T

GTe

PS

a

Zb

c

RA

oUT

0d

ransactions of the Royal Society of Tropical Medicine and Hygiene (2008) 102, 194—199

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ierhea l th .com/ journa ls / t rs t

astrointestinal pathology in the Universityeaching Hospital, Lusaka, Zambia: review ofndoscopic and pathology records

aul Kellya,b,c,∗, Mwamba Katemaa, Beatrice Amadia, Lameck Zimbaa,ylvia Aparicioa, Victor Mudendaa, K. Sridutt Babooa, Isaac Zulua,b

Tropical Gastroenterology and Nutrition Group, University of Zambia School of Medicine, University Teaching Hospital, Lusaka,ambiaInstitute of Cell and Molecular Science, Barts & The London School of Medicine, London, UKDepartment of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, Keppel Street, London, UK

eceived 28 March 2007; received in revised form 16 October 2007; accepted 16 October 2007vailable online 3 December 2007

KEYWORDSEndoscopy;Gastrointestinaldiseases;Gastrointestinalhaemorrhage;Gastrointestinalcancer;Cancer;Africa

Summary There is a shortage of information on the epidemiology of digestive disease indeveloping countries. In the belief that such information will inform public health priorities andepidemiological comparisons between different geographical regions, we analysed 2132 diag-nostic upper gastrointestinal endoscopy records from 1999 to 2005 in the University TeachingHospital, Lusaka, Zambia. In order to clarify unexpected impressions about the age distributionof cancers, a retrospective analysis of pathology records was also undertaken. No abnormalitywas found in 31% of procedures, and in 42% of procedures in children. In patients with gastroin-testinal haemorrhage, the common findings were oesophageal varices (26%), duodenal ulcer(17%) and gastric ulcer (12%). Gastrointestinal malignancy was found in 8.8% of all diagnosticprocedures, in descending order of frequency: gastric adenocarcinoma, oesophageal squamous

carcinoma, Kaposi’s sarcoma, oesophageal adenocarcinoma. Data from endoscopy records andpathology records strongly suggest that the incidence in adults under the age of 45 years ishigher than in the USA or UK, and pathology records suggest that this effect is particularlymarked for colorectal carcinoma.© 2007 Royal Society of Tropicareserved.

1

Symptoms of digestive disorders are among the most com-

∗ Corresponding author. Present address: Tropical Gastroenterol-gy and Nutrition Group, University of Zambia School of Medicine,niversity Teaching Hospital, P.O. Box 50398, Lusaka, Zambia.el.: +260 1 252269; fax: +260 1 252269.

E-mail address: guts@coppernet.zm (P. Kelly).

mfmi

035-9203/$ — see front matter © 2007 Royal Society of Tropical Medicinoi:10.1016/j.trstmh.2007.10.006

l Medicine and Hygiene. Published by Elsevier Ltd. All rights

. Introduction

on causes of consultation to health workers worldwide, butrom sub-Saharan Africa there are few data on morbidity andortality relating to digestive disease. Such data would be

mportant not only to inform health-care planning, but also

e and Hygiene. Published by Elsevier Ltd. All rights reserved.

wicubdlaTssfiscsat‘p

3

A6tcdliwt(af3

3

M(b(eg(

3t

Nslower in patients with dysphagia (13%), caustic ingestion(0%) or gastrointestinal haemorrhage (16%) than in patients

Endoscopy findings in Zambia

to answer specific questions, such as the following, aboutthe epidemiology of digestive diseases in Africa: What is cur-rently the commonest cause of gastrointestinal bleeding? Isthere really an ‘African enigma’ relating to peptic ulcer-ation? Is functional dyspepsia, commonly diagnosed whenendoscopy findings are normal, less common in Africa than inindustrialised countries? Does the pattern of gastrointestinalcancer differ markedly from that in industrialised countries?Furthermore, there are to our knowledge no data availableon upper gastrointestinal disease in children in Africa.

Gastrointestinal bleeding in any population is usuallydue to peptic ulceration, gastrointestinal malignancy, orto oesophageal varices. Mallory-Weiss tears at the gastro-oesophageal junction also frequently cause bleeding, but itis unusual for this sort of bleeding to be life-threatening.Studies from different parts of Africa give different resultsregarding the dominant cause of bleeding. In Kenya andMalawi, oesophageal varices were the commonest cause(Harries and Wirima, 1989; Lule et al., 1994), but inZimbabwe and Cameroon, peptic ulceration was domi-nant (Ndjitoyap Ndam et al., 1990; Wicks et al., 1975).Development of effective guidelines for management of gas-trointestinal haemorrhage should be informed by reliable,recent data on the dominant causes of bleeding.

