are urban zones. The six borderingzones contributed 57% of the cases while theurban zones recorded 38% cases. The NDMC,Railways and Cantonment areas contributed3.7%, 0.9% and 0.4% cases respectively.

were the most preferred sites,followed by cement water tanks,flowers vases and overhead watertanks (OHT). High breeding in watercoolers is facilitated by thecommunity’s lax behaviour towardsweekly cleaning which is essential toprevent breeding. Undergroundcement tanks remained covered, butwhenever broken, they do not findreplacement. Overhead water tanks,mostly with plastic bodies, get heatedup in summer and do not supportbreeding.

Table 1: Disease incidence pertaining toNCT areas and other agencies (2003–2008)

S. Area PercentageNo.1 MCD area

(a) Zones borderingHaryana andUttar Pradesh (6 zones) 57.0

(b) Urban zones (6 zones) 38.02 NDMC area 3.73 Railway area 0.94 Cantonment area 0.4

Vector surveillance: The Ae. aegyptisurveillance was carried out throughout theyear, with special vigilance being maintainedin the zones reporting active cases. The searchfor the breeding habitats of Ae. aegypti tomonitor the entomological indices, viz. HouseIndex (HI), Container Index (CI) and BreteauIndex (BI), was carried out (Table 2).

From Table 2 it is evident that the Ae.aegypti breeding in NCT Delhi started risingfrom April, reaching the peak in August(BI=7.3), and then the decline in September(BI=5.8).

Breeding potential of Ae. aegypti(indoors and outdoors): The break-up of thepreferred breeding sites of Ae. aegypti, bothindoors and outdoors is given in Figure 4.

(1) Indoor breeding: Indoor watercoolers (cooling by water evaporation)

Table 2: Cumulative month-wise HI, CI andBI in NCT Delhi (2003–2008)

Month HI CI BIJan 0.0 0.0 0.0Feb 0.0 0.0 0.0Mar 0.2 0.1 0.2Apr 0.3 0.3 0.3May 1.2 0.9 1.2Jun 2.6 2.5 3.0Jul 4.1 4.3 6.1Aug 5.4 5.2 7.3Sep 4.6 4.4 5.8Oct 2.2 2.0 2.6Nov 0.6 0.5 0.6Dec 0.1 0.1 0.1

Figure 4: Breeding potential of Ae. aegypti inNCT Delhi

0.05.0

10.015.020.025.030.035.040.045.0

Cooler CementedTanks

FlowerVases

OHT PlasticContainer

EarthernPot

PlasticStorage

Containers

IronScraps

Tyre

Per

cen

tage

Indoor = 54.3% Outdoor = 45.7%



(2) Outdoor breeding: During the rainyseason, trash comprising of packagedconsumer items, broken pots, ironscrap and used tyres are the mostcommon items holding rain-water,which supports breeding and multiplythe vector population.

DF transmission season: NCT Delhi hasfour distinct seasons – (i) hot and dry (April–June); (ii) rainy season (July–September); (iii)autumn (October–November); and (iv) winter(December–March). The transmission seasonstarts in the rainy season and extends into theautumn because of congenial temperature andhumidity combination. However, possibilitiesof DF transmission during the hot and dryseason cannot be ruled out. Room watercoolers, which is the cheapest product availablelocally for the cooling of human dwellings, isaffordable even by the people belonging tolower socioeconomic groups living in slums.Room coolers not only cool the houseseffectively but also provide high humidity. Ae.aegypti breeds prolifically in water troughs andrests indoors in humid dark places. Extendedperiods of winter rains and the combination oflow temperature and high humidity in summermonths permit completion of the extrinsicincubation cycle of the virus when activetransmission can take place.

Discussion

Entomological surveillance: The first-evercomprehensive survey of Ae. aegypti populationwas carried out in 1964 to assess the potentialthreat to Delhi after the Kolkata episode in1963. The study revealed that Ae. aegypti wasconfined to the central part of city and theperipheral areas were free of the infestation.In the walled (old) city, the breeding index wasreported to vary from 50 to 100, thus indicatingvery high receptivity of the vector species[6].

(1) High endemicity of Ae. aegypti: Thelongitudinal studies by the NationalAnti-Malaria Programme (NAMP),now called NVBDCP, indicated thatthe vector species had now fullyestablished itself in all zones of MCD,NDMC, Cantonment and Railways.The monthly cumulative indices fullysynchronized with the diseaseincidence[7] (Figure 5).

(2) Receptivity of schools/hospitals: Aspatial study carried out in 1998showed high receptivity of schools andhospitals in view of the highvulnerability in a randomly selected 12schools/hospitals. The Container Indexin schools varied from 2.5 to 28.3,while in hospitals it varied from 2 to45.1[8]; these are the most commonplaces of DF transmission.

(3) Epidemiological impact: A graphicintegration of the epidemiological andentomological data is presented inFigure 5. It is evident that August toOctober were the most crucialmonths epidemiologically.Subsequently, with the onset of winter,the transmission ceased by December.

Conclusion

From the forgoing, it is evident that the NCTDelhi region carries a high receptivity and

Figure 5: Seasonal incidence of dengue feverand dynamics of Ae. aegypti (2003–2008)

2 3 4 1 3 2 25 204

1610

5485

2220

1780.0 0.0 0.2 0.3

1.2

3.0

6.1

7.3

5.8

2.6

0.6

0.10

1000

2000

3000

4000

5000

6000


Cas

es

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

BI

Cases BI



vulnerability to Ae. aegypti because of highinternational traffic as well as from thebordering satellite towns of Gurgaon (Haryana)and Noida (Uttar Pradesh), which are highlyendemic for DF/DHF. The transmission ofdengue here, unlike in other parts of thecountry where it occurs only during the rainyseason, has an extended transmission periodcovering both the hot and dry and rainy seasons.The transmission may start in summer (hot anddry season) in the years of extended winterrainfall up to April, followed by the rainy season.This requires round-the-year vector controlmeasures to ensure effective control of thevector. To meet these challenges, strengtheningof the health infrastructure of all agenciesinvolved in the prevention and control of

dengue is essential. These activities require awell-organized and coordinated effort to controlthe incidence of dengue in the NCT region.

Acknowledgement

The authors thankfully acknowledge the ableguidance received from Mr N.L. Kalra, formerDeputy Director of NVBDCP, for improving thecontents of this manuscript. The sincere andhard work of CCCO staff under NVBDCP ismuch appreciated. The authors are grateful toMr Anil Negi, Mr Raj Kumar, Mr DhirenderSingh and Mr V.K. Sood for their contributionin the preparation of figures, tables and thecompilation of field data.

References

[1] Kumar M, Pasha ST, Mittal V, Rawat DS, AryaSC, Agarwal N, Bhattacharya D, Lal S and RaiA. Unusual emergence of Guate 98 likemolecular subtype of DENV-3 during 2003dengue outbreak in Delhi. Dengue Bulletin.2004, 28: 161-167.

[2] Kabra SK, Verma IC, Arora NK, Jain Y, Kalra V.Dengue haemorrhagic fever in children inDelhi. Bull World Health Organ. 1992; 70(1):105-8.

[3] Kaul SM, Sharma RS, Sharma SN, PanigrahiN, Phukan PK, Shiv Lal. Preventing dengueand DHF – the role of entomologicalsurveillance. J Comm Dis. 1998; 30: 187-92.

[4] Kalra NL, GK Sharma. Malaria control in India– past, present and future. J Com Dis. 1987,19(2): 91-116.

[5] World Health Organization, Regional Officefor South-East Asia. Prevention and control ofdengue and dengue haemorrhagic fever :comprehensive guidelines. WHO RegionalPublications. South-East Asia Series No. 29.New Delhi: WHO-SEARO, 1999.

[6] Krishnamurthy BS, Kalra NL, Joshi GC, SinghNN. Reconnaissance survey of Aedesmosquitoes in Delhi. Bull Ind Soc Mal ComDis. 1965 Mar; 2(1): 56-67.

[7] Sharma RS, Panigrahi N, Kaul SM, Shilval,Barura K, Bhardwaj M. Status report on DF/DHF during 1998 in NCT Delhi, India. DengueBulletin. 1999; 23: 109-12.

[8] Sharma RS, Panigrahi N, Kaul SM. Aedesaegypti prevalence in hospitals and schools,the Priority Sites for DHF Transmission inDelhi, India. Dengue Bulletin. 2002, 25:107-108.


Increased utilization of treatment centre facilitiesduring a dengue fever outbreak in Kolkata, India

Shanta Duttaa, Jacqueline L. Deenb#, Dipika Sura, Byomkesh Mannaa,Suman Kanungoa, Barnali Bhaduria, Anna Lena Lopezb, Lorenz von Seidleinb,

John D. Clemensb, Sujit K. Bhattacharyaa

aNational Institute of Cholera and Enteric Diseases, P-33 CIT road, Scheme XM, Beliaghata,Kolkata 700 010, India

bInternational Vaccine Institute, San 4-8 Bongcheon-7-dong, Kwanak-gu, Seoul 151-818, Korea

Abstract

An outbreak of febrile illness occurred in Kolkata (formerly Calcutta), India, which led to an increasedutilization of treatment centre facilities during August – September 2005. The etiological agent wasconfirmed to be dengue by analysing 308 acute-phase clinical specimens for virus-specific IgMantibodies.

Keywords: Dengue fever; Outbreak, Treatment centres.

#E-mail: [email protected]; Tel: 82-2-872-2801; Fax: 82-2-872-2803

Background

In dengue-hyperendemic countries such asIndonesia, Thailand, the Philippines and VietNam, the circulation of multiple virus serotypesis well-established, regular dengue outbreaksoccur, and the severe form of the disease is acommon problem readily recognized byexperienced clinicians. In contrast, in SouthAsian countries such as India, Bangladesh, andSri Lanka, dengue is considered as an emergingdisease and epidemics have been morerecently recognized[1]. Clinicians have lessexperience with dengue, and laboratoryconfirmation of the etiology of the viral disease

is urgently needed to highlight this problem.Recently, an increasing number of outbreakshave been reported from various parts of India[2-

9]. We report a laboratory-confirmed outbreakof dengue fever in an urban slum communityof Kolkata during the course of a community-based fever surveillance study.

Methods

Kolkata is the third largest city in India and isone of the world’s most densely populatedcities; 13 million residents live within an areaof 1450 sq. kms. Kolkata has three seasons,

A dengue fever outbreak in Kolkata, India


the cool dry months from November toFebruary, the hot dry period from March toMay, and the monsoon season from June toOctober. As part of a typhoid fever vaccinetrial, we conducted a community-based passivesurveillance for febrile illness in a well-definedurban slum population of about 60 000individuals living in Wards 29 and 30. Bloodsamples were collected from patients residingin the study area who presented to treatmentfacilities in the study area with fever of 3 daysor longer[10]. The blood samples were used toinoculate Bactec Plus Aerobic bottles (BectonDickinson, New Jersey, USA) for bacterialculture, to make thick and thin blood films formalaria diagnosis, and preserved as sera forserological testing. A detailed description ofthe methods is presented elsewhere[10]. Thestudy was approved by the Institutional EthicsCommittee of the National Institute of Choleraand Enteric Diseases, the Ministry of Health

Screening Committee of the Government ofIndia, and the Institutional Review Board ofthe International Vaccine Institute.

An alarming increase in febrile episodeswas noted in 2005 (Figure). The number offever episodes evaluated in August andSeptember 2005 was 1637, nearly double the935 cases evaluated in August and September2004. The possibility of a dengue outbreak wasinvestigated using a Commercial PathozymeDengue IgM kit (Omega Diagnostics Limited,Omega House, Carsebridge Court, WhinsRoad, Scotland, UK)[11]. This is an in vitrodiagnostic test based on an indirect enzymeimmunoassay for screening dengue IgMantibody in infections caused by all fourserotypes. Briefly, diluted sera (after absorptionof IgG antibody) were added to the wellscoated with dengue-specific antigen. After athorough wash, peroxidase conjugated

Figure: The number of fever episodes evaluated and confirmed malaria, typhoid andparatyphoid fever from January 2004 to December 2005, Ward 29 and 30, Kolkata, India

358

291

332

807830

558563563

391

332323

236215

318

350

535

400389

244

181

256

194

351

248

0

10

20

30

40

50

60

70

80

90

100

Jan-

04

Feb-

04

Mar

-04

Apr-0

4

May

-04

Jun-

04

Jul-0

4

Aug-0

4

Sep-

04

Oct-0

4

Nov-04

Dec-04

J an-

05

Feb-0

5

Mar

-05

Apr-0

5

May

-05

Jun-

05

Jul-0

5

Aug-0

5

Sep-

05

Oct-0

5

Nov-05

Dec-05

Nu

mb

er

of

mala

ria,

typ

ho

id,

an

dp

ara

typ

ho

idfe

ver

cases

0

100

200

300

400

500

600

700

800

900

1000

Nu

mb

er

of

fever

ep

iso

des

evalu

ate

dP. vivax malaria

S. typhi

S. paratyphi

fever episodes



antihuman IgM followed by specific substratewere added. A colour development indicatedthe presence of human anti-dengue antibody.The reaction was stopped by the addition ofdiluted sulphuric acid and absorbance wasmeasured. Positive and negative controlssupplied with the kit were used during eachrun. We compared the characteristics ofpatients with and without dengue IgM. Weused the chi-square test for comparison ofcategorical variables and the Wilcoxon rank sumtest for comparison of medians. Statisticalsignificance was designated as a p-value lessthan 0.05 (2-tailed).

Results

From 1 August to 30 September 2005, a totalof 1637 residents in the study area presentedto a treatment centre with fever of 3 days ormore. Of these 1637 fever cases, 471 (29%)presented with fever of 5 days or longer, and308 (65%) were tested for dengue. 87/308(28%) were positive for dengue IgM antibodies,suggestive of primary dengue infection. Theages ranged from 3 years to 60 years for thosewho tested positive and 1 year to 77 years forthose who were negative. The characteristicsof the patients with a positive and negative

test for dengue IgM were compared (Table).The patients who had a positive dengue IgMtest had a slightly lower median age and weremore likely to have vomiting. There were nosignificant differences in the othercharacteristics between the two groups.

Discussion

Our surveillance detected an outbreak ofdengue fever during the rainy season in adensely-populated area of India, wheredengue has not traditionally been considereda local cause of fever. The principal vector fordengue fever is the female Aedes aegypti whichbreeds around human dwellings, in watercontainers, vases, cans, old tyres and otherdiscarded objects[12], which are common in thestudy site. The presence of this vector inKolkata has been documented. Ae. aegyptiprevalence coincides with the rainy seasonwhich sets in Kolkata from July to September[13],the same months when this dengue outbreakoccurred.

We confirmed acute dengue fever in ourstudy area with no evidence of severemanifestations (i.e. plasma leakage,haemorrhage, shock). There are four dengue

Table: Characteristics of febrile patients presenting for treatment with a positive andnegative test for dengue IgM, Ward 29 and 30, Kolkata, India

Positive for Negative fordengue IgM dengue IgM p value

(n = 87) (n = 221)Females (%) 48 (55%) 101 (46%) N.S.Median age in years 17 20 0.05Median number of days fever 5 6 N.S.With continuous fever (%) 24 (28%) 56 (25%) N.S.With nausea (%) 27 (31%) 56 (25%) N.S.With vomiting (%) 17 (20%) 18 (8%) 0.01With abdominal pain (%) 15 (17%) 21 (10%) N.S.

N.S.: Not significant



serotypes. Infection with one provides life-longimmunity against the same serotype, but notagainst the other serotypes. The risk of severedisease is increased about 15-fold during repeatinfection due to a serotype different from aprevious dengue infection[14]. Thus, populationspreviously infected by one or more dengueserotypes are at an increased risk for moresevere manifestation during subsequentdengue episodes due to other serotypes.

The observations made in Kolkata in 2005are consistent with a population with little tono pre-existing, anti-dengue immunity andwhere dengue is an emerging disease.Although the illnesses in the current outbreakwere self-limiting, the sudden increase inconsultations was an unexpected burden to theexisting treatment facilities. Furthermore, thereis the potential for more severe diseasemanifestations in future outbreaks. Theproportion of febrile episodes caused bydengue during the coming years warrantsfurther investigation. It would also be importantto follow any changes in signs and symptomsof the disease, particularly the occurrence ofsevere manifestations.

We did not perform virologicalconfirmation of the disease and therefore couldnot evaluate the circulating dengue serotype(s).We were not able to check for other etiologies,especially other flaviviruses with potentialdiagnostic cross-reaction. Our serologicaldiagnosis of dengue infection relied on the

presence of IgM antibody. By day five of illness,80% of dengue cases had detectable IgMantibody, and by day six to ten, 93 to 99% ofcases have detectable IgM that may persist forover 90 days[15]. We were unable to check fordengue IgG antibodies in the acute sera, nordid we collect convalescent sera to asses a risein IgG antibody titre. However, we believe thatthe test for IgM antibodies among those withfever of 5 days or longer is appropriate toconfirm an outbreak, particularly in thisrelatively dengue-naïve community.

The realization that dengue fever is anincreasing cause of febrile disease in South Asiasuggests the urgent need for preventiveinterventions. A safe, highly protective, long-lasting vaccine at an affordable price for largepopulations at risk in the tropical regions of Asiaand the Americas would be the ideal controlstrategy. Meanwhile, preventive activities haveto focus on vigorous vector control.

Acknowledgement

This work was supported by the Diseases of theMost Impoverished Program, funded by the Billand Melinda Gates Foundation and coordinatedby the International Vaccine Institute. We aregrateful to the residents of the study area whomade this work possible. We thank all technicalstaff and research assistants associated with thestudy. We acknowledge Dr Dianna M. Edgil forher useful advice.

References

[1] World Health Organization. Strengtheningimplementation of the global strategy for denguefever / dengue hemorrhagic fever prevention andcontrol: report of an Informal Consultation, 18-20 October 1999. Geneva: WHO, 2000.

[2] Gupta E, Dar L, Narang P, Srivastava VK, BroorS. Serodiagnosis of dengue during an outbreakat a tertiary care hospital in Delhi. Indian Journalof Medical Research. 2005; 121: 36-8.



[3] Kabilan L, Balasubramanian S, Keshava SM,Thenmozhi V, Sekar G, Tewari SC, et al.Dengue disease spectrum among infants in the2001 dengue epidemic in Chennai, TamilNadu, India. Journal of Clinical Microbiology.2003; 41: 3919-21.

[4] Narayanan M, Aravind MA, Thilothammal N,Prema R, Sargunam CSR, Ramamurty N.Dengue fever epidemic in Chennai - a studyof clinical profile and outcome. IndianPediatrics. 2002; 39: 1027-33.

[5] Agarwal R, Kapoor S, Nagar R, Misra A, TandonR, Mathur A etal. A clinical study of the patientswith dengue hemorrhagic fever during theepidemic of 1996 at Lucknow, India.Southeast Asian Journal of Tropical Medicineand Public Health. 1999; 30: 735-40.

[6] Dar L, Broor S, Sengupta S, Xess I, Seth P. Thefirst major outbreak of dengue haemorrhagicfever in Delhi, India. Emerging InfectiousDiseases. 1999; 5: 589-90.

[7] Aggarwal A, Chandra J, Aneja S, Patwari AK,Dutta AK. An epidemic of dengue hemorrhagicfever and dengue shock syndrome in childrenin Delhi. Indian Pediatrics. 1998; 35: 727-32.

[8] Pushpa V, Venkatadesikalu M, Mohan S,Cherian T, John TJ, Ponnuraj EM. An epidemicof dengue haemorrhagic fever/dengue shocksyndrome in tropical India. Annals of TropicalPediatrics. 1998; 18: 289-93.

[9] Bhattacharya N, Neogi DK, Hati AK, PramanikN, Banerjee A, Mukherjee KK. An outbreak ofdengue fever in a rural area of West Bengal.Indian Journal of Medical Microbiology. 1997;15: 139-41.

[10] Sur D, von Seidlein L, Manna B, Dutta S, DebA, Sarkar BL, Kanungo Suman, Deen JL, Ali M,Kim DR, Gupta VK, Ochiai RL, Tsuzuki A,Clemens J, Bhattacharya SK. The malaria andtyphoid fever burden in the slums of Kolkata,India: Data from a prospective, community-based study. Transaction of the Royal Society ofTropical Medicine and Hygiene. 2006; 100:725-33.

[11] Ha DQ, Thang CM, Ton T, Huong VTQ, LoanHTK, Dao HTN, Tam TTH. Evaluation ofcommercial pathozyme dengue IgM and IgGtests for serodiagnosis of dengue virus infection.Dengue Bulletin. 2000; 24: 97-102.

[12] Tsai T. Flaviviruses. In: Mandel GL, Bennet JE,Dolin R, eds. Principles and Practice of InfectiousDiseases. 5th ed. Philadelphia, PA: ChurchillLivingston, 2000: pp. 1714-36.

[13] Kalra NL, Wattal BC, Raghvan NGS.Distribution pattern of Aedes aegypti in India.Some ecological considerations. Bulletin ofIndian Society of Malaria and CommunicableDiseases. 1968; 5: 307-34.

[14] Halstead SB. Immunological parameters oftogavirus disease syndromes. In: SchlesingerRW, ed. The togaviruses: biology, structure,replication. New York: Academic Press, 1980:pp. 107-73.

[15] Pan American Health Organization. Dengueand dengue hemorrhagic fever in the Americas:guidelines for prevention and control. ScientificPublication No. 548. Washington: WHO-PAHO, 1994.


Entomological survey of dengue vectors as basisfor developing vector control measures in

Barangay Poblacion, Muntinlupa City, Philippines, 2008

Estrella Irlandez Cruza#, Ferdinand V. Salazara, Elizabeth Porrasb,Remigio Mercadob, Virginia Oraisb, Juancho Bunyic

aResearch Institute For Tropical Medicine, Alabang, Muntinlupa City, Philippines

bInstitute of Community and Family Health Inc., Graduate School of Public Health Banaue, Quezon City,Metro Manila, Philippines

cMuntinlupa City Health Office, Muntinlupa City, Philippines

Abstract

An entomological survey of dengue vectors was carried out in Barangay Poblacion Muntinlupa City inPhilippines. The survey revealed the presence of only one Aedes species, i.e. Aedes aegypti. The HouseIndex (HI), Container Index (CI) and Bretaeu Index (BI) were estimated as 23%, 23.4% and 63 respectively.Drums, used tyres and soft-drink cases were the major breeding habitats. Although communities keptthe water storage containers covered, but this failed to prevent the breeding.

Keywords: Aedes aegypti; Drums; Used tyres; Aedes indices; Philippines.


Introduction

Dengue fever (DF) and dengue haemorrhagicfever (DHF) was first recognized in thePhilippines in 1953[1]. Dengue virus istransmitted by Aedes aegypti, which aboundsin all cities and towns in the country.

In recent years, a major outbreak wasrecorded in 1998, with 35 100 cases and 500deaths for all regions (Department of Health,2001). Seventy per cent of those admitted werechildren less than 15 years of age.

Dengue control in Philippines is acommunity-based programme. Hence, theparticipation of all members of the communityas well as all sectors of society is a must forthe success of the programme.

The present entomological study wasundertaken in Barangay Poblacion Muntinlupacity during 2008 to detect the prevalence ofAedes species and their breeding habitats fordeveloping a control strategy for the same.


Entomological survey of dengue vectors in the Philippines

Study area

The entomological survey was conducted inBarangay Poblacion, Muntinlupa City. BarangayPoblacion has a population of 33 096 [CityHealth Office (CHO), 2008]. According to therecords of the CHO Surveillance Unit,Barangay Poblacion ranks number one in 2007with 87 dengue cases and three deaths in thenine barangays of the city. Clustering of thedengue cases was recorded in Magdaong,hence it was chosen to be the study site.

Magdaong is located in the National BilibidPrison (NBP) Reservation area in BarangayPoblacion, Muntinlupa City. It has a totalpopulation of 3312 with 552 households. Mostof the houses are constructed with lightmaterials while others are made of mixed woodand hollow blocks. There is no piped watersupply in the area. The community gets waterfor washing/bathing purposes from the NBPdelivery truck and free water delivery projectof Mayor Aldrin San Pedro. People buy waterfor drinking purposes. There is no artesian wellin the area.

Methodology

A total of 100 households were chosen forAedes larval survey. The sampling unit is thehousehold defined as one unit ofaccommodation sharing a common pot,irrespective of the number of persons residingtherein.

