A

tTeewvi©

K

1

edSa(tntai

0d

Available online at www.sciencedirect.com

Vaccine 25 (2007) 7848–7857

Typhoid fever vaccines: Systematic review and meta-analysis ofrandomised controlled trials

Abigail Fraser a,∗, Mical Paul b,c, Elad Goldberg d,Camilo J. Acosta e, Leonard Leibovici c,d

a Department of Social Medicine, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8 2PR, UKb Unit of Infectious Diseases, Rabin Medical Center, Beilinson Campus, Petah-Tikva 49100, Israel

c Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israeld Department of Medicine E, Rabin Medical Center, Beilinson Campus, Petah-Tikva 49100, Israel

e International Vaccine Institute (IVI), Seoul, Republic of Korea

Received 7 February 2007; received in revised form 31 May 2007; accepted 9 August 2007Available online 4 September 2007

bstract

We undertook a systematic review and meta-analysis of randomised controlled trials comparing a typhoid fever vaccine with any alternativeyphoid fever vaccine or inactive agent. Trials evaluating killed whole-cell vaccines were excluded. The cumulative efficacy at 3 years for they21a and the polysaccharide Vi vaccine were similar: 51% (95%CI 36%, 62%), and 55% (95%CI 30%, 70%), respectively. The cumulativefficacy of the Vi-rEPA vaccine at 3.8 years was higher, 89% (95%CI 76%, 97%), but this vaccine has not yet been licensed for use and wasvaluated in only one trial. Adverse events were mild in nature and for most, not significantly more frequent in any of the vaccine groupshen compared with placebo. Both the currently licensed Ty21a and Vi vaccine, are safe and efficacious for preventing typhoid fever. Neither

accine is currently registered for administration to children below 2 years of age. Given the recent finding that typhoid fever also affectsnfants, development of a conjugate vaccine is warranted.

2007 Elsevier Ltd. All rights reserved.

eywords: Salmonella typhi; Typhoid fever; Vaccine; Systematic review; Meta-analysis

wb

raaaicb

. Background

Typhoid fever remains a major cause of morbidity with anstimated global incidence of 22 million cases and 200,000eaths per year. It is a substantial public health problem inouth and East Asia (>100 cases per 100,000 people annu-lly), Africa, Latin America, the Caribbean and Oceania10–100 cases per 100,000 people annually), except for Aus-ralia and New Zealand [1]. Typhoid is rare in industrialisedations, though travellers to endemic countries may acquire

he disease [2]. It is now recognised that typhoid fever, besidesffecting school-age children and young adults, is also anmportant cause of morbidity among infants and toddlers in

∗ Corresponding author. Tel.: +44 117 3313803; fax: +44 117 3313815.E-mail address: Abigail.fraser@bristol.ac.uk (A. Fraser).

inc

dc

264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2007.08.027

hom high rates of complications and hospitalisation haveeen documented [3,4].

Humans are the only source of infection and due to theoute of transmission, improved sanitation and food hygienere important control measures. However, these are associ-ted with affluence and are slow to improve in most endemicreas. Typhoid fever is treated with antibiotics, but increas-ng resistance rates to the primary agents used (ampicillin,hloramphenicol, co-trimoxazole as well as quinolones) haveeen associated with complications and increased severity ofllness [5]. Therefore, if effective and cost-effective, vacci-ation against typhoid fever could constitute an important

ontrol measure.

The first inactivated whole-cell typhoid vaccine was intro-uced in 1896. Its efficacy was established only in 1960 inontrolled trials in Yugoslavia, the Soviet Union, Poland, and

cine 25

Geqwaawr

aapmhusma

coytavS

iitirAcaCtgecha

tv

2

amcota

ie

CT2tatttoia

iafd

qwttacustoi

fAeapett5Ae

ciattFfa

A. Fraser et al. / Vac

uyana. Two doses of this type of vaccine resulted in 73%fficacy over 3 years (95%CI: 65–80%) [6,7]. Given the fre-uency of side events (10% of vaccinnees missed school orork after vaccination), these vaccines are considered unsuit-

ble for use as a public health vaccine and although licensed,re no longer available for use [8]. Consequently, killedhole-cell vaccines will not be evaluated in this systematic

eview.Two typhoid vaccines are currently commercially avail-

ble, Ty21a (an attenuated strain of S. enterica var. Typhidministered orally) and Vi (based on the purified capsularolysaccharide S. enterica var. Typhi Vi antigen, given intra-uscular. Neither has been widely adopted as a routine public

ealth tool in any country with endemic typhoid fever. These of Vi vaccine, as a public health tool, has been limited inome typhoid endemic regions in China and Vietnam whereost of the burden still occurs among school-age children

nd young adults [9,10].The live oral vaccine Ty2la is available in enteric-coated

apsule or liquid formulation. It is given in three doses everyther day and is approved for use in children aged at least 5ears. It elicits protection starting from 10 to 14 days afterhe third dose. Travellers should be revaccinated annually,nd those living in disease endemic areas every 3 years. Thisaccine is licensed in 56 countries in Africa, Asia, Europe,outh America, and the USA [11].