The ‘African enigma’ is a theory that peptic ulceration isuncommon in Africa despite a high prevalence of Helicobac-ter pylori (around 80%) (Holcombe, 1992). The idea thatpeptic ulceration is uncommon in Africa was, however, care-fully refuted by Cook as early as 1980 (Cook, 1980), beforethe aetiological role of H. pylori was identified. However,the theory continues to be discussed. In an earlier study, weshowed that the prevalence of peptic ulceration in Lusakais not low (Fernando et al., 2001), and a systematic reviewof studies from Africa confirms that peptic ulcer disease isas common in Africa (24.5%) as in industrialised countries(12—25%) (Agha and Graham, 2005). A recent text of gas-troenterology in the tropics confirms that the prevalenceof non-ulcer dyspepsia in the tropics is poorly documented(Holcombe, 1995). Likewise, there are few data on the inci-dence of gastrointestinal malignancies in Africa (Okobia,2003), which would be needed to formulate managementpolicies. If gastrointestinal malignancies are as uncommonin patients under the age of 40 years in Africa as they are inpatients in industrialised countries, then symptomatic treat-ment of dyspepsia in patients under 40 years of age, asrecommended by Holcombe (1995), is justified. If not, thispolicy needs revision and endoscopy needs to be offeredirrespective of age.

In the belief that an analysis of endoscopic findings wouldbegin to answer many of these questions, we carried out aretrospective audit of 2132 diagnostic upper gastrointestinalendoscopies performed in the University Teaching Hospital(UTH), Lusaka, Zambia. When it became apparent that theage distribution of digestive cancers followed an unexpectedpattern, we also reviewed all gastrointestinal cancer diag-noses from pathology records.

2. Materials and methods

The endoscopy unit in the UTH was commissioned in 1987.The unit serves the hospital wards and outpatient clinics,

waaht

195

ith some referrals from other hospitals and private clin-cs in Zambia. Copies of all reports are kept in the unit. Weonducted a retrospective review of all reports relating topper gastrointestinal endoscopy over the period Septem-er 1999—September 2005. All records were entered onto aatabase in Epi6 (CDC, Atlanta, GA, USA and WHO, Switzer-and, Geneva) by one of the authors (MK) and analysedfter conversion into Stata 8 (Stata Corp., College Station,X, USA). The distribution of pathologies was analysed byex and by gastrointestinal bleeding; the results are pre-ented as odds ratio (OR) with CI and P-values estimatedrom Fisher’s exact test. In order to clarify unexpected find-ngs regarding age and sex of patients with malignancies, aecond retrospective audit was conducted of all histologi-al analysis of suspected gastrointestinal cancers over theame 5-year period. Pathology records were reviewed andny record from oesophagus, stomach, small or large intes-ine or rectum that mentioned ‘cancer’ or ‘malignancy’ orinvasive’ was reviewed by a gastroenterologist (PK) and aathologist (VM) together.

. Results

ll 2187 upper gastrointestinal endoscopy reports from theyears of the audit were reviewed; of these, 55 were

herapeutic and 2132 were diagnostic. Therapeutic pro-edures included 40 variceal ligation and 15 oesophagealilatation procedures for fixed strictures (n = 9) or for acha-asia (n = 6); these are not included in the analysis. Activityncreased gradually over time as endoscopes were replacedith Pentax FG-29W gastroscopes with a video camera, and

he number of endoscopists increased from three to fivenot including visiting endoscopists). Of the patients whosege and sex were recorded, 1100 were male and 941 wereemale, and their mean ages by sex were 41.5 (SD 18) and9.7 (SD 19) years, respectively.

.1. Indications

ajor indications for endoscopy were epigastric pain37%), non-localised abdominal pain (28%), gastrointestinalleeding (11%), dysphagia (8%), dyspepsia (8%), vomiting4%), heartburn (0.7%) or anaemia (0.5%). Indications forndoscopy did not differ by sex, except that as expectedastrointestinal haemorrhage was more likely among menrelative risk 1.45; 95% CI 1—2.1; P = 0.05).