Every fifth house was included in thesurvey. The larval survey and computation ofentomological indices, viz. House Index (HI),Container Index (CI) and Breteau Index (BI),were carried out as per WHO guidelines[2].

Results

Out of the 100 houses searched, 23 houseswere found infested with immatures for Aedesmosquitoes. There were 269 water-holdingcontainers, out of which 63 containers werebreeding Aedine mosquitoes (Table 1).Identification of both the larval and pupal stagesrevealed the presence of only one Aedes species,i.e. Ae. aegypti, in the study area. Computedlarval indices are presented in Table 1.

From the table it is apparent that thethreshold levels of House Index (HI), ContainerIndex (CI) and Bretaeu Index (BI) were high.The BI of 63 containers, which establishesdirect relationship between positive containersand houses, indicate a high transmissionpotential for dengue.

Key breeding sites of Ae. aegypti

The distribution of breeding habitats is givenin Table 2.

From the table it is apparent that metaldrums, used tyres and soft-drink cases werethe most preferred breeding sites, while pails,jars and jugs were the least attractive.

Table 1: Computed larval indices of Ae. aegypti as observed in Magdaong Poblacion,Muntinlupa City

Magdaong 100 23 23 269 63 23.4 63

Study area Total housessearched

House Index (HI)+ve Index %

Breteau Index(BI) per

100 houses

Container Index (CI)Wet Containers Index

containers +ve %searched



Classification by coverage ofbreeding containers

To know the impact of ‘covering’ the breedingcontainers as practised by the localcommunities, 63 positive containers wereclassified into (i) covered; and (ii) without cover(Table 3).

From the table it is clear that a majority ofthe drums and, to some extent, pails, jars andjugs, were kept covered. However, a majorityof them failed to prevent breeding. Metaldrums, which were the main storage containersand were kept covered, were found breedingto the extent of 92.5 per cent. Hence, thereis a need to render the covers more airtight toprevent the entry of mosquitoes.

The results showed that the study areacarries a high potential for dengue outbreaks.Since dengue control in Philippines iscommunity-based, there is an urgent need (i)to strengthen the health infrastructure, (ii) foradvocacy to sensitize local populations; and (iii)for intersectoral coordination to improvedependable water supply and professionalmanagement of solid waste disposal.

Acknowledgement

The authors thank the Officer-In-Charge CityHealth Officer, Dr Edilinda Patac and AssistantCity Health Officer of Muntinlupa City, DrJuancho Bunyi, for permission to conduct andsupport of this study. Thanks are also due to

Table 2: Number of positive containers by habitats inspected in Magdoang Poblacion,Muntinlupa City

Type of Total number of Number of containers Percentage positivecontainer containers inspected positive containersDrums 107 40 37.3%Used tyres 43 17 40%Soft-drink cases 2 2 100%Pails 69 2 2.9%Jars 36 1 2.8%Jugs 12 1 8.3%Total 269 63

Table 3: Number of containers with cover and without cover positive for Ae. aegypti larvae atMagdoang Poblacion, Muntinlupa City

Type of container With cover Without cover Total Percentage +vewith cover

Drums 37 3 40 92.5Used tyres 0 17 17 0.0Soft-drink cases 0 2 2 0.0Pails 2 0 2 100Jars 1 0 1 100Jugs 1 0 1 100Total 41 22 63 65



Dr Remigio Olveda, Director of ResearchInstitute for Tropical Medicine (RITM), Dr RoseCapeding (RITM), Barangay Poblacion HealthCentre and the Barangay health workers of thecity of Muntinlupa, Ms Erlinda Servillon,Majhalia Torno, Jo Geronimo, Dr Fidel Malbas,

Krisstoffere Joy I. Cruz, Darwin Charles I. Cruz,Katherine Rose Ann I. Cruz and Ma. FrancesKristine I. Cruz, Edward Brian Pineda, PrinceJohn Albert Wilson ll and the MedicalEntomology staff of RITM for their invaluableassistance.

References

[1] World Health Organization. Denguehaemorrhagic fever : diagnosis, treatment,prevention and control. 2nd edn. Geneva:WHO, 1997.

[2] World Health Organization, Regional Officefor South-East Asia. Prevention and control ofdengue and dengue hemorrhagic fever. RegionalPublication No. 29. New Delhi : WHO-SEARO, 1999.


Aedes survey of selected public hospitals admittingdengue patients in Metro Manila, Philippines

Estrella Irlandez Cruz#, Ferdinand V. Salazar, Wilfredo E. Aure,Elizabeth P. Torres

Department of Medical Entomology, Research Institute for Tropical Medicine, Department of Health,Filinvest Corporate City, Alabang, Muntinlupa City, Metro Manila, Philippines

Abstract

Entomological investigations were carried out in five selected public hospitals admitting dengue patientsin Metro Manila, Philippines. The results revealed the presence of only one species, i.e. Aedes aegypti,mostly breeding in fresh water plant vases, drums, basins, plastic cups, tin cans, and empty paint cans.The water plant vase/bowl was found to be the most preferred container for Ae. aegypti breeding. Thepercentage positive rates of fresh water plant vases of the five public hospitals were: RMC (40.69%),TMC (76.62%), SLH (80.60%), NCH (64.06%), and EAMC (40.40%). An analysis of data revealed thatthe Premise, Container, and Breteau indices varied from 0.0 to 4.0, 0.0 to 41.1 and 0.0 to 11.0respectively, indicating thereby the high receptivity of the area to DF/DHF transmission. The egg densityranged from 0.0 to 48.5 which showed the presence of Ae. aegypti vector in the five public hospitals.The presence of productive breeding sites indoors and outdoors in the study area revealed that outbreakscould possibly occur in the future if no vector control plan is adopted and implemented.

Keywords: Aedes aegypti; Hospitals; Receptivity; Ovitrap index; Philippines.


Introduction

The Philippines is one of the dengue endemiccountries in Asia. Since the first outbreak ofdengue fever (DF) and dengue haemorrhagicfever (DHF) in 1953[1], Philippines has beenexperiencing an increasing incidence of denguecases. Based on the data of the NationalEpidemiology Centre of the Department ofHealth[2], there were 12 900 cases in 1997,35 600 in 1998, 9221 in 1999, 8761 in 2000and 25 050 cases in 2001.

Entomological surveys of communities havebeen conducted in the Philippines since thedevelopment of the Dengue Prevention andControl Programme (DPCP) of the Departmentof Health was initiated. The increase and peakin the number of dengue cases were observedto coincide with the rainy months of August toNovember, and the Aedes aegypti density wasthe highest during this period, while the monthsof February to April are low-density months. In1993, Schultz[3], in his study on the seasonalabundance of dengue, confirmed that Ae.


Aedes survey of public hospitals in Metro Manila, Philippines

aegypti was the major container breeder inresidential areas while Ae. albopictuspredominated in cemeteries in Manila.

In view of the high vulnerability of majorhospitals, which admit dengue patients, nostudy has been carried out to assess thereceptivity for Ae. aegypti.

The present study was conducted in fivemajor hospitals during September 2001 to April2002 and the results are incorporated in thepresent communication.

Study area

An entomological survey of dengue vectors wasconducted in five selected public hospitals inMetro Manila, which had the highest numberof dengue admissions for five years (1995–2000) based on the Department of Health(DOH) – National Capital Region (NCR)Surveillance data 1995–2000[2]. These were:San Lazaro Hospital (SLH), Rizal Medical Centre(RMC), Tondo Medical Centre (TMC), EastAvenue Medical Centre (EAMC) and NationalChildren’s Hospital (NCH).

Methodology

Aedes larval/pupal survey

A total of 500 rooms were surveyed monthlyfor Ae. aegypti in the five selected hospitals.Larval/pupal survey was done monthly for sixmonths. The survey for Ae. aegypti mosquitobreeding was carried out using the single larvatechniques. The larvae/pupae were collectedfrom all infested containers and identified inthe laboratory for species identification.Subsequently, water was removed to destroythe foci of breeding.

The level of infestation of Ae. aegyptimosquitoes, i.e. Premise (House) Index (PI),Container Index (CI) and Breteau Index (BI),was estimated as per WHO guidelines[4].

Ovitrap survey

The ovitrap was made from a tin can 76.2 mmin diameter, 101.6 mm in height, and paintedblack inside and outside. Water was added toa level of 76.2 mm. A wooden paddle (25.4mm x 152.4 mm) was placed diagonally insidethe tin can with the rough surface on top. Thepaddle served as oviposition substrate. Thenumber of ovitraps used was calculated throughsequential sampling (Lee, 1987)[5]. Thirtyovitraps were set up indoors of hospitals and30 ovitraps were placed outdoors. “Indoors”refers to the interior of the room while“outdoors” refers to the outside of the hospitalbuilding but within the hospital compound. Theovitraps were set up at different strategic placesindoors – under the sink, in the corner ofrooms, under the cabinet, under the bed,inside the comfort room, in the hospital chapel,and the nurse station; while outdoors, thesewere placed in corners of the hospital building,in the parking area, near vegetation, near pilesof wood/hollow blocks, and plants/trees in thegarden. Lost or damaged ovitraps werereplaced regularly.

The ovitraps were collected after seven dayswith the water and paddle placed separately inplastic bags. The eggs were counted in all positiveovitraps. Plastic bags of positive ovitraps weresubmerged into a pan for larval emergence. Thelarvae were counted and identified for speciesconfirmation. Similarly, the collected paddles weresubmerged into a pan with water for larvalemergence. Fish flakes were used as larval food.The third and fourth instar larvae were identifiedfor confirmation of species.



Analysis of data

The PI, CI and BI were computed using theWHO guidelines. At the same time, theOvitrap Index and Ovitrap Density Index werecomputed.


The monthly averages of Premise Index (PI),Container Index (CI) and Breteau Index (BI) ofSan Lazaro Hospital (SLH) varied from 0 to 2, 5to 25 and 2.0 to 8 respectively; in the TondoMedical Centre (TMC) these were 1.0 to 3, 2.7to 30.55 and 1 to 11; in the National Children’sHospital (NCH) 1 to 3, 7.14 to 27.27 and 1 to4; in the East Avenue Medical Centre (EAMC)these were 0 to 3; 0 to 13.04 and 0-4; and in

the Rizal Medical Centre (RMC) the averageswere 0 to 2, 5.12 to 25 and 2 to 8 respectively,thereby indicating the low receptivity of the fivepublic hospitals to DF/DHF transmission(Table 1). The results revealed that the breedingof Ae. aegypti occurred throughout the six monthsin the five public hospitals in Metro Manila. Thebreeding index was very low during the coolmonths of January and February and the hotmonths of March and April in all the hospitalssurveyed. The proportion of breeding containersof Ae. aegypti as presented in Table 2 showedthat fresh water plant vases were the major indoorcontainers inspected in the five public hospitals.

The results revealed that Ae. aegypti bredin both indoor and outdoor breeding sites. Theindoor containers searched and found infestedwith larvae were fresh water plant vases, drums

Table 1: Monthly Ae. aegypti indices of five public hospitals in Metro Manila duringSeptember 2001 to April 2002

San Lazaro Hospital (SLH)

MonthIndices

PI ICI OCI CI BIOctober 4.0 0 36 4.0 4November 2.0 31.0 75 41.1 7December 4.0 8.3 100 15.38 4January 1.0 4.76 33.3 8.3 2February 2.0 4.5.0 33.3 8.0 2March 0 0 0 0 0

Tondo Medical Centre (TMC)

MonthIndices

PI ICI OCI CI BIOctober 3 18.75 40.0 30.55 11November 2 28.57 27.77 28.0 7December 1 5.55 25.0 14.70 5January 1 18.75 40.0 30.55 11February 1 11.11 0 2.70 1March 1 11.11 0 5 1



East Avenue Medical Centre (EAMC)

MonthIndices

PI ICI OCI CI BINovember 1 2.4 50 8.33 4December 2 5.26 33.3 9.52 4January 1 1.92 0 1.72 1February 3 12 0 10.7 3March 3 14.2 0 13.04 3April 0 0 0 0 0

Rizal Medical Centre (RMC)

MonthIndices

PI ICI OCI CI BISeptember 1.0 1.5 25.0 5.12 4October 1.0 25.0 53.3 17.02 8November 1.0 5.8 14.28 5.88 2December 2.0 11.76 7.4 14.28 3January 0 0 22.22 15.38 3February 0 27.27 22.22 25.0 2

PI = Premise Index (%), ICI= Indoor Container Index (%), OCI = Outdoor Container Index (%),CI = Container Index (%), BI = Breteau Index (per 100 rooms)

National Children’s Hospital (NCH)

MonthIndices

PI ICI OCI CI BIOctober 2 21.4 0 20 3November 3 30 0 27.27 3December 1 9.09 8.33 11.11 1January 1 12.5 0 11.11 1February 3 0.4 0 36.36 4March 1 9.09 0 7.14 1

and basins, while the outdoor containerspositive for Aedes larvae were fresh water plantvases, empty paint cans, discarded plastic cups,drums, used automobile tyres, empty bottlesand abandoned toilet bowls. These receptaclesprovided breeding in both dry and wet seasons(Table 3).

Ovitrap survey

The Ovitrap Index indoors was also observedto build up from October to January (Table 4).The findings were contrary to reports that thebreeding was only confined to the wet season.The Mean Ovitrap Index varied from 0.0 to 48.5



Table 2: Percentage of indoor containers inspected in five public hospitals

Type of containers RMC TMC NCH EAMC SLHFresh water plant vase 59 59 41 90 166

(40.69%) (76.62%) (64.06%) (46.40%) (80.60%)Basin 1 0 0 0 0

(0.69%)Pails 43 13 9 98 9

(29.65%) (16.88%) (14.06%) (50.51%) (4.36%)Plastic drums (orocan) 42 3 14 6 29

(28.97%) (3.90%) (21.88%) (3.09%) (14.07%)Plant saucer 0 0 0 0 2

(0.97%)Total 145 77 64 194 206

(100%) (100%) (100%) (100%) (100%)

Table 3: Number of indoor and outdoor containers positive for Ae. aegypti in five hospitals inMetro Manila, during September 2001 to April 2002

Type of containers Indoors OutdoorsNo. searched Positive No. searched Positive

Freshwater plant vase 356 50 (14) 11 4 (360)Drums 1 1 (100) 0 0Basins 1 1 (100) 0 0Empty coconut shells 0 0 1 1 (100)Plastic cups 0 0 20 16 (80)Tin cans 0 0 14 6 (42.8)Used tyres 0 0 38 4 (10.5)Empty paint cans 0 0 15 12 (80)Abandoned toilet bowls 0 0 32 10 (31.5)Decorative jars 0 0 2 1 (50)

Figures in brackets indicate percentage

and showed the presence of Ae. aegypti vectorin the five public hospitals. The Mean OvitrapDensity Index ranged from 24.98 to 32.58 perpositive ovitrap (Table 5). There were 8464 eggscounted and 10% were reared to Aedes larvae,90% were Ae. aegypti and 10% were Ae.albopictus. This showed that Ae. aegypti bredin the hospitals predominantly over Ae.albopictus. Our study showed that the ovitraptechnique could be a reliable tool in vector

surveillance. The ovitrap detected density ofAedes mosquito in hospitals even during hotmonths.

Conclusion

Our study established the high receptivity offive hospitals in Metro Manila (breedingpotential) for dengue vectors. The presence



Table 4: Ovitrap Index

MonthIndices

Oli OIo OImSeptember 17 10 13.5October 37 27 32November 7 7 7December 20 20 20January 17 23.3 40.3February 3.3 3.3 3.3

Oli = Ovitrap Index indoor (%), OIo = OvitrapIndex outdoor (%), Olm = Mean Ovitrap Index

Rizal Medical Centre (RMC)

MonthIndices

Oli OIo OImOctober 27 7 17November 3.33 6.66 5December 3.33 6.7 5January 7 10 8.5February 10 17 13.5March 13.3 6.6 10

National Children's Hospital (NCH)

MonthIndices

Oli OIo OImOctober 3.3 10 6.7November 10 23.3 17December 13.3 40 27January 20 57 39February 3.4 3.4 3.4March 0 3.3 2

Tondo Medical Centre (TMC)

MonthIndices

Oli OIo OImOctober 23.3 23.3 23.3November 3.3 10 6.6December 6.6 16.6 11.6January 20 7 13.5February 37 17 27March 0 0 0

San Lazaro Hospital (SLH)

MonthIndices

Oli OIo OImNovember 26.6 70 48.5December 16.6 10 13.5January 6.7 6.7 7February 17 37 27March 13.3 53 3.3April 0 13 6.5

East Avenue Medical Centre (EAMC)

Name of No. of eggs No. of eggs Total No. of (+) Ovitraphospital (indoor) (outdoor) ovitrap Density

IndexEAMC 898 1741 2639 81 32.58SLH 891 690 1581 49 32.26RMC 613 811 1424 57 24.98NCH 296 792 1088 35 31.08TMC 523 1209 1732 56 30.92

Table 5: Ovitrap Density Index



References

[1] Halstead SB. Epidemiology of dengue anddengue haemorrhagic fever. In: Gubler DJ andKuno G, eds. Dengue and dengue haemorrhagicfever. New York: CAB International, 1997. pp.23-44.

[2] Department of Health - National EpidemiologyCentre (DOH-NEC), Manila. Unpublishedreport, 1998.

[3] Shultz GW. Seasonal abundance of denguevectors in Manila, Republic of the Philippines.Southeast Asian Journal of Tropical Medicineand Public Health. 1993 Jun; 24(2): 369-75.

[4] World Health Organization, Regional Officefor South-East Asia. Prevention and control ofdengue/dengue haemorrhagic fever:comprehensive guidelines. New Delhi : WHO-SEARO, 1999.

[5] Lee HL. Sequential sampling: its applicationin ovitrap surveillance of Aedes (Diptera:Culicidae) in Selangor, Malaysia. TropicalBiomedicine. 1992; (9): 29-34.

of key breeding receptacles indoors andoutdoors confirmed the presence of denguevectors, Ae. aegypti and Ae. albopictus, in allpublic hospitals surveyed. Hospitals are apriority area for surveillance and control of DF/DHF as they are highly vulnerable, i.e. sourceof virus through patients. As a result of thestudy, guidelines for vector control in hospitalswere prepared to prevent mosquito breedingin all potential breeding sites of dengue vectorsin hospital premises.

Acknowledgements

The authors extend their gratitude to Dr KevinPalmer, Regional Adviser of Malaria and Vector-

borne Diseases, WHO/WPRO Manila, DrRemigio Olveda, Director (RITM), Dr Rose Z.Capeding, Dr Ma. Nerissa N. Dominguez, MrsNorma Joson, Research Institute for TropicalMedicine (RITM) Department of MedicalEntomology staff, hospital staff and chiefs of RizalMedical Centre, San Lazaro Hospital, TondoMedical Centre, National Children's Hospital andEast Avenue Medical Centre for their invaluablehelp in this study. Thanks are also due toKrisstoffere Joy I. Cruz, Darwin Charles I. Cruz,Katherine Rose Ann I. Cruz , Ma. France KristineI. Cruz and Edward Brian Pineda, for theirinvaluable support to this study.


Epidemiological and entomological aspects of anoutbreak of chikungunya in Lakshadweep Islands,

India, during 2007

R.S. Sharmaa#, M.K. Showkath Alib, G.P.S. Dhillona

aNational Vector-Borne Disease Control Programme, Delhi – 110 054, India

bNational Institute of Communicable Diseases, Kozhikode, Kerala, India

Abstract

Since 2006, the Indian state of Kerala has reported outbreaks of chikungunya (CHIK). During July-August 2007, an unusual increase in the incidence of fever was noticed in Kadmat, Amini and KavarattiIslands in the Union Territory of Lakshwadeep, a group of Indian islands adjacent to the Kerala coast inthe Arabian Sea.

The populations affected as per the primary health centre (PHC) records of three islands, viz. Kadmat,Amini and Kavaratti, was 85%, 1.4% and 0.15% respectively. Entomological surveys revealed very highlarval indices of Aedes albopictus only in the three surveyed islands. Aedes aegypti, the classical vectorof dengue, was not detected. The maximum breeding of Ae. albopictus was found in coconut shells(57%), tyres (9%), metal containers (9%) and plastic containers (8%). The breeding was also detected intree holes and rat-bitten coconuts on top of the trees. The House Index for Ae. albopictus rangedbetween 95.4% in Kavaratti to 79% in Amini. Kadmat island which was the worst affected, recording themaximum Container Index of 90%, compared with 40% in Amini island. The CHIK outbreak seemedto have been caused by importation of the virus from Kerala, because of heavy movement of theislanders to the mainland.

Keywords: Chikungunya; Aedes albopictus; Lakshadweep Islands.

#E-mail: [email protected]; Tel: 011-23972884; Fax: 011-23986329

Introduction

In India, a chikungunya (CHIK) outbreak wasreported for the first time in Calcutta (nowKolkatta) in 1963[1]. Subsequently, epidemicsof chikungunya fever were reported inPondicherry (now Puducherry), Chennai,

Rajahmundry, Vishakpatnam and Kakinada in1965. Chikungunya reappeared in India afterquiescence of four decades in 2005[2]. A largenumber of cases were reported from the statesof Andhra Pradesh, Kerala, Karnataka, TamilNadu, Gujarat and Maharashtra[2,3]. As per theNational Vector-Borne Diseases Control

Epidemiological and entomological aspects of an outbreak of chikungunya in Lakshadweep, 2007


Programme (NVBDCP) records, a total of 1828CHIK virus-specific IgM confirmed cases wererecorded from 210 districts of 12 states afterits re-emergence in 2007. In India, both dengueand chikungunya have been reported to betransmitted by Aedes aegypti[4].

During July–August 2007, an outbreak ofchikungunya was recorded in Kadmat, Aminiand Kavaratti islands of the Union Territory ofLakshwadeep in the Arabian Sea. The presentstudy was carried out from 2 August to 7September 2007 to investigate theepidemiological and entomological aspects ofthe CHIK outbreak. The results of this studyare incorporated in the present communication.

Study area

Lakshwadeep comprises 36 islands, 11 ofwhich are inhabited. The islands arescattered in the Arabian Sea betweenlongitude 71°–74° east and latitude 8°–12°.The total area of Lakshwadeep is 32 sq. kmsand the population as per the 2001 censusis 60 650 (31 131 males and 29 519 females).The population density is 2255 per sq. km.The people have regular contact withErnakulam in the state of Kerala on themainland. Ferry services are frequent fromKochi (earlier Cochin) (Kerala) to differentislands of Lakshwadeep (Figure 1).

The climate is tropical, humid and warm.The average annual rainfall is more than 1500mm, mainly during the monsoon season (Mayto September). The relative humidity is alwaysabove 70%–75%. The average maximumtemperature varies from 29.5 °C to 33.2 °Cand the minimum from 23.6 ° to 27 °C.

Kadmat is one of the Lakshadweep groupof islands, with an area of 3.2 sq. km. with a

Figure 1: Study area showing differentchikungunya-affected islands in Lakshadweep

and movement of people from Kochi(Cochin), Kerala

population of 5319. Kavaratti island has 10 113population with an area of 4.22 sq. km. Aminiis a small island with 2.60 sq. km. area and7340 population. Coconut cultivation is themain occupation of the people here.


Epidemiological data

House-to-house surveys to find out the attackrate of fever with joint pains were conductedin the affected islands. These surveys werecombined with entomological surveys.Epidemiological information was also collectedfrom the primary heath care (PHC) centres inthe affected islands.



Entomological data

The entomological studies were carried out inKadmat, Amini and Kavaratti islands as perWHO guidelines[5]. Larval collections werereared into adults to determine the Aedesspecies. After the identification of the species,the emerging adults were destroyed. Thesearches were carried out in domestic andperidomestic habitats, viz. coconut shells,grinding stones and metal containers, tyres,plastic containers, cement tanks, glass bottles,mud pots, plastic sheets, etc. The House Index(% houses positive for Aedes breeding),Container Index (% containers positive forAedes breeding) and Breteau index (positivecontainers per 100 houses) were calculated tomake an assessment of the vector populationin the affected islands.

Results

Epidemiological

Signs and symptoms: In Kadmat island thefever cases started reporting on 2 July 2007and the daily trend was increasing. The fevercases reported to the PHC centres were mainlywith high fever, malaise, headache andarthralgia, particularly pain in small gouts, andvomiting. Skin rashes were also seen in someof the patients. Twenty-three blood sampleswere collected from eligible patients from PHCand 10 were found positive for CHIK-specificIgM antibody.