The Vi polysaccharide vaccine is given in a single dosentramuscular. Protection begins 7 days after injection, max-mum protection being reached 28 days after injection whenhe highest antibody concentration is obtained. The vaccines approved for persons aged over 2 years. Revaccination isecommended every 3 years. The Vi vaccine is licensed inustralia and in more than 92 countries in Africa, the Ameri-

as, Asia, and Europe [11]. Vi is not a patent protected vaccinend therefore its production in developing countries such ashina, Vietnam and Indonesia has been possible through

ransfer of technology. A new modified Vi vaccine conju-ated to a nontoxic recombinant Pseudomonas aeruginosaxotoxin A (rEPA) has also been evaluated in a randomisedontrolled trial among children aged 2–5 years. This vaccineas the potential of being immunogenic in infants under thege of 2 [12].

We undertook a systematic review and meta-analysiso evaluate the efficacy and safety of typhoid feveraccines.

. Methods

We searched for randomised controlled trials comparingtyphoid fever vaccine (live oral vaccine Ty2la or geneticodifications of this strain, Vi polysaccharide vaccine, Vi

onjugate vaccine) with any alternative typhoid fever vacciner inactive agent (placebo or vaccine for a disease other thanyphoid fever), using the terms “typhoid” and salmonella”nd similar as well as appropriate medical subject head-

ttfi

(2007) 7848–7857 7849

ngs. Trials evaluating killed whole-cell vaccines werexcluded.

We searched the following electronic databases: Theochrane Central Register of Controlled Trials (CEN-RAL), published in The Cochrane Library (Issue 2,006), MEDLINE (1966 to April 2006), EMBASE (1974o April 2006), LILACS (1982 to April 2006). Welso searched the metaRegister of controlled clinicalrials (http://www.controlled-trials.com/), the National Insti-utes of Health database (http://clinicaltrials.gov/), andhe web for new drug application (NDA) documentsf the US Food and Drug Administration, which maynclude unpublished studies. No language restrictions werepplied.

Two reviewers independently inspected titles and abstractsdentified by the literature and potentially relevant articles,ccording to at least one reviewer, were obtained in full textormat. Both reviewers applied inclusion criteria for the finalecision regarding inclusion of trials (Fig. 1).

Two independent reviewers extracted data and performeduality assessment. In cases of disagreement, a third revieweras consulted. We assessed the quality of the following

rial characteristics: allocation concealment, generation ofhe allocation sequence generation, blinding and whethernalysis was performed by intention-to-treat analysis. Allo-ation concealment and generation were graded as adequate,nclear, or inadequate [13]. We recorded whether trials usedingle, double or no blinding and whether all randomised par-icipants were included in the results. We classified inclusionf all randomised participants as adequate if at least 90% andnadequate if less than 90%.

The primary outcome was of the occurrence of typhoidever, defined by a clinical disease with isolation of S. typhi.dverse events were a secondary outcome. Trial results were

xpressed as relative risks with 95% confidence intervalsnd pooled using a random effects model. We also inter-reted the results as efficacy, defined as 1-(relative risk) andxpressed as a percentage. Heterogeneity was assessed usinghe Chi-squared test with a p-value of <0.10 indicating sta-istical significance, and by the I2 statistic with a value of0% used to denote moderate levels of heterogeneity [14].

funnel plot was inspected for evidence of small studyffects.

Four cluster randomised efficacy trials that randomisedlassrooms and did not adjusted for clustering are includedn this review [15–20]. When results of two these trials werenalysed by comparing the proportion of classrooms withyphoid fever in the study arms, they were essentially the sameo those obtained if analysing individual level data [16,18].urther, clusters were small, and the outcomes rare, there-ore their results were included in the meta-analysis withoutdjusting for clustering.

For the primary analysis we evaluated the 3-year cumula-ive efficacy of three doses of the Ty21a vaccine, one dose ofhe Vi and two doses of the Vi-rEPA vaccine. We also strati-ed analyses by number of doses, and year of follow-up (post

7850 A. Fraser et al. / Vaccine 25 (2007) 7848–7857

of inclu

vayet

t(iic

3

Aindaceq

3

3

(E8gocfi0

iwoocf

Fig. 1. Flow

accination). When trials included follow-up for only part ofyear, we rounded to the nearest year to decide for which

ear’s estimate to use the data. If cumulative data on vaccinefficacy were available for longer than 3 years of follow-up,hey were also recorded.

