Dtifo

PLOS NEGLECTED TROPICAL DISEASES
RESEARCH ARTICLE
Protection conferred by typhoid fever against

recurrent typhoid fever in urban Kolkata
Justin Im ID1☯*, Md. Taufiqul Islam2☯, Deok Ryun Kim1, Faisal Ahmmed2, Yun Chon1,
K. Zaman2, Ashraful Islam Khan2, Mohammad Ali3, Dipika Sur4, Suman Kanungo4,
Shanta Dutta4, Sujit K. Bhattacharya4, Gordon Dougan5, Kathryn E. Holt6,7,
Florian Marks1,5, Jerome H. Kim1, Firdausi Qadri2, John D. Clemens2,8
1 International Vaccine Institute, Seoul, Republic of Korea, 2 International Centre for Diarrheal Disease
Research, Bangladesh, Dhaka, Bangladesh, 3 Johns Hopkins Bloomberg School of Public Health, Baltimore,
Maryland, United States of America, 4 National Institute of Cholera and Enteric Diseases, Kolkata, West
Bengal, India, 5 Department of Medicine, University of Cambridge, Cambridge, United Kingdom,
6 Department of Infectious Diseases, Central Clinical School, Monash University, Melbourne, Victoria,
a1111111111
Australia, 7 London School of Hygiene and Tropical Medicine, London, United Kingdom, 8 Fielding School of
a1111111111 Public Health, University of California at Los Angeles, Los Angeles, California, United States of America
a1111111111
a1111111111 ☯ These authors contributed equally to this work.
a1111111111 * [email protected]
Abstract
OPEN ACCESS We evaluated the protection conferred by a first documented visit for clinical care of typhoid
Citation: Im J, Islam M.T, Kim DR, Ahmmed F, fever against recurrent typhoid fever prompting a visit. This study takes advantage of multi-
Chon Y, Zaman K, et al. (2020) Protection
year follow-up of a population with endemic typhoid participating in a cluster-randomized
conferred by typhoid fever against recurrent
typhoid fever in urban Kolkata. PLoS Negl Trop Dis control trial of Vi capsular polysaccharide typhoid vaccine in Kolkata, India. A population of
14(8): e0008530. https://doi.org/10.1371/journal. 70,566 individuals, of whom 37,673 were vaccinated with one dose of either Vi vaccine or a
pntd.0008530 control (Hepatitis A) vaccine, were observed for four years. Surveillance detected 315 first
Editor: David Joseph Diemert, George Washington typhoid visits, among whom 4 developed subsequent typhoid, 3 due to reinfection, defined
University School of Medicine and Health Sciences, using genomic criteria and corresponding to -124% (95% CI: -599, 28) protection by the ini-
UNITED STATES
tial illness. Point estimates of protection conferred by an initial illness were negative or negli-
Received: August 5, 2019 gible in both vaccinated and non-vaccinated subjects, though confidence intervals around
Accepted: June 27, 2020 the point estimates were wide. These data provide little support for a protective immunizing
Published: August 17, 2020 effect of clinically treated typhoid illness, though modest levels of protection cannot be
excluded.
Copyright: © 2020 Im et al. This is an open access
article distributed under the terms of the Creative
Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in
any medium, provided the original author and
source are credited.
Author summary
Data Availability Statement: All relevant data are Recurrent episodes of typhoid fever can occur in individuals, particularly those living in
within the manuscript and its Supporting typhoid endemic regions. Natural typhoid infection should, in theory, offer a degree of
Information files. The data underlying the results protection against future typhoid, however, this protection has only been assessed in a
presented in the study were accessed through a limited number of studies, none of which were conducted recently nor in the context of a
data sharing agreement between icddr,b and IVI.
carefully controlled clinical trial. The findings from our study, which were based on data
Data are available for researchers who meet the
criteria for access to confidential data. Contact
from a cluster-randomized trial of Vi polysaccharide vaccine in Kolkata, India, suggest
Armana Ahmed, Head, Research Administration, that previous typhoid fever does not provide an appreciable level of protection against
icddr,b (E-mail address: [email protected]). future typhoid disease. This finding is important for informing mathematical modelers
PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 1 / 11

PLOS NEGLECTED TROPICAL DISEASES Natural protection of typhoid disease
Funding: The research was supported by the

