https://orcid.org/0000-0003-4093-4054
Figures
Abstract
Background
Cholera surveillance relies on clinical diagnosis of acute watery diarrhea. Suspected cholera case definitions have high sensitivity but low specificity, challenging our ability to characterize cholera burden and epidemiology. Our objective was to estimate the proportion of clinically suspected cholera that are true Vibrio cholerae infections and identify factors that explain variation in positivity.
Methods and findings
We conducted a systematic review of studies that tested ≥10 suspected cholera cases for V. cholerae O1/O139 using culture, PCR, and/or a rapid diagnostic test. We searched PubMed, Embase, Scopus, and Google Scholar for studies that sampled at least one suspected case between January 1, 2000 and April 19, 2023, to reflect contemporary patterns in V. cholerae positivity. We estimated diagnostic test sensitivity and specificity using a latent class meta-analysis. We estimated V. cholerae positivity using a random-effects meta-analysis, adjusting for test performance. We included 119 studies from 30 countries. V. cholerae positivity was lower in studies with representative sampling and in studies that set minimum ages in suspected case definitions. After adjusting for test performance, on average, 52% (95% credible interval (CrI): 24%, 80%) of suspected cases represented true V. cholerae infections. After adjusting for test performance and study methodology, the odds of a suspected case having a true infection were 5.71 (odds ratio 95% CrI: 1.53, 15.43) times higher when surveillance was initiated in response to an outbreak than in non-outbreak settings. Variation across studies was high, and a limitation of our approach was that we were unable to explain all the heterogeneity with study-level attributes, including diagnostic test used, setting, and case definitions.
Conclusions
In this study, we found that burden estimates based on suspected cases alone may overestimate the incidence of medically attended cholera by 2-fold. However, accounting for cases missed by traditional clinical surveillance is key to unbiased cholera burden estimates. Given the substantial variability in positivity between settings, extrapolations from suspected to confirmed cases, which is necessary to estimate cholera incidence rates without exhaustive testing, should be based on local data.
Author summary
Why was this study done?
- Cholera surveillance typically relies on the clinical diagnosis of acute watery diarrhea (i.e., “suspected cholera”), but this definition has a low specificity for cholera.
- Our goal was to estimate the proportion of suspected cholera cases that are true Vibrio cholerae infections and identify factors that contribute to variation in observed positivity.
What did the researchers do and find?
- We conducted a systematic review of studies from 2000 to 2023 that tested suspected cholera cases for V. cholerae infection using one of 3 different laboratory tests.
- We included 119 studies from 30 countries and found that, on average, half of suspected cholera cases represented true V. cholerae infections, after accounting for laboratory test accuracy.
- We also found high variability between studies and that the odds of a suspected case being a true infection were higher during outbreaks compared to non-outbreak settings.
What do these findings mean?
- Our findings suggest that burden estimates based solely on suspected cases may overestimate the incidence of medically attended cholera by 2-fold.
- The high variability across studies suggests also that local testing data should be used to inform assumptions about positivity when exhaustive testing is not feasible.
- A limitation of our approach was that we could not account for cases missed by clinical surveillance, which is crucial for unbiased overall cholera burden estimates and an important area for future work.
Citation: Wiens KE, Xu H, Zou K, Mwaba J, Lessler J, Malembaka EB, et al. (2023) Estimating the proportion of clinically suspected cholera cases that are true Vibrio cholerae infections: A systematic review and meta-analysis. PLoS Med 20(9): e1004286. https://doi.org/10.1371/journal.pmed.1004286
Received: October 25, 2022; Accepted: August 25, 2023; Published: September 14, 2023
Copyright: © 2023 Wiens 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.
Data Availability: All input data and analytical code are available at https://github.com/HopkinsIDD/cholera_positivity.
Funding: This work was supported by the Bill and Melinda Gates Foundation (https://www.gatesfoundation.org/) [grant number OPP1171700 to A.S.A.] and the National Institute of Allergy and Infectious Disease (https://www.niaid.nih.gov/) [grant number AI135115-01A1 to A.S.A.]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: JL is a paid statistical advisor for PLOS Medicine.