In trials with several vaccine arms and a single con-rol arm, we divided the control arm into equal partsaccording to the number of vaccine arms) assuming equalncidence in these groups. This was done in order to avoidncluding data for controls more than once in the sameomparison.

. Results

Characteristics of included trials are presented in Table 1.ll efficacy trials and all but two safety trials were performed

n settings in which typhoid fever is endemic. We foundo trials evaluating vaccine efficacy among travellers fromeveloped countries. Most trials included children 5 years old

nd older. The single trial evaluating the Vi-rEPA includedhildren 2–5 years old. None of the trials compared differ-nt types of typhoid vaccines. Details on the methodologicaluality of included trials are presented in Table 2.

dcr

ded studies.

.1. Efficacy

.1.1. Ty21a vaccineCumulative 3-years efficacy of three doses of the Ty21a

all formulations) was 51% (95%CI 36%, 62%, Fig. 2).fficacy of the liquid formulation was 74% (95%CI 38%,9%); of enteric capsules 47% (95%CI 33%, 58%); and ofelatin capsules 25% (95%CI −10%, 49%). When resultsf two trials comparing the liquid preparation with entericapsules were pooled, the 3 years cumulative risk of typhoidever favoured the liquid formulation, however confidencentervals were wide and included par (RR = 0.54, 95%CI:.23–1.23).

Efficacy results stratified by dose and year are presentedn Table 3. Long-term (>3 years) cumulative efficacy dataas also available for three doses of enteric capsules andf the liquid formulation. Seven years cumulative efficacyf the enteric capsules was 62% (95%CI 48%, 73%, oneomparison) and 5-years cumulative efficacy of the liquidormulation was 79% (95%CI 65%, 87%, one comparison).

Three trials provided results per age group however veryifferent age strata were used [16,18,26]. One trial includedhildren 6–7 years old [19]. Three years cumulative efficacyesults stratified by age (5–9 and 10–44 years old) were sim-

A.F

raseretal./Vaccine

25(2007)

7848–78577851

Table 1Trial characteristics

Study ID Country Arm: numberof doses

Vaccine arm: vaccine formulation,interval between doses (days)

Ages Surveillance forefficacya

Length of follow-up(years)

Outcomes

Ty21aBlack et al., 1990 [15,21] Chile I: 2 I: Enteric capsules, 7 5–22 Intermediate 5 Incidence of typhoid fever

II: 1 II: Enteric capsules

Levine et al. 1986 A [22] Chile 3 Enteric capsules Adults Intermediate Not applicable Advese events

Levine et al. 1986 B [22] Chile I: 3 I: Enteric capsules, not reported Children Intermediate Not applicable Advese eventsII: 3 II: In milk with NaHCO3, not

reportedLevine et al., 1987 [16,17,23,24] Chile I: 3 I: Enteric capsules, 21 6–21 Intermediate 3 Incidence of typhoid fever

II: 3 II: Enteric capsules, 2III: 3 III: Gelatin capsules, 21IV: 3 IV: Gelatin capsules, 2

Levine et al., 1999 [17] Chile I: 3 I: Liquid, 2 5–19 Intermediate I: 3 Incidence of typhoid feverII: 3 II: Enteric capsules, 2 II: 5

Olanratmanee et al., 1992 [25] Thailand 3 Liquid, 2 4–6 Not applicable Not applicable Adverse events

Simanjuntak et al., 1991 [26] Indonesia I: 3 I: Liquid, 7 3–44 Intermediate 2.5 Incidence of typhoid fever, adverse eventsII: 3 II: Enteric capsules, 7

Wahdan et al., 1980 [19,20] Egypt 3 Liquid, 7 6–7 Intermediate 3 Incidence of typhoid fever, adverse eventsWahdan et al., 1980 [19,20] Egypt 3 Liquid, 7 6–7 Intermediate 3 Adverse events

ViKeitel et al., 1994 [27] USA 1 Not applicable 8–40 Not applicable Not applicable Advese eventsKlugman et al., 1987 [28,29] South Africa 1 Not applicable 5–15 Active 3 Incidence of typhoid feverAcharya et al., 1987 [30] Nepal 1 Not applicable 5–44 Active 1.4 Incidence of typhoid fever, adverse eventsWang et al., 1997 A [31,32] China 1 Not applicable 5–55 Passive 1 Incidence of typhoid feverWang et al., 1997 B [32] China 1 Not applicable 5–55 Passive 1 Adverse eventsYang et al., 2001 [31,33] China 1 Not applicable 3–50 Passive 1.6 Incidence of typhoid fever, adverse events

Vi-rEPALin et al., 2001 [34,35] Vietnam 2 Not applicable, 42 2–5 Active (27 months) 3.8 Incidence of typhoid fever, adverse events

Passive (additional19 months)

a Active: staff going into the field to identify cases; intermediate: relying on existing clinics and encouragement to evaluate patients for typhoid fever; passive: no increase of surveillance.