.2. Endoscopic diagnoses: upper gastrointestinalract

o abnormality was found in 31% of all procedures. Noturprisingly, the proportion of normal endoscopies was

ith dyspepsia or abdominal pain (54%). Diseases usuallyttributable to H. pylori, i.e. gastric ulcer, duodenal ulcernd macroscopic ‘gastritis’, were common (Table 1). Theigh prevalence of oesophageal candidiasis (12.8%) reflectshe severity of the HIV epidemic in Zambia.

196 P. Kelly et al.

Table 1 Upper gastrointestinal endoscopy findings inpatients attending the University Teaching Hospital, Lusaka,Zambia

Finding Frequency Percentageof 2132

Normal 658 31Gastric ulcer 158 7.4Duodenal ulcer 231 10.8Oesophageal candidiasis 272 12.8Gastritis 436 20.4Reflux oesophagitis 124 5.8Hiatus hernia 11 0.5Oesophageal varices 79 3.7Gastric varices 5 0.2Caustic damage 4 0.2Oesophageal cancer 51 2.3Gastric cancer 113 5.3Gastric outlet obstruction 26 1.2Kaposi’s sarcoma 26 1.2Gastric erosions 15 0.7Oesophageal stricture 30 1.4Duodenitis 186 8.7Duodenal growth/polyp 10 0.5Oesophageal ulcer(s) 7 0.3Worms (Ascaris) 3 0.1Gastric heterotopic pancreas 2 0.1Gastric and duodenal ulcers 6 0.3Duodenal obstruction 6 0.3Oesophageal achalasia 4 0.2

3

Sgt

Table 3 Gastrointestinal cancers in pathology recordsaccording to age below or above 45 years in patients attend-ing the University Teaching Hospital, Lusaka, Zambia

Site Age(years)

Total

<45 >45

Oesophagus—squamouscarcinoma

12 (27.9) 31 43

Oesophagus—adenocarcinoma 0 2 2Stomach (type not

specified)11 (25.0) 33 44

Small intestine 2 0 2Colon 47 (57.3) 35 82Intestine, site not 1 1 2

crr

3

Gqfst1ct

Total 2463 in 2132patients

.3. Gastrointestinal haemorrhage

ignificant findings in patients who presented with upperastrointestinal bleeding are shown in Table 2. In contrasto experience in industrialised countries, the commonest

pueit

Table 2 Findings in patients presenting with upper gastrointesZambiaa

Finding Frequency (%) in 179 patients w

Varices 46 (25.7)Duodenal ulcer 31 (17.3)Normal 27 (15.1)Gastric ulcer 21 (11.7)Gastric cancer 10 (5.6)Gastritis +/− erosions 16 (8.9)Oesophageal candidiasis 13 (7.3)Kaposi’s sarcoma 4 (2.2)Reflux oesophagitis 3 (1.7)Mallory-Weiss tear 2 (1.1)Oesophageal candidiasis 1 (0.6)Other 5 (2.7)

NA: not applicable; NS: not significant.a Only 1627 patients analysed, as the indication was not always record

specifiedTotal 73 102 175

Values in parentheses are percentages.

ause was variceal bleeding, but all common diagnoses wereepresented. Oesophageal candidiasis should probably beegarded as an incidental finding.

.4. Gastrointestinal malignancy

astric cancer was diagnosed endoscopically twice as fre-uently as oesophageal cancer (Table 1). Both cancers wereound more commonly in men, but the difference was nottatistically significant. In view of the age distribution ofhese cancers, we carried out a retrospective analysis of75 surgical pathology records relating to gastrointestinalancer over the same period. Both sources showed a consis-ent pattern, with a surprising proportion of cases in young

eople (Table 3; Figure 1). Oesophageal cancer in patientsnder 45 years of age represented 16% of all those diagnosedndoscopically and 28% of pathology reports. Gastric cancern patients under 45 years constituted 33 and 25%, respec-ively. Surgical pathology records in addition showed that

tinal bleeding at the University Teaching Hospital, Lusaka,

ith bleeding Association with bleeding

Odds ratio (95% CI) P-value

30.1 (16—60) <0.0011.7 (1.1—2.7) 0.01NA2.0 (1.2—3.4) 0.01NS0.51 (0.3—0.8) 0.004NSNSNSNSNSNS

ed.