Diseases of the Most Impoverished Program of the who predict transmission dynamics within a population and influence the design of effec-
Bill & Melinda Gates Foundation. This publication tive vaccine intervention strategies.
was made possible through a grant from the Bill &
Melinda Gates Foundation (OPP1171432) awarded
to JC. GD was supported by Wellcome. The
funders had no role in study design, data collection
and analysis, decision to publish, or preparation of
the manuscript. Introduction
Competing interests: The authors have declared Typhoid fever is a systemic bacterial infection caused by Salmonella enterica serovar Typhi (S.
that no competing interests exist. Typhi) and is responsible for significant levels of morbidity and mortality in low- and middle-
income countries. Global burden of disease studies have estimated that 14.3 million cases of
typhoid fever occur annually [1], and case-fatality rates have dropped to <1% since the era of
antibiotics ushered in a series of highly effective treatment options [1, 2]. Better understanding
of the genetic elements conferring resistance against antibiotics in emergent S. Typhi lineages
has been made possible by whole-genome sequencing, and the potential for resistant clades to
spread locally and globally has been well-documented [3–5]. Consequently, antimicrobial ther-
apy as a means to limit severe disease and death may become critically compromised.
Although the typhoid endgame will ultimately be shaped by widespread and lasting access to
clean water and improved standards of sanitation, strategic use of existing vaccines, including
typhoid conjugate vaccine, must be made a priority as a shorter-term solution [6].
The parenteral Vi capsular polysaccharide vaccine and the oral live-attenuated Ty21a strain
are safe and effective but will soon be supplanted by a promising, parenteral Vi-tetanus toxoid
conjugate vaccine (Vi-TT), which the WHO prequalified in 2018 and recommended for use in
children 6 months to adults 45 years of age in endemic regions [6]. This conjugate vaccine has
several potential advantages over earlier vaccines, including extending protection to younger
children and inducing long-lasting protection in all age groups [7], and documented protective
efficacy of 82% in children between 9 months and 16 years of age after one year of follow-up
[8].
The Vi-TT vaccine has the potential to dramatically reduce the burden of typhoid fever in
vulnerable populations. However, the optimal strategy for vaccine introduction will be greatly
assisted by predictions of dynamic transmission models of typhoid incidence that capture the
interplay between natural and vaccine-induced protection. Such predictions will allow policy-
makers to make rational decisions on such issues as age at vaccination, the need for catch up
vaccination, and the frequency of boosting. To date, predictions by these models have been
limited by uncertainty about several assumptions, including the protective effect of natural
typhoid infection against subsequent typhoid infection in typhoid-endemic settings [9, 10].
Herein, we present an evaluation of the protection conferred by an initial clinically treated
typhoid fever illness against subsequent typhoid fever due to reinfection, taking advantage of
the multi-year follow-up of a population with endemic typhoid participating in a cluster-ran-
domized control trial of Vi capsular polysaccharide typhoid vaccine.
Methods
Ethics statement
All subjects or their guardians provided written informed consent. The protocol was approved
by the Institutional Review Boards of the International Vaccine Institute (Republic of Korea),
the National Institute of Cholera and Enteric Diseases (India), and the Indian Council of Med-
ical Research.

Study site
We conducted the evaluation in urban slums of Kolkata, India, where a cluster-randomized
trial of Vi polysaccharide vaccine against typhoid was conducted. Two years before the onset
of the trial, in January 2003, a population of approximately 60,000 persons was censused,
recording individual and household level demographic and socioeconomic data. This informa-
tion was subsequently updated at regular intervals in a demographic surveillance system that
continued through the vaccine trial, culminating in a close-out census at the end of surveil-
lance, two years after vaccination [11]. Typhoid fever surveillance commenced in the study
population from the onset of the two-year lead-in period before vaccination and continued for
two years following vaccination.
Vi typhoid vaccine trial

The cluster-randomized trial of Vi capsular polysaccharide typhoid vaccine was initiated in
2004 (ClinicalTrials.gov reference number: NCT00125008). A total of 37,673 individuals,
excluding pregnant and lactating women, aged �2 years were vaccinated with one dose of
either: 1) Vi vaccine (Typherix, GlaxoSmithKline) containing 25μg of Vi polysaccharide; or 2)
inactivated hepatitis A (HepA) vaccine (Havrix, GlaxoSmithKline). In the surveillance, partici-
pants from the study area who presented at one of five study clinics were examined by a study
physician, and for each fever visit, defined as a patient who presented with a history of fever
lasting at least 3 days, a blood culture was performed. We defined a typhoid visit as a fever visit
in which S. Typhi was isolated from a blood culture, with the onset of the illness as the date of
fever onset.
Defining typhoid relapse and reinfections