Disease incidence: The island-wise CHIK-affected population as per PHC centresrecorded from 2 July to 7 September 2007 isincluded in Table 1.

From Table 1 it is evident that the outbreakwas centred around Kadmat island where 85%of the population was affected. Similarly, inAmini and Karavatti islands, 1.4% and 0.15%respectively of the population was affected.

Trend of outbreak build-up: The dailyfever cases reported at Kadmat PHC centrefrom 2 July to 7 September 2007 are shown inFigure 2.

From Figure 2, it may be seen that during2 July to 30 July 2007, the number of casesreported ranged between 10 and 50, but on31 July, it climbed to 200 cases. Thereafter, itshowed a declining trend, and, by 7September, the figure had come down tobelow 100 cases.

House-to-house survey: In a house-to-house survey to assess the attack rate of CHIK,the numbers of sick persons who had feverwith joint pains in each house on the threeaffected islands were recorded. The results arepresented in Table 2.

From Table 2, it is evident that the attackrate in Kadmat island of 31.3% showed adeclining trend, while in Amini and Kavaratti,each showing 1.9% attack rate, showed a risingtrend.

Affected PHC centres Area in sq. km. Population No. affected with CHIKKadmat 3.20 5319 4565 (85%)Amini 2.60 7340 1025 (1.4%)Kavaratti 4.22 10113 16 (0.15%)

Table 1: Island-wise CHIK-affected population as per PHC records(2 July to 7 September 2007)

Figures in brackets indicate percentage of people affected.



Figure 2: Daily reported CHIK fever cases at PHC, Kadmat island(2 July to 7 September 2007)

0

50

100

150

200

250

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 1 2 3 4 5 6 7

No

.o

ffe

ve

rca

ses

Table 2: Attack rate of CHIK as determined during house-to-house survey onKadmat, Amini and Kavaratti islands

Date of survey No. of houses Total no. of persons Total no. of personsvisited in the houses visited affected at the time of visit

Kadmat27 August 2007 100 647 203 (31.3%)Amini29-30 August 2007 200 1278 25 (1.9)Kavaratti31 August -1 September 2007 200 1292 25 (1.9)

Figures in brackets indicate percentage of people affected clinically

Entomological

Aedes species prevalence: The entomologicalinvestigations on all the three islands revealedthe presence of only one species, Ae.albopictus. Ae. aegypti, the classical vector ofdengue, could not be detected.

Larval indices of Ae. albopictus: Thelarval indices as worked out on the threeaffected islands are included in Table 3.

From Table 3, it may be seen that thethreshold levels of HI, CI and BI on all thethree islands were very high. However, the



maximum BI (380) which establishes a directrelationship between positive containers andhouses was recorded on Kadmat island,indicating a high transmission potential asevidenced by the population affected (85%).

Key breeding habitats: The maximumbreeding of Ae. albopictus was recorded intender coconut shells (57%), grinding stones(3%) and metal containers (9%), tyres (9%),

plastic containers (8%), cement tanks (2%),glass bottles (2%), mud pots (2%) and plasticsheets (2%) (Figure 3). The indoor larval habitatswere: coconut shells, plastic containers, metalcontainers, grinding stones, cement tanks, mudpots and plastic sheets. Outdoor breedingplaces were: tree holes, tyres, broken glassbottles, clogged roof gutters, etc., on all thethree islands (Figure 4).

Figure 3: Key breeding habitats of Ae. albopictus observed inKadmat, Amini and Kavaratti islands

Coconut shells

57%

Tyres

9%

Grinding stones

3%

Metal containers

9%

Plastic containers

8%

Others

3%

Plastic sheets

2%

Cement

tanks/Cisterns

2%

Glass bottles

2%

Leaf axils

3%

Mud pots

2%

Table 3: Larval indices for Aedes albopictus in different islands

Vector indicesHouse Index Container Index Breteau

(%) (%) Index

Date of survey No. of housessurveyed

Kadmat27 August 2007 100 88 90 380Amini29–30 August 2007 200 79 40 146Kavaratti31 August –1 September 2007 200 95.4 67.1 222.7



Discarded coconut shells

Discarded cups

Earthen pot and plastic tank

Water-holding container for areca nut

Discarded tyres

Rat-bitten coconut

Figure 4: Key breeding places for Aedes albopictus in Kadmat, Amini and Kavaratti islands

Discussion

During 2007, an outbreak of chikungunya wasreported from Lakshawadeep island. Theoutbreak was serologically confirmed (CHIK-specific IgM) to be due to chikungunya virusand this was the first recorded outbreak in thehistory of Lakshadweep. All the vector indices

in all islands were very high and conducive forthe transmission of the disease (Table 2).

Out of the three islands surveyed as perPHC centre records, Kadmat island was worstaffected, where 85% of the population hadsuffered from the infection. In Amini andKaravatti islands, only 1.4% and 0.15%



respectively of the population were affected,despite the fact that all three islands had veryhigh vector indices. These variable results couldbe explained by high vulnerability (introductionof virus) at Kadmat island as compared to thetwo other islands. The high build-up of theCHIK epidemic can be explained by the shorterextrinsic incubation period (3–5 days) ascompared to dengue virus (8–14 days).Consequently, even if the longevity of thevector is adversely affected by climatic factors,the transmission still continues unabated[6]. TheCHIK transmission continues till all thepopulation develops immunity, which is long-lasting. During the inter-epidemic periods thevirus enters into wild animals. Lakshadweephas high rodent and bird populations whichpossibly act as a reservoir of infection duringthe inter-epidemic period.

Data collected during the house-to-housesurvey explains the declining trend in Kadmat,as 85% of the population had already developedimmunity due to infection, whereas on theother two islands, the population affected waslow and the susceptible population was stillhigh, as reflected by higher indices (1.9%).

In India, Ae. aegypti has been incriminatedas the principal vector of CHIK virus in all urbanand rural areas[4]. However, in Lakshaweep, Ae.albopictus is the CHIK vector in the absenceof Ae. aegypti. In La Réunion island in theIndian Ocean, Ae. albopictus was alsoincriminated as the CHIK vector[4]. In nature,Ae. albopictus is assumed to have a low vectorialcapacity (i.e. efficacy as a vector) because bloodmeals taken from non-susceptible hosts do notcontribute to the transmission[7]. However, inLakshadweep, Ae. albopictus, in the absence

of monkeys and other domestic animals,maintains a high degree of contact withhumans. However, this aspect needs to befurther assessed.

Kerala on the Indian mainland was one ofthe endemic states for chikungunya during2006. In all probability the virus was introducedin these islands through frequent movementof the islanders to the mainland. A similarphenomenon was observed in Italy in 2007when the virus was again introduced fromKerala[8,9].

The control of Ae. albopictus is most difficultas the species occupies both natural habitats,viz. leaf axils and tree holes as well as man-made domestic/peridomestic receptacles whichcan hold rainwater. The control strategy wouldrequire source reduction/larvicidal applicationin domestic/peridomestic habitats. Professionalmanagement of solid waste material outdoorsis essential. Coconut shells should be removedor stored under sheds to prevent breeding. Allthese activities require active participation ofthe community. Advocacy programmes needto be developed for providing information tothe community as well as inter-sectoralcooperation.

Acknowledgement

The authors thank Mr N.L. Kalra, WHOConsultant, for his valuable suggestions andhelp in the preparation of the manuscript. Thehelp and cooperation extended by theLakshadweep Administration is gratefullyacknowledged.



References[1] Shah KV, Gibbs CJ Jr, Banerjee G. Virological

investigation of the epidemic of haemorrhagicfever in Calcutta: isolation of three strains ofchikungunya virus. Indian Journal of MedicalResearch. 1964 Jul; 52: 676-83.

[2] Mohan A. Chikungunya fever: clinicalmanifestations & management. Indian Journalof Medical Research. 2006 Nov; 124(5): 471-4.

[3] Sharma RS. Epidemiological studies duringoutbreak of chikungunya in MarathwadaRegion of Maharashtra, (India). Indian Journalof Practising Doctor. 2006; 3(5): 34-6.

[4] Outbreak and spread of chikungunya. WeeklyEpidemiological Record. 2007 Nov 23; 82(47):409-15.

[5] World Health Organization, Regional Officefor South-East Asia. Prevention and control ofdengue and DHF. Regional Publication No. 29.New Delhi: WHO-SEARO, 1999.

[6] Mourya DT, Yadav P. Vector biology of dengue& chikungunya viruses. Indian Journal ofMedical Research. 2006 Nov; 124(5): 475-80.

[7] Rodhain F, Rosen L. Mosquito vectors anddengue virus – vector relationship. In: GublerDJ, Kuno G, eds. Dengue and denguehaemorragic fever. New York: CAB Publishing,1997. pp. 45-60.

[8] Rezza G, Nicoletti L, Angelini R, Romi R,Finarelli AC, Panning M, Cordioli P, FortunaC, Boros S, Magurano F, Silvi G, Angelini P,Dottori M, Ciufolini MG, Majori GC, CassoneA; CHIKV study group. Infection withchikungunya virus in Italy: an outbreak in atemperate region. Lancet. 2007 Dec 1;370(9602): 1840-6.

[9] Yergolkar PN, Tandale BV, Arankalle VA, SathePS, Sudeep AB, Gandhe SS, Gokhle MD,Jacob GP, Hundekar SL, Mishra AC.Chikungunya outbreaks caused by Africangenotype, India. Emerging Infectious Diseases.2006 Oct; 12(10): 1580-3.


Effect of pyriproxyfen in Aedes aegypti populationswith different levels of susceptibility to the

organophosphate temephos

Maria Teresa Macoris Andrighetti, Fernanda Cerone, Marcelo Rigueti,Karen Cristina Galvani, Maria de Lourdes da Graça Macoris#

Superintendência de Controle de Endemias – SUCEN, Serviço Regional 11, Av. Santo Antônio, 1627. 17506 040 Bairro Somenzari, Marília, São Paulo, Brazil

Abstract

Vector control with larvicides is an important component in dengue control programmes. In Brazil, theextensive use of temephos has led to the evolution of resistance in Aedes aegypti populations in manyparts of the country. One of the strategies proposed for managing temephos resistance is the use of theinsect-growth regulator – pyriproxyfen. This study evaluated the lethal concentration for this product inmosquito populations with different profiles of susceptibility to temephos and semi-field residual responseto a commercial product. The results suggest the possibility of cross-resistance between temephos andpyriproxyfen.

Keywords: Aedes aegypti; Susceptibility to insecticides; Cross-resistance; Insect-growth regulators.

#E-mail: [email protected]; Tel./Fax 55 14 34331080

Introduction

It is estimated that about 975 million peoplearound the world live in areas with denguetransmission risk[1]. In Brazil, all states areinfested by Aedes aegypti[2], the main denguevector in the tropical region[3].

The organophosphate temephos has beenused as larvicide since the beginning of the1980s for controlling Aedes aegypti in Brazil,and its use has been indicated by the NationalProgram for Dengue Control (PNCD)[4]. Theprolonged use of this larvicide has selected

resistant populations in many parts of theworld[5,6]. In Brazil, there are many reports onAe. aegypti resistance to temephos in severalstates[7-15].

The first strategy for managing temephosresistance adopted by the Brazilian governmentwas the substitution of larvicides by Bacillusthuringiensisi var. israelensis (Bti)[16]. Under fieldconditions the use of biolarvicide presents thedisadvantage of a shorter residual effect thanthe chemical one. Bti residual effect evaluatedunder field conditions varied 30–36 days[17] to7–12 weeks[18,19]. In Brazil, a residual effect of

Effect of pyriproxyfen in Aedes aegypti populations with different levels of temephos resistance


larvicide lasted nearly 60 days for the controlof Ae. aegypti, and this coincides with thefrequency of visits of vector control teams inthe context of PNCD[4].

In order to manage the temephosresistance in Ae. aegypti, besides Bti-basedlarvicides, products that are classified as insectgrowth regulators (IGRs) have been pointedout as an alternative by the Ministry of Healthin Brazil[20]. The World Health Organization(WHO) recommends pyriproxyfen (IGR)technical grade ingredient for the control ofAe. aegypti population. Many studies indicatedthat pyriproxyfen will not adversely affect anon-target species when applied at rates usually<50 ppb in mosquito control programmes[21].Among IGRs, pyriproxyfen, which is a mimicof juvenile hormone, is a potent inhibitor ofembryogenesis, metamorphosis and adultformation and shows a long residual effect[22,23].Although resistance to pyriproxyfen has notbeen reported in Ae. aegypti populations, it is

important to evaluate its effect on insects thatare resistant to organophosphates as they wouldbe the target for the management of resistanceto temephos.

The objectives of this study were toestimate lethal concentrations of pyriproxyfenand evaluate the residual effect of onecommercial formulation of this product on Ae.aegypti populations with different susceptibilitylevels to the organophosphate temephos.

Methods

Origin of Ae. aegypti populations

Field populations were collected throughovitraps according to the sampling methodologyused in the Brazilian network for the evaluationof resistance of Ae. aegypti to insecticides[16].Ae. aegypti populations were collected fromtwo different regions of Brazil. From the south-

Figure: Brazil map with localization of Brazilian municipalitieswhere mosquito populations were sampled



east, four populations were collected from SãoPaulo state, cities of Araçatuba (AT), Bauru(BR), Marília (MA) and Santos (SA) and onefrom Parana state, Maringa (MG). The secondregion, the north-east, was represented bythree states: Salvador (SS) and Barreiras (BA)from Bahia state, Recife (RE) from Pernambucostate and Fortaleza (FO) from Ceará state. Thegeographical distribution of these cities isillustrated in the Figure. The eggs collectedwere used to rear laboratory colonies. TheRockefeller susceptible strain, provided by theCenters of Disease Control, Puerto Rico, wasused for comparison.

Laboratory assays

Pyriproxyfen technical grade 98.5%, batch2006018 [4-phenoxyphenyl (RS)-2-(2-pyridyloxy) propylether]* was evaluated for fieldtests. The stock solution (250 mg/L) wasprepared in deionized water and stored at 4 °C.The work solution (2.5 mg/L) was preparedimmediately before each test.

The effect of the insect growth regulatorwas evaluated by the estimation ofconcentrations that caused inhibition of adultemergence, according to the World HealthOrganization methodology[24]. This estimationwas done through dose-response bioassays.Four Ae. aegypti populations were evaluated:Rockefeller, an insecticide-susceptiblereference strain, and three field populations,on the second generation reared in laboratory(F2). The field populations used had beenclassified according to their temephos-resistance ratio (R.R.) which were calculatedat the lethal concentration 95% (LC95). Thepopulation Salvador (SS), highly resistant[25,26]

(R.R. 11), the population Barrreiras (BA),moderately resistant (R.R. 6.9) and the

population Bauru (BR) with low level ofresistance (R.R. 3.8) were compared to thesusceptible Rockefeller strain. For eachpopulation, three bioassays were performed,each one with 720 third instar larvae exposedto eight pyriproxyfen concentrations. Eightylarvae were used per dose and for the controlgroup, in four replicates with 20 larvae each.All larvae were fed every other day duringobservation period with 0.5 ml of solution madeof 10% of fish food (Tetra Marine Granules®).Evaluations were done 48 hours after exposureand after each remaining 24 hours, throughquantification of live and dead larvae, pupaeand adult. The exposure lasted till the lastindividual died or emerged as adult. Bioassaydata were pooled by doses and the percentageof inhibition of adult emergence at each dosewas calculated dividing the percentage of adultemergence by the percentage of adultemergence at the control replicates[24]. Afterthat, the estimation of doses that caused 50%and 95% of emergence inhibition (EI) wasobtained by the software Polo-PC[27]. The EIdoses obtained with Rockefeller strain wereused for the calculation of resistance ratio ofpyriproxyfen for field populations.

The comparison of populations wasperformed by analysing the overlapping of EIdoses confident intervals.

Simulated field trial

The persistence of two products, Sumilarv®

0.5G (pyriproxyfen, SUMITOMO batch5099X31) and Temefós Fersol 1G (FersolIndústria e Comercio Ltda., Product batch287SP0214000), was evaluated in two kindsof containers, i.e. glass vases and tyres, whichare important breeding sites in Sao Paulo state.Glass vases are commonly used for maintaining

* Technical material was provided by Sumitomo Chemical Co., Ltd. First, Tokyo, Japan



plants (cut or live plants) and contribute toaround 30% of Ae. aegypti foci in many citiesof Sao Paulo state[28].

The Rockefeller strain and six Ae. aegyptifield populations with different susceptibilityto temephos were used in the test. Amongthe field populations one was consideredsusceptible MA (R.R. 1.8), two with low levelof resistance; MG, and AT (R.R. 2.7 and 2.6respectively) and three moderately-resistantpopulations: SA, FO and RE (R.R. 4.6, 8.4 and9.0 respectively).

For each population five breeding sites ofeach type were used, one for control and twofor each product. Glass vases were filled with4 l of water and tyres with 800 ml.

Pyriproxyfen at 0.05 ppm and 1 ppmtemephos were applied as per themanufacturers’ instructions. All containers,vases and tyres, were kept in a shaded area inan open garage.

Insecticide application was performed atday “zero” and 30 third instar larvae wereexposed one day later. New larvae were addedat fifteen-day intervals in a period of twomonths. Mortality was recorded every 24 hoursafter exposure and daily until the emergenceof adults. Pupae were transferred and observeddaily to register adult emergence or death. Thetemperature and pH of treated containers wererecorded once a week along with the test.

One-third of the water of each vase wasreplaced at 15-days interval, before a newexposition of larvae, in order to simulate thedomestic situation. There was no waterreplacement in tyres. In this case the watervolume was just filled to the original levelbefore each new larvae exposure.

The control group was used for Abbottcorrection when the observed mortality wasbetween 5% and 20%.

As pyriproxyfen – the candidate IGR acton different stages of mosquito development,the results were expressed as inhibition of adultemergence following the scheme presentedby Pinzon et al.[29] where EI (emergenceinhibition) = 1-[(Ad/Lexp)], where Lexp =larvae exposed and Ad = adults. The resultwas multiplied by 100 to express it aspercentage of inhibition.

The effect of candidate IGR was evaluatedthrough their capacity for providing at least 95%mortality along the time after treatment. As inBrazil, the cycle of visits for vector controlproposed by the National Programme (PNCD)[4]

is of 60 days, the expected effect of a larvicideshould suit this period of time.

The results obtained in Ae. aegyptipopulations were compared aftertransformation of the mortality data into arcsinvalues. Data were pooled by group accordingto resistance status of populations being MG,MA and AT pooled at one group for presentinglow level of resistance (R.R. below 4) andpopulations SA, RE and FO pooled in a secondgroup for presenting moderate resistance totemephos (R.R. between 6 and below 10).Comparisons were made between the twogroups and between each group and Rockefellerstrain with Student t-test.

Biochemical assays

The activity of metabolic enzymes alpha andbeta esterases, mixed function oxidases (MFO)and glutathione-S-transferase (GST) wereevaluated in larvae according to the Centersof Disease Control protocol[30,31]. The enzymaticactivity obtained by each larvae was correctedby the respective protein values. The resultswere analysed as proposed by Montella et al.[15],which is the standard method for analysis atthe Brazilian Network for the evaluation ofresistance of Ae. aegypti to insecticides[32]. The



enzyme activity of field populations wascompared with the susceptible Rockefellerstrain. The percentage of individuals withenzyme activity higher than the Rockefellerpercentage 99 classifies that activity as “normal”when it is below 15%, “altered” when it isbetween 15% and 49%, and “highly altered”when it is more than 50%.

Results

Estimation of adult emergenceinhibition concentrations

Pyriproxyfen exhibited a major effect on adultemergence with high mortality at pupal stage(97.2%) and very low mortality of larvae (2.8%).The estimated EI 50 and EI 955 of pyriproxyfenfor the susceptible Rockefeller strain and the

three field Ae. aegypti populations arepresented in Table 1 where data of temephoslethal concentrations and resistance ratios ofboth products are compared.

By the analysis of confidence interval ofEI concentrations, all the populations differedfrom the Rockefeller strain and also amongeach other. SS population presented the higherR.R. for both temephos and pyriproxyfen.Population BA, moderately resistant totemephos, had a lower R.R. for pyriproxyfenthan population BR which had the smallest R.R.for temephos.

Enzyme activity

The enzyme activity was evaluated for allpopulations except the field RE and MFOenzyme for MA population. The number of

Table 1: Estimated adult emergence inhibition concentrations of pyriproxyfen and lethalconcentrations of temephos (confidence interval). Resistance ratios based on Rockefeller

concentrations.

Pyriproxyfen Population EI 50(ppb) EI95(ppb) R.R. 50 R.R. 95 SS 3.37 6.44 6.5 5.5

(3.10 – 3.60) (5.90 – 7.20)BA 0.74 2.70 1.4 1.5

(0.64 – 0.84) (2.31 – 3.33)BR 1.86 4.13 3.6 3.5

(1.70 – 2.00) (3.75 – 4.63)Rockefeller 0.52 1.17 – –

(0.48 – 0.55) (1.06 – 1.32)Temephos Population LC 50 LC 95 R.R. R.R.

(ppm) (ppm) 50 95SS 0.028 0.053 11.0 12.6

(0.027 – 0.029) (0.05 – 0.056)BA 0.013 0.029 5.2 6.9

(0.012 – 0.013) (0.026 – 0.033)BR 0.0051 0.0110 2.0 2.6

(0.0048 – 0.0053) (0.014 – 0.016)Rockefeller 0.0025 0.0042 – –

(0.0024 – 0.0026) (0.004 – 0.0045)

ppb: parts per billion; ppm: parts per million.



Alp

ha

este

rase

(nm

ole

s/m

gp

tn/m

in)

Bet

aes

tera

se(n

mo

les/

mg

ptn

/min

)G

luta

thio

n-S

-tra

snfe

rase

(mm

ole

s/m

gp

tn/m

in)

MFO

(nm

ole

s/m

gp

tn/m

in)

Po

pu

lati

on

s

nM

edsd

evp

99

nm

edsd

evp

99

nm

edsd

evp

99

nm

edsd

evp

99

Rock

efel

ler

150

11.1

2.6

21.7

118

14.3

2.4

20.6

149

1.0

0.3

1.6

142

26.6

2.9

33.0

Fiel

dp

op

ula

tio

ns

nM

edsd

evp

%>

p9

9n

med

sdev

p%

>p

99

nm

edsd

evp

%>

p9

9n

med

sdev

p%

>p

99

Mar

ília

(MA

)120

11.5

1.8

12.6

150

16.8

2.6

6.0

83

1.2

0.4

––

––

Bau

ru(B

R)

117

16.1

2.7

6.0

147

15.1

2.1

1.4

89

1.3

0.4

119

21.5

4.7

4.2

Mar

ingá

(MG

)1

19

17

.92

.80

12

01

9.5

2.9

37

.51

15

1.4

0.3

11

62

2.1

3.9

3.4

Bar

reir

as(B

A)

14

61

9.4

4.4

30

.11

48

19

.23

.13

3.1

90

1.6

0.4

47

.81

50

21

.44

.44

.7

Ara

çatu

ba

(AT)

15

01

9.5

3.7

26

.71

47

16

.34

.42

1.8

14

81

.40

.63

7.2

14

72

8.0

5.4

27

.2

Santo

s(S

A)

147

24.7

16.5

120

21.1

6.8

150

1.7

0.4

147

30.1

16.0

Fort

alez

a(F

O)

150

21.0

3.0

30.0

120

21.5

4.8

120

1.5

0.3

47.5

117

42.8

9.4

Salv

ador

(SS)

150

27.1

3.3

115

53.4

13.4

100

2.2

1.6

82

49.1

35.3

15.7

18.0

25.2

67.3

55.0

58.7

51.7

54.2

64.1

93.3

99.1

78.0

75.6

n:

nu

mb

erof

larv

aeas

saye

d;

med

:m

edia

n;

std

ev:

stan

dar

dd

evia

tion;

p9

9:

per

centa

ge9

9p

%>

p9

9:

per

centa

geof

ind

ivid

ual

sw

ith

enzy

me

activi

tyhig

her

than

Rock

efel

ler

per

centa

ge9

9ac

tivi

ty:

Not

det

erm

ined

Norm

alen

zym

eac

tivi

tyA

lter

eden

zym

eac

tivi

tyV

ery

alte

red

enzy

me

activi

ty-

Tabl

e 2:

Qua

ntifi

catio

n of

enz

yme

activ

ity in

Ae.

aeg

ypti

popu

latio

ns



larvae assayed, median activity of enzymes,standard deviation and percentage of individualswith activity higher than the percentage 99 ofRockefeller strain are given in Table 2. A higheractivity of all four metabolic enzymes wasobserved on population SS, followed by SA andFO, which were characterized as resistant totemephos. The enzyme GST was altered in allfield populations but at a higher level ofalteration in SS and SA. Populations BA andMG, with respectively moderate and low levelof resistance to temephos, presented normalactivity of MFO and alteration on beta esteraseand GST. Populations with the lowest R.R. totemephos, MA and BR, presented normalactivity for all enzymes, except for GST withthe smallest percentage of individuals withalteration on that enzyme activity (15.7% and18.0% respectively).