Abbreviations: CI, confidence interval; CrI, credible interval; GTFCC, Global Task Force on Cholera Control; IQR, interquartile range; JAGS, Just Another Gibbs Sampler; PCR, polymerase chain reaction; RDT, rapid diagnostic test
Introduction
Current estimates of cholera burden rely on clinical diagnosis of individuals with acute watery diarrhea (i.e., suspected cholera cases) [1,2]. It is unclear how many Vibrio cholerae O1/O139 (serogroups that cause current epidemics) infections get missed due to mild symptoms and other barriers to care-seeking or how many get overcounted due to nonspecific suspected case definitions. In Bangladesh, previous studies estimated that asymptomatic and unreported infections account for at least half of V. cholerae infections [3–5]. Meanwhile, the proportion of suspected cholera cases that represent laboratory-confirmed infections varies widely between studies, from 6% of those tested during routine surveillance in Bangladesh [6] to 72% of those tested during the initial phase of the 2017 outbreak in Yemen [7].
This wide variation in positivity may be caused by differences between sites in V. cholerae epidemiology [8], epidemiology of non-cholera diseases causing the same clinical symptoms [9–12], and variations in diagnostic tests and case definitions [13–15]. Typical suspected cholera case definitions have been shown to have high sensitivity but low specificity [14] for detecting true cholera and can vary by location across seasons [13]. Culture-based methods or polymerase chain reaction (PCR) are the gold standards to confirm cholera in clinical samples and generally have high specificity. Lateral flow rapid diagnostic tests (RDTs) may also be used and can be as sensitive as PCR [16]. Although recommended by the Global Task Force on Cholera Control (GTFCC) [17], systematic microbiological confirmation in surveillance is not always implemented, particularly during outbreaks when resources are limited [8]. To our knowledge—based on a literature review and discussion with experts—no study had yet systematically synthesized these data to estimate overall V. cholerae positivity and identify sources of this variation.
Understanding V. cholerae positivity among clinical cases could provide insights needed to improve laboratory testing strategies and allow for better estimates of cholera burden and risk, which are often used to allocate cholera resources, including oral cholera vaccines. Starting in 2023, the GTFCC has recommended using a combination of suspected cholera incidence, persistence, mortality, and cholera test positivity data across multiple years to identify priority areas for multisectoral interventions [18], which is particularly relevant in cholera endemic areas. As described above, the V. cholerae positivity data are often not available. We sought to address this knowledge gap by modeling the relationship between clinically suspected and laboratory confirmed cholera. Specifically, we aimed to estimate the proportion of suspected cholera cases that represent true V. cholerae O1/O139 infections and identify factors that explain variability in positivity across settings.
Methods
Ethics
This study was approved by the Johns Hopkins University Institutional Review Board and Temple University Institutional Review Board.
Terminology
We focused on V. cholerae O1 and O139 because these are the serogroups that are responsible for the current seventh pandemic and the only ones known to lead to large outbreaks in humans [19]. These are also the serogroups that are targeted by each of the commonly used V. cholerae diagnostic tests (culture, PCR, and RDT). Throughout this manuscript, we refer to the proportion of suspected cholera cases that represent true V. cholerae O1/O139 infections as “V. cholerae positivity” or “cholera positivity.” In addition, since the available data did not allow us to evaluate the performance of multiple RDTs, we refer to RDT as any rapid diagnostic test for V. cholerae O1/O139 and do not distinguish between different RDT manufacturers or whether the RDT is enriched/direct swab RDT or stool RDT.
Systematic review
This study is reported as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline (S1 Checklist). The review was not preregistered, and a formal public protocol was not prepared, although all study methods can be found in the Methods below and the Methods in S1 Appendix.