We defined typhoid relapse as a second typhoid visit caused by the infecting strain from the
first visit and can be understood as a continuation of the first infection. A typhoid reinfection
was defined as an independent typhoid visit resulting from a new infection from an external
source. Because there is no gold standard for distinguishing relapses from reinfections in all
instances of recurrent typhoid visits, we defined typhoid reinfection in several alternative
ways.
Our primary approach was based on genomic criteria. S. Typhi isolates cultured from the
first and all recurrent visits were subjected to whole genome sequencing and single nucleotide
polymorphism (SNP) analysis as previously described [12]. Primary analysis considered a
recurrent typhoid visit to be a reinfection if there were >5 pairwise SNP differences between
isolates from the two visits, a definition that has previously been used to distinguish relapse
from reinfection for invasive non-typhoidal salmonellosis [13]. Secondary analyses considered
two alternative criteria for defining recurrent typhoid as reinfections. First, we considered the
susceptibility pattern across 11 tested antibiotics (chloramphenicol, ampicillin, amoxicillin, tri-
methoprim-sulfamethoxazole, ciprofloxacin, ceftriaxone, nalidixic acid, tetracycline, ofloxacin,
aztreonam, amikacin), where the detection of at least one discordant susceptibility result
between isolated pairs indicated infection with a new strain. Second, we considered a recurrent
typhoid visit to be a reinfection if a period of >30 days intervened between the date of the first
typhoid visit and the date of onset of the next typhoid visit. Each definition was considered
irrespective of classification in alternative definitions. It should be noted that each of these def-
initions gave a conservative measure of the risk of reinfection, since second typhoid attacks
not meeting the most stringent definition above, given by genomic criterion, could be caused
by reinfection of genetically identical S. Typhi strains, given that S. Typhi evolution is very
slow, less than 1 SNP per year [4, 14].

Analysis
Our primary analysis, which considered the entire dynamic cohort followed from January 1,
2003 until December 31, 2006, estimated the hazard ratio (HR) for the incidence rate of recur-
rent typhoid due to reinfection, defined by genetic criteria (vide supra), in persons in whom an
earlier typhoid visit was observed (Cohort 2), relative to the incidence of a first typhoid visit in
persons without an earlier detected visit (Cohort 1). All other analyses were considered sec-
ondary. Ratios <1 reflected a protective relationship.
For Cohort 1, we measured the incidence of first typhoid visits for the study population
residing in the study area at the study start, January 1, 2003, as well as for persons who moved
into the area via in-migration or birth thereafter. Person-days of follow-up for each member of
Cohort 1 were calculated from zero time (January 1, 2003, or date of entry into the cohort via
birth or in-migration, if after January 1, 2003) until the date of onset of the first typhoid visit
date, or one of the right censoring events of out-migration, death or the end of follow-up
(December 31, 2006), whichever came first. For Cohort 2, we measured the incidence of recur-
rent typhoid due to reinfection among those in whom a first visit was recorded, with zero time
for follow-up beginning on the visit date for the first visit, and continuing until the date of
onset of a reinfection visit, or right censoring event, whichever came first (Fig 1A).
To estimate the HRs relating the rate of recurrent typhoid due to reinfection to the rate of a
first typhoid visits in subgroups of Vi vaccinated versus non-vaccinated individuals, we
employed a different strategy for defining Cohort 1 and Cohort 2, since vaccination took place
Fig 1. CONSORT diagram showing culture-positive S. Typhi first and second visits in A) the total population during the study
period of January 1, 2003 until December 31, 2006, and B) sub-populations of Vi-vaccinated and non-dosed individuals from zero
time to December 31, 2006. Note: In- and out-migratory populations are not reflected in this diagram. � Given as date of vaccination
occurring between November 27, 2004 and December 31, 2004. § Based on pairwise SNP difference >5.
https://doi.org/10.1371/journal.pntd.0008530.g001

between November 27, 2004 and December 31, 2004. Among Vi vaccinees, for Cohort 1, zero
time was taken as the date of dosing, with follow-up continuing until the date of onset of the
first typhoid visit or the first right censoring event, whichever came first. Cohort 2 measured
the incidence of recurrent typhoid due to reinfection among Vi vaccinees in whom a prior
typhoid visit was detected, with zero time taken as the date of the first visit and follow-up con-
tinuing until the date of onset of a typhoid reinfection visit, or a right censoring event, which-
ever came first. For the non-Vi vaccinated group, followed during a secular interval
comparable to Vi vaccinees, Cohort 1 zero time was the date of vaccination with HepA vac-
cine, or, among non-dosed individuals, the median date of dosing with either vaccine for other
members of the same cluster in the trial if residing in the cluster at the time of vaccination, oth-
erwise birth date or date of migration into the cluster. Zero time for Cohort 2 was the date of
the first typhoid visit. Follow-up continued until the date of onset of the first visit (for Cohort
1), the date of onset of the recurrent visit due to reinfection (for Cohort 2), or a right censoring
event, whichever came first (Fig 1B).
We fitted generalized Cox models (Andersen-Gill [AG] model), which allow for recurrent
events along the timeline for an individual, to estimate HRs, crude and adjusted for potential
covariates, and 95% confidence intervals [15]. The AG model is able to evaluate subsequent
events which are mediated through time-varying covariates and gives an output indicating the
occurrence intensity of a recurrent event. For the primary analysis, the AG HR was adjusted
for Vi vaccination status (vaccinated or non-vaccinated) as a time varying covariate and age at
zero time, which were independently associated with time to event at p-value <0.10, and pro-
tection was expressed as [(1 –adjusted HR) X 100%]. For the subgroup analyses, we fitted mod-
els for Vi vaccinees and non-vaccinees separately. We considered p-value <0.05 (two-tailed)
as the margin of statistical significance and all modeling was done using the PHREG procedure
in SAS/STAT software (SAS Institute Inc., Cary, NC, USA).
Results
There were 70,566 individuals residing in the study area at any point in time from the study
start date to study end (Fig 1A). A total of 315 first typhoid visits were detected in Cohort 1
during the study period, four of which were followed by a second typhoid visit (Fig 1A). Of
315 first typhoid visits, 313 were discharged as outpatients and two were admitted for a dura-
tion of 20 and 28 days, respectively. All second typhoid visits were discharged as outpatients.
Second typhoid visits occurred in patients ranging from 10 to 33 years of age. Three of four
isolate pairs belonged to the H58 haplotype. By SNP criterion, three second typhoid visits were
considered to be reinfections with a different strain (Table 1). One second visit fit the reinfec-
tion criterion based on antibiotic susceptibility and another based on intervening period defi-
nition (Table 1). None of the second visits were classified as reinfection by all defining criteria
and one infection pair was not classified as reinfection according to any criteria.
The crude incidence of first visits during the entire surveillance period was 0.36 per 100,000
person-days and the crude incidence of all second visits was 0.95 per 100,000 person-days
(Table 2). In adjusted analyses, individuals detected with a first typhoid visit were two times
more likely to have a subsequent typhoid visit compared to individuals without a prior typhoid
visit detected during the study period (Adjusted Protection: -200% (-708, -12)) (Table 2).
Using our primary genomic definition of reinfection gave an adjusted protection effect con-
ferred by a first typhoid visit of -124% (-599, 28) (Table 2). Significant protection conferred by
first typhoid visit against reinfection, after adjustment for potential confounders, was not
detected regardless of which of the three criteria was used to define reinfection, although the
antibiotic susceptibility and intervening period criteria gave an estimate that was positive, but