Residual effect of commercialproducts

The adult emergence of all non-treatedcontainers was higher than 90%. Temephos

induced mortality at larval stage whilepyriproxyfen treatment affected mainly thepupal stage. Larval mortality with pyriproxyfenwas 2.3% on average. Exceptions wereobserved only in MA and AT populations, withlarval mortality higher than 10.8%.

The residual effect of both products variedaccording to the containers. Both treatmentspresented a shorter residual effect in tyres.

Treatment of glass vases with temephos(Table 3) resulted in 100% inhibition of adultemergence during the whole test period,except for the FO population where inhibitionended in 92%. In tyres the effect of temephosdecreased over time, especially for populationsFO (13 days), RE and SA (29 days), which hadthe higher RRs for temephos. The populationsMA and Rockefeller were the only ones toarrest 100% inhibition of adult emergence for44 days.

On the vases treated with pyriproxyfen(Table 4), 100% of inhibition of adultemergence was observed at the 58-days period

Table 3: Adult emergence inhibition observed in vases and tyres treated with temephos(Temefós Fersol 1G)

Days*Ae. aegypti population

Rockefeller MA AT MG SA FO RE1 100 100 100 100 100 100 100

13 100 100 100 100 100 100 10029 100 100 100 100 100 100 10044 100 100 100 100 100 100 10058 100 100 100 100 100 92 1001 100 100 100 100 100 100 100

13 100 100 100 100 100 98 10029 100 100 100 100 95 30 8444 100 100 78 51 56 28 1658 58 84 9 0 7 2 2

Tyre

sVa

ses

* Days after treatment



Table 4: Percentage of adult emergence inhibition observed in vases and tyres treated withpyriproxyfen (Sumilarv® 0.5 G)

Days*Aedes aegypti populations

Rockefeller MA AT MG SA FO RE1 100 100 100 100 100 100 100

13 100 100 100 100 100 100 10029 100 98 100 100 97 100 9844 100 97 100 100 83 98 10058 100 100 73 63 86 82 781 100 100 100 100 100 100 100

13 100 100 86 100 96 62 10029 100 98 90 98 73 52 9344 100 85 91 100 40 50 6558 97 61 43 80 39 20 22

Tyre

sVa

ses

* Days after treatment

only for Rockefeller and the field populationMA. The shorter effect was observed in theSA population (29 days), and in all the otherpopulations inhibition of adult emergence wasobserved to be higher than 95% for 44 days.In tyres, only Rockefeller showed inhibitionduring the whole test period. While ATpresented a shift in inhibition (86% to 91%)during the observation period, MG presentedthe longer effect among field populations (44days) followed by MA and RE (29 days). Again,the populations FO and SA presented theshorter residual effect, i.e. 1 and 13 daysrespectively.

The mortality data were converted intoarcsin values and analysed by Student-t test tocompare the response of groups of populations.The analysis of data from treatment in vasesshowed that while temephos treatment did notcause significant difference between the twogroups of populations, treatment withpyriproxyfen presented a significantly lowereffect on the moderately resistant populationgroups when compared to the Rockefellerstrain (p=0.01), and a significantly shorter

effect when the results of all populations werecompared with the treatment results withtemephos (p=0.002).

The treatment with temephos in tyresshowed a significantly shorter residual effectonly on the group of moderately resistantpopulations (p=0.02), while treatment withpyriproxyfen presented significant differencebetween the two groups of populations(p=0.02) and between each group andRockefeller strain (p<0.05).

Discussion

The pyriproxyfen doses that caused adultemergence inhibition estimated in this studywere higher than the lethal concentrationsobserved by Hatakoshi et al.[33] (L.C. 50 of0.023 ppb); Itoh et al.[34] (L.C. 50 of 0.056ppb) and Henrick[35] (L.C. 50 of 0.0039 ppb).Our estimations are closer to the lethalconcentrations described by Estrada andMulla[36] (L.C. 50 of 0.33 ppb and L.C. 95 of2.6 ppb).



Population SS, which presents the highestR.R. to temephos, also presented the highestR.R. to pyriproxyfen. No field populationevaluated in this study had been previouslyexposed to pyriproxyfen, so they had not beenunder selection for resistance to this IGR.

Although there is no recorded evidenceof pyriproxyfen resistance to Ae. aegypti, thepossibility of cross-resistance betweenconventional insecticides and IGRs has beenreported for other insects like Triboliumcasteneum[37] and houseflies[38]. Oxidase activityseems to be involved on resistance to juvenilehormone in houseflies[39-42]. In the presentstudy, differences on pyriproxyfen IEconcentrations among susceptible and resistantorganophosphate populations were alsoobserved, suggesting the possibility of cross-resistance between the IGR and temephos.

Braga et al.[43] discuss the possibility ofcross-resistance between temephos andmethoprene, another juvenile hormoneanalogue, in Ae. aegypti populations thatpresented high esterase and monoxigenaseactivity. The metabolism of endogenousjuvenile hormones is associated to both classesof enzymes in other insects [40-43].

The role of the studied metabolic enzymeson the observed resistance to temephos isnot easy to define since all four enzymes werehighly altered on resistant populations (SS, FOand SA). It is possible that the multiplemetabolic alterations are responsible fortemephos resistance and also for the higherpyriproxyfen observed on population SS.Nevertheless, the alteration on beta esterasesobserved for populations BA and MG, lesssusceptible than BR and MA, which hadnormal activity, also indicate a possible roleof this class of enzymes. The enzyme GSTmight also play an important role on temephosresistance as it is highly altered on resistantpopulations, especially on population SS.

Although esterases have been previouslyrelated to temephos resistance in Ae.aegypti[13,44-46] this class of enzyme is not thesole enzyme to be more active in resistantpopulations. GST was also characterized inCuba[44]. Braga et al.[43] relates alterations notonly in esterases activity but also in MFO andGST enzymes on temephos-resistant BrazilianAe. aegypti populations, making it difficult toascribe temephos resistance to only one classof enzymes. The same could be said forpyriproxyfen, as strain SS presented the higherR.R. and a high activity of all enzymes,although MFO presented the higher alterationon that population (75.6% of individuals) andthe role of this enzyme on IGR resistance iswell-documented in literature[47-50].

Data from the simulated field trial test invases indicate that the commercial productpyriproxyfen showed a significant shortereffect on adult emergence when comparedwith temephos (p=0.02). Temephos-treatedvases exhibited 100% inhibition compatiblewith the 60-day treatment cycle proposed bythe PNCD[4] while treatment with pyriproxyfenwas effective for this period only forRockefeller and MA (temephos-susceptiblepopulation).

On tyres, pyriproxyfen treatmentpromoted a 100% inhibition of adult emergenceon temephos-susceptible populationRockefeller. For the three resistant populations,temephos lasted more for two of them (SAand FO) and less for RE. The duration of effectof inhibition of adult emergence above 95% inboth products for tyres is not compatible withthe cycle of treatments as they lasted half ofthe expected period in many populations.

The lack of temephos effectiveness for 60days against the susceptible Rockefeller straincontrasts with previous studies[51] and raisessuspicion on the quality of commercialproducts.



References

Melo-Santos et al.[52] tested pyriproxyfenin water-storage cement boxes and plasticbuckets at the same concentration 0.05 ppmwith water replaced three times a week, andfound a complete inhibition of adultemergence of 160 days at shaded area and 46days at sunlight exposure. Resende[53], withpyriproxyfen trials found total inhibition of adultemergence of 90 days at 0.05 ppm dose and45 days at 0.01 ppm dose for the Rockefellerstrain in cement boxes and glass vases; but inplastic buckets it was 30 days. Vythilingam[54]

testing an Ae. aegypti population from Malaysia,resistant to temephos, observed pyriproxyfencomplete inhibition of adult emergence of 160days at 0.02 ppm even with water repositionevery fifteen days in earthen jars and plastictubs and 100% EI was obtained for 10 weeksin earthen jars where water was replaced daily.

The variation found in literature for thetime of complete inhibition of adult emergenceof products might be explained by thedifference in the surface of containers used intests and also by the variation in climateconditions. Also, bigger volumes of water tendto promote a more stable situation for larvicideaction. This might also play a role in persistenceeffect.

The choice of containers tested in this studywas based on their distinct surface ofabsorption, aiming to reach the best and worstavailability of larvicides, respectively, for glassvases and tyres. We do believe that thetreatment of those kinds of recipients should

not be encouraged. More sustainable actionsshould be encouraged. In this respect,community participation has shown verysatisfactory results in controlling Ae. aegypti fociin plant-related containers[28].

Conclusions

The development of resistance to temephosin many Ae. aegypti populations makes thechange for alternative larvicides for denguecontrol programmes most desirable.

Commercial products based on the IGRpyriproxyfen are one of the alternatives listedby the Brazilian Health Ministry for substitutingtemephos.

A temephos-resistant Ae. aegyptipopulation showed higher pyriproxyfen adultinhibition concentrations than susceptible ones.Besides, in semi-field assays, pyriproxyfenexhibited a lower residual effect againstpopulations characterized as temephos-resistant, suggesting interference of temephosresistance with pyriproxyfen action. Themechanism by which this interference acts isnot clear since MFOs, cited on literature asinvolved, are not the only class of enzymesaltered in temephos-resistant populations.

Data presented here should be consideredto make further evaluations and studies aboutthis IGR action as well as on the choice of thisproduct for use in management strategies.

[1] Nathan MB, Dayal-Drager R. Recentepidemiological trends, the global strategy andpublic health advances in dengue. In: UNDP,UNICEF, World Bank, WHO. Report of theScientific Working Group Report on Dengue.

Geneva: World Health Organization, 2007.pp 30-34. Document TDR/SWG/08. (http://w w w. w h o . i n t / t d r o l d / p u b l i c a t i o n s /publications/pdf/swg_dengue_2.pdf –accessed 18 June 2009).



[2] Ministério da Saúde. Secretaria de Vigilânciaem Saúde. Departamento de análise desituação de saúde. Vigilância em saúde: dadose indicadores selecionados – ano 3, n 3, Nov.2005.Brasília: Ministério da Saúde; 2003.

[3] Gluber DJ. Aedes aegypti and Aedes aegypti-borne diseases control in the 1990s. Topdown or bottom up. American Journal ofTropical Medicine and Hygiene. 1989; 40:571-578.

[4] Fundação Nacional da Saúde. ProgramaNacional de Controle da Dengue (PNCD).Brasília: Funasa, 2002.

[5] Rawlins SC. Spatial distribution of insecticideresistance in Caribbean populations of Aedesaegypti and its significance. Pan AmericanJournal of Public Health. 1998; 4: 243-251.

[6] World Health Organization. Vector resistanceto pesticides. Technical Report Series No. 818.Geneva: WHO. 1992.

[7] Beserra EB, Fernandes CRM, Queiroga MFC,Castro Jr FP. Resistência de populações deAedes aegypti. (Diptera: Culicidae) aoorganofosforado temefós na Paraíba.Neotropical Entomology. 2007; 36: 303-7.

[8] Braga IA, Lima JBP, Soares SS, Valle D. Aedesaegypti resistance to temephos during 2001in several municipalities in the states of Rio deJaneiro, Sergipe and Alagoas, Brazil. Memóriasdo Instituto Oswaldo Cruz. 2004; 99:199-203.

[9] Lima JB, Pereira da Cunha M, Carneiro daSilva R, Galardo AKR, Soares SS, Braga IA,Ramos RP, Valle D. Resistance of Aedes aegyptito organophosphates in several municipalitiesin the state of Rio de Janeiro and Espirito Santo,Brazil. American Journal of Tropical Medicineand Hygiene. 2003; 68: 329-33.

[10] Lima EP, Oliveira Filho AMO, Oliveira LimaJW, Ramos Júnior ANR, Cavalcanti LPG, PontesRJS. Resistência do Aedes aegypti ao temefósem municípios do estado do Ceará. Revistada Sociedade Brasileira de Medicina Tropical.2006; 39: 259 -63.

[11] Luna JED, Martins MF, Anjos AF, Kuwabara EF,Navarro-Silva MA. Susceptibilidade de Aedesaegypti aos inseticidas temephos ecipermetrina, Brasil. Revista de Saúde Pública.2004; 38: 842-843.

[12] Macoris MLG, Andrigheti MTM, Takaku I,Glasser CN, Garbeloto VC, Cirino VCD.Alteração de resposta de susceptibilidade ainseticidas organofosforados em municípiosdo estado de São Paulo, Brasil. Revista deSaúde Pública. 1999; 33(5): 86-7.

[13] Macoris MLG, Andrighetti MTM, Takaku L,Glasser CM, Garbeloto VC, Bracco JE.Resistance of Aedes aegypti from the state ofSão Paulo, Brazil to organophosphatesinsecticides. Memórias do Instituto OswaldoCruz. 2003; 98: 703-8.

[14] Macoris MLG, Andrighetti MTM, Otrera VCG,Carvalho LR, Caldas-Júnior AL, Brogdon WG.Association of insecticide use and alterationon Aedes aegypti susceptibility status.Memórias do Instituto Oswaldo Cruz. 2007;102(8): 895-900.

[15] Montella Montella IR, Martins AJ, Viana-Medeiros PF, Lima JBP, Braga IA, Valle D.Insecticide resistance mechanisms of brazilianAedes aegypti populations from 2001 to 2004.American Journal of Tropical Medicine andHygiene. 2007; 77(3): 467-77

[16] Braga IA, Valle D. Aedes aegypti: vigilância,monitoramento da resistência e alternativasde controle no Brasil. Epidemiologia e Serviçosde Saúde, Brasília. 2007; 16(4): 295-302.

[17] Lima JBP; MELO NV de; Valle D. Persistenceof Vectobac WDG and Metoprag S-2G againstAedes aegypti larvae using a semi-field bioassayin Rio de Janeiro, Brazil. Revista do Instituto deMedicina Tropical. 2005; 47: 1.

[18] Mulla MS, Thavara U, Tawatsin A, ChonpoosriJ. Procedures for the evaluation of field efficacyof slow release formulations of larvicidesagainst Aedes aegypti in water-storagecontainers. Journal of American MosquitoControl Association. 2004; 20: 64-73.



[19] Vilarinhos PTR, Monnerat R. Lavicidalpersistence of formulations of Bacillusthuringiensis var. israelensis to control larvalAedes aegypti. Journal of American MosquitoControl Association. 2004; 20(3): 311-14.

[20] Ministério da Saúde, Secretaria de Vigilânciaem Saúde. Avaliação da eficácia de análogosde hormônio juvenil e inibidores da síntesede quitina no controle de Aedes aegypti.Documento interno da Secretaria de Vigilânciaem Saúde. Brasília. 2005; 44p.

[21] World Health Organization. Report of thefourth WHOPES working group meeting, 4-5December 2000. Geneva: WHO, 2000.

[22] Ishaaya I, Horowitz AR. Novel phenoxyhormone analog (pyriproxyfen) suppressesembryogenesis and adult emergence of sweetpotato whitefly. Journal of Economic Entomology.1992; 85: 2113-7.

[23] Slama K, Romanuk M, Sorm F. Insect hormonesand bioanalogues. New York: Springer Verlag,1974.

[24] World Health Organization. Guidelines forlaboratory and field testing of mosquitolarvicides. Geneva: WHO, 2005. DocumentWHO/CDS/WHOPES/GCDPP/2005.13.

[25] Brown AWA. Insecticide Resistance inmosquitoes. A pragmatic review. Journal ofAmerican Mosquito Association. 1986; 2:123-40.

[26] Mazarri MB, Georghiou GP. Characterizationof resistance to organophosphate, carbamateand pyretroid insecticides in field populationsof Aedes aegypti from Venezuela. Journal ofMosquito Control Association. 1995; 11:315-322.

[27] Polo-PC Probit or Logit analysis [software].Berkeley, Leora Software, 1987.

[28] Macoris MLG, Mazine CAB, Andrighetti MTM,Yasumaro S, Silva ME, Nelson MJ, Winch PJ.Factors favoring houseplant containerinfeswtation with Aedes aegypti larvae inMarilia, Sao Paulo, Brazil. Pan American Journalof Public Health. 1997; (4): 280-6

[29] Pinzon, MLQ et al. Evaluación de la eficacia ypersistencia del regulador de crecimientopyriproxyfen (S-31183) bajo condiciones delaboratorio contra larvas de Aedes aegypti,Culex quinquefasciatus y Anopheles albimanusen Colombia. Informe final (mimeo) Medellín,Junio de 2003.

[30] Brogdon WG. Biochemical resistancedetection: an alternative to bioassay.Parasitology Today. 1998; 5: 56-60.

[31] Centers of Diseases Control. Microplate assays.Insecticide Resistance Workshop. Atlanta: GAS,1998.

[32] Ministério da Saúde, Fundação Oswaldo Cruz.Metodologia para quantificação de atividadesde enzimas relacionadas a resistência ainseticidas em Aedes aegypti. Ministério daSaúde, Série A. Normas e Manuais Técnicos.2006, 128p.

[33] Hatakoshi M, Kawada H, Nishida S, Kisida H,Nakayama I. Laboratory evaluation of 2-[1-m e t h y l - 2 - ( 4 - p h e n o x y p h e n o x y ) -ethoxy]pyridine against larvae of mosquitoesand housefly. Japanese Journal of SanitaryZoology. 1987; 38: 271-4.

[34] Itoh T. Utilization of blood fed females of Aedesaegypti as a vehicle for the transfer of the insectgrowth regulator, pyriproxyfen to larvalhabitats. Journal of the American MosquitoControl Association. 1994; 36: 243-8.

[35] Henrick CA. Juvenoides. In: Agrochemicalsfrom Natural Products. Godfrey, C.R.A, ed.New York-Basel-Hong Kong: Marcel Dekker,Inc., 1995. pp 147-213.

[36] Estrada JG, Mulla MS. Evaluation of two newinsect growth regulators against mosquitoes inthe laboratory. Journal of the AmericanMosquito Control Association. 1986; 2: 57-60.

[37] Dyte CE. Resistance to synthetic juvenilehormone in a strain of the flour beetle,Tribolium castaneum. Nature. 1972; 238:48-9.



[38] Cerf DC, Georghiou GP. Evidence of cross-resistance to a juvenile hormone analogue insome insecticide-resistant houseflies. Nature.1972; 329: 401-2.

[39] Cerf DC, Georghiou GP. Cross resistance tojuvenile hormone analogues in insecticide-resistant strains of Musca domestica L. PesticideScience. 1974; 5: 759-67.

[40] Brown TM, Brown AWA. Experimentalinduction of resistance to a juvenile hormonemimic. Journal of Economic Entomology. 1974;67: 799-801.

[41] Plapp FW, Vinson SB. Juvenile hormoneanalogs: toxicity and cross-resistance in thehousefly. Pesticide Biochemistry and Physiology.1973; 3: 131-6.

[42] Vinson SB, Plapp Jr. FW. Third generationpesticides: the potential for the developmentof resistance by insects. Journal of Agriculturaland Food Chemistry. 1974; 22(3): 356-60.

[43] Braga IA, Mello CB, Montella Jr, Lima JBP, JúniorAJM, Medeiros PFY, Valle D. Effectiveness ofmethoprene, an insect growth regulator,against temephos-resistant Aedes aegyptipopulations from different Brazilian localities,under laboratory conditions. Journal ofMedical Entomology. 2005; 43(5): 830-7.

[44] Rodriguez MM, Bisset JA, Mila LH, Calvo E,Diaz C, Alain Soca L. Levels of insecticideresistance and its mechanisms in a strain ofAedes aegypti of Santiago de Cuba. RevistaCubana de Medicina Tropical. 1999; 51: 83-8.

[45] Gokhale MD, Jacob PG, Mourya DT. Denguevirus and insecticide susceptibility status ofAedes aegypti mosquitoes from Belagolavillage, Mandya District, Karnakata state:during and post epidemic investigations. Journalof Communicable Diseases. 2000; 32: 247-53.

[46] Hemingway J. The molecular basis of twocontrasting metabolic mechanisms ofinsecticide resistance. Insect Biochemistry andMolecular Biology. 2000; 30: 1009-15.

[47] Bergé JB, Feyereisen R, Amichot M.Cytochrome P450 monooxygenases andinsecticide resistance in insects. PhilosophicalTransactions of the Royal Society BiologicalScience. 1998; 353: 1701-5.

[48] Sutherland TD, Unnithan GC, Andersen JF,Evans PH, Murataliev MB, Szabo LZ, MashEA, Bowers WS, Feyereisen R. A cytochromeP450 terpenoid hydrolxylase linked to thesuppression of insect juvenile hormonesynthesis. Proceedings of the National Academyof Sciences. 1998; 95: 12884-9.

[49] Feyereisen R. Insect P450 enzymes. AnnualReview of Entomology. 1999; 44: 507-33.

[50] Lu HK, Bradfield JY, Keeley LL. Juvenilehormone inhibition of gene expression forcytochrome P4504Cl in adult females of thecockroach, Blaberus discoidalis. InsecticideBiochemistry and Molecular Biology. 1999; 29:667-73.

[51] Macoris MLG, Andrighetti MTM, Takaku L.Efeito Residual de Temephos em larvas deAedes aegypti. Revista da Sociedade Brasileirade Medicina Tropical. 1995; 28(4): 375–7.

[52] Melo-Santos MAV, Regis L. Avaliação deprodutos para o controle de Aedes aegypti.Relatório final (mimeo). Recife, FIOCRUZ/CPqAM/Depto de Entomologia. 2003. 23p.

[53] Resende MC, Gama RA. Persistência e eficáciado regulador de crescimento Pyriproxyfen emcondições de laboratório para Aedes aegypti.Revista da Sociedade Brasileira de MedicinaTropical. 2006; 39(1): 72-5.

[54] Vythilingam I, Luz BM, Hanni R, Beng TS, HuatTC. Laboratory and field evaluation of theinsect growth regulator pyriproxyfen (Sumilarv0.5G) against dengue vectors. Journal of theAmerican Mosquito Control Association. 2005;21(3): 296-300.


Effectiveness of pyriproxyfen-controlled release blockagainst larvae of Aedes (Stegomyia) aegypti in Kuala

Lumpur, Malaysia

C.D. Chena,b#, W.A. Andy-Tana,b, S.R. Lokea,b, H.L. Leea,A.R. Yasminb, M. Sofian-Azirunb

aMedical Entomology Unit, WHO Collaborating Center for Vectors, Infectious Diseases Research Center,Institute for Medical Research, Jalan Pahang, 50588 Kuala Lumpur, Malaysia

bInstitute of Biological Sciences, Faculty of Science, University of Malaya, 50603 Kuala Lumpur, Malaysia

Abstract

This study was conducted to evaluate the effect of a commercially available pyriproxyfen, an insectgrowth regulator (IGR) on the larvae of a dengue vector, Ae. aegypti. The study site was the surroundingarea of the Medical Entomology Unit, Institute for Medical Research (IMR), Kuala Lumpur (N03°10.167’,E101°41.919’). Pyriproxyfen-controlled release blocks with dosages of 10% w/w and 20% w/w wereused to treat a set of earthen jars placed outdoors. Untreated jars were also set up as controls. Fiftylaboratory-bred 2nd instar larvae were introduced into each earthen jar and observed daily. Thenumber of adults that emerged was recorded and the larval mortality was calculated. The indicators ofeffectiveness of IGR for these studies were: (i) duration of effectiveness, and (ii) percentage of emergenceinhibition (EI). There was a significant difference in the number of emerged adults obtained from theuntreated and treated earthen jars up to 25 weeks (p<0.05). The duration of effectiveness ofpyriproxyfen caused 80% emergence inhibition in earthen jars treated with 10% w/w and 20% w/wpyriproxyfen up to 22 and 25 weeks, respectively. Pyriproxyfen-controlled release block is an effectivemethod of controlling mosquito larvae for several months. The method of application of the block issimple and straightforward and can therefore be used easily.