7852 A. Fraser et al. / Vaccine 25 (2007) 7848–7857

Table 2Methodological quality of included trials

Study ID, arms Allocationconcealment

Generation of theallocation sequence

Blinding Inclusion of allrandomised participants

Unit of randomisation

Acharya et al., 1987 [30] Adequate Adequate Double blind Adequate IndividualsBlack et al., 1990 [15] Adequate Unclear Double blind Adequate ClassroomsKeitel et al., 1994 [27] Unclear Unclear Double blind Adequate IndividualsKlugman et al., 1987 [29] Adequate Adequate Double blind Adequate IndividualsLevine et al., 1986 A [22] Unclear Unclear Double blind Unclear IndividualsLevine et al., 1986 B [22] Unclear Unclear Double blind Unclear IndividualsLevine et al., 1987 [16] Adequate Unclear Double blind Inadequate ClassroomsLevine et al., 1990 [18] Adequate Unclear Double blind Adequate ClassroomsLin et al., 2001 [34] Adequate Adequate Double blind Adequate IndividualsOlanratmanee et al., 1992 [25] Unclear Unclear Double blind Adequate IndividualsSimanjuntak et al., 1991 [26] Adequate Adequate Double blind Adequate IndividualsWahdan et al., 1980 A [20] Adequate Unclear Double blind Adequate ClassroomsWahdan et al., 1980 B [20] Unclear Unclear Double blind Adequate ClassroomsWWY

irtm

ang et al., 1997 A [32] Adequate Adequateang et al., 1997 B [32] Adequate Adequateang et al., 2001 [33] Adequate Adequate

lar (60%, 95%CI: 17%, 81% and 61%, 95%CI: 42%, 70%,espectively). We found no association between the risk ofyphoid fever in the control group and vaccine efficacy by

etaregression.

3

5c

Fig. 2. Three years cumulative risk o

Double blind Adequate IndividualsDouble blind Adequate IndividualsDouble blind Adequate Individuals

.1.2. Vi polysaccharide vaccineCumulative efficacy for 3 years based on a single study was

5% (95%CI 30, 70%, Fig. 2). The Vi polysaccharide vaccineonferred statistically significant protection in the first 2 years

f typhoid fever (RR, 95%CI).

A. Fraser et al. / Vaccine 25 (2007) 7848–7857 7853

Table 3Vaccine efficacy (95%CI) and number of arms, by number of doses and year

Vaccine Year 1 Year 2 Year 3 Year 4 Year 5

Ty21aOne dose 25 (−9, 49) 35 (−8, 61) 1 (−87, 48) −6 (−77, 37) −10 (−113, 43)No. of arms 1 1 1 1 1Two doses 52 (24, 69) 71 (44, 85) 22 (−54, 60) 19 (−41, 53) 7 (−84, 53)No. of arms 1 1 1 1 1Three doses 49 (16, 70) 60 (44, 71) 59 (32, 75) 78 (35, 93) 47 (−24, 78)No. arms 4 4 4 1 1

ViOne dose 68 (50, 80) 60 (31, 76) 50 (−11, 78)No. of arms 3 2 1

Vi-rEPATwo doses 94 (75, 99) 87 (56, 94)No. of arms 1 1

Fig. 3. Risk of fever (RR, 95%CI).

7854 A. Fraser et al. / Vaccine 25 (2007) 7848–7857

Table 4Relative risks (95%CI) of adverse events

Local swelling Local erythema Vomiting Diarrhea

Ty21a 0.72 (0.38, 1.37) 1.16 (0.50, 2.69) 0.72 (0.38, 1.37)No. of vaccine arms 5 7 7Vi polysaccharide 8.13 (1.34, 63.66) 2.30 (0.22, 24.23)No. of vaccine arms 2 2V No eve

or placeN 1

brw(0

3

2uvmb(

3

vo

3

cwcR(mtdfl

3

ocnc2

3

s

gif0

3

yIWbciff

ocv(i[aocacy of 47% (95%CI 33%, 58%). The calculated NNV forthe lowest incidence in a control group (289/100,000 [18])was 737 (95%CI 598, 1050) and for the highest incidence(2031/100,000 [26]) the NNV was 105 (95%CI 85, 150).