Table 1. Classification of recurrent typhoid pairings as due to reinfection based on genomic, antibiotic susceptibility, and intervening period criteria.
ID NCBI/ENA accession Haplotype SNPs§ (No. pairwise SNP Antibiotic susceptibilityk (No. discordant Intervening period¶ (No.
differences) pairs) days)
C03551 ERR279346 H58 Reinfection (11) Relapse (0) Reinfection (49)
C03891 ERR279347 H58
C03495 ERR279348 H58 Reinfection (10) Reinfection (1) Relapse (19)
C03634 ERR279349 H58
E02889 ERR279350 H58 Reinfection (9) Relapse (0) Relapse (30)
E02990 ERR279351 H58
E01240 ERR279352 H14 Relapse (5) Relapse (0) Relapse (14)
E01303 ERR279353 H14
§ Pairwise SNP difference >5

k Discordance in antibiotic susceptibility between infection pair >0
¶ Intervening period >30 days
https://doi.org/10.1371/journal.pntd.0008530.t001
low, with wide confidence limits (Adjusted Protection: 26% (-427, 90) and 26% (-429, 90),
respectively) (Table 2).
Among the four second typhoid visits, two occurred in the Vi vaccine arm and one was in
the non-Vi vaccine arm (includes HepA vaccinated and non-dosed) (Fig 1B). One second
typhoid visit (E01303) occurred before the period of vaccination and was therefore excluded
from vaccination subgroup analyses. Analyses of these subgroups estimated that the crude
incidence of first typhoid visit was 0.25 and 0.45 per 100,000 person-days in the Vi-vaccinated
group (Table 3) and non-Vi vaccinated group (Table 4), respectively. We identified a very high
Table 2. Protective effect of a first typhoid visit against recurrent typhoid due to reinfection in the total population (n = 70,566).
No. of individuals No. of typhoid visits Person-days † Incidence per 100,000 person-days Crude HR ‡ Adjusted Protective Effect (95% CI) ‡
Cohort 70566 315 86369890 0.36 Ref
1 Ref
All recurrent typhoid visits
Cohort 315 4 420965 0.95 5.37 (2.00, -200% (-708, -12)
2 14.43)
Recurrent typhoid due to reinfection defined by SNP difference§
Cohort 315 3 421740 0.71 4.00 (1.28, -124% (-599, 28)
2 12.50)
Recurrent typhoid due to reinfection defined by antibiotic susceptibilityk
Cohort 315 1 422584 0.24 1.33 (0.19, 9.46) 26% (-429, 90)
2
Recurrent typhoid due to reinfection defined by intervening time¶
Cohort 315 1 422653 0.24 1.33 (0.19, 9.46) 26% (-427, 90)
2
†In cohort 1, person-days calculated from study start date or date when person entered the study to end date, which is defined as date of study end, out-migration, or
onset of the first typhoid visit. In cohort 2, person-days are calculated from the date of first typhoid visit to the date of study end, out-migration, death, or onset of the
subsequent typhoid visit.
‡ Fitted a generalized proportional hazard model (AG model). The number of previous infections (0 or 1) was included in the model as a dummy variable; the reference
group was the group without a previously detected typhoid visit. The risk was adjusted for Vi vaccination status as a time varying covariate and age at zero time.