We searched PubMed, Embase, Scopus, Google Scholar, and medRxiv on October 16, 2021, using search provided in the Supplementary Methods in S1 Appendix. We updated PubMed, Embase, and Scopus searches on April 19, 2023. We included studies that (1) collected human samples; (2) reported the number of suspected and confirmed cholera cases in the sampling frame; (3) used culture, PCR, and/or RDT to test suspected cases for cholera; and (4) had at least one suspected case sample collected on or after January 1, 2000, to reflect contemporary patterns in cholera positivity. We excluded studies that (1) used a case definition not specific for suspected cholera (i.e., we accepted non-bloody watery diarrhea, acute watery diarrhea, or simply suspected cholera but not diarrhea, acute diarrhea, or acute gastroenteritis); (2) sampled only special populations (i.e., people living with HIV or cancer); (3) selected suspected cases based on epidemiological link to other cases or environmental sources; (4) tested fewer than 10 suspected cases; and (5) were reported in languages other than English, French, Spanish, and Chinese (languages our study team had proficiency in). We did not exclude studies based on study type or sampling method. Although we originally included preprints in our screening and extracted one preprint, we excluded this study at the time of the updated search because the published version of the manuscript no longer included positivity data.
Titles, abstracts, and full texts were uploaded to Covidence, a web-based screening tool (https://www.covidence.org/), and were assessed independently by two of the reviewers (ASA, ECL, HX, KEW, KZ, MND) for inclusion. Conflicts were resolved either by a third reviewer or through consensus/discussion. Data were extracted from included studies in a shared spreadsheet (S1 Data) by a single reviewer. The key extracted items included study timeframe and location, surveillance type (routine, outbreak, post-vaccination, or hybrid), case definition of suspected cholera (including age constraint and whether dehydrated or hospitalized, if provided), test method(s), sampling strategy for the test (all suspected cases, systematic or random sampling, convenience sampling, or unreported), number of tested and confirmed suspected cases, among other sample characteristics, if included. If only the proportion positive and total number tested were reported, the number of confirmed cholera cases was calculated by hand and rounded to the nearest whole number. If the surveillance contained multiple timeframes, tested samples with multiple tests, or reported stratified results, we extracted the data separately into different rows in the spreadsheet.
To identify overlapping samples, we manually reviewed all studies with overlapping timeframes by country. We excluded studies that had shorter timeframes, fewer suspected cases tested, less representative sampling methods, fewer confirmation tests, or reported positive results by 2 tests but did not disaggregate. Within studies, when suspected and confirmed cases were stratified multiple ways, we included the stratification by surveillance type if available, followed by age, antibiotic use, dehydration status, year, geography, or sex, in that order. When studies used multiple RDTs, we included results for Crystal VC (Arkray Healthcare, Gujarat, India) and direct rapid tests (as opposed to rapid tests performed after an enrichment step) because these were the most common.
To identify any mistakes and ensure quality of the extracted data, we performed data quality checks using a series of automated functions in R to identify implausible values (e.g., start date of study after end date, more cases positive than tested, lower age limit larger than upper age limit) and missing required data. If impossible or missing values were found, the entire extraction was double checked for accuracy and corrected by a single reviewer.
To assess whether different studies used methodologies that may have biased our results, we plotted cholera positivity in the raw data by (1) diagnostic test used; (2) sampling method quality; and (3) suspected cholera case definition. In addition, we plotted the relationship between cholera positivity in the raw data and (1) estimated suspected cholera incidence [2]; (2) the proportion of cases severely dehydrated; and (3) the proportion on antibiotics. We quantified the correlation between these variables using Spearman’s rank correlation coefficient using the spearman.ci function of the RVAideMemoire package in R [20]. Since these continuous variables were only available in a subset of studies, we did not adjust for them in final analyses. All data visualization was conducted using the ggplot2 package in R [21].
Data analysis
Estimating sensitivity and specificity of cholera confirmation tests.
We constructed a latent-class model to assess sensitivity and specificity of culture, PCR, and RDT, assuming none had perfect performance. We fit a hierarchical conditional dependence model, similar to that proposed by Wang and colleagues, which takes into account potential pairwise dependence between the tests that could occur if the tests have reduced performance for similar reasons [22]. We performed inference in a Bayesian framework using Just Another Gibbs Sampler (JAGS) through the rjags package in R [23,24]. We pooled estimates across 4 published studies that reported cholera confirmation results for all 3 test methods [16,25–27].