Endoscopy findings in Zambia 197

Figure 1 Age distribution of oesophageal (A), gastric (B) and colorectal (C) cancers in the University Teaching Hospital, Lusaka,of edige

nosc

(oda

4

Tc

Zambia. Two data sources were used: (1) retrospective analysisanalysis of pathology records for all diagnoses of cancer of thecancer, only pathology records were available, as very few colo

an even higher proportion of colorectal carcinomas werein adults younger than 45 years (47 of 82, 57%), althoughvery few colonoscopies were performed over this period andthese records were not reviewed.

3.5. Findings in children

Altogether, 55 diagnostic procedures were performed onchildren (under the age of 16 years; 20 boys, 33 girls, 2 sex

not recorded), and 23 (42%) were normal. Significant findingswere duodenitis or gastritis (6), gastric ulceration (5), refluxoesophagitis (5), candidiasis (4), oesophageal varices (4),benign oesophageal stricture (3), caustic injury (1), duode-nal tumour (1), gastric tumour (1) and achalasia of the cardia

wsupp

Table 4 Estimation of incidence of gastrointestinal cancers inpopulation distribution in 2001—2002

Age group(years)

Population inZambia

Age-specific incidUSA per 100 000 p

Oesophageal

20—24 887 400 0.025—29 795 600 0.030—34 591 600 0.135—39 459 000 0.340—44 346 800 1.0

Total annual incidenceTotal cancers expected

ndoscopy records over the period 1999—2005 (black bars); (2)stive tract over the same period (white bars). For colorectal

opy procedures were carried out over this time period.

1). All the varices and gastric ulcers were found in childrenver the age of 10 years. In addition to diagnostic proce-ures, variceal ligation was performed on four occasions,nd stricture dilatation was performed once.

. Discussion

here is a severe shortage of data on the profile of non-ommunicable disease in Africa, so even hospital-based data

ith all their inherent bias relating to patient access and

election can still contribute usefully to epidemiologicalnderstanding. Our data suggest that in this population dys-epsia is commonly functional (i.e. not due to demonstrableathology) as in industrialised countries, that peptic ulcer is

Zambia in 6 years, using US incidence rates and Zambian

ence iner year

Number of cancers per yearexpected in age group

Gastric Colon Oesophageal Gastric Colon

0.0 0.2 0 0 1.770.4 0.4 0 3.04 3.040.8 1.0 0.59 4.73 5.921.3 2.0 1.38 5.97 9.182.4 4.0 3.47 8.32 13.87

5.44 22.06 33.7833 132 203

1

chdutw1sfao2hdtol

tcutbfi(suip(ttd13aaa5tptwLpuauowFtataceeoma

rswcbs

tedrrdsm

ottotodteitsw

AMceSraa

AStZC

FB

C

Eait

98

ommon (18.2% of all endoscopies) and that gastrointestinalaemorrhage is most frequently due to varices. Most varicealisease in our patients is due to schistosomiasis (Paul Kelly,npublished observations), and the high burden of schis-osomiasis in Africans is an important point for physiciansho treat migrants and travellers from Africa (Obeid et al.,983). In Kenya in 1982, 30% of variceal bleeds were due tochistosomiasis (De Cock et al., 1982), 35% in a later studyrom Kenya (Lule et al., 1994) and 45% in Malawi (Harriesnd Wirima, 1989). Variceal ligation is effective for controlf variceal bleeding in schistosomiasis (El-Saify and Mourad,005), and these patients, having preserved liver function,ave a better outcome than patients with variceal bleedingue to cirrhosis. In UTH we have only had one fatality dueo gastrointestinal bleeding from varices in 15 years, and tour knowledge no patients have developed decompensatediver failure following gastrointestinal bleeding.

Our data suggest that Zambian patients with gastroin-estinal cancer are surprisingly young. Is this simply aonsequence of the age distribution of a typical African pop-lation, in which around half of the population is underhe age of 16 years of age? Could it be due to referralias, with younger people being more likely to be referredor tertiary care? In the absence of sound age-standardisedncidence rates derived from formal cancer registry dataOkobia, 2003), we cannot exclude these possibilities, butome simple calculations suggest that the first of these isnlikely. Using age-specific incidence rates for these cancersn the USA for 2001—2002 (Ries et al., 2006) and the knownopulation distribution in Zambian adults for 2001—2002Central Statistical Office, Zambia, 2003) we can estimatehe expected incidence for these cancers if the rate werehe same as in the USA (Table 4). We would expect the inci-ence of these cancers in Zambia, with a total population of0.3 million (Central Statistical Office, Zambia, 2003), to be3 cases of oesophageal cancer in 6 years, and 132 gastricnd 203 colorectal cancer in adults between the ages of 20nd 44 years. We diagnosed 32 cases of oesophageal cancernd 72 cases of gastric cancer in this age group over theseyears just in one endoscopy unit, in a health care sys-