Table 3. Protective effect of a first typhoid visit against recurrent typhoid due to reinfection in the Vi-vaccinated population (n = 18,869).
No. of individuals No. of typhoid visits Person-days † Incidence per 100,000 person- Crude HR ‡ Adjusted Protective Effect (95% CI)
days ‡
Cohort 18869 34 13705796 0.25 Ref Ref
1
Cohort 34 2 24624 8.12 74.08 (17.52, -3549% (-15473, -755)
2 313.21)
Recurrent typhoid due to reinfection defined by SNP difference§
Cohort 34 2 24624 8.12 74.08 (17.52, -3549% (-15473, -755)
2 313.21)
Cohort 34 0 25468 0 -- --
2
Recurrent typhoid due to reinfection defined by intervening time¶
Cohort 34 1 24909 4.01 35.56 (4.81, 262.83) -1640% (-12835, -134)
2
†In cohort 1, person-days calculated from zero time (date of vaccination) to end date, which is defined as date of study end, out-migration, or onset of the first typhoid
visit. In cohort 2, person-days are calculated from the date of first typhoid visit to the date of study end, out-migration, death, or onset of the subsequent typhoid visit.
group was the group with no previously detected typhoid visit. The hazard ratio was adjusted for age at date of entry.
Table 4. Protective effect of a first typhoid visit against recurrent typhoid due to reinfection in the non-Vi vaccinated population including in-migrants post-vacci-
nation (n = 47,995).
No. of individuals No. of typhoid visits Person-days † Incidence per 100,000 person- Crude HR ‡ Adjusted Protective Efficacy (95% CI)
days ‡
Cohort 47995 149 32868952 0.45 Ref Ref
1
Cohort 149 1 108912 0.92 4.21 (0.59, -88% (-1246, 74)
2 30.13)
Recurrent typhoid due to reinfection defined by strain§
Cohort 149 1 108912 0.92 4.21 (0.59, -88% (-1246, 74)
2 30.13)
Cohort 149 1 108912 0.92 4.21 (0.59, -88% (-1246, 74)
2 30.13)
Recurrent typhoid due to reinfection defined by interval ¶
Cohort 147 0 109540 0 -- --
2
†In cohort 1, person-days calculated from zero time (date of vaccination with placebo or for non-dosed individuals, the median zero time for the cluster of residence at
the time of vaccination, or date of entry for in-migrants post-vaccination) to end date, which is defined as date of study end, out-migration, or onset of the first typhoid
visit. In cohort 2, person-days are calculated from the date of first typhoid visit to the date of study end, out-migration, death, or onset of the subsequent typhoid visit.
group was the group with no previously detected typhoid visit. The hazard ratio was adjusted for age at date of entry.

HR of reinfection (by genomic criterion) in the Vi-vaccinated group with evidence that those
with prior typhoid visits were at increased risk for typhoid relative to those without a previ-
ously detected typhoid (Adjusted Protection: -3549% (-15473, -755)) (Table 3). In the non-Vi
vaccinated group all point estimates for reinfection reflected no protection by an initial
typhoid visit (Adjusted Protection: -88% (-1246, 74)), although the wide confidence interval
did not exclude a moderate level of protection (Table 4).
Discussion
In this study, we assessed the level of protection associated with typhoid illnesses in patients
seeking clinical care against a subsequent visit due to reinfection. Point estimates failed to find
a significant level of protection by an initial typhoid illness against typhoid illnesses defined as
reinfections using multiple alternative definitions of reinfection. Moreover, our analyses sug-
gested the relationship was different for those who had earlier received Vi polysaccharide vac-
cine and those who had not, with the former group exhibiting negative protective
relationships. The 95% confidence interval upper boundaries were below 0% not only suggest-
ing that there was an absence of protection but that the risk of subsequent typhoid reinfection
was actually higher in those with an initial infection. In contrast, non-recipients of Vi polysac-
charide, who also had negative values for point estimates of adjusted protection, displayed very
wide confidence intervals whose upper boundaries did not exclude moderate protection by an
initial visit.
In aggregate, these findings are compatible with previous reports that a natural typhoid
infection confers, at best, moderate or incomplete protective effects against subsequent attack
[16, 17]. We also observed that the hazard ratios for reinfection were higher in Vi vaccinated
compared to non-vaccinated individuals. These data suggest that persons who develop typhoid
fever despite earlier Vi vaccination represent a subgroup of non-responders who are especially
vulnerable to typhoid, for immune or non-immune reasons. The earlier studies evaluated risk
of typhoid in adults from non-endemic settings who were exposed to typhoid during sequen-
tial, concentrated outbreaks occurring in a military unit [16] or in the setting of an experimen-
tal typhoid volunteer challenge study [17]. These studies did not address the issue of greatest
relevance to deployment of new generation vaccines to populations experiencing endemic
typhoid: what level of natural protection is conferred by typhoid infection against recurrent
infection in a typhoid-endemic population of children and adults followed over a period of
several years. This issue which is evaluated in our study, is a key gap in evidence for construct-
ing dynamic transmission models designed to predict the population impact of alternative
introduction strategies for new generation typhoid vaccines into populations with endemic
typhoid, who account for the vast majority of the world’s typhoid morbidity and mortality.
It is important to discuss the study’s limitations. First, we lacked a gold standard criterion
for differentiating typhoid due to reinfection from typhoid due to relapse. However, the con-
sistency of our findings despite use of multiple alternative definitions for reinfection supports
the credibility of our findings. In this evaluation, we considered that findings based on geno-
mic criteria, which were more discriminatory, were likely more credible. Secondly, we did not
randomly allocate and concurrently follow the two groups under comparison, namely those
with previously detected typhoid and those who had not yet been shown to have typhoid.
Therefore, those who developed typhoid fever may have been at systematically higher risk of
typhoid or were more likely to seek medical care for fever. This may be especially true of those
who were vaccine failures, and may not have been completely controlled for in our multivari-
able models. Counterbalancing this potential bias, however, is the fact that all of our criteria
for defining typhoid reinfection were likely underestimates, making our estimates of increased