I-i-rEPA No events in either vaccineor placebo arms

o. of vaccine arms 1

ut not on the 3rd year (Table 3). Vaccine efficacy by age waseported in a single trial [32]. In this trial vaccine efficacyas 46% (95%CI: 0%, 90%) in 5–9 year old children, 85%

95%CI: 77%, 97%) in 10–14 year olds, and 60% (95%CI:%, 92%) in 15–19 year old adolescents.

.1.3. Vi-rEPA vaccineA single trial evaluated the Vi-rEPA vaccine in children

–5 years old. The cumulative efficacy for 3.8 years of follow-p (46 months) was 89% (95%CI: 76%, 97%) The Vi-rEPAaccine was efficacious in each of the first 2 years (up toonth 27 after vaccination, Table 3). Vi-rEPA continued to

e effective for an additional 1.6 years (19 months) 82.4%95%CI 22.3%, 99.1%).

.2. Adverse events

To evaluate safety we included trials comparing a typhoidaccine to placebo. We did not identify any trials comparingne type of typhoid vaccine to another.

.2.1. Ty21a vaccineResults for fever are presented in Fig. 3 and for other out-

omes in Table 4. When results for different preparationsere examined separately, the liquid formulation was asso-

iated with more nausea and abdominal pain than placebo,R = 3.18 (95%CI 1.15, 8.83) and vomiting, RR = 2.48

95%CI 1.56, 3.95). Enteric capsules were associated withore mild adverse events (i.e. any adverse events) in one trial

hat reported this outcome RR = 1.78 (95%CI 1.08, 2.95). Noifference in the incidence of any mild adverse events wasound in a single trial that compared this outcome for theiquid formulation versus enteric capsules [26].

.2.2. Vi polysaccharide vaccineNo difference was found in the incidence rate of fever

r erythema between the Vi polysaccharide vaccine groupompared with placebo. Swelling at the injection site was sig-ificantly more common among Vi vaccine recipients whenompared with placebo, however the incidence was low,.9%.

.2.3. Vi-rEPA vaccineIn the single study evaluating the Vi-rEPA vaccine no

welling or erythema occurred in either the vaccine or placeboFt

nts in either vaccinebo arms

roups. Fever was more frequent in the Vi-rEPA group but thencidence rate in the vaccine group was only 1.3%. No dif-erence was found for all-cause mortality (RR = 0.71, 95%CI.34, 1.48).

.3. Other analyses

As most comparisons included few trials, sensitivity anal-sis by trial methodological quality could not be performed.ncluding all controls in the analyses did not alter results.

e explored the possibility of small study effect and biasy examining a funnel plot of trials that assessed the 3-yearumulative efficacy of the Ty21a vaccine, as this comparisonncluded a relatively large number of studies. As can be seenrom Fig. 4 no strong evidence suggesting such an effect wasound.

We calculated the numbers needed to vaccinate (NNV) inrder to prevent one case of typhoid fever for the two commer-ially available vaccines. The liquid formulation of the Ty21aaccine has a 3-year cumulative protective efficacy of 74%95%CI 38, 89%). The lowest and highest incidence ratesn control groups were 138/100,000 [19] and 2021/100,00026]. The corresponding NNVs are: 977 (95%CI 812, 1902)nd 67 (95%CI 56, 131). The enteric capsules formulationf the Ty21a vaccine has a 3-year cumulative protective effi-

ig. 4. Funnel plot of trials assessing the 2.5–3 years cumulative efficacy ofhree doses of the Ty21a vaccine.

cine 25

Tcs1(

4

vTv(t7umiItveHwtvctwsvon

vttrpofr

tcbwp

e2iltB

k7

asafotfiewacwvWnc

4

cWftc

atccmh

aviicTttbtafhdi

A. Fraser et al. / Vac

he Vi vaccine has a 2.5–3-year cumulative protective effi-acy of 55% (95%CI 30, 70%) and the incidence rate in theingle study that assessed 2.5–3-year cumulative efficacy was160/100,000 [29]. Therefore, the NNV is estimated at 19295%CI 124, 288).