Keywords: Aedes aegypti; Pyriproxyfen; Controlled release block; Duration of effectiveness; Emergence inhibition.


IntroductionAedes aegypti is a principle vector of denguein many parts of the world. It is one of themajor domestic groups of mosquitoes that arepests of man as well as a vector of disease. In

many areas of the world, this speciescommands considerable attention in term ofits management and control[1]. The number ofoptions available for mosquito control at thepresent time are also limited. The control ofthis mosquito is still dependent on the use ofchemical insecticides.


Effectiveness of pyriproxyfen-controlled release block against larvae of Aedes aegypti

Although insecticides are invaluable inpreventing and controlling damage toagricultural products and to the health of manand animals, they are not without side-effectson the environment and its biota[2]. There is acritical need to find and develop new agentsand products for the control of this and otherimportant species of mosquitoes. Insect growthregulators (IGRs) are now increasingly used tocontrol Aedes and other mosquito larvae. Thesecompounds have unique modes of action, andare often selective and do not persist in theenvironment. Such attributes are desirablewhen dealing with the problem of pestresurgence, secondary pest outbreaks andinsecticides resistance[3].

Pyriproxyfen, 2-[1-methyl-2-(4-phenoxyphenoxy) ethoxyl] pyridine, is a newgeneration of IGR. It is a juvenile hormoneanalogue and a relatively stable aromaticcompound. It functions as an insecticide byoverloading the hormonal system of the targetinsect, ultimately affecting its egg production,brood care and other social interactions, andinhibiting its growth[4]. Pyriproxyfen works wellagainst public health insects like houseflies andmosquitoes[5]. Pyriproxyfen is reported to exhibit95% inhibition of the emergence of mosquitolarvae and its effects on mosquito larvae havinglasted for two months after application[6].Although the treated mosquito larvae continueto pupate, however, their emergence is inhibitedby the action of pyriproxyfen[7].

The controlled release block used in thisstudy was impregnated with 10% w/w and 20%w/w a.i. (active ingredient) of pyriproxyfengranules. Pyriproxyfen-controlled release blockis claimed to be an easy method applicable inareas such as drains, ponds, lakes, etc., wheremosquitoes breed.

This study was conducted to evaluate thecommercially available pyriproxyfen-controlledrelease block used for the control of Ae. aegyptilarvae in earthen jars.


Test site

The study was conducted in the areasurrounding the Medical Entomology Unit,Institute for Medical Research (IMR), JalanPahang , Kuala Lumpur (N03°10.167’,E101°41.919’).

Insect growth regulator

A formulation of insect growth regulator,pyriproxyfen-controlled release block, was usedin this study. Two concentrations of controlledrelease block were provided, each containingpyriproxyfen at 10% w/w a.i. and 20% w/wa.i. The formulation was provided by Zero-MozJapan Sdn. Bhd.

Test containers

Earthen jars were used as mosquito breedingcontainers in this study. Earthen jars each withan opening of 52 cm in diameter, basediameter of 35 cm and 47 cm in height wereprepared and placed outdoors. Three replicatesof each were used in each research arm of thestudy (Table 1). Each earthen jar held 60 L tapwater. Before initiating the study, all containerswere washed with tap water and tested forthe presence of contaminant, such asinsecticides, by introduction of 50 Ae. aegypti2nd instar larvae. The larvae were observed untilcomplete emergence as adult.



Earthen Chemical Number ofjar (active ingredient) replicatesUntreated None 3Treated (with controlled Granular pyriproxyfen 10% w/w 3release block) Granular pyriproxyfen 20% w/w 3

Table 1: Set-up of earthen jars for testing

Trial procedures

Each pyriproxyfen-controlled release block wasplaced into earthen jar (3 replicates) andlabelled. Three earthen jars withoutpyriproxyfen-controlled release block served asuntreated control. In each test, 50 laboratory-bred 2nd instar larvae were introduced into eachearthen jar and observed daily. Pupae werecollected, recorded and transferred into papercups covered with net. The total number ofadults emerged was recorded and the larvaemortality rates were calculated. A total of 50%of water (30 L) was removed and added intothe earthen jars every alternate day. The sameprocedure was repeated by adding fresh batchof larvae (50 larvae) into each earthen jarweekly.

Data analysis

The indicators of effectiveness of pyriproxyfen-controlled release block for these studies were:

(1) duration of effectiveness of eachdosage, and

(2) percentage of emergence inhibition(EI) =

Number of larvae introduced – Number of adult emergedx 100%

Number of larvae introduced

A cut-off point of EI ≥ 80% was consideredto be effective.

If percentage of untreated EI was > 5%,the percentage of treated EI was corrected byAbbott’s formula:

% treated EI – % untreated EIX 100%

100 - % untreated EI


Table 2 shows the number of pupae, adultemergence and emergence inhibition obtainedfrom earthen jars treated with 10% w/w and20% w/w pyriproxyfen-controlled release blockthat were impregnated with 10% w/w and 20%w/w pyriproxyfen granules. The result showeda significant difference on the number of pupaecollected from all treated (10% w/w and 20%w/w) and untreated earthen jars (p<0.05).However, there was no significant differenceon the number of pupae collected from the6th week onwards (p>0.05), indicating thatpyriproxyfen exhibited low larvicidal activityagainst Ae. aegypti. This is similarly reportedby Lee et al.[8] in which studies were carriedout on the bio-efficacy and duration of theeffectiveness of pyriproxyfen (Sumilarv 0.5%)as direct applications for the control of larvaeof Ae. aegypti and Ae. albopictus. At 79.5 mg/L and 159.0 mg/L, pyriproxyfen showed lowlarvicidal activity but provided very effectivecontrol of adult emergence against larvae ofAe. aegypti and Ae. albopictus[8].

A significant difference in the number ofadult emergence was observed in both the



Table 2: Number of pupae, emergence of adult and emergence inhibition obtained fromearthen jars treated with 10% w/w and 20% w/w pyriproxyfen-controlled release block

Mean ± SE of collected pupae Mean ± SE adult emerged

Treated TreatedWeek Test period

Untreated

10% w/w 20% w/w

One wayANOVA

Untreated

10% w/w 20% w/w

One wayANOVA

1 5 Feb – 11Feb

48.00 ±1.00

15.67 ±10.73

37.33 ±3.28

F = 6.42p = 0.032

48.00 ±1.00

0.00 ±0.00

0.00 ±0.00

F = 2304.00p = 0.000

2 12 Feb – 18Feb

43.18 ±3.18

21.67 ±11.79

9.33 ±1.20

F = 5.85p = 0.039

43.18 ±3.18

0.00 ±0.00

0.00 ±0.00

F = 184.38p = 0.000

3 19 Feb – 25Feb

40.67 ±2.03

2.00 ±0.58

1.33 ±0.33

F = 333.26p = 0.000

40.67 ±2.03

0.00 ±0.00

0.00 ±0.00

F = 401.38p = 0.000

4 26 Feb – 4Mar

39.00 ±2.65

10.33 ±5.90

4.33 ±2.03

F = 22.41p = 0.002

39.00 ±2.65

0.00 ±0.00

0.00 ±0.00

F = 216.59p = 0.000

5 5 Mar – 11Mar

20.50 ±7.50

0.33 ±0.33

0.00 ±0.00

F = 7.34p = 0.024

20.50 ±7.50

0.00 ±0.00

0.00 ±0.00

F = 7.47p = 0.024

6 12 Mar –18 Mar

45.33 ±1.76

39.33 ±1.33

36.33 ±4.10

F = 2.91p = 0.131

45.33 ±1.76

0.00 ±0.00

0.00 ±0.00

F = 663.36p = 0.000

7 19 Mar –25 Mar

41.47 ±0.88

35.33 ±2.85

33.67 ±1.86

F = 8.44p = 0.018

41.33 ±1.20

0.00 ±0.00

0.00 ±0.00

F = 1186.23p = 0.000

8 26 Mar – 1Apr

30.33 ±14.40

0.67 ±0.33

0.67 ±0.67

F = 4.23p = 0.134

44.00 ±5.00

0.00 ±0.00

0.00 ±0.00

F = 77.44p = 0.003

9 2 Apr – 8Apr

49.67 ±0.33

47.00 ±3.00

40.00 ±2.65

F = 4.64p = 0.061

49.00 ±0.58

0.00 ±0.00

0.00 ±0.00

F = 7137.34p = 0.000

10 9 Apr – 15Apr

49.33 ±0.67

42.33 ±3.38

25.00 ±6.56

F = 8.57p = 0.017

48.67 ±0.88

0.00 ±0.00

0.00 ±0.00

F = 3058.84p = 0.000

11 16 Apr – 22Apr

47.33 ±2.19

46.33 ±2.73

43.33 ±3.38

F = 0.55p = 0.604

46.33 ±2.33

0.00 ±0.00

0.00 ±0.00

F = 395.38p = 0.000

12 23 Apr – 29Apr

44.00 ±2.08

36.33 ±2.40

41.67 ±2.03

F = 3.26p = 0.110

40.00 ±2.52

0.00 ±0.00

0.00 ±0.00

F = 251.95p = 0.000

13 30 Apr – 6May

42.00 ±2.08

38.67 ±2.40

40.67 ±3.18

F = 0.42p = 0.677

40.00 ±1.53

0.00 ±0.00

0.00 ±0.00

F = 683.50p = 0.000



Mean ± SE of collected pupae Mean ± SE adult emerged

Treated TreatedWeek Test period

Untreated

10% w/w 20% w/w

One wayANOVA

Untreated

10% w/w 20% w/w

One wayANOVA

14 7 May – 13May

43.33 ±2.73

38.00 ±3.21

34.00 ±4.73

F = 1.64p = 0.271

41.67 ±2.40

0.00 ±0.00

0.00 ±0.00

F = 301.46p = 0.000

15 14 May –20 May

41.67 ±2.91

39.67 ±2.96

35.33 ±3.76

F = 1.00p = 0.420

38.33 ±1.67

0.00 ±0.00

0.00 ±0.00

F = 526.80p = 0.000

16 21 May –27 May

44.33 ±3.48

37.67 ±2.33

33.67 ±6.33

F = 1.51p = 0.294

43.67 ±3.76

0.00 ±0.00

0.00 ±0.00

F = 134.89p = 0.000

17 28 May – 3June

46.00 ±2.65

35.33 ±4.91

33.33 ±5.04

F = 2.46p = 0.166

41.67 ±1.67

0.00 ±0.00

0.00 ±0.00

F = 622.61p = 0.000

18 4 June – 10June

41.33 ±1.86

34.67 ±3.84

33.00 ±3.21

F = 2.04p = 0.210

39.67 ±2.19

0.00 ±0.00

0.00 ±0.00

F = 328.12p = 0.000

19 11 June –17 June

37.67 ±2.19

36.00 ±1.53

36.00 ±3.61

F = 0.14p = 0.874

36.67 ±1.76

0.00 ±0.00

0.00 ±0.00

F = 1302.32p = 0.000

20 18 June –24 June

36.67 ±5.78

42.00 ±2.08

40.67 ±2.60

F = 0.52p = 0.620

30.00 ±8.33

1.67 ±0.08

0.00 ±0.00

F = 12.29p = 0.008

21 25 June – 1July

Not available

22 2 July – 8July

44.00 ±1.15

37.67 ±1.45

26.33 ±10.68

F = 2.05p = 0.210

44.00 ±1.15

3.00 ±0.58

0.00 ±0.00

F = 1092.89p = 0.000

23 9 July – 15July

46.50 ±1.50

27.00 ±5.51

16.33 ±1.86

F = 13.21p = 0.010

45.50 ±2.50

12.33 ±3.53

0.33 ±0.33

F = 71.58p = 0.000


Not available


48.00 ±1.00

35.00 ±14.50

41.67 ±4.91

F = 0.35p = 0.719

47.67 ±0.33

31.67 ±12.86

9.33 ±6.98

F = 5.19p = 0.049

26 31 July – 6Aug

36.67 ±1.20

37.67 ±2.85

34.33 ±5.17

F = 0.24p = 0.792

33.00 ±3.06

33.67 ±2.40

30.67 ±5.36

F = 0.17p = 0.848

27 7 Aug – 13Aug

34.33 ±12.20

46.00 ±1.53

43.67 ±2.33

F = 0.73p = 0.520

31.33 ±11.17

42.00 ±0.58

19.33 ±9.33

F = 1.82p = 0.241

SE = standard error; p>0.05 = no significant difference; p<0.05 = significant difference;p<0.01 = highly significant difference



treated and untreated earthen jars up to 25weeks (p<0.05). The result indicated that inthe untreated jar, not all the pupae successfullyemerged as adults throughout the trial period.In the earthen jars treated with 10% w/w and20% w/w pyriproxyfen granules, although somelarvae pupated successfully, none of thesepupae could emerge as adults up to 19 and 22weeks, respectively. This finding was similarto that reported by Sihuincha et al.[7], wherelarvae continued to pupate but failed toemerge. After this, both earthen jars (treatedwith 10% and 20% pyriproxyfen granules)exhibited e”80% emergence inhibition foranother 3 weeks, indicating that both 10% and20% pyriproxyfen were able to inhibit theemergence of adult Ae. aegypti for 22 weeks(5 months) and 25 weeks (6 months),respectively (Figure).

In Cambodia, the inhibition of adultemergence of Ae. aegypti in simulated domesticwater storage containers by using controlled-release formulation of pyriproxyfen showedthat at target dosages of 18, 27 and 36 µg/L ofa.i., inhibition of adult emergence remainedabove 95% for at least two months. After threemonths at 18 µg/L a.i., the residual efficacywas significantly lower than for the higherdosages (p<0.05)[9]. At the higher dosages,inhibition of adult emergence was > or = 87%for six months[9].

The persistence and efficacy ofpyriproxyfen were evaluated in two finalconcentrations of 0.01 and 0.05 mg/L againstAe. aegypti larvae in laboratory conditions usingthree types of containers, i.e. cement box, glassbottle and plastic bucket, in the University of

Figure: Percentage emergence inhibition of Ae. aegypti exposed to 10% w/w and 20% w/wpyriproxyfen-controlled release block in earthen jars

36.66

72.90

94.43

93.18

30.33

7.06

38.30

81.34

99.27

0

10

20

30

40

50

60

70

80

90

100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 22 23 25 26 27 29

Number of week

Pe

rce

nta

ge

of

em

erg

en

cein

hib

itio

n,

%

10% w/w pyriproxyfen 20% w/w pyriproxyfen

Cut-off point of adult emergence inhibition =80%



Federal de Minas Gerais. The study indicatedthat a persistency of 45 and 90 days by using0.01 and 0.05 mg/L final concentrations ofpyriproxyfen respectively was observed[10].

In another study, pyriproxyfen was testedagainst Ae. aegypti at 0.01 and 0.02 mg of activeingredient (a.i.) per litre of water in 60 litreearthen jars. Both concentrations provided100% control for four months. Moreover, infield trial condition, pyriproxyfen at a dosageof 0.02 mg a.i. per litre provided 100% controlfor 10 weeks against Ae. albopictus[11].

In Japan, blood-fed female Ae. aegyptiwere exposed to a surface treated withpyriproxyfen at 1.0 g/m2. The results showedthat transmission of pyriproxyfen from femalesto the water was revealed[12]. Pyriproxyfenaffected the egg maturation of females treatedbefore blood meals, as the number of eggsdeposited decreased concurrently with thenumber of days before the blood meal[12].

The use of pyriproxyfen-treated ovipositioncontainers to achieve horizontal transfer ofpyriproxyfen to mosquito oviposition sites alsocan be a field management technique basedon mosquito biology and behaviour. A studyconducted in the North Carolina State

University showed that horizontal transfer ofpyriproxyfen by Ae. albopictus from a containerwith a treated ovistrip (0.3 or 0.4 mg/cm2) toan untreated microcosm resulted in 14% – 38%inhibition[13].

Besides controlling Aedes, manyresearchers have also reported that pyriproxyfenwas able to control Culex mosquitoes in Israel[14],Egypt[15], Florida[16] and Bangladesh[17]; andAnopheles mosquitoes in Sri Lanka[18] andSolomon Island[19].

The current study showed thatpyriproxyfen-controlled release block is aneffective method of controlling mosquito larvaefor several months. The method of applicationof the block is simple and straightforward, andcan therefore be used easily. This method canbe applied in areas such as drains, ponds andlakes where mosquitoes breed and in which along-term control is desired.

Acknowledgements

The authors thank the Director, Institute forMedical Research, Kuala Lumpur, for thepermission to publish this paper.

References

[1] Mulla MS, Thavara U, Tawatsin A,Chompoosri J, Zaim M, Su T. Laboratory andfield evaluation of novaluron, a new acylureainsect growth regulator, against Aedes aegypti(Diptera: Culicidae). Journal of Vector Ecology.2003 Dec; 28(2): 241-54.

[2] Julius JM, Morton B. Insect juvenile hormones:chemistry and action. New York: AcademicPress, 1972.

[3] Sparks TC, Hammock BD. Induction andregulation of juvenile hormone esterases duringthe last larval instar of the cabbage looper,Trichoplusia ni. Journal of InsecticidePhysiology. 1979; 25: 511-60.

[4] Glancey BM, Reimer N, Banks WA. Effects ofIGR Fenoxycarb and Sumitomo S-31183 onthe queens of two myrmicine ant species. In:Robert K, Meer V, Jaffe K, Cedeno A, eds.Applied Myrmecology: A world Perspective.Boulder: Westview Press, 1990. pp. 604-13.



[5] Worthing, CR and Hence, RJ, eds. The britishcrop protection council. The pesticide manual:a world compendium. 9th edn. London: UnwinBrothers Limited, 1991.

[6] Miyamoto J, Hirano M, Takimoto Y, HatakoshiM. Insect growth regulators for pest controlwith emphasis on juvenile hormone analogs:present status and future prospects. Pest Controlwith Enhanced Environmental Safety. 1993;524: 144-68.

[7] Sihuincha M, Zamora-Perea E, Orellana-RiosW, Stancil JD, Lopez-Sifuentes V, Vidal-Ore C,Devine GJ. Potential use of pyriproxyfen forcontrol of Aedes aegypti (Diptera: Culicidae)in Iquitos, Peru. Journal of Medical Entomology.2005; 42(4): 620-30.

[8] Lee WY, Zairi J, Yap HH, Adanan, CR.Integration of Bacillus thuringiensis H-14formulations and pyriproxyfen for the controlof larvae of Aedes aegypti and Aedesalbopictus. Journal of American MosquitoControl Association. 2005; 21(1): 84-9.

[9] Chang MS, Setha T, Chanta N, Socheat D,Guillet P, Nathan MB. Inhibition of adultemergence of Aedes aegypti in stimulateddomestic water-storage containers by using acontrolled-release formulation ofpyriproxyfen. Journal of American MosquitoControl Association. 2006; 22(1): 152-4.

[10] Resende MC, Gama RA. Persistence andefficacy of growth regulator pyriproxyfen inlaboratory conditions for Aedes aegypti. Revistada Sociedade Braseleira de Medicina Tropical.2006; 39(1): 72-5.

[11] Vythilingam I, Luz BM, Hanni R, Tan SB, TanCH. Laboratory and field evaluation of theinsects growth regulator pyriproxyfen (Sumilarv0.5G) against dengue vectors. Journal ofAmerican Mosquito Control Association. 2005;21(3): 296-300.

[12] Itoh T, Kawada H, Abe A, Eshita Y, RongsriyamY, Igarashi A. Utilization of bloodfed femalesof Aedes aegypti as a vehicle for the transfer ofthe insect growth regulator pyriproxyfen tolarval habitats. Journal of American MosquitoControl Association. 1994; 10(3): 344-7.

[13] Dell Chism B, Apperson CS. Horizontal transferof the insect growth regulator pyriproxyfen tolarval microcosms by gravid Aedes albopictusand Ochlerotatus triseriatus mosquitoes in thelaboratory. Medical and Veterinary Entomology.2003; 17(2): 211-20.

[14] Schwartz L, Wolf D, Markus A, Wybraniec S,Wiesman Z. Controlled-release systems for theinsect growth regulator pyriproxyfen. Journalof Agricultural and Food Chemistry. 2003;51(20): 5985-9.

[15] El-Shazly MM, Refaie BM. Larvicidal effect ofthe juvenile hormone mimic pyriproxyfen onCulex pipiens. Journal of American MosquitoControl Association . 2002; 18(4): 321-8.

[16] Nayar JK, Ali A, Zaim M. Effectiveness andresidual activity comparison of granularformulations of insect growth regulatorspyriproxyfen and s-methoprene againstFlorida mosquitoes in laboratory and outdoorconditions. Journal of American MosquitoControl Association. 2002; 18(3): 196-201.

[17] Ali A, Chowdhury MA, Hossain MI, Mahmud-Ul-Ameen Habiba DB, Aslam AF. Laboratoryevaluation of selected larvicides and insectgrowth regulators against field-collected Culexquinquefasciatus larvae from urban Dhaka,Bangladesh. Journal of American MosquitoControl Association. 1999; 15(1): 43-7.

[18] Yapabandara AM, Curtis CF. Laboratory andfield comparisons of pyriproxyfen, polystyrenebeads and other larvicidal methods againstmalaria vectors in Sri Lanka. Acta Tropica.2002; 81(3): 211-23.

[19] Okazawa T, Bakote’e B, Suzuki H, Kawada H,Kere N. Field evaluation of insect growthregulator, pyriproxyfen, against Anophelespunctulatus on north Guadalcanal, SolomonIsland. Journal of American Mosquito ControlAssociation. 1991; 7(4): 604-7.


Laboratory evaluation of Mesocyclops aspericornis as abiocontrol agent of Aedes aegypti

R. Ramanibai#, Kanniga S.

Unit of Biomonitoring, Department of Zoology, University of Madras,Guindy Campus, Chennai–600 025, India

Abstract

Mesocyclops aspericornis abounds in village ponds. Hence, the predatory capacity of M. aspericorniswas considered for use as a biological control agent for Aedes aegypti mosquitoes. In laboratoryexperiments, M. aspericornis consumed 33–50 mosquitoe larvae within 24-hours time period. M.aspericornis preyed upon only the first instar larvae of Ae. aegypti within a few seconds after theirintroduction. It started feeding on the tail portion first and ended with the head capsule. The meanvalue (triplicate) showed that the predatory capacity was 45.76 against the control 1.2. M. aspericornisprefers only the first instar mosquitoe larvae and feeds on them voraciously. When the Aedes larvaeattained the second instar stage, M. aspericornis attacked and killed them.

Keywords: Biological control; Aedes aegypti; Mesocyclops apericornis; Predatory capacity.


Introduction

In India, particularly in the state of Tamil Nadu,dengue and chikungunya have been reportedfrom many places. The National Vector-BorneDisease Control Programme (NVBDCP)recommended the Integrated VectorManagement (IVM) approach. This includesbiocontrol agents. It has been proved thatlarvicidal measures sustain mosquito populationfor a short period and require repeatedapplications of chemicals and eventuallydevelop resistance against that chemical[1]).Therefore, search for an effective biocontrol

agent to control mosquito population hasbecome top priority among researchers.Predatory fishes and zooplankton have beenwidely used as a biocontrol method to controlvector population[2,3,4]. Integration of thesemethods can be a low-cost andenvironmentally-friendly approach in controllingmosquito vectors[5,6]. Cyclopoid copepods(planktonic microcrustaceans) have beenextensively used as biocontrol agents in theSouth-East Asian countries for container-breeding mosquito species like Ae. aegypti[7].The copepod, Mesocyclops aspericornis, is aneffective predator of Ae. aegypti. It is known


Laboratory evaluation of Mesocyclops aspericornis as a biocontrol agent of Aedes aegypti

for its wide distribution and predatory efficiencyagainst several species of mosquito larvae. Thepresent study was conducted as a brieflaboratory experiment designed to understandthe mode of destruction of mosquito larvae byM. aspericornis.

Methodology

Out of a few preliminary surveys carried out inthe nearby environs of Chennai, Capital ofTamil Nadu, two ponds were identified for thecollection of cyclopoid copepods. The planktonmesh size used for the collection was 100 µm.Mesocyclops were isolated from the sampleand identified up to species level with the helpof standard keys[8,9]. M. aspericornis, once itsspecies identity was confirmed, was selectedfor experimental studies and reared in thelaboratory. Females with egg sacs collected fromthe stock were placed on a petridish and wereexamined under the dissection microscope.These were transferred into 600 ml beakerwhere 50 newly-hatched Ae. aegypti larvaewere introduced. We sacrificed the second-generation Mesocyclops collected from ourstock for experimental purposes. Fully-fed Ae.aegypti females were collected from the houseand kept in a small cloth cage for egg-laying.The mosquitoes were provided a small dish,half filled with water, and a paper strip for egg-laying. The eggs were hatched in a Petri dishand used for experiment.