risk conservative, since identical organisms may have caused secondary attacks, given that the
low rate of nucleotide substitution in S. Typhi would have corresponded to an accumulation of
very few SNPs throughout the study [4, 14]. As well, when we controlled for typhoid seasonal-
ity to adjust for unequal exposure to typhoid, we found that our results were unaffected. Addi-
tionally, this approach of comparing disease incidence in naturally infected versus non-
infected individuals in cohort studies has been successfully used to determine protection natu-
rally conferred by infection in other infectious diseases such as cholera [18] and rotavirus [19].
Thirdly, the number of typhoid reinfections, regardless of the defining criteria, were low in
number, and the resulting estimates of protection had very wide confidence intervals. However,
confidence intervals for most of our estimates exclude even moderate levels of protection.
Fourthly, some typhoid visits that were classified as first visits in our surveillance may in fact
have been reinfections, as an artifact of the limited time period for our surveillance, though exclu-
sion of these cases would have lowered the incidence of first visits, again making our analyses
conservative. Fifthly, our findings reflect the protection conferred by clinical typhoid, excluding
asymptomatic or mildly symptomatic typhoid infections, though it seems counterintuitive that
these milder typhoid infections could confer greater protection than clinical typhoid severe
enough to prompt patients to seek care. Sixthly, all confirmed typhoid cases were treated with
antibiotics according to WHO guidelines, considering antibiotic-susceptibility results. Treatment
may have reduced the course of infection thereby impacting immunological development and
diminishing the measurable protective effect of natural infection. Lastly, these findings reflect the
protection conferred by the particular S. Typhi strains circulating in the population under obser-
vation during the period of study. Importantly, our analysis does not make provisions for unde-
tected typhoid due to limitations of diagnostic sensitivity of blood cultures, although presumably
sensitivity is not impacted by the serial order of visits occurring in an individual.
The biological plausibility of our findings is supported by several observations. Supportive
evidence includes the intrinsic ability of S. Typhi to down-regulate its own virulence genes to
facilitate evasion of human immune signaling pathways [20] and to disrupt host transcription
regulation effectively disabling components of the host immune response [21]. Additionally,
host characteristics, such as genetic determinants [22, 23], prior pathogen exposure [24], and
baseline immune cell presence [25] can lead to a muting of immune responses against S. Typhi
invasion. In contrast, a murine model using S. Typhimurium has shown that immunity devel-
oped after infection does indeed offer protection from a second attack, though this relationship
may be specific to the murine host or to the non-typhoid Salmonella pathogen under study
[26]. In a human challenge model, elicited immunity provided, at best, only moderate protec-
tion against second attack [17].
In view of the dearth of other contemporary studies of natural protection conferred by
typhoid disease in human populations with endemic typhoid disease, our analyses should be
considered when parameterizing dynamic transmission models of typhoid fever for the pur-
pose of designing programs for and predicting the impact of alternative introduction strategies
for Vi conjugate vaccines in endemic settings. Protection conferred by natural typhoid disease
against recurrent disease due to typhoid reinfection was not detected in our cohort analyses.
However, in view of the limitations of our study (vide supra) additional studies of the protec-
tive impact of typhoid fever are needed.
Author Contributions
Conceptualization: John D. Clemens.
Data curation: Deok Ryun Kim, Faisal Ahmmed, Mohammad Ali.

Formal analysis: Deok Ryun Kim, Faisal Ahmmed, John D. Clemens.