. Discussion

Both the Ty21a (three doses) and the Vi polysaccharideaccine were safe and efficacious in preventing typhoid fever.he cumulative efficacy at 3 years for the Ty21a and the Viaccine were similar: 51% (95%CI 36%, 62%), and 55%95%CI 30%, 70%), respectively. The cumulative efficacy ofhe Vi-rEPA vaccine at 3.8 years was higher, 89% (95%CI6%, 97%), but this vaccine has not yet been licensed forse and was evaluated in only one trial. Adverse events wereild in nature and for most, not significantly more frequent

n any of the vaccine groups when compared with placebo.n comparison, according to a previous meta-analysis [6],he 3-year cumulative efficacy of two doses of the whole-cellaccines was 73% (95%CI: 65%, 80%), i.e. higher than thefficacy of both the Ty21a and the polysaccharide Vi vaccines.owever, the frequency of adverse effects associated with thehole-cell vaccines was much higher than with the Ty21a and

he polysaccharide Vi vaccines. For example, 2% of Ty21aaccinnees and 1.1% of Vi polyssacharide the whole-cell vac-ines had a fever associated with the vaccination comparedo 15.7% of whole-cell vaccinnees. The whole-cell vaccinesere also associated with frequent swelling at the injection

ite (20%) compared to the Vi vaccine (3.7%), and 10% ofaccinnees missed school or work after vaccination. Becausef these frequent adverse effects, the whole-cell vaccines areo longer used.

The majority of trials evaluating the efficacy of the Ty21aaccine were cluster randomised trials that did not account forhe design in the analysis. However, we did not expect thiso have a meaningful impact on results. Two of the clusterandomised trials obtained similar results to those used in theooled analysis when analysing outcomes as the proportionf clusters (classrooms) with and without cases of typhoidever. Furthermore, clusters were small, and the outcomesare.

When the 3-year cumulative efficacy of three doses ofhe Ty21a liquid formulation and the enteric capsules wereompared, the point estimate favoured the liquid formulationut the result was not statistically significant. No differenceas found in the risk of adverse events between the tworeparations.

The Ty21a liquid formulation has an advantage over thenteric capsules as it is licensed for use in children fromyears of age while enteric capsules are licensed for use

n 5-year-old children [36]. This difference may be of pub-ic health importance due to findings that the incidence ofyphoid fever is elevated in children under the age of 5 [3,4].oth the liquid formulation and enteric capsules should be

4

f

(2007) 7848–7857 7855

ept refrigerated. The first is stable for 48 h and the latter fordays if not refrigerated [37].Efficacy per year and cumulative efficacy were analysed

s they provide different information. Analyses per year willhow whether the effect of the vaccine decreases over time,nd the cumulative efficacy demonstrates efficacy overall,or a given period, regardless whether changes over timeccurred within this period. Evidence available from fourrials demonstrates that the Vi vaccine is efficacious for therst 2 years. Only one trial provided data for year 3, the pointstimate for efficacy was 50%, however confidence intervalsere wide (95%CI −11%, 78%). The new Vi-rEPA vaccine

lso seems to be a promising candidate, however it is noturrently licensed for use. The 3.8 years cumulative efficacyas 89% (95%CI 76%, 97%) in children 2–5 years old. Thisaccine is now undergoing further evaluation through theorld Health Organisation Expanded Programme on Immu-

ization (EPI) in infants in Vietnam (Szu Shouson, personalommunication).

.1. Implications for practice

Both the currently licensed Ty21a and Vi vaccine, are effi-acious for preventing typhoid fever. As recommended by theorld Health Organization both vaccines can be considered

or use as disease control tools in high-risk areas, to reducehe burden in populations older that 5 years of age, schoolhildren in particular [11].

We provide updated estimates as to their efficacy, tollow local calculations of the cost to prevent one case ofyphoid fever. Additional considerations (e.g. consistent effi-acy results, acceptance and delivery, cost and availability,old chain requirements and convenience of administration),ay determine if and which vaccine is chosen for public

ealth use in typhoid endemic areas.Despite a continuing high burden of typhoid and the

vailability of safe and protective modern typhoid vaccines,accination is still not being used routinely to control typhoidn most endemic populations. There are many reasons for this,ncluding residual biases against vaccinating due to the unac-eptable reactogenicity of older generation typhoid vaccines.he cost of vaccines also undoubtedly contributed to hesi-

ations about introducing them into public health programs,hough the uptake of Vi polysaccharide vaccine productiony several manufacturers in developing countries has beguno yield vaccine at affordable prices. Even with an afford-ble vaccine, Ministries of Health in developing countries areaced with many competing demands upon already limitedealthcare budgets. Use of existing vaccines and continuedevelopment of improved vaccines provide the prospect formproved control of typhoid in affected countries.

.2. Implications for research

Neither the Vi vaccine nor the Ty21a vaccine is licensedor children under 2 years of age. Given the recent finding that

7 cine 25

tv

pScaiactw

A

Dt

d&mfD

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

856 A. Fraser et al. / Vac

yphoid fever also affects infants, development of a conjugateaccine suitable for this age group is warranted.