We used flat prior distributions on sensitivity and specificity of each test with a lower bound set based on plausible values from the literature [15,16,25–27] (Table A in S1 Appendix). We assumed that culture had lower sensitivity than PCR and RDT because it depends on successful growth of viable V. cholerae in the laboratory. We assumed that RDT had lower specificity than culture and PCR because it may have cross-reactivity with other antigens in the stool or defects that lead to false positive results. For each prior, we selected a wider range than had been reported in previous studies to allow for greater variation. We ran 4 chains of 100,000 iterations and assessed convergence through visual inspection of traceplots and with the Gelman-Rubin R-hat statistic.
Estimating V. cholerae positivity and sources of heterogeneity.
We pooled estimates of V. cholerae positivity across all studies using a generalized linear model with a study level random intercept, which allowed us to adjust for sensitivity and specificity of the diagnostic tests as well as examine the contributions of study methodology (i.e., whether the study used low- versus high-quality sampling, and whether or not the study set a minimum age in the suspected cholera case definition) and setting (whether surveillance was routine or post-vaccination versus initiated in response to an outbreak) on variation in positivity. To estimate the proportion positive, overall and by strata, we marginalized over study-level random effects. See Supplementary Methods in S1 Appendix for the full statistical model. We performed inference in a Bayesian framework using CmdStanR version 0.5.2 as an interface to Stan for R [24,28]. We additionally performed a sensitivity analysis where we shifted the prior set on the global intercept (see Methods in S1 Appendix). The odds of a suspected cholera case having a true V. cholerae infection given each covariate were calculated as odds ratios by taking the mean and 95% credible interval (CrI) of 8,000 draws from the posterior distribution of each covariate’s exponentiated coefficient. Odds ratios with 95% CrIs that did not cross the value 1 were considered statistically significant.
To estimate the proportion of the variance in positivity attributable to true differences between studies, beyond simple sampling error, we calculated the I2 statistic [29] as
where τ2 was between-study heterogeneity or the variance of the random effect by observation. We calculated the within-study variance, υ, [30] as
where k was the number of studies or observations included in the meta-analysis, and ωi = 1/vi where vi was the variance of the proportion positive by culture, PCR, or RDT within each study/observation. When multiple tests were used in a study, we used the maximum variance estimate across the tests.
Results
Study characteristics
We identified 131 studies that met our inclusion criteria (Fig 1). Of these, 119 studies contained nonoverlapping samples and were included in our analysis dataset [6,7,9,10,12–14,16,25–27,31–131] and 12 were excluded from analysis due to overlaps [8,11,132–141] (Fig 1). Of the 119 studies included in our analysis dataset, one reported data for more than one sampling method [7], one for both outbreak and non-outbreak surveillance [37], and one for outbreak and non-outbreak surveillance in 6 different countries [13]. We defined each of these as separate entries in the dataset for a total of 132 observations. Extracted data including detailed individual study information can be found in S1 Data.
Diagram illustrating literature selection process, including databases searched, literature screened, and full texts reviewed for eligibility. Reasons for exclusion are indicated along with the number of studies that fell within each category. PCR, polymerase chain reaction; RDT, rapid diagnostic test.
The nonoverlapping observations in our analysis dataset came from 30 countries and were reported at different geographic levels, including the country level (n = 16 observations) and first (n = 25), second (n = 66), and third administrative levels (n = 25) (Fig A in S1 Appendix). Twelve studies reported data for multiple administrative units, and 3 reported across multiple administrative divisions within a country; the numbers above reflect the largest administrative division reported per observation. Data were collected from 1992 through 2022 with most observations from studies that completed sampling during 2015 to 2022 (n = 53 observations), followed by 2010 to 2014 (n = 32), 2005 to 2009 (n = 21), and 1997 to 2004 (n = 17) (Fig B in S1 Appendix). Nine studies were missing sampling end dates. Most studies were conducted in South Asia and West, Central, and East Africa, with additional studies from Haiti, Yemen, Iraq, Iran, Laos, Vietnam, Papua New Guinea, Algeria, and the Philippines (Fig A in S1 Appendix).