Fifty newly-hatched first instar larvae wereintroduced into a 600 ml beaker containing 500ml dechlorinated water where a single M.aspericornis was introduced. The experimentlasted for 24 hours. The number of larvae thatsurvived at the end of 24 hours was recorded.Triplicates were maintained simultaneously at26±1 °C under photoperiod 12L:12D alongwith the control without the introduction ofM. aspericornis.

Results

M. aspericornis preyed upon the first instar ofAe. aegypti larvae within 24 hours, which wasrecorded. In general, they attacked the tailregion of the mosquito larvae and consumedthem. On a few occasions they left out thehead capsule of the mosquito larvae, and, attimes they killed the mosquito larvae withoutconsuming them. Ten experiments wereconducted for 10 days in the laboratory on relaybasis. M. aspericornis consumed about 33 to50 first instar larvae of Ae. aegypti within 24hours time period. The number of mosquitolarvae left inside the experimental beakersranged from nil to 17 nos. On the whole, themean predatory capacity of a single M.aspericornis was calculated at 45.75 (see Table).

Discussion and conclusions

According to Nam et al. [10], the dailyconsumption/killing average of a single M.aspericornis ranged between 16 to 41 larvae.Through continuous observations, M.aspericornis attacked the first instar larvae withina few seconds. They mainly consumed thecentral portion, leaving the head capsule.Occasionally, they just killed the larvae withoutconsuming it. Using their strong mandible theypierced and crammed the larvae into pieces.According to Lardeux et al.[11], M. aspericornisserved as a good biocontrol agent against Ae.aegypti within a three-weeks time period. Inthe present study, 10 experiments wereconducted simultaneously to know thepredatory capacity of M. aspericornis underlaboratory conditions. The maximum predatorycapacity of M. apericornis was found to be 49.3(mean value) and the minimum 39.3 (Table).Our results demonstrate that M. aspericornisis an efficient predator of Ae. aegypti underlaboratory conditions.



Table: The mean value of Ae. aegypti larvae consumed by M. aspericornis

No. of Ae. aegypti larvae consumed byS. no. Experiment no. M. aspericornis Mean value

TriplicatesA B C

1 1 42 42 45 432 2 33 40 45 39.33 3 50 50 48 49.34 4 50 44 46 46.65 5 46 50 46 47.36 6 47 47 48 47.37 7 46 46 48 46.68 8 50 48 48 48.69 9 45 47 45 45.610 10 44 44 44 44

Total mean: 45.75

References

[1] Gratz NG. Emerging and resurging vector-borne diseases. Annual Review of Entomology.1999; 44: 51-75.

[2] Russell BM, Wang J, Williams Y, Hearnden MN,Kay BH. Laboratory evaluation of two nativefishes from tropical North Queensland asbiological control agents subterranean Aedesaegypti. Journal of American Mosquito ControlAssociation. 2001; 17: 124-6.

[3] Kay BH, Nam VS, Tien TV, Yen NT, Phong TV,Diep VTB, Ninb TV, Bektas A, Aaskov JG.Control of Aedes vectors of dengue in threeprovinces of Vietnam by use of Mesocyclops(copepoda) and community-based methodsvalidated by entomologic, clinical andserological surveillance. American Journal ofTropical Medicine and Hygiene. 2002; 6: 40-8.

[4] Micieli MV, Garcia JJ, Andreadis TG.Epizootiological studies of Amblyosporaalbifasciati (Microsporidiida: Amblyosporidae)in natural populations of Aedes albifasciatus(Diptera: Culicidae) and Mesocyclops annulatus(Copepoda: Cyclopidae) in a transientfloodwater habitat. Journal of InvertebratePathology. 2001 Jan; 77(1): 68-74.

[5] Tietze NS, Hester PG, Snaffer KR, Prescott ST,Schreiber ET. Integrated management of wastetire mosquito utilizing Mesocyclops longisetus(Copepoda: Cyclopidae), Bacillus thuringiensisvar, Israelenisis, Bacillus sphaericus andmethoprene. Journal of the American MosquitoControl Association. 1994; 10: 363-73.

Acknowledgements

We thank the University Grants Commision(UGC) for financial support (grant no F,no.33-

362/2007(SR) and the anonymous reviewersfor their critical comments that helped inimproving the manuscript.



[6] Wang CH, Chang NT, Wu HH, Ho CM.Integrated control of the dengue vector Aedesaegypti in Liu-Chiu village, Ping-Tung County,Taiwan. Journal of the American MosquitoControl Association. 2000 Jun; 16(2): 93-9.

[7] Marco F, Marten G, Clark G. A simple methodfor cultivating freshwater copepods used inbiological control of Aedes aegypti. Journal ofthe American Mosquito Control Association.1992; 8: 4.

[8] Battish SK. Freshwater zooplankton of India.New Delhi : Oxford and IBH Publishing Co.Pvt. Ltd., 1992.

[9] Edmondson WT. Freshwater Biology. 2nd edn.New York : John Wiley and Sons Inc., 1959.pp 420-94.

[10] Larduex F, Loncke S, Sechan Y, Kay BH, RiviereF. Potentialities of Mesocyclops aspericornis(Copepoda) for broad scale control of Aedespolynesiensis and Aedes aegypti in FrenchPolynesia. Arbovirus Research in Australia.1992; 5:154-9.

[11] Nam VS, Yen NT, Holynska M, Reid JW, KayBH. National progress in dengue vector controlin Vietnam: survey for Mesocyclops(Copepoda), Micronecta (Corixidae), and fishas biological control agents. American Journalof Tropical Medicine and Hygiene. 2000 Jan;62(1): 5-10.


Viewpoint

Management dilemmas in the treatment ofdengue fever*

Kolitha H Sellahewa#

National Hospital of Sri Lanka, Regent’s Street, Colombo 8, Sri Lanka

Abstract

This paper is aimed at highlighting some of the dilemmas faced by clinicians in the management of adultpatients with dengue and my views in resolving these issues.

Even though early diagnosis and prompt fluid therapy are central to reduce morbidity and mortality indengue, achieving these goals are contentious issues and are often hampered by the limited access toexpensive laboratory data in most developing countries which would enable a rapid and accuratediagnosis. My viewpoint on overcoming these dilemmas is to make an early diagnosis on the clinicalfeatures, and apply clinical predictors of disease severity in selecting patients for interventions. In thisregard, diffuse blanching erythema in a patient with features of a viral fever during dengue epidemicswould suffice to diagnose and treat the patient as a dengue case. Laboratory confirmatory data areexpensive, not readily available and could delay treatment. Fluid therapy and intervention modalitiesfor thrombocytopaenia should be judged clinically on an individual basis rather than the blind, strictadherence to theoretical fluid regimens with the potential risk of fluid overloading. Capillary refill time,pulse pressure, cervical lymphadenopathy, and changes in the sensorium are useful clinical parametersfor selection of patients for intervention as well as subtle adjustments and termination of fluid therapy.A practically feasible step-wise approach to dengue management is described in this paper.

Keywords: Dengue; management; disease severity predictors.

* The conclusions and recommendations in this paper are based on the personal experiences of the authorwhile treating adult patients with dengue. The views expressed here are entirely of the author and aresubject to technical analysis or confirmation by other experts. These do no necessarily reflect theopinions or decisions of WHO. – Editor#E-mail: [email protected]

Introduction

The management dilemmas faced by busyclinicians in developing countries where denguehas reached epidemic proportions are centredon:

• Early diagnosis of dengue;

• Prediction of disease severity;

• Selection of patients for aggressiveinterventions;

• Implementation of the concept ofjudicious fluid therapy on an individualbasis to prevent fluid overloading andits attendant complications;


Management dilemmas in the treatment of dengue fever

• Concerns on thrombocytopaenia anddecisions on when to intervene, andhow to intervene.

It has been widely recognized that earlydiagnosis and prompt appropriate treatment ofdengue prevents both morbidity and mortality[1].The vast majority of adult patients with denguerecover completely without specific aggressiveinterventions (personal experience). It is thusnecessary to select the minority of patients earlyin the disease for close monitoring andappropriately timed fluid therapy to preventprogression to dengue shock syndrome (DSS).Achieving this objective, which is a fundamentalrequirement for proper management, requiresdengue to be diagnosed early and then be ableto predict disease severity. Dengue-specific IgMis positive in only 55% of patients in 4th to 7th

days while 94% positivity is evident after the7th day. No IgM is detected in 1 to 3 days afterinfection[2]. PCR amplification for dengue RNAprovides a fast diagnosis but is expensive andfalse positive reports are seen. In numerousacute dengue fever patients an early diagnosiswill be obtained only by combining IgMantibody detection with detection of virus orvirus RNA using RT-PCR[2,3].

These facilities are also not available to mostclinicians at the point of delivery of care.Consequently, they need to rely on clinical skillsto arrive at an early diagnosis. The diagnosticdilemma, however, is that the classic clinicalfeatures of dengue such as saddle-back feverand break-bone pain are not seen in all thepatients, and varying degrees of headache,myalgia, arthralgia and vomiting are common tomost non-specific viral fevers often encounteredin the community and among inpatients in amedical unit. However, what is encounteredvery often in dengue patients is diffuse blanchingerythema (personal experience)[4,5,6]. Any patientwith features of a viral fever having diffuseblanching erythema should be treated as denguefever during epidemics.

The management dilemma iscompounded by the current WHOclassification of grading dengue haemorrhagicfever (DHF) from I to IV, as there is an overlapof DHF grades III and IV with dengue shocksyndrome (DSS). Also, the case definition ofDHF requires there to be evidence of plasmaleakage[7]. However, there are many patientswith haemorrhage, particularly skinhaemorrhages, who do not have evidence ofplasma leakage and therefore cannot beclassified as DHF despite the presence ofhaemorrhages. This can confuse theinexperienced clinician and lead to unnecessaryand inappropriate overzealous fluid therapy.

A more practical classification based on theclinical findings would be more appropriate andsimpler to determine. Dengue patients withoutany haemorrhages or shock are classified asdengue fever. Patients who have diffuseblanching erythema or blanching papularerythema would also fall into this category.Blanching papular erythema should not bemistaken for petechiae. Petechiae will not blanchon pressure and implies skin haemorrhages.Dengue patients with haemorrhages,irrespective of its magnitude or site, are classifiedas DHF. Dengue patients in shock are classifiedas DSS. Intervention decisions at all levels ofcare could then be based on accurate and easilydetermined clinical parameters such as capillaryrefill time and pulse pressure.

Accurate prediction of disease severityrequires the analysis of specific data related topathogenesis and plasma leakage such as viralserotype, serum levels of non-structural protein1 (NS1), immunoglobulin G (IgG) subclass,dengue-specific and total immunoglobulin E(IgE), serum concentration of antiplateletantibodies, and levels of cytokines such as TNFα, IFN α and IL-10[8,9,10,11,12]. Non-availability ofsuch information to clinicians at the point ofdelivery of care leaves no option but to rely onclinical parameters to predict disease severity.



Thrombocytopaenia is a common problemin dengue, which causes a lot of concern notonly to the patient but also to the relations aswell as the attending physician. No clearguidelines exist for its management. The naturaltendency is to transfuse platelets.Thrombocytopaenia in dengue is primarilyimmune-mediated. It can therefore besurmised that platelet transfusions by presentinga strong antigenic stimulus can aggravate thethrombocytopaenia by an exalted immuneresponse. Besides, the short life span oftransfused platelets would result only in atransient non-sustained elevation of the plateletcount. Additionally, platelet transfusions canevoke hypersensitivity reactions and fluidoverloading with attendant complications ofpleural effusions, such as ascites, andpulmonary oedema.

Clearly, prophylactic platelet transfusionsfor dengue are baseless and appear to be anirrational and inappropriate intervention.However, transfer of patients from peripheralhospitals to tertiary care hospitals primarily forplatelet transfusions reflect the dilemmasconfronting clinicians in managingthrombocytopaenia in patients with dengue.

Management

From a practical point of view, reduction inmorbidity and mortality of dengue rests firstlyin the early diagnosis of dengue and then onthe ability to identify the minority of patientswith a propensity to develop severe diseaseearly in the disease course. Individually-tailoredparameters designed to detect potentialcomplications should be monitored in suchpatients. Critically-timed appropriateinterventions based on changes in themonitoring parameters could thereby thwartdisease progression and an adverse outcome.

Diffuse blanching erythema is a very usefulsign in the early clinical diagnosis of dengue inadults during epidemics[1,13]. This is recognizedas a generalized flushed appearance in the skin,particularly in the posterior aspect of the chestwhen observed in a good light (Figure 1). Theerythematous areas have no clear boundariesand blanch on light pressure applied with thefingers. Some patients have blanching papularerythema, particularly in the limbs (Figure 2).These appear as minute erythematous papuleswhich blanch when light pressure is appliedwith the finger tip. It should not be mistakenfor petichiae which do not blanch on pressure.Serological confirmation is not required formanagement and should be requested only if

Figure 1: Diffuse blanching erythema

Figure 2: Blanching papular erythema



there is a doubt in the diagnosis, specially todifferentiate from other febrile illnesses withmyalgia and thrombocytopaenia likeleptospirosis.

Patients with a normal sensorium, goodperipheral circulation with warm extremities,bounding pulse, capillary refilling time <2seconds, posterior cervical lymphadenopathyand platelet counts over 50 000/μL are boundto make an uneventful recovery. Patients witha pulse pressure of <20 mmHg and poorcapillary refilling will require aggressive fluidtherapy to prevent progression to DSS.Interventions in such patients should be early,decisive and aggressive. The quantum andquality of the fluid infused should be dictatedby clinical judgement and changes in themonitoring parameters, particularly pulsepressure, haematocrit and platelet count.

Peripheral pulse, capillary refill time, andpulse pressure are the most consistent andimportant parameters to base decisions oninterventions at all levels of care. However, indifficult and ambiguous situations, seekingadditional information on predictors of diseaseseverity and capillary leakage facilitatesdecision-making. These include cervicallympadenopathy, acute right hypochondrialpain and tenderness, retro-orbital pain, alteredsensorium, pleural effusions, ascites,oedematous gall bladder on ultrasonography,positive tourniquet test, platelet count, andaspartate amino transferase (AST) levels. Itshould be noted that in this context thepresence of cervical lymphadenopathy[4] andnormal levels of AST[13] are strong negativepredictors of dengue fever progressing to DHFor DSS. Such patients very often will notrequire aggressive fluid therapy. On thecontrary, the presence of any one or more ofthe other predictors of disease severitymentioned above should alert the clinician tothe potential probability of an adverse outcomeand the need for close monitoring of pulse

pressure, capillary refill time and haemotocritand to be more liberal on fluid therapy. Lessthan optimal care, both with regard tomonitoring and fluid therapy early in the diseasecourse in such patients, could result in DSS[14].

Clinicians are cautioned against aggressiveand overzealous fluid therapy in the face ofstable haemodynamics even in patients withextensive blotchy erythema and the above-mentioned predictors of disease severity, whichimply incipient increase in vascularpermeability, and increased vulnerability to fluidoverloading and attendant mortality. A dynamicapproach to management and subtleadjustments in fluid therapy, based on astuteclinical judgement, requires one to strike thecorrect balance between augmentation of fluidtherapy to offset a drop in the pulse pressureand capillary filling on the one hand anddecrementation of fluids withoutcompromising the haemodynamics on theother, when increased capillary permeabilityhas shifted the balance towards aggravation ofpleural effusions and pulmonary oedema. Theentire quantum of intravenous fluid ascalculated per guidelines need not necessarilybe given to patients with pleural effusionsprovided haemodynamic parameters aresatisfactorily maintained, usually with 1 to 1.5litres of isotonic saline infused over 24 hours.Any deterioration in the haemodynamics shouldprompt the immediate administration of anintravenous saline bolus (10 ml/kg). Satisfactorycirculation is reflected by a pulse pressure ofover 20 mmHg, capillary refill time <2 seconds,and hourly urine output of >0.5 ml/kg bodyweight. These are the parameters to bemonitored and utilized to optimize fluidtherapy.

Platelet transfusions are hardly everrequired even with counts as low as 10 000/μL because the circulating platelets arehaematologically active and sufficient to preventbleeding by thrombocytopaenia per se.



Besides, the survival of transfused platelets isvery short in cases with DSS[15]. In general,platelet transfusions are given only when thereare serious haemorrhagic manifestations.Transfusion requirements correlate with theoccurrence of bleeding in the gastrointestinaltract but not with the platelet count[16]. Thereis no place for prophylactic platelettransfusions[17].

Fresh frozen plasma is an useful and safertherapeutic option than platelet transfusions forpatients with severe thrombocytopaenia. Itsuse, however, should be reserved only forhighly selected patients with severethrombocytopaenia early in the disease[18].

In summary, the management dilemmascan be resolved by applying the following stepswhen treating a febrile patient with suspecteddengue in an adult medical ward.

Step I: Early diagnosis of dengue

It should be a clinical diagnosis based on thepresence of diffuse blanching erythema in apatient with features of a viral fever.

Step II: Classify the clinical type asDF, DHF or DSS

Step III: Risk stratification andselection of patients for specificinterventions

(1) Minor disease: Patients with a normalsensorium who do not look too ill,have a good appetite, warmextremities, pulse pressure of 20mmHg, bounding peripheral pulse,capillary refill time <2 seconds andenlarged posterior cervicallymphnodes. A majority of patients fall

into this category (personalexperience). Intravenous fluids are notmandatory and can be managedeffectively with oral fluids[7,17].Intravenous fluids will be required ifthere are excessive fluid losses dueto undue vomiting.

(2) Severe disease: Patients with alteredsensorium, sustained righthypochondrial pain, absence of cervicallymhpadenopathy, cold extremities,pulse pressure <20 mmHg, capillaryrefill time >2 seconds or any otherpredictors of severe disease should beselected for monitoring and judiciousfluid therapy.

Step IV: Monitoring and critical care

Patients with clinical markers of severe diseaseor predicted to develop severe disease shouldbe shifted to a high dependency area in theward where close observation and monitoringis feasible with the prioritized utilization oflimited facilities in a resource-poor setting.Monitoring charts in these selected patientsshould be simple and practical. The monitoringparameters should be individualized and shouldhave predictive utility to base interventionaldecisions. These include the pulse rate, pulsepressure, capillary refill time, respiratory rate,and hourly urine output. Serial estimation ofhaematocrit and platelet count in the first 24to 48 hours where facilities are available willprovide additional and useful information tobase decisions on interventions.

Judicious fluid therapy for this category ofpatients is aimed at maintaining circulatorystability in the face of continuing plasmaleakage and thereby preventing progression toDSS. The approach is less aggressive and isdictated by clinical judgement which takes intoconsideration the pulse pressure, urine outputand haematocrit. This is achieved by infusing



the minimal quantum of intravenous fluid tomaintain a pulse pressure of 20 mmHg or moreand an hourly urine output of 0.5 ml/kg bodyweight. This usually amounts to approximately1000 to 1500 ml of isotonic saline over 24hours. Necessarily the fluid intake has to beadjusted appropriately for those patients withexcessive fluid losses due to undue vomiting.Plasma leakage continues for about 24 to 48hours and good pulse volume, wide pulsepressure diuresis and stable haemotocrit areindicators to stop fluid therapy[17]. Over-treatment with fluids at this stage will have anadverse outcome and will lead to pleuraleffusions and pulmonary oedema. Progressiveincrease in the respiratory rate in the monitoringchart should alert the clinician to the possibilityof this complication, prompting a carefulevaluation of the lungs clinically, radiologicallyand measurements of oxygen saturation as wellas the need to withhold intravenous fluidsprovided minimal haemodynamic stability ismaintained.

Step V: Aggressive fluid therapy

Urgent and aggressive fluid therapy is requiredfor those with a pulse pressure <20 mmHg,cold clammy skin, rapid weak pulse andrestlessness. Aggressive intervention entails anintravenous bolus of isotonic saline orHartman’s solution at a dose of 10 ml/kg bodyweight. If the pulse pressure remains less than20 mmHg the same dose is repeated twice. Ifthere are still no signs of improvement, up totwo doses of colloid (plasma or dextran) at adose of 10 ml/kg body weight should begiven[17].

Step VI: Consider intervention forthrombocytopaenia

Platelet transfusions are hardly ever indicatedeven with counts as low as 10 000. It should

be considered only if there is significantbleeding attributable to thrombocytopaenia[17].

Consider fresh frozen plasma as atherapeutic option for thrombocytopaenia inhighly selected patients early in the diseasecourse[18].

Conclusion

The management of dengue patients,particularly during epidemics, should be basedon arriving at a clinical diagnosis early in thedisease course. Recognition of clinicalpredictors of disease severity at the time ofpresentation, avoiding unnecessary platelettransfusions, and judicious fluid therapy dictatedby clinical judgement early in the diseasecourse are of crucial importance. Properutilization of simple parameters for monitoringsuch as pulse pressure, capillary refill time andhaematocrit play a central role in the selectionof patients for interventions as well asdetermination of end points and optimizationof fluid therapy to reduce morbidity andmortality due to dengue.

Recommendations

(1) Utilize a simple and practical six-stepapproach to manage adult denguepatients in a ward setting . Thisapproach is particularly applicable tobusy overcrowded hospitals withlimited resources and little access tointensive care units, as is the case withmost rural and district hospitals indeveloping countries where dengueis common.

(2) Develop clear consensus guidelines onthe management of thrombocytopaeniain dengue.



(3) Current classification of the grading ofdengue needs to be reviewed.

(4) Aggressive educational programmestargeting care-providers at all levels to

avoid unnecessary transfers, irrationalplatelet transfusions and fluidoverloading.

References

[1] Malavige GN, Fernendo S, Fernando DJ,Seneviratne SL. Dengue viral infections.Postgrad Med J. 2004; 80: 588-601.

[2] Schilling S, Ludolfs D, Van An, Schmitz H.Laboratory diagnosis of primary andsecondary dengue infection. J. Clin.Virol. 2004Nov; 31(3): 179-84.

[3] Ole Wichmann, Klaus Stark, Pei-Yun Shu,Matthias Niedrig, Christina Frank, Jyh-HsiungHuang, Tomas Jelinek. Clinical features andpitfalls in the laboratory diagnosis of denguein travelers. BMC Infect Dis. 2006; 6: 120.

[4] Sellahewa KH, Samaraweera DN, ThusitaKPGD, Vallipuranathan M. Clinical predictorsof disease severity in dengue patients atNational hospital of Sri Lanka. Ceylon MedicalJournal. 2007; 52 (Suppl1): 28.

[5] Salem K, Shaikh I. Skin lesions in hospitalizedcases of dengue fever. J College of Physicians.&Surgeons Pakistan. 2008 Oct; 18(10): 608-11.

[6] Premaratna R, Pathmeswaran A, AmarasekaraND, Motha MB, Perera KV, de Silva HJ. Aclinical guide for early detection of denguefever and timing of investigations to detectpatients likely to develop complications. TransR Soc Trop Med Hyg. 2009 Feb; 103(2):127-31.

[7] World Health Organization, Regional Officefor South-East Asia. Prevention and control ofdengue and dengue haemorrhagic fever:comprehensive guidelines. WHO regionalpublication, SEARO, No. 29. New Delhi:WHO SEARO, 1999.

[8] Libraty DH, Young PR, Pickering D, Endy TP,Kalayanarooj S, Green S, Vaughn DW, NisalakA, Eniss F A, Rothman AL, High circulating levelsof the dengue virus non-structural protein NS1 early in dengue illness correlate with thedevelopment of dengue haemorrhagic fever. JInfect Diseases. 2002; 186(8): 1165-8

[9] Koraka P, Suharti C, Setiati TE, Mairuhu AT,Van Gorp E, Hack CE, Juffrie M, Sutaryo J, VanDer Meer GM, Groen J, Osterhaus AD. Kineticsof dengue virus specific serumimmunoglobulin classes and subclassescorrelate with clinical outcome of infection. JClinical Microbiol. 2001; 39: 4332-8.

[10] Thein S, Aaskov J, Myint TT, Shew TN, Saw TT,Zaw A. Changes in levels of anti-dengue virusIgG subclasses in patients with disease ofvarying severity. J Med Virol. 1993; 40: 102-6.