None of the trials included in the present review com-ared different types of vaccines to prevent typhoid fever.uch future comparisons may be helpful in drawing directonclusion regarding the relative efficacy of the vaccines,lthough such evidence would not necessarily promote thentroduction of vaccines against typhoid fever to new settingsnd would require substantial resources. Future trials shouldonduct analyses suited their design, i.e. cluster randomisa-ion has to be accounted for in sample size calculations asell as analyses of results.

cknowledgements

Funding: Work was supported by the Cochrane Infectiousiseases Group. The funding source had no involvement in

he work undertaken.Conflict of interest: C.J. Acosta is currently Director, Epi-

emiology North America, Clinical Research, DevelopmentMedical Affairs, Vaccines, N.A. GlaxoSmithKline. The

ajority of his typhoid related work was done while workingor The International Vaccine Institute, Seoul Korea as theO MI-Typhoid Project Coordinator.

eferences

[1] Crump JA, Luby SP, Mintz ED. The global burden of typhoid fever.Bull World Health Organ 2004;82:346–53.

[2] Bennish ML. Immunization against Salmonella typhi. Infect Dis ClinPrac 1995;4:114–22.

[3] Sinha A, Sazawal S, Kumar R, Sood S, Reddaiah VP, Singh B,et al. Typhoid fever in children aged less than 5 years. Lancet1999;354:734–7.

[4] Saha SK, Baqui AH, Hanif M, Darmstadt GL, Ruhulamin M, NagatakeT, et al. Typhoid fever in Bangladesh: implications for vaccinationpolicy. Pediatr Infect Dis J 2001;20:521–4.

[5] Bhutta ZA. Current concepts in the diagnosis and treatment of typhoidfever. BMJ 2006;333:78–82.

[6] Engels EA, Falagas MA, Lau J, Bennish ML. Typhoid fever vac-cines: a meta-analysis of efficacy and toxicity studies. Br Med J1998;316:110–6.

[7] Engels EA, Lau J. Vaccines for preventing typhoid fever. CochraneDatabase Syst Rev 1998.

[8] Garmory HS, Brown KA, Titball RW. Salmonella vaccines for usein humans: present and future perspectives. FEMS Microbiol Rev2002;26:339–53.

[9] Yang J, Acosta CJ, Si GA, Zeng J, Li CY, Liang DB, et al. A mass vacci-nation campaign targeting adults and children to prevent typhoid feverin Hechi; expanding the use of Vi polysaccharide vaccine in SoutheastChina: a cluster-randomised trial. BMC Public Health 2005;5:49.

10] Thiem VD, Meal D-H. The feasibility of a school-based Vi polysaccha-ride vaccine mass immunization campaign in Hue city, central Vietnam:streamlining a typhoid fever preventive strategy. Southeast Asian J Trop

Med Public Health 2006;37:515–22.

11] WHO. Background document: the diagnosis, treatment and preventionof typhoid fever; 2003. www.who.int/vaccines documents/.

12] Parry CM, Hien TT, Dougan G, White NJ, Farrar JJ. Typhoid fever.New Engl J Med 2002;347:1770–82.

[

(2007) 7848–7857

13] Alderson P, Green S, Higgins JPT, editors. Cochrane reviewers? Hand-book 4.2.2. The Cochrane Library, no. 1. Chichester, UK: John Wiley& Sons, Ltd.; 2004 [updated March 2004].

14] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsis-tency in meta-analyses. Br Med J 2003;327:557–60.

15] Black RE, Levine MM, Ferreccio C, Clements ML, Lanata C, Rooney J,et al. Efficacy of one or two doses of Ty21a Salmonella typhi vaccine inenteric coated capsules in a controlled field trial. Vaccine 1990;8:81–4.

16] Levine MM, Ferreccio C, Black RE, Germanier R. Large-scale fieldtrial of Ty21a live oral typhoid vaccine in enteric-coated capsule for-mulation. Lancet 1987;1:1049–52.

17] Levine MM, Ferreccio C, Abrego P, Martin OS, Ortiz E, Cryz S. Dura-tion of efficacy of Ty21a, attenuated Salmonella typhi live oral vaccine.Vaccine 1999;17:S22–7.

18] Levine MM, Ferreccio C, Cryz S, Ortiz E. Comparison of enteric-coated capsules and liquid formulation of Ty21a typhoid vaccine inrandomised controlled field trial. Lancet 1990;336:891–4.

19] Wahdan MH, Serie C, Cerisier Y, Sallam S, Germanier R. A controlledfield trial of live Salmonella typhi strain Ty 21a oral vaccine againsttyphoid: three year results. J Infect Dis 1982;145:292–5.