Funding acquisition: John D. Clemens.
Investigation: Justin Im, Md. Taufiqul Islam, Deok Ryun Kim, Yun Chon, K. Zaman, Ashraful
Islam Khan, Mohammad Ali, Dipika Sur, Suman Kanungo, Shanta Dutta, Sujit K. Bhatta-
charya, Gordon Dougan, Kathryn E. Holt, Florian Marks, Jerome H. Kim, Firdausi Qadri,
John D. Clemens.
Methodology: Justin Im, Deok Ryun Kim, Faisal Ahmmed, Mohammad Ali, Gordon Dougan,
John D. Clemens.
Supervision: K. Zaman, Ashraful Islam Khan, Mohammad Ali, Gordon Dougan, Kathryn E.
Holt, Florian Marks, Jerome H. Kim, Firdausi Qadri, John D. Clemens.
Writing – original draft: Justin Im, Md. Taufiqul Islam, John D. Clemens.
Writing – review & editing: Justin Im, Md. Taufiqul Islam, Deok Ryun Kim, Faisal Ahmmed,
Yun Chon, K. Zaman, Ashraful Islam Khan, Mohammad Ali, Dipika Sur, Suman Kanungo,
Shanta Dutta, Sujit K. Bhattacharya, Gordon Dougan, Kathryn E. Holt, Florian Marks,
Jerome H. Kim, Firdausi Qadri, John D. Clemens.
References
1. Collaborators GTaP. The global burden of typhoid and paratyphoid fevers: a systematic analysis for the
Global Burden of Disease Study 2017. Lancet Infectious diseases. 2019; 19:369–81 https://doi.org/10.
1016/S1473-3099(18)30685-6 PMID: 30792131
2. Bhutta ZA. Current concepts in the diagnosis and treatment of typhoid fever. BMJ. 2006; 333:78–82.
https://doi.org/10.1136/bmj.333.7558.78 PMID: 16825230
3. Park S, Pham DT, Boinett C, Wong VK, Pak GD, Panzner U, et al. The phylogeography and incidence
of multi-drug resistant typhoid fever in sub-Saharan Africa. Nat Commun 2018; 5094.
4. Wong VK, Baker S, Pickard DJ, Parkhill J, Page AJ, Feasey NA, et al. Phylogeographical analysis of
the dominant multidrug-resistant H58 clade of Salmonella Typhi identifies inter- and intracontinental
transmission events. Nature genetics. 2015; 47:632–9. https://doi.org/10.1038/ng.3281 PMID:
25961941
5. Klemm EJ, Shakoor S, Page AJ, Qamar FN, Judge K, Saeed DK, et al. Emergence of an Extensively
Drug-Resistant Salmonella enterica Serovar Typhi Clone Harboring a Promiscuous Plasmid Encoding
Resistance to Fluoroquinolones and Third-Generation Cephalosporins. mBio 2018; 9.
6. World Health Organization. Typhoid vaccines: WHO position paper. Wkly Epidemiol Rec 2018; 13:
153–172.
7. Mohan VK, Varanasi V, Singh A, Pasetti MF, Levine MM, Venkatesan R, Ella KM. Safety and immuno-
genicity of a Vi polysaccharide-tetanus toxoid conjugate vaccine (Typbar-TCV) in healthy infants, chil-
dren, and adults in typhoid endemic areas: a multicenter, 2-cohort, open-label, double-blind,
randomized controlled phase 3 study. Clin Infet Dis 2015; 61: 393–402.
8. Shakya M, Colin-Jones R, Theiss-Nyland K, Voysey M, Pant D, Smith N, Liu X, et al. Phase 3 Efficacy
Analysis of a Typhoid Conjugate Vaccine Trial in Nepal. N Engl J Med 2019; 381 2209–2218. https://
doi.org/10.1056/NEJMoa1905047 PMID: 31800986
9. Pitzer VE, Bowles CC, Baker S, Kang G, Balaji V, Farrar JJ, et al. Predicting the impact of vaccination
on the transmission dynamics of typhoid in South Asia: a mathematical modeling study. PLOS NTD
2014; 8:e2642.
10. Saul A, Smith T, Maire N. Stochastic simulation of endemic Salmonella enterica serovar Typhi: the
importance of long lasting immunity and the carrier state. PLOS ONE 2013; 8:e74097. https://doi.org/
10.1371/journal.pone.0074097 PMID: 24040177
11. Sur D, Ochiai RL, Bhattacharya SK, Ganguly NK, Ali M, Manna B, et al. A cluster-randomized effective-
ness trial of Vi typhoid vaccine in India. N. Engl. J. Med. 2009; 361:335–44. https://doi.org/10.1056/
NEJMoa0807521 PMID: 19625715
12. Holt KE, Dutta S, Manna B, Bhattacharya SK, Bhaduri B, Pickard DJ, et al. High-resolution genotyping
of the endemic Salmonella Typhi population during a Vi (typhoid) vaccination trial in Kolkata. PLOS
NTD 2012; 6:e1490.