[11] Koraka P, Murgue B, Deparis X, Setiati TE,Suharti C, van Gorp EC, Hack CE, OsterhausAD, Groen J. Elevated levels of total anddengue virus-specific immunoglobulin E inpatients with disease of varying severity. J MedVirol. 2003; 70: 91-8.

[12] Mustafa AS, Elbishbishi EA, Agarwal R,Chaturvedi UC. Elevated levels of interleukin-13 and IL-18 in patients with denguehemorrhagic fever. FEMS Immunol MedMicrobiol. 2001 Apr; 30(3): 229-33.

[13] Jacobs M. Dengue. Medicine International.2005; 5(3): 46-8.



[14] Sellahewa KH. Dengue fever – predictors ofdisease severity and their influence onmanagement. Ceylon Medical Journal. 2008;53: 75–78.

[15] Isarangkura P, Tuchinda S. The behavior oftransfused platelets in dengue haemorrhagicfever. Southeast Asian J Trop Med Public Health.1993; 24 Suppl 1: 222-4.

[16] Chuansumrit A, Phimolthares V, Tardtong P,Tapaneya-Olarn C, Tapaneya-Olarn W,Kowsathit P, Chantarojsiri T. Transfusionrequirements in patients with denguehaemorrhagic fever. South East Asian J Trop MedPublic Health. 2000; 31: 10-4.

[17] Ministry of Health, Sri Lanka. Guidelines onclinical management of dengue fever/denguehaemorrhagic fever. Colombo: EpidemiologyUnit, 2005.

[18] Sellahewa KH, Samaraweera N, Thusita KP,Fernando JL. Is fresh frozen plasma effectivefor thrombocytopaenia in adults with denguefever? Ceylon Med J. 2008 Jun; 53(2): 36-40.


Short note

An outbreak of dengue in Moreh: A small rural town inManipur near Indo-Myanmar border

Kalpana Baruaha#, K. Indra Singhb, C.M. Agrawala, G.P.S. Dhillona

aDirectorate of National Vector-Borne Disease Control Programme (Directorate-General of Health Services,Ministry of Health and Family Welfare, Govt. of India), 22 Shamnath Marg, Delhi – 110 054, India

bState Vector-Borne Disease Control Programme, Medical Directorate, Lamphalpet, Imphal West,Manipur, India


Introduction

Dengue is endemic in most countries of south-east Asia. As of December 2007, out of the11 countries in the WHO South-East AsiaRegion, 10 countries (Bangladesh, Bhutan,India, Indonesia, Maldives, Myanmar, Nepal,Sri Lanka, Thailand and Timor-Leste) have beenreporting the incidence of dengue every year.Thailand, Indonesia and Myanmar continue tobe hyperendemic countries in the Region(Source: WHO-SEARO, New Delhi, 2007)[1].

In India, out of the 35 states/UnionTerritories in the country, 23 have reportedcases and deaths due to dengue. All the fourserotypes, i.e. DENV-1 to -4 have been isolatedin the country (Source. www: nvbdcp.gov.in)[2].Aedes aegypti is the most efficient vector ofdengue in India[2]. Ecology is a greatdeterminant of the occurrence of dengue. Dueto the man-made ecological and lifestylechanges, DF/DHF has now spread to rural areasas well[3,4,5,6].

In November 2007, Moreh town inChandel district of Manipur state reported thedeath of an 11-year-old girl due to denguehaemorrhagic fever. The girl was admitted withhigh fever with haemorrhagic manifestationsand was serologically confirmed for dengue-specific IgM. Following this reported denguedeath, a sero-surveillance study was carried outin the area. This note incorporates the resultsof the investigations.

Study area

Moreh is a small border town adjacent toMyanmar having a municipal council of ninewards and is a border trade centre. Thepopulation of the town is about 20 000, ofwhich 4–5000 people comprise the floatingpopulation. The area is suptropical, hot andhumid, with variable temperature ranging froma minimum of 10 °C to a maximum of 35 °C.The annual rainfall varies from 1000 mm to1200 mm.


An outbreak of dengue in Moreh

Results

Sero-surveillance

Blood samples were collected from 281clinically-suspected dengue cases with durationof ≥5 days for serological confirmation of DENV.Out of the 281 samples tested, 51 were foundpositive for DENV by IgM Mac ELISA. Foranalyses of the cases, all the 281 cases weredivided into three age groups 0 to 5 years, 5.1to 14 years, and above 14 years. The resultsare given in the Table. There was no significantdifference in the prevalence of the diseaseamong the three age groups (p=0.07797 at5%), indicating fresh introduction of the virusinfecting all age groups in the absence ofimmunity. Furthermore, when the cases wereanalysed for disease prevalence among children0 to 14 years against those aged 14.1 yearsand above, the findings revealed significantdifferences in the prevalence among both theage groups (p=0.02645 at 5%). The prevalencewas significantly higher in the age group above14 years. Thus, the probability of extra-domiciliary transmission of dengue at the worksite (border areas, trade centres) in the affectedareas cannot be ruled out. Both sexes wereequally affected. There was no significantdifference in the prevalence of the diseaseamong males and females (p=0.2415 at 5%).

Among the 51 serologically-positive cases,haemorhagic manifestation was observed in 2%of the cases.

This is the first report of DF/DHF outbreakfrom Manipur, a state in the north-eastern part ofIndia, with 51 serologically-confirmed denguecases and 9.6% DHF cases. However, dengue isrampant in the bordering country of Myanmar[1].

Since the area is endemic for both Ae.aegypti and Ae. albopictus, it is desirable toundertake comprehensive entomologicalstudies to establish the breeding habitats ofboth the species indoors and outdoors and toidentify the areas of overlap and/or co-breeding.This information is essential for developing anyvector control strategies in the state.

Acknowledgements

The authors are grateful to the state healthauthorities of Manipur and the Regional Officeof Health and Family Welfare, Imphal, foreffectively controlling the DF/DHF outbreak.The help of the Director, Regional MedicalResearch Centre (NE), Indian Council ofMedical Research, is gratefully acknowledgedfor sero-surveillance and laboratory confirmationof the febrile cases as apex referral laboratory.

0-5 years 9 1 6 0 15 15-14 years 35 5 36 3 71 8Above14 years 84 21 111 21 195 42Total 128 27 153 24 281 51

Age groupMale

Samples IgM positivetested for dengue

FemaleSamples IgM positivetested for dengue

TotalSamples IgM positivetested for dengue

Table: Results of sero-surveillance of febrile cases

An outbreak of dengue in Moreh


References

[1] Editorial. Dengue Bulletin. 2007 Dec; 31.

[2] National Vector Borne Disease ControlProgramme (NVBDCP), Directorate-Generalof Health Services, Ministry of Health & FamilyWelfare, Govt. of India. Web site: http://www.nvbdcp.gov.in/ - accessed 20 June 2009

[3] Abdul Kader MS, Kandaswamy P, AppavooNC, Anuradha. Outbreak and control ofdengue in a village in Dharamapuri, TamilNadu. Journal of Communicable Diseases.1967; 29 (1): 69-71.

[4] Ilkal MA, Dhanda V, Hassan MM, Mavale M,Mahadev PV, Shetty PS, Guttikar SN, BanerjeeK. Entomological investigations duringoutbreaks of dengue fever in certain villages inMaharashtra state. Indian Journal of MedicalResearch. 1991; 93: 174-8.

[5] Katyal R, Gill KS, Kumar K. Breeding of Aedesaegypti and its impact on dengue/DHF in ruralareas. Dengue Bulletin. 1997; 21: 93-5.

[6] Baruah K, Kumar A, Meena VR. Entomologicalinvestigations for DF/DHF in Alwar district,Rajasthan, India. Dengue Bulletin. 2004; 28:213-5.


Short note

Imported dengue fever cases in Gunma prefecture,Japan

Yukio Moritaa#, Tomoyuki Suzukia, Kunihisa Kozawaa, Masahiro Nodab,Nobuhiko Okabec, Hirokazu Kimurac

aGunma Prefectural Institute of Public Health and Environmental Sciences, 378 Kamioki, Maebashi,Gunma 371-0052, Japan

andTokyo Kasei University, 1-18-1 Kaga, Itabashi, Tokyo 173-8602, Japan

bDepartment of Virology III, National Institute of Infectious Diseases, Musashimurayama-shi, Tokyo, Japan

cInfectious Disease Surveillance Center, National Institute of Infectious Diseases,Musashimurayama-shi, Tokyo, Japan

#E-mail: [email protected], [email protected]; Tel: +81-27-232-4881; Fax: +81-27-234-8438

Dengue virus (DENV), which is classified intofour serotypes designated as DENV-1 to DENV-4, is endemic in more than 100 countries inboth tropical and subtropical regions, withsouth-east Asia and the western Pacific beingthe most seriously affected[1]. According to theWorld Health Organization (WHO), the globalprevalence of dengue fever (DF) and denguehaemorrhagic fever (DHF) has dramaticallyincreased in recent decades (http://www.who.int/mediacentre/factsheets/fs117/en/). Concern is therefore growing over DF/DHF as one of the most important mosquito-borne human infectious diseases[2]. In Japan,relatively large epidemics of DF occurredbetween 1942 and 1944 in Nagasaki, Kobeand Osaka, originating from persons repatriatingfrom the tropics during World War II; theseepidemics were eliminated in 1946[1]. In recentyears, between 10 to 70 cases of DF/DHF havebeen reported annually in Japan, all of whichwere imported as a traveller’s disease[3-6]. Thenumber of the reported DF/DHF cases hasshown an increasing trend since the late 1990s

(http://idsc.nih.go.jp/idwr/kansen/k04/k04_50/k04_50.html [in Japanese]). In 2007, a total of89 cases of DF/DHF were reported, the highestnumber recorded in Japan thus far.

Gunma prefecture, located in the north-west corner of the Kanto region on Honshuisland, has a population of approximately 2million, and 10% of its residents travel overseas.In 2007, we experienced 2 cases of DF: onepatient (Case 1) diagnosed with DENV-3 wasplausibly infected in Ho Chinh City, Viet Nam[7],and the other patient (Case 2) diagnosed withDENV-2 was infected in Kingston, Jamaica. Ingeneral, when hospital doctors in Japan suspecta case of DF, specimens are sent to a localpublic health institute for confirmation of DENVinfection. In the present two cases, specimenswere sent to our institute (Gunma PrefecturalInstitute of Public Health and EnvironmentalSciences) and virological analysis was performedfor each case. In Case 1, from a blood samplecollected on hospital day 3, we detected theDENV gene by reverse transcription-

Imported dengue fever cases in Gunma prefecture, Japan


polymerase chain reaction (RT-PCR) using theDENV-3-specific primers covering E-NS1 gene(GenBank accession number: AB362210)[8]

(Figure 1), although we could not isolate thevirus using various cell lines (Vero, RD, HEp-2,and HEL cells). In Case 2, from serial bloodsamples collected on hospital days 3 and 4,

Figure 1: Phylogenetic tree based on the E-NS1 sequences of DENV-3

[Phylogenetic distance was calculated using Kimura’s two-parameter method, and the tree wasplotted using the neighbor-joining method. Numbers at each branch indicate the bootstrapvalues of the clusters supported by that branch. Inscriptions indicate the country where the

dengue virus gene was detected, GenBank accession numbers, and collection year.]

DENV was not detected by RT-PCR, butspecific IgG and IgM antibodies against DENVwere significantly raised. In addition, DENV waspropagated and isolated from Vero cell cultures.From the isolate, we confirmed DENV-2 by RT-PCR and phylogenetic analysis for the E gene(AB470342) of the virus (Figure 2).

Thailand, AY923865, 1994

Thailand,AY912458, 1998 East Timor,AB214879,2005

Taiwan, DQ675532, 1998

Thailand,AY676353, 1997

Thailand, AY676348, 1998

Case 1 strain (AB362210)

Viet Nam,AF400029, 1996

Viet Nam,AF400034, 1996

Viet Nam,AF400035, 1996

Viet Nam,AF400033, 1996

Viet Nam,AF400027, 1996

Viet Nam,AF400031, 1996

Viet Nam,AF400030, 1996

Viet Nam,AF400026, 1996

Viet Nam,AF400038, 1996

Viet Nam, AF400036, 1996

464409

168

433

813532

119

650

252

0.01

,

,

N

am,

N

am,

Distance



As mentioned above, DF is an importantinfectious disease not only in tropical andsubtropical regions but even in countries intemperate zones such as Japan. There iscurrently no vaccine available for human use,and as in the present cases, there is thepotential for relatively severe clinicalmanifestations such as continuous high fever,hepatic disorder, leukocytopenia andthrombocytopenia to develop[2]. There is noevidence of domestic DENV transmission in

Figure 2: Phylogenetic tree based on the E sequences of DENV-2

[Phylogenetic distance was calculated using Kimura’s two-parameter method, and the tree wasplotted using the neighbor-joining method. Numbers at each branch indicate the bootstrapvalues of the clusters supported by that branch. Inscriptions indicate the country where the

dengue virus gene was detected, GenBank accession numbers, and collection year.]

Dominican Republic, AB122022, 2001

Case 2 stain (AB470342)

Puerto Rico, DQ364518, 1994

Mexico, AY449683, 2002

Puerto Rico, EU482580,1989

Nicaragua, FJ226066, 2005

Cuba, AY702059, 1997

Puerto Rico, EU687216, 2005



698

371

344

411

485

692

0.002

Distance

Japan; however, Aedes albopictus, one of themain vectors, is widely distributed across thecountry, with the exception of Hokkaido[9], andan outbreak of DF/DHF is possible in Japanonce the virus enters the country[1]. This isespecially important to consider in the light ofthe large numbers of overseas travellers whoare at risk of DENV infection. Thus, attentionto trends in DENVs and the incidence of DF/DHF in Japan is required.



References

[1] Hotta S. Dengue fever and dengue virus–achallenge to tropical medicine. JapaneseJournal of Tropical Medicine and Hygiene. 2000;28(4): 369-81.

[2] Malavige GN, Fernando S, Fernando DJ,Seneviratne SL. Dengue viral infections.Postgraduate Medical Journal. 2004; 80:588-601.

[3] Takasaki T, Kotaki A, Nishimura K, Sato Y,Tokuda A, Lim CK, Ito M, Tajima S, Nerome R,Kurane I. Dengue virus type 2 isolated froman imported dengue patient in Japan: Firstisolation of dengue virus from Nepal. Journalof Travel Medicine. 2007; 14: 445-8.

[4] Itoda I, Masuda G, Suganuma A, Imamura A,Ajisawa A, Yamada K, Yabe S, Takasaki T, KuraneI, Totsuka K, Negishi M. Clinical features of 62imported cases of dengue fever in Japan.American Journal of Tropical Medicine andHygiene. 2006; 75: 470-4.

[5] Yamada KI, Takasaki T, Nawa M, NakayamaM, Arai YT, Yabe S, Kurane I. The features ofimported dengue fever cases from 1996 to1999. Japanese Journal of Infectious Diseases.1999; 52: 257-9.

[6] Kurane I, Takasaki T, Yamada K. Trends inflavivirus infections in Japan. Emerging InfectiousDiseases. 2000; 6(6): 569-71.

[7] Morita Y, Kogure H, Sandoh M, KawashimaG, Sato Y, Nanba S, Shoda Y, Suzuki T, ShionoM, Shiobara M, Kato M, Kozawa K, Noda M,Okabe N, Kimura H. An imported dengue fevercase by dengue virus 3 (DENV-3) Infection inGunma, Japan. Japanese Journal of InfectiousDiseases. 2008; 61: 90-2.

[8] Ito M, Takasaki T, Yamada K, Nerome R, TajimaS, Kurane I. Development and evaluation offluorogenic TaqMan reverse transcriptase PCRassays for detection of dengue virus types 1 to4. Journal of Clinical Microbiology. 2004; 42:5935-7.

[9] Kobayashi M, Nihei N, Kurihara T. Analysis ofnorthern distribution of Aedes albopictus(Diptera: Culicidae) in Japan by geographicalinformation system. Journal of MedicalEntomology. 2002; 39: 4-11.


Short note

Concurrent dengue fever and bacterial septicemiaduring the 2008 dengue outbreak in Delhi

Subhash C. Arya#, Nirmala Agarwal

Sant Parmanand Hospital, Delhi-110054, India


There have been reports of unusual clinicalpresentations associated with dengue virusinfection with sporadic reports of cardiacinvolvement[1], altered consciousness withelectroencephalographic changes[2] and liverpathology[3]. During 2008, there was a suddenspurt in dengue cases in the Indian Capital,Delhi, and its adjoining areas[4]. We foundconcurrent dengue virus infection and bacterialsepticemia in three hospitalized cases at SantParmanand Hospital, a 140-bedded tertiary-care hospital catering to the population of Delhiand adjoining townships.

During the period September toNovember 2008, 125 suspected cases ofdengue were hospitalized. Of these, 114 wereconfirmed cases of dengue; three of them werealso blood culture positive: two Staphylococcusaureus and one Salmonella typhi/paratyphi A,B group. S. aureus was isolated from twofemales, aged 28 and 30 years. The femalepatient, aged 30 years, platelet count 0.98x103/µl, total leukocyte count 9600/mm3, positivefor anti-dengue virus IgG and IgM, was not onhand for any antibiotic treatment. The otherfemale, with platelet count 0.22x103/µl,leukocyte count 2000/mm3, positive for anti-dengue virus IgM, was prescribed parenteral

amoxicillin-clavulanic acid during hospitalizationand ofloxacin during her convalescence.

The S. typhi/paratyphi group A, B was isolatedfrom a 30-year-old male, platelet count 0.38x103/µl, total leukocyte count 2500/mm3, and positivefor anti-dengue virus IgG and IgM. Isolate wassusceptible to amoxicillin-clavulanic acid,amikacin, piperacillin, ceftizoxime, cefepime,aztreonam, chloramphenicol and tetracycline, butresistant to ampicillin-sulbactam, ceftazidime,cefaclor, cefotaxime, ceftriaxone, cefuroxime andtrimethoprim-sulphamethoxazole. The patientresponded to parenteral amoxicillin-clavulanic acidand amikacin.

Reports on concurrent bacterial infectionshave been meagre, with not many denguepatients with concurrent typhoid fever[5].Among 5000 cases with symptomatic dengueinfection during an outbreak in Taiwan, China,there were only seven cases with becteremiaat Chang Gung Memorial Hospital, Kaohsiung,in southern Taiwan[6]. Furthermore, during the1990s, only one of the 19 serologicallyconfirmed infants at Chon Buri RegionalHospital, Thailand, had Staphylococcus aureussepsis, who recovered with appropriatemanagement and treatment[7].

Dengue and bacterial septicemia


Dengue patients, upon a significant diseaseresolution, would by and large be asymptomaticwith normal platelet count and haematocritvalues during a short hospital stay. Theinvolvement of several organs[1-3] would implya prolonged stay. There are not many reportson concurrent dengue infection andsepticemias, which is really not all that odd.Initial symptoms of fever would camouflageany septicemia. Dengue outbreaks are knownto overwhelm the limited outpatient andinpatient facilities available in hospitals in

countries like India. Furthermore, during suchepisodes, the medical personnel there areoverworked and completely exhausted[8].

The recommended essential laboratoryinvestigations during a dengue outbreak includeprothrombin time, partial thromboplastin time,thrombin time, electrolytes and liver functiontests[8]. Addition of blood culture would bedesirable for detection and treatment of anyconcurrent septicemia.

References

[1] Kularatne SA, Pathirage MM, Kumarasiri PV,Gunasena S, Mahindawanse SI. Cardiaccomplications of a dengue fever outbreak inSri Lanka, 2005. Transactions of the RoyalSociety of Tropical Medicine and Hygiene.2007; 101(8) : 804-5.

[2] Seneviratne SL, Malavige GN, de Silva HJ.Pathogenesis of liver involvement duringdengue viral infections. Transactions of the RoyalSociety of Tropical Medicine and Hygiene.2006; 100(7): 608-14.

[3] Kularatne SAM, Pathirage MMK, Gunasena S.A case series of dengue fever with alteredconsciousness and electroencephalogramchanges in Sri Lanka. Transactions of the RoyalSociety of Tropical Medicine and Hygiene.2008; 102(10): 1053-4.

[4] Café Sentido. New Delhi sees rise in denguefever cases, to 1,008 this year. Posted on 24October 2008. Web site: http://www.casavaria.com/cafesentido/2008/10/24/675/new-delhi-sees-rise-in-dengue-fever-cases-to-1008-this-year/ - accessed 20 June2009).

[5] Sudjana P, Jusuf H. Concurrent denguehemorrhagic fever and typhoid fever infectionin adult: Case report. Southeast Asian Journalof Tropical Medicine and Public Health. 1998;29 (2): 370-2.

[6] Lee IK, Liu JW, Yang KD. Clinical characteristicsand risk factors for concurrent bacteremia inadults with dengue hemorrhagic fever.American Journal of Tropical Medicine andHygiene. 2005 Feb; 72(2): 221-6.

[7] Hongsiriwon S. Dengue hemorrhagic fever ininfants. Southeast Asian Journal of TropicalMedicine and Public Health. 2002; 33(1):49-55.

[8] World Health Organization. Denguehaemorrhagic fever: diagnosis, treatment,prevention and control. 2nd ed. Geneva: WHO,1997.


Short note

Fibre-glass drums as the key containers of Aedes aegyptibreeding in apartments occupied by expatriates in

Jeddah, Saudi Arabia

A.A.N. Aljawia, T. Mariappanb#, A. Abo-Khatwaa, Khalid M. Al-Ghamdia,Hani M. Aburasa

aKing Abdul Aziz University, Jeddah, 21589, Kingdom of Saudi Arabia

bVector Control Research Centre, Medical Complex, Indira Nagar, Puducherry 605006, India

#E-mail: [email protected]; Tel.: 91-413-2272219

Dengue fever (DF) and its severe forms denguehaemorrhagic fever (DHF) and dengue shocksyndrome (DSS) are endemic in many parts ofthe tropical world. The disease is caused byfour antigenically-related dengue viruses(DENV-1 to 4) belonging to the genus Flavivirusand transmitted by Aedes (Stegomyia) aegypti.

In recent times DF made its firstappearance in 1994 with 289 cases in Jeddah[1]

(21°29’31”N 39°11’24”E), a cosmopolitan citywhich is situated on the coast of the Red Seaand divided into 14 sub-municipalities coveringa total of 95 districts. The Ministry of Health(MoH) initiated dengue surveillance as perWHO guidelines[2] without effective vectorcontrol measures. However, noticing theupward trend of DF cases during 2004-2005[3],the Government of Saudi Arabia establishedthe dengue crisis management and mosquitocontrol programme under the Department ofPest Control and Public Health of Jeddahmunicipality. The main focus of the programmewas on vector control activities equipped withmen and materials. Control of dengue vectorswas started by the municipality through reputed“company contractors” to cover the outside

building areas which included open areas,building under construction, blocks and bricksfactories, gardens, shopping complexes, etc.The contractors applied various insecticides inthe form of larvicides and adulticides and space-spray applications with ULV thermal fogging.Mosquito surveillance activities were carriedout using light traps and ovitraps within theoperational area. This operation was assistedby the Geographical Information System (GIS)unit, in preparing baseline maps and updatingrelevant data.

For indoor activities student volunteers(SVs) studying in various colleges, schools,universities and other educational institutionswere involved in house-to-house campaign indifferent operational areas of sub-municipalitiesto carry out source reduction measures. Undereach operational area, 22 to 25 SVs wereengaged on a day-to-day control operationbetween 4.30 PM and 10.00 PM everydayexcept Friday. Two SVs formed a team andminimum nine such teams were involved inthe campaign under the supervision of a teamleader of the respective operational areas. OneSV involved recorded all potential Aedes

Fibre-glass drums as the key containers of Aedes aegypti in Jeddah, Saudi Arabia


breeding sources with the help of hand-heldtorch lights while the other SV treated allbreeding sources with required larvicide orInsect Growth Regulator (IGR) granules. Oneof the authors (T. Mariappan) who was recruitedas a consultant from the Vector ControlResearch Centre (VCRC), Puducherry, India,provided all SVs a practical hand in training in

the field with regard to identification of Aedesmosquito breeding habitats[4], collection ofdata, sampling methods of larvae fromdifferent water storage containers, and utilityof both larvicides and adulticides and handlingof spray application equipment, viz.compression sprayers, including hand-heldthermal fogging machines.