20] Wahdan MH, Serie C, Germanier R, Lackany A, Cerisier Y, Guerin N,et al. A controlled field trial of live oral typhoid vaccine Ty21a. BullWorld Health Organ 1980;58:469–74.

21] Black R, Levine MM, Young C, Rooney J, Levine S, Clements ML,et al. Immunogenicity of Ty21a attenuated “Salmonella typhi” givenwith sodium bicarbonate or in enteric-coated capsules. Dev Biol Stand1983;53:9–14.

22] Levine MM, Black RE, Ferreccio C, Clements ML, Lanata C, RooneyJ, et al. The efficacy of attenuated Salmonella typhi oral vaccine strainTY21A evaluated in controlled field trials. In: Holmgren J, LindbergA, Mollby R, editors. Development of vaccines and drugs againstdiarrhoea. 11th Noble Conference, Stockholm, 1985. Lund, Sweden:Studentlitteratur; 1986. p. 90–101.

23] Levine MM, Tacket CO, Herrington D, Losonsky G, Murphy J, Ferrec-cio C. The current status of typhoid vaccine development and clinicaltrials with typhoid vaccines. Southeast Asian J Trop Med Public Health1988;19:459–69.

24] Levine MM, Taylor DN, Ferreccio C. Typhoid vaccines come of age.Pediatr Infect Dis J 1989;8:374–81.

25] Olanratmanee T, Levine M, Losonsky G, Thisyakorn V, Cryz Jr SJ.Safety and immunogenicity of Salmonella typhi Ty21a liquid for-mulation vaccine in 4- to 6-year-old Thai children. J Infect Dis1992;166:451–2.

26] Simanjuntak CH, Paleologo FP, Punjabi NH, Darmowigoto R, Soepra-woto, Totosudirjo H, et al. Oral immunisation against typhoid fever inIndonesia with Ty21a vaccine. Lancet 1991;338:1055–9.

27] Keitel WA, Bond NL, Zahradnik JM, Cramton TA, Robbins JB. Clinicaland serological responses following primary and booster immunizationwith Salmonella typhi Vi capsular polysaccharide vaccines. Vaccine1994;12:195–9.

28] Klugman KP, Koornhof HJ, Robbins JB, Le-Cam NN. Immunogenic-ity, efficacy and serological correlate of protection of Salmonella typhiVi capsular polysaccharide vaccine three years after immunization.Vaccine 1996;14:435–8.

29] Klugman KP, Gilbertson IT, Koornhof HJ, Robbins JB, SchneersonR, Schulz D, et al. Protective activity of Vi capsular polysaccharidevaccine against typhoid fever. Lancet 1987;ii:1165–9.

30] Acharya IL, Lowe CU, Thapa R, Gurubacharya VL, Shrestha MB,Cadoz M, et al. Prevention of typhoid fever in Nepal with the Vi capsularpolysaccharide of Salmonella typhi: a preliminary report. N Engl J Med1987;317:1101–4.

31] Acosta CJ, Hong HY, Ning W, Qion G, Qun D, Xiaolei M, et al. Efficacy

of a locally produced, Chinese Vi polysaccharide typhoid fever vaccineduring six years of follow-up. Vaccine 2005.

32] Wang ZG, Zhou WZ, Shi J. Efficacy and side effects following immu-nization with Salmonella typhi Vi capsular polysaccharide vaccine.Zhonghua Liu Xing Bing Xue Za Zhi 1997;18:26–9.

cine 25

[

[

[

A. Fraser et al. / Vac

33] Yang HH, Wu CG, Xie GZ, Gu QW, Wang BR, Wang LY, et al.Efficacy trial of Vi polysaccharide vaccine against typhoid fever

in south-western China. Bull World Health Organ 2001;79:625–31.

34] Lin FY, Ho VA, Khiem HB, Trach DD, Bay PV, Thanh TC, et al. Theefficacy of a Salmonella typhi Vi conjugate vaccine in two-to-five-year-old children. N Engl J Med 2001;344:1263–9.

[

[

(2007) 7848–7857 7857

35] Mai NL, Phan VB, Vo AH, Tran CT, Lin FY, Bryla DA, et al. Persistentefficacy of Vi conjugate vaccine against typhoid fever in young children.

N Engl J Med 2003;349:1390–1.

36] WHO. Typhoid vaccines. WHO position paper. Wkly Epidemiol Rec2000;75:257–64.

37] Public-Health-Agency-of-Canada. Typhoid; 2005. http://www.phac.aspc.gc.ca/im/vpd.mev./typhoid. e.html.