Most of the observations were from surveillance studies (93/132, 70.5%), followed by diagnostic test accuracy studies (28/132, 21.2%) and vaccine effectiveness studies (10/132, 7.6%) (Table 1). High-quality sampling methods (i.e., tested all suspected cases, a random sample, or systematically selected every nth suspected case) were used in 28% (37/132) of observations, while the remaining 72% (95/132) used convenience sampling or did not report the sampling approach (Table 1). Even though most studies did not include V. cholerae positivity disaggregated by individual-level characteristics, 24.2% (32/132) reported the proportion of suspected cases under age 5, 8.3% (11/132) reported the proportion severely dehydrated, 7.6% (10/132) reported the proportion on antibiotics, and one study reported all 3 (Table B in S1 Appendix).
Number of observations included in the analysis dataset with each study characteristic. There is more than one observation per study when the study reported data for more than one sampling method, surveillance type, and/or country.
V. cholerae positivity in unadjusted data
We found that reported V. cholerae positivity varied greatly across studies with an interquartile range (IQR) of 30% to 60% (N = 165 observations of positivity; 25 of the 131 observations had positivity results for multiple tests) (Table 1). As expected, positivity varied by diagnostic test used with a median positivity of 36% by culture (IQR, 27% to 55%; N = 121), 37% by PCR (IQR, 34% to 55%; N = 11), and 49% by RDT (IQR, 38% to 67%; N = 33), with substantial overlap between distributions (Fig 2A). Positivity was higher across studies that used low-quality or convenience sampling methods (median of 43%; N = 117; IQR, 33% to 62%) compared to those that used high-quality or representative sampling (median of 35%; IQR, 14% to 51%) (Fig 2B). Positivity increased with higher minimum ages in suspected cholera case definitions (Fig 2C), and we found a modest negative correlation between positivity and the proportion of suspected cases under 5 years old (Spearman r = −0.60; 95% confidence interval (CI): −0.81, −0.32; p < 0.001) (Fig Ca in S1 Appendix).
Proportion of suspected cholera cases that were confirmed positive by (A) diagnostic test type, (B) quality of sampling methods, where “high” includes all suspected cases or a random or stratified sample and “low” includes convenience or unreported sampling methods, (C) age minimum in suspected case definition, where “0” indicates that no minimum age was set, and (D) whether surveillance was initiated in response to an outbreak or whether it was routine surveillance or non-outbreak. Each point is an observation included in the analysis dataset. There is more than one observation per study when the study reported data for more than one sampling method, surveillance type, and/or country. Boxes represent the median and IQR of positivity for each group. Lines extend from the top and bottom of box to the largest positivity value no further than 1.5 * IQR from the box. IQR, interquartile range; PCR, polymerase chain reaction; RDT, rapid diagnostic test.
Unadjusted positivity was higher when surveillance was initiated in response to an outbreak (median of 47%; IQR, 33% to 66%; N = 80) compared to situations where surveillance was routine or post-vaccination (median of 35%; IQR 17% to 49%; N = 85) (Fig 2D). We found limited evidence for differences in positivity by the 2010 to 2016 estimated mean annual suspected case incidence rate in countries where these estimates were available (Fig Cb in S1 Appendix; [2]).
We found a modest positive correlation between positivity and the proportion of suspected cases severely dehydrated (Spearman r = 0.64; 95% CI: 0.22, 0.90; p = 0.001) (Fig Cc in S1 Appendix). While not statistically significant, we found a weak negative correlation between positivity and the proportion of suspected cases that had received antibiotics prior to testing (Spearman r = −0.46; 95% CI: −0.83, 0.09; p = 0.07) (Fig Cd in S1 Appendix).
Adjusted underlying V. cholerae positivity
Since different imperfect diagnostic tests were used to confirm V. cholerae O1/O139, we adjusted positivity estimates from each study to account for test performance. To estimate a median performance of each type of diagnostic test, we pooled estimates of sensitivity and specificity across 4 studies that reported detailed results for all 3 tests (see Methods). This included data from Bangladesh [27], South Sudan [16], Kenya [25], and Zambia [26]. We estimated a median sensitivity of 82.0% (95% CrI: 37.5, 98.7) and specificity of 94.3% (95% Crl: 81.5, 99.6) for culture, a median sensitivity of 85.1% (95% CrI: 53.6%, 98.9%) and specificity of 94.2 (95% CrI: 81.8, 99.7) for PCR, and a median sensitivity of 90.4% (95% CrI: 55.2, 99.5) and specificity of 88.9% (95% CrI: 54.9, 99.4) for RDT (Fig 3A and Table C in S1 Appendix).