13. Okoro CK, Kingsley RA, Quail MA, Kankwatira AM, Feasey NA, Parkhill J, et al. High-resolution single
nucleotide polymorphism analysis distinguishes recrudescence and reinfection in recurrent invasive
nontyphoidal Salmonella typhimurium disease. Clin. Infect. Dis. 2012; 54:955–63. https://doi.org/10.
1093/cid/cir1032 PMID: 22318974
14. Britto CD, Dyson ZA, Duchene S, Carter MJ, Gurung M, et al. Laboratory and molecular surveillance of
paediatric typhoidal Salmonella in Nepal: Antimicrobial resistance and implications for vaccine policy.
PLOS Negl. Trop. Dis 12(4): e0006408. https://doi.org/10.1371/journal.pntd.0006408 PMID: 29684021
15. Amorim LD, Cai J. Modelling recurrent events: a tutorial for analysis in epidemiology. Int. J. Epidemiol.
2015; 44:324–33. https://doi.org/10.1093/ije/dyu222 PMID: 25501468
16. Marmion DE, Naylor GR, Stewart IO. Second attacks of typhoid fever. Epidemiol. Infect. 1953; 51:260–
7.
17. Dupont HL, Hornick RB, Snyder MJ, Dawkins AT, Heiner GG, Woodward TE. Studies of immunity in
typhoid fever. Protection induced by killed oral antigens or by primary infection. Bull. World Health
Organ. 1971; 44:667–72. PMID: 5316750
18. Ali M, Emch M, Park JK, Yunus M, Clemens J. Natural cholera infection-derived immunity in an endemic
setting. J Infect. Dis. 2011; 204:912–8. https://doi.org/10.1093/infdis/jir416 PMID: 21849288
19. Gladstone BP, Ramani S, Mukhopadhya I, Muliyil J, Sarkar R, Rehman AM, et al. Protective effect of
natural rotavirus infection in an Indian birth cohort. N. Engl. J. Med. 2011; 365:337–46. https://doi.org/
10.1056/NEJMoa1006261 PMID: 21793745
20. Winter SE, Winter MG, Poon V, Keestra AM, Sterzenbach T, Faber F, et al. Salmonella enterica Sero-
var Typhi conceals the invasion-associated type three secretion system from the innate immune system
by gene regulation. PLOS Pathog. 2014; 10:e1004207. https://doi.org/10.1371/journal.ppat.1004207
PMID: 24992093
21. Nickerson KP, Senger S, Zhang Y, Lima R, Patel S, Ingano L, et al. Salmonella Typhi Colonization Pro-
vokes Extensive Transcriptional Changes Aimed at Evading Host Mucosal Immune Defense During
Early Infection of Human Intestinal Tissue. EBioMedicine 2018; 31:92–109. https://doi.org/10.1016/j.
ebiom.2018.04.005 PMID: 29735417
22. Dunstan SJ, Stephens HA, Blackwell JM, Duc CM, Lanh MN, Dudbridge F, et al. Genes of the class II
and class III major histocompatibility complex are associated with typhoid fever in Vietnam. J Infect.
Dis. 2001; 183:261–8. https://doi.org/10.1086/317940 PMID: 11120931
23. Majumder PP. Genomics of immune response to typhoid and cholera vaccines. Philos. Trans. R. Soc.
B. 2015; 370.
24. McArthur MA, Fresnay S, Magder LS, Darton TC, Jones C, Waddington CS, et al. Activation of Salmo-
nella Typhi-specific regulatory T cells in typhoid disease in a wild-type S. Typhi challenge model. PLOS
Pathog. 2015; 11:e1004914. https://doi.org/10.1371/journal.ppat.1004914 PMID: 26001081
25. Fresnay S, McArthur MA, Magder L, Darton TC, Jones C, Waddington CS, et al. Salmonella Typhi-spe-
cific multifunctional CD8+ T cells play a dominant role in protection from typhoid fever in humans. J.
Transl. Med. 2016; 14:62. https://doi.org/10.1186/s12967-016-0819-7 PMID: 26928826
26. Hobson D. Resistance to reinfection in experimental mouse typhoid. Epidemiol. Infect. 1957; 55
(3):334–43.

Dtifo

Uploaded by

Copyright:

Available Formats

Dtifo

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dtifo

Uploaded by

Copyright:

Available Formats

PLOS NEGLECTED TROPICAL DISEASES

Protection conferred by typhoid fever against

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 1 / 11

Funding: The research was supported by the

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 2 / 11

Vi typhoid vaccine trial

Defining typhoid relapse and reinfections

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 3 / 11

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 4 / 11

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 5 / 11

§ Pairwise SNP difference >5

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 6 / 11

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 7 / 11

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 8 / 11

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 9 / 11

Formal analysis: Deok Ryun Kim, Faisal Ahmmed, John D. Clemens.

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 10 / 11

PLOS Neglected Tropical Diseases | https://doi.org/10.1371/journal.pntd.0008530 August 17, 2020 11 / 11

You might also like