Table: Number of houses surveyed and Aedes breeding status

Drums >100 litres

No. ofhouses

surveyed(+ve)

No. ofdrums

with lidsand

locked(+ve)

No. ofdrumswithout

lock(+ve)

+No. ofdrumsopen(+ve)

Total no(+ve)

Drums<100 litres

No./+veS.

No.Street name

1 2 3 4 5 6

1. Shawas Lane 2/16(12.5)

0 1/22(4.6)

1/18(5.6)

2/40(5)

0/86

2. Shaman Lane 3/9(33.3)

0 3/14(21.4)

0/15(0)

3/29(10.3)

0/51

3. Al-Zahab Street(3000)

1/5(20)

0 1/11(9.1)

0/8(0)

1/19(5.3)

0/45

4. SurorBaa’seellane (4) 4/12(33.3)

0 3/17(17.7)

1/29(3.5)

4/46(8.7)

0/59

5. Amnai Lane (21) 4/15(26.7)

0 4/19 (21.1) 0/24(0)

4/43(9.3)

0/85

6. Marash Street (2) 0/18(0)

0 0/8(0)

0/6(0)

0/14(0)

0/142

7. Muhaymed Lane (6) 4/10(40)

0 3/12(25)

1/28(3.6)

4/40(10)

0/77

8. Al-Shurakari Street(109)

1/8(12.5)

3/14(21.43)

0/6(0)

0/7(0)

3/27(11.1)

0/28

9. Al-Badia Lane (59) 0/14(0)

0 0/21(0)

0/8(0)

0/29(0)

0/238

Total 19/107(17.76)

3/14(21.43)

15/130(11.5)

3/143(2.1)

21/287(7.32)

0/811

Figures in brackets in columns 1 2, 3, 4 and 5 indicate percentage



The present communication incorporatesthe field collection data collected on 15 July2007 (Table) at Al-Balad district from apartmentsin nine lanes occupied by non-Saudiexpatriates. Larvae and pupae collected fromthe field were reared in the laboratory forconfirmation of species.

A total of 107 houses searched for Aedesbreeding and 19 were found positive.Computation of House Index (HI) wasestimated at 17.76. A total of 1098 drums/containers were grouped into two categories{i.e. drums/containers with >100 litres(columns 2, 3, 4) and <100 litres (column 6)}.The proportion of drums/containers of >100litres (i.e. 287 numbers) versus <100 litres (i.e.811 numbers) were 1: 2.83. The drums/containers with >100 litres further classifiedinto (i) L-Ds {i.e. drum covered with lid alongwith independent lock (column 2)} (ii) drumcovered with lid without lock (column 3) and(iii) drum without any lid (open). Thepercentages of contribution of L-Ds, drum withlid without lock and drum without any lid (open)were 4.9%, 45.3% and 49.8% respectively. Thedrums/containers with <100 litres wereunclassified ones due to absence of Ae. aegyptibreeding. Other water storage containersincluded ornamental plant jars/bottles, flowerpots, draining water from air-conditioners andtheir contribution of Ae. aegypti breeding wasnil. It is apparent that only drum with >100litres only contributed Aedes breeding. Thoughan overall 7.32% drums supported Aedesbreeding , however the percentage ofcontribution by L-Ds, drums covered withoutlock and open ones were 21.43% and 11.5%and 2.1%. The study has indicated that drumswith lids have contributed Ae. aegypti

significantly higher than open ones. This canbe explained by negative phototropism of Ae.aegypti for sunlight. The overall calculatedContainer Index (CI) was 7.32.

The absence of breeding in othercontainers as outlines above may largely bedue to scarcity of rains. Locked drums had acover where the design of cover was notmosquito proof, hence allowed breeding.

Conclusion

There is a need for more extensive field surveysand statistically significant data on samplingbasis covering both Saudi and non-Saudiapartments to unravel the varied types ofstorage containers used by the civil society toassess the Aedes breeding potential to furtherstrengthen the vector control programme andto raise material for Information, Education andCommunication (IEC) material.

Acknowledgements

The authors are thankful to His Excellency theMayor of Jeddah Municipality, Jeddah, Eng.Adel M. Fakeih for his encouragement andcontinuous support of this programme and wishto thank the Deputy Mayor Eng. Khalid F. Akeelfor his efforts on services and valuablesuggestions. Thanks also due to Mr. TurkiAhamed Midad, Manager and other fieldSupervisors and SVs involved at Al-Balad areafor their technical assistance in field activities.Also thanks are due to Mrs. SundarammalRajendran Senior Librarian and InformationOfficer, VCRC towards library assistance.



References

[1] Fakeeh M, Zaki AM. Dengue in Jeddah, SaudiArabia, 1994-2002. Dengue Bulletin. 2003;27: 13-18.

[2] World Health Organization. Denguehaemorrhagic fever: diagnosis, treatment,prevention and control. 2nd edn. Geneva:WHO, 1997.

[3] Muhammad Ayyub, Adel M Khazindar, EmanH Lubbad, Shahid Barlas, Adnaan Y Alfi,

Sawsan Al-Ukayli. Characteristics of denguefever in a large Public Hospital, Jeddah, SaudiArabia. Journal of Ayub Medical College,Abbottabad. 2006; 18(2): 9-13.

[4] Barraud PJ. The fauna of British India, includingCeylon and Burma. Diptera, Vol V : FamilyCulicidae. Tribes Megarhinini and Culicini.London: Taylor and Francis, 1934.


Book review

Asia-Pacific Dengue Strategic Plan(2008–2015)

The Asia-Pacific Dengue Strategic Plan (2008-2015) has been prepared in response to theincreasing threat from dengue, which isspreading to new geographical areas andcausing high mortality during the early phaseof the outbreaks. Among an estimated 2.5billion people at risk globally, about 1.8 billion(more than 70%) reside in the Asia-Pacificregion. The development of this Strategic Planis also important to meet the requirements ofthe International Health Regulations (IHR)2005. The goal is to reverse the rising trend ofdengue in countries of the Asia-Pacific region.

Countries of the region vary in terms oftheir preparedness, their capacity to respondand in the allocation of financial resources inthe prevention and control of dengue. TheStrategic Plan provides genericrecommendations to allow local adaptation ofstrategies.

Dengue does not respect internationalboundaries. Effective dengue control is notpossible if control efforts are limited to onecountry or a few. It requires the adoption of aregional approach through collaboration

between countries and sustained partnershipsto enable countries to implement evidence-based interventions and the use of bestpractices.

The Asia-Pacific Strategic Plan would assistcountries to enhance their preparedness;enable them to promptly detect, characterizeand contain outbreaks; and limit epidemics ofdengue for effective prevention and control.This plan should be implemented in harmonywith the Strategic Framework for Asia-PacificPartnership for Dengue Prevention and Control(APDP).

The Strategic Plan should be used forelaborating national operational plans; todevelop capacity and strengthen the healthsystem; establish networking; harmonize itselfwith the APDP Strategic Framework formobilization of resources and sustain ongoinginformation exchange; and be able to advocatefor prevention and control of dengue. It wouldalso assist in increasing access to innovationssuch as with tools for the diagnosis, preventionand treatment of dengue.


Book review

Dengue prevention and control

Resolution of the WHO Regional Committee for South-East Asia (SEA/RC61/R5)

The Regional Committee,

Concerned with the emergence and re-emergence of dengue as a serious publichealth threat in countries of the Region,

Understanding that global climate changehas resulted in the emergence andreemergence of dengue in the Region with anincrease in outbreaks,

Recognizing that dengue has far-reachingcross-border and international implications,

Noting that most Member States in theRegion need to strengthen denguesurveillance, prevention and control systems,

Appreciating efforts made by the South-East Asia and Western Pacific regions fordeveloping the Asia-Pacific Dengue StrategicPlan, 2008-2015 focusing on reversing theincreasing trend of dengue,

Recognizing the importance of communityownership and multisectoral interventions askey strategies in prevention and control ofdengue, and

Having considered the report andrecommendations by the Meeting of theAdvisory Committee held in the RegionalOffice, New Delhi from 30 June – 3 July 2008,

1. URGES Member States:

(1) to implement the biregional Asia-Pacific Dengue Strategic Plan,2008-2015;

(2) to implement the primary healthcare approach to promotecommunity ownership,intersectoral collaboration andcoordination among relevantministries for effectiveimplementation of dengueprevention and control, as wellas intercountry activities,

(3) to strengthen national and cross-border surveillance to assess theburden of dengue;

(4) to implement an integratedvector management strategy as amajor preventive strategy, and

(5) to strengthen clinical competencyin diagnosis and management ofcases, and

2. REQUESTS the Regional Director:

(1) to advocate for and mobilizeadditional financial resources forstrengthening the dengueprevention and controlprogrammes in Member States;

(2) to provide technical support toMember States in implementingthe bioregional Asia-PacificDengue Strategic Plan (2008-2015);


Dengue prevention and control

(3) to facilitate the acceleration ofdengue vaccine research anddevelopment;

(4) to support Member States inassessing and monitoring theimpact of climate change ondengue, and

(5) to provide technical support toMember States in prioritizingoperations research in order tosupport evidence-based policydecisions and effective preventioninterventions.


Book review

Six countries test dengue interventions

TDRnews, March 2009 (No.82)

Eco-bio-social research

A multi-country research effort in Asia designedto study dengue transmission and then testsocial and ecosystem-based interventions islaunching its Phase II. Following the completionof the Phase I situation analysis, all six siteswill initiate tests of various community-basedmanagement approaches in April. These aredesigned to address locally identified factorsin disease transmission.

The initiative, funded by the EcoHealthProgramme of the Canadian InternationalDevelopment Research Centre (IDRC),involves multi-disciplinary research teams atuniversities and research centres in India,Indonesia, Myanmar, the Philippines, Sri Lankaand Thailand.

The initiative has been designed to improvedengue prevention through betterunderstanding of its ecological, biological andsocial (“eco-bio-social”) determinants.

Eco-bio-social research is a trans-disciplinaryresearch concept that integrates research onenvironmental, vector-epidemiological(entomological) and social factors that makecommunities vulnerable to vector bornediseases such as dengue. The aim of suchresearch is to develop inter-sectoral approachesto disease control, addressing issues thatextend beyond traditional boundaries of health-sector activities.

In the current study, research teams areexamining both effectiveness and communityacceptance of locally developed vector controlmeasures. For instance, in Yangon, Myanmar,teams will examine how use of naturalpredators and biological larvicides such asdragonfly nymphs and Bacillus thuringiensisserovar israelensis (Bti), as well as water covers,window curtains and waste control measures,may reduce vector densities. Stakeholderalliances, community partner groups andvolunteers will be involved in these activities.

In Muntinlupa City, Philippines, part of themetropolitan Manila region, the Phase II workalso aims to evaluate relative acceptance bycommunities and local governments ofcommunity-based vector control effortsinvolving solid waste management, lid coversfor key containers, larvicides and healtheducation measures. The social researchelement will include stakeholder meetings,key informant interviews and focus groupdiscussions.

Similar projects are taking place inGampaha District, Sri Lanka; Chennai, India;Yogyakarta, Indonesia; and ChachoengsaoProvince, Thailand. The project in Thailand aimsto measure the efficiency of insecticideimpregnated window curtains and watercontainer covers on the reduction of vectordensity measured by standard entomologicalindices. EcoHealth volunteers and students willbe engaged in applying the vector controlmethodologies at the community level.


Six countries test dengue interventions

“The aim is to show that denguetransmission can be reduced with anappropriate management of ecosystems, indifferent ecological environments.” said Dr OlafHorstick, a TDR technical officer involved in

TDR’s eco-bio-social efforts. “If this conceptis successful the studies should underline theimportance of inter- and intrasectoralapproaches to the control of vector-borne, butalso other communicable diseases.”


Book review

Dengue in Africa: Emergence of DENV-3,Côte d'Ivoire, 2008

Weekly Epidemiological Record, No. 11/12, 2009, 84, 85-96

Dengue in Africa

Dengue fever is an emerging pandemic thathas spread globally during the past 30 yearsas a result of changes in human ecology.Around 2.5 billion people live in the >100countries in tropical and subtropical areaswhere the dengue virus is transmitted. Morethan 70% of the disease burden occurs in Asiaand the Pacific, followed by the Americas, theMiddle East and Africa. Dengue is caused by4 serologically distinct, but closely related,viruses: dengue virus (DENV) 1, 2, 3 and 4 ofthe Flaviviridae family. This family of virusesalso includes yellow fever virus and West Nilevirus. There is good evidence that sequentialinfection with the different serotypes ofdengue virus increases the risk of more severedisease that can result in shock syndrome anddeath. The increase in incidence and thegeographical expansion of dengue have beenfacilitated by the rapid movement of all 4viruses and mosquito vectors throughinternational air travel and trade; populationincreases; global urbanization; and theabundance of disposable, non-degradablecontainers that serve as breeding sites in theperidomestic environment for the principalvector, Aedes aegypti, which maintains theurban dengue transmission cycle amonghumans[1]. Additionally, a sylvatic transmissioncycle has been documented in west Africa,where DENV-2 has been found circulating

among Erythrocebus patas monkeys andvarious sylvatic Aedes species, including Ae.taylori, Ae. furcifer, and Ae. luteocephalus[2].

Map 1 shows instances of dengue occurringin Africa during 1948–2008; the data are basedon published reports of outbreaks andserosurveys, and reports of dengue diagnosisin suspected cases of travellers returning fromAfrica[3,4]. These reports indicate there was asubstantial increase in epidemic dengue activityin Africa during the 1980s. However, becauseepidemic dengue activity in Africa has mostlybeen classical dengue fever caused by DENV-1 and DENV-2 without associated mortality,dengue has not been seen as a disease thatshould be given priority when compared withmalaria, HIV/AIDS and other diseases thatcause high morbidity and mortality in Africa.Because of this, dengue surveillance data fromAfrica are sparse, and cases and outbreaks arenot reported to WHO. During 1984–1985, thefirst major outbreak of DENV-3 in Africa wasdocumented in Pemba, Mozambique. Duringthis outbreak, most patients experiencedsecondary infections, and 2 deaths wereattributed to dengue haemorrhagic fever andshock. DENV-3 was documented again in amixed outbreak caused by DENV-2 and DENV-3 in Somalia in 1993. During the 2000s, DENV-3 caused major outbreaks with high incidenceand severity in many countries in the Americas,Asia and the Middle East[5].


Dengue in Africa: Emergence of DENV-3, Côte d'Ivoire, 2008

Although many arboviruses – such asdengue, chikungunya, Crimean-Congohaemorrhagic fever, Rift Valley fever, yellowfever and West Nile virus – affect human healthin west Africa, surveillance programmes andreference diagnostics are not consistentlyavailable except for yellow fever, for whichthere is an effective vaccine and a programmefor implementing mass vaccination followingan outbreak alert.

DENV-3 alert, Côte d’Ivoire,2008

In April 2008, an international alert was issuedwhen 3 cases of yellow fever in urban areaswere confirmed in Abidjan through the activesurveillance system for yellow fever in westAfrica. In response to this alert, WHO, theGlobal Outbreak Alert and Response Networkand the Ministry of Health conducted a yellow

fever vaccination campaign, finalized aseroprevalence survey and completed anenvironmental survey to investigate yellowfever in the mosquito vector and the monkeyreservoir[6].

A second alert was issued following thediagnosis of an acute infection caused byDENV-3 in 1 Japanese tourist and 1 Frenchexpatriate, both of whom were hospitalized(in Tokyo and Marseille, respectively) upon theirreturn from Abidjan; their visits occurredbetween May and July 2008. The virus isolatewas similar to the DENV-3 genotype that hadbeen circulating in Saudi Arabia in 2004.Additionally, the Centre National de Référencedes Arbovirus in Paris also identified positiveanti-dengue immunoglobulin M (IgM) in 7 of14 specimens collected from suspecteddengue cases occurring in travellers returningfrom Abidjan between 1 May and 31 August2008. Further analysis by reverse-transcriptasepolymerase chain reaction (RT-PCR) identified

Map 1: Published reports of dengue outbreaks and cases in Africa, 1948–2008



additional cases who were positive for DENV-3. One case, returning to France after a 2-month stay in Yamoussoukro, was RT-PCRpositive for West Nile virus. In 2008, reportsfrom the Institut de Veille Sanitaire in Paris alsoidentified a sharp increase in imported casesof dengue in travellers returning from Côted’Ivoire; of 12 cases reported between 1January 2006 and 31 August 2008, 8 werereported in 2008, 3 in 2007 and 1 in 2006.

In 2008, Côte d’Ivoire was faced with theco-circulation of yellow fever virus and DENV-3 in Abidjan city and the surrounding suburbanareas. This raises questions of whether therewas an urban transmission cycle for dengue inAbidjan and whether it was caused by anemerging serotype since DENV-3 has neverbeen reported in west Africa. To better assessthe extent of the epidemic and the risk ofdengue in Abidjan, a team from WHO andthe Ministry of Health followed up with aninvestigation in Abidjan during 2–16 September2008. The yellow fever surveillance systemhelped the followup team to obtain specimensfrom humans, mosquitoes and monkeys. Theinvestigation team also collected mosquito andserum specimens from around the residencesof confirmed dengue cases (where thisinformation was available) and established anactive case-finding system in health-carecentres in districts reporting confirmed casesof yellow fever.

The surveillance systems established inhealth-care centres by the Ministry of Healthidentified many cases of suspected classicaldengue but did not detect any severe cases ordeaths. However, specimens were notcollected and laboratory diagnosis was notavailable. DENV-3 was confirmed by RT-PCRin 2 specimens collected as part of routinesurveillance during week 19 (1 from CocodyBingerville and 1 from Adzopé). The vectorand larval indices were high and above thethreshold of 5% in these areas.

Transfer of technology forlaboratory diagnosis

In response to this concurrent epidemic andthe need for differential laboratory diagnosisof dengue and yellow fever, the InstitutPasteur in Dakar and in Paris, in coordinationwith WHO, deployed a task force to transferthis technology to the Institut Pasteur in Côted’Ivoire during 2 successive missions in lateSeptember and October 2008. Severaldiagnostic platforms were put in place.Serological tools for detecting IgG and IgMantibodies to dengue and yellow fever wereintroduced. Confirmatory platforms, such asthe plaquereduction neutralization test anddengue serotype-specific enzyme-linkedimmunosorbent assays (ELISAs), were alsointroduced. Direct detection platforms (suchreal-time RT-PCR) for detecting yellow feverand dengue, and for dengue serotypedetermination, were implemented for humandiagnosis and entomological investigations.Cell cultures for virus isolation from humanand mosquito specimens were established.Laboratory protocols and assays werevalidated on site by testing specimens fromthe 2008 epidemic. Analyses are in progress.ELISA and RT-PCR platforms were providedby the Institut Pasteur in Paris. Recombinantantigens, monoclonal antibodies, cell linesand other diagnostic and reference reagentshave been provided by the Institut Pasteurin Paris and in Dakar. An external qualitycontrol schedule has been organized for allmethods. The continuous exchange betweenthe 3 institutes in Côte d’Ivoire, Dakar andParis will likely sustain the diagnosticcapability for rapid diagnosis and confirmationof dengue virus in Côte d’Ivoire in 2009.However, there is a need to plan to assessthe risk and prepare for an appropriateresponse to dengue in Africa should anoutbreak occur.



References[1] Farrar J et al. Towards a global dengue research

agenda. Tropical Medicine and InternationalHealth. 2007; 12: 695-9.

[2] Vasilakis N et al. Evolutionary processes amongsylvatic dengue type 2 viruses. Journal ofVirology. 2007; 81: 9591-5.

[3] Sang RC. Dengue in Africa. Nairobi, Kenya.Arbovirology/Viral Haemorrhagic FeverLaboratory, Centre for Virus Research, KenyaMedical Research Institute, 2007. (http://w w w. t r o p i k a . n e t / r e v i e w / 0 6 1 0 0 1 -Dengue_in_Africa/article.pdf accessed 18 June2009).

[4] Dengue in Africa: Dengue Digest. Singapore:Novartis Institute of Tropical Diseases, 2006.(http://www.dengueinfo.org/NITDupload/docs/Publication/vol3no2.pdf; accessed 15June 2009).

[5] Messer WB et al. Emergence and global spreadof a dengue serotype 3, subtype III virus.Emerging Infectious Diseases. 2003; 9: 800-9.

[6] World Health Organization, Regional Officefor Africa. Note for the Press: 2 million peoplesuccessfully vaccinated in Abidjan following anurban yellow fever outbreak. Brazzaville: WHOAFRO, 2008. (http://www.afro.who.int/press/2008/pr20080908_1.html - accessed 19 June2009).


Instructions for contributors

Dengue Bulletin welcomes all original researchpapers, short notes, review articles, letters tothe Editor and book reviews which have adirect or indirect bearing on dengue fever/dengue haemorrhagic fever prevention andcontrol, including case management. Papersshould not contain any political statement orreference.

Manuscripts should be typewritten inEnglish in double space on one side of whiteA4-size paper, with a margin of at least oneinch on either side of the text and should notexceed 15 pages. The title should be as shortas possible. The name of the author(s) shouldappear after the title, followed by the nameof the institution and complete address. Thee-mail address of the corresponding authorshould also be included and indicatedaccordingly.

References to published works should belisted on a separate page at the end of thepaper. References to periodicals shouldinclude the following elements: name andinitials of author(s); title of paper or book inits original language; complete name of thejournal, publishing house or institutionconcerned; and volume and issue number,relevant pages and date of publication, andplace of publication (city and country).References should appear in the text in thesame numerical order (Arabic numbers inparenthesis) as at the end of the article. Forexample:

(1) Nimmannitaya S. Clinical spectrumand management of denguehaemorrhagic fever. The Proceedingsof the International Conference onDengue Haemorrhagic Fever, KualaLumpur, September 1-3, 1983:16-26.

(2) Gubler DJ. Dengue and denguehaemorrhagic fever: Its history andresurgence as a global public healthproblem. In: Gubler DJ, Kuno G (ed.),Dengue and dengue haemorrhagicfever. CAB International, New York,NY, 1997, 1-22.

(3) Nguyen Trong Lan, Nguyen ThanhHung, Do Quang Ha, Bui Thi MaiPhuong, Le Bich Lien, Luong AnhTuan, Vu Thi Que Huong, Lu Thi MinhHieu, Tieu Ngoc Tran, Le Thi Cam andNguyen Anh Tuan. Treatment ofdengue haemorrhagic fever atChildren’s Hospital N.1, Ho Chi MinhCity, 1991-1996. Dengue Bulletin.1997; 22: 150-161.

Figures and tables (Arabic numerals), withappropriate captions and titles, should beincluded on separate pages, numberedconsecutively, and included at the end of thetext with instructions as to where they belong.Abbreviations should be avoided or explainedat the first mention. Graphs or figures shouldbe clearly drawn and properly labelled,preferably using MS Excel, and all data clearlyidentified.

Articles should include a self-explanatoryabstract at the beginning of the paper of notmore than 300 words explaining the need/gapin knowledge and stating very briefly the areaand period of study. The outcome of theresearch should be complete, concise andfocused, conveying the conclusions in totality.Appropriate keywords and a running titleshould also be provided.

Articles submitted for publication shouldbe accompanied by a statement that they havenot already been published, and, if acceptedfor publication in the Bulletin, will not be


Instructions for contributors

submitted for publication elsewhere withoutthe agreement of WHO, and that the right ofrepublication in any form is reserved by theWHO Regional Offices for South-East Asia(SEARO) and the Western Pacific (WPRO).

One hard copy of the manuscript withoriginal and clear figures/tables and acomputer diskette/CD-ROM indicating thename of the software should be submitted to:

The EditorDengue BulletinWHO Regional Office for South-EastAsia

Indraprastha EstateMahatma Gandhi RoadNew Delhi 110002, IndiaTelephone: 91-11-23370804Fax: 91-11-23379507, 23370972E-mail: [email protected]

Manuscripts received for publication aresubjected to in-house review by professionalexperts and are peer-reviewed by experts inthe respective disciplines. Papers are acceptedon the understanding that they are subject toeditorial revision, including, where necessary,condensation of the text and omission oftabular and illustrative material.

Original copies of articles submitted forpublication will not be returned. The principalauthor will receive 10 reprints of the articlepublished in the Bulletin. A pdf file can besupplied on request.

D B Dengue - apps.searo.who.intapps.searo.who.int/pds_docs/B4325.pdf · attachment to the Editor at...

Documents

Transcript of D B Dengue - apps.searo.who.intapps.searo.who.int/pds_docs/B4325.pdf · attachment to the Editor at...