em characterised by inadequate diagnostic facilities andoor referral flows from rural areas. It is highly unlikelyhat such a high proportion of all young Zambian patientsith gastrointestinal malignancies would find their way tousaka, be referred appropriately, make all the requisiteayments of fees and get endoscopy in our one endoscopynit. It is therefore probable that incidence rates in thisge group are considerably higher than in the USA. If wesed incidence rates in the UK (Cancer Research UK, 2006),ur estimates of expected incidence of cancers in Zambiaould be somewhat higher, but the point remains the same.urthermore, the high proportion of our cases in adults inhe third and fourth decades of life (Figure 1) suggest thatge-specific incidence does not follow the pattern in indus-rialised countries, in which incidence in these age groups islmost negligible and is strongly predictive of familial can-er. The high prevalence of H. pylori, which we recently

stimated to be 81% (Fernando et al., 2001), would not itselfxplain why gastric cancers should occur in the third decadef life. It has been suggested that cancer of the oesophagusay be related to diet (Mengesha and Ergete, 2005), andrecent paper suggests that oesophageal cancer may be

R

A

P. Kelly et al.

elated to human papilloma virus (Sitas et al., 2007). In atudy from Nigeria, 73% of patients with colorectal cancerere under the age of 50 years (Edino et al., 2005), andolorectal cancer in exceptionally young Nigerians has alsoeen reported elsewhere (Seleye-Fubara and Gbobo, 2005),o it is likely that this is not an isolated finding.

Availability of endoscopy for children is a valuable fea-ure of our service. Oesophageal varices can be treatedndoscopically by ligation, and strictures and achalasia byilatation, thus reducing morbidity and mortality withoutecourse to surgery. The aetiology of varices in childrenemains to be determined. Most were found in older chil-ren, which would be consistent with schistosomiasis, but inome the possibility remains that neonatal umbilical sepsisay have precipitated portal vein thrombosis.Much remains to be learned about the epidemiology

f digestive disease in Africa. There seems little doubthat H. pylori-related peptic ulcer disease is common andhat schistosomiasis-related portal hypertension (leading toesophageal varices) is also common. Much more work needso be done on digestive cancer in Africa, which on the basisf our preliminary data and comparison with good registryata seems to have a higher incidence in younger adultshan in the UK or USA. Our data indicate that a policy of notndoscoping patients under the age of 40 years, as practisedn industrialised countries and recommended in a gastroen-erology test for the tropics (Holcombe, 1995), may not beafe, as a significant proportion of gastrointestinal cancersill go undetected.

uthors’ contributions: PK, KSB and IZ designed the study;K and PK carried out the primary analysis; BA, PK, IZ and LZollected data and reviewed the analysis in the light of theirxperience; VM and PK carried out the pathology review;A analysed the UK and US cancer statistics. All authorseviewed and amended the manuscript during preparationnd read and approved the final version. PK and IZ are guar-ntors of the paper.

cknowledgements: We are grateful for the hard work oftayner Mwanamakondo, Rosemary Soko and Janet Sakala ofhe endoscopy unit, University Teaching Hospital, Lusaka,ambia. In commemoration of our colleague, the late Druster Mulia.

unding: PK is supported by The Wellcome Trust of Greatritain.

onflicts of interest: None declared.

thical approval: As the work reported was a retrospectiveudit of clinical data, it was cleared as not raising ethicalssues by the University of Zambia Research Ethics Commit-ee, Lusaka, Zambia.

eferences

gha, A., Graham, D.Y., 2005. Evidence-based examination of theAfrican enigma in relation to Helicobacter pylori infection.Scand. J. Gastroenterol. 40, 523—529.

M

N

O

O

R

S

S

Endoscopy findings in Zambia

Cancer Research UK, 2006. CancerStats. http://info.cancerresear-chuk.org/cancerstats/ [accessed July 2007].

Central Statistical Office, Zambia, 2003. Zambia Demographic andHealth Survey 2001-2002. Central Statistical Office, Lusaka,Zambia; Central Board of Health, Lusaka, Zambia; and ORCMacro, Calverton, MD, USA.

Cook, G., 1980. Tropical Gastroenterology. Oxford University Press,Oxford, UK, pp. 44—47.

De Cock, K.M., Awadh, S., Raja, R.S., Wankya, B.M., Lucas, S.B.,1982. Esophageal varices in Nairobi, Kenya: a study of 68 cases.Am. J. Trop. Med. Hyg. 31, 579—588.

Edino, S.T., Mohammed, A.Z., Ochicha, O., 2005. Characteristics ofcolorectal carcinoma in Kano, Nigeria: an analysis of 50 cases.Niger. J. Med. 14, 161—166.

El-Saify, W.M., Mourad, F.A., 2005. Use of the six-shooter ligationdevice in the management of bleeding esophageal varices: adeveloping country experience. Dig. Dis. Sci. 50, 394—398.

Fernando, N., Holton, J., Zulu, I., Vaira, D., Mwaba, P., Kelly, P.,2001. Helicobacter pylori infection in an urban Zambian popu-lation. J. Clin. Micro. 39, 1323—1327.

Harries, A.D., Wirima, J.J., 1989. Upper gastrointestinal bleeding inMalawian adults and value of splenomegaly in predicting sourceof haemorrhage. East Afr. Med. J. 66, 97—99.

Holcombe, C., 1992. Helicobacter pylori: the African enigma. Gut33, 429—431.

Holcombe, C., 1995. Upper abdominal pain, in: Watters, D., Kiire,

C. (Eds), Gastroenterology in the Tropics and Subtropics. Macmil-lan, London, pp. 269—295.

Lule, G.N., Obiero, E.T., Ogutu, E.O., 1994. Factors that influencethe short-term outcome of upper gastrointestinal bleeding atKenyatta National Hospital. East Afr. Med. J. 71, 240—245.

W

199

engesha, B., Ergete, W., 2005. Staple Ethiopian diet and cancerof the oesophagus. East Afr. Med. J. 82, 353—356.

djitoyap Ndam, E.C., Koki Ndombo, P.O., Fouda, O.A., Mougnutou,R.S., Nguemne, T.A., Behle, A., Tzeuton, C., Malonga, E.,Essomba, R., 1990. Upper digestive system hemorrhages inCameroon (apropos of 172 cases examined via endoscopy). Med.Trop. (Mars) 50, 181—184.

beid, F.N., Smith, R.F., Elliott, J.P., Reddy, D.J., Hageman, J.H.,1983. Bilharzial portal hypertension. Arch. Surg. 118, 702—708.

kobia, M.N., 2003. Cancer care in sub-Saharan Africa — urgentneed for population-based cancer registries. Ethiop. J. HealthDev. 17, 89—98.

ies, L.A.G., Harkins, D., Krapcho, M., Mariotto, A., Miller, B.A.,Feuer, E.J., Clegg, L., Eisner, M.P., Horner, M.J., Howlader,N., Hayat, M., Hankey, B.F., Edwards, B.K. (Eds.), 2006. SEERCancer Statistics Review 1975—2003. National Cancer Institute,Bethesda, MD, http://seer.cancer.gov/csr/1975 2004/. Basedon November 2005 SEER data submission, posted to the SEERwebsite 2006 [accessed July 2007].

eleye-Fubara, D., Gbobo, I., 2005. Pathological study of colorectalcarcinoma in adult Nigerians: a study of 45 cases. Niger. J. Med.14, 167—172.

itas, F., Urban, M., Stein, L., Beral, V., Ruff, P., Hale, M., Patel, M.,O’Connell, D., Yu, X.Q., Verzijden, A., Marais, D., Willliamson,A.L., 2007. The relationship between anti-HPV-16 IgG seropos-itivity and cancer of the cervix, anogenital organs, oral cavity

and pharynx, oesophagus and prostate in a black South Africanpopulation. Infect. Agent Cancer 2, 6 (Open Access).

icks, A.C., Thomas, G.E., Clain, D.J., 1975. Comparison of fibre-optic endoscopy in acute upper gastrointestinal haemorrhage inAfricans and Europeans. Br. Med. J. 4, 259—260.