(A) Posterior distributions of pooled percent sensitivity and specificity of culture (top), PCR (middle), and RDT (bottom) for detecting V. cholerae O1/O139 infections in suspected cholera cases. Dashed lines represent median values of each distribution. (B) The “Unadjusted” dot is mean V. cholerae positivity (lines represent 95% CrI) from random effects meta-analysis without adjustments for test performance. The “Adjusted for test performance” and “Stratum: …” dots are estimated mean V. cholerae positivity (lines represent 95% CrIs), adjusted for sensitivity/specificity of the tests. High-quality stratified estimates correspond to post-stratified estimates of V. cholerae positivity for studies that use high quality sampling methods and whether an age minimum was set in the suspected case definition, as well as whether surveillance was initiated in response to an outbreak. CrI, credible interval; PCR, polymerase chain reaction; RDT, rapid diagnostic test.
After adjusting for diagnostic test performance, we estimated that 53% (95% CrI: 24%, 80%) of suspected cases tested were true V. cholerae O1/O139 infections across all studies (Fig 3 and Fig D in S1 Appendix and Table D in S1 Appendix). These estimates remained similar in sensitivity analysis with an alternative prior distribution (Table D in S1 Appendix).
With additional adjustments for study methodology (i.e., sampling quality and whether an age minimum was set in suspected case definition), we estimated that V. cholerae positivity for studies with high-quality sampling methods was 46% (95% CrI: 19%, 76%) when no age restriction was used and 68% (95% CrI: 33%, 98%) when a minimum age (typically 1 or 5 years old) was incorporated into the case definition (Fig 3, Table D in S1 Appendix). After adjusting for sampling quality and whether or not surveillance was initiated in response to a cholera outbreak, we estimated that V. cholerae positivity for studies with high-quality sampling methods was 42% (95% CrI: 12%, 77%) in non-outbreak settings and 78% (95% CrI: 40%, 99%) in outbreak settings (Fig 3, Table D in S1 Appendix).
We found substantial heterogeneity between studies (I2 = >99.99% (95% CrI: >99.99%, >99.99%; I2 = 0.96 (95% CrI: 0.94, 0.98)) (Fig 4). Adjusted underlying positivity rates ranged from 0.008% (95% CrI: 0.0004%, 0.04%) for a high-quality study conducted during routine surveillance in Bangladesh to 99.8% (95% CrI: 98.7%, 100.0%) for a “low-quality” study conducted during a cholera outbreak in Uganda (Fig 4).
Black points indicate mean study-level underlying positivity and 95% CrI. Teal, orange, and purple points indicate the proportion positive reported by study for culture, PCR, and RDT, respectively, and corresponding error bars indicate 95% CI for a binomial probability using the normal approximation [147]. Studies are labeled by country ISO3 code, quality of sampling methods, (high or low), and whether a minimum age was set in the suspected cholera case definition, (yes or no). Studies are split into outbreak and non-outbreak for ease of interpretation. CI, confidence interval; CrI, credible interval; PCR, polymerase chain reaction; RDT, rapid diagnostic test.
Factors associated with variation in V. cholerae positivity
We then examined factors that could explain variation in V. cholerae positivity. After adjusting for test performance, sampling quality, and outbreak setting, we found that setting any minimum age in the case definition (i.e., 1, 2, 5, or 10) was associated with 2.33 (95% CrI: 0.54, 6.40) times higher odds of a suspected cholera case having a true infection (Table E in S1 Appendix).
We estimated that the odds of a suspected cholera case having a true V. cholerae O1/O139 infection were 5.71 (95% CrI: 1.53, 15.43) times higher when surveillance was initiated in response to a cholera outbreak compared to non-outbreak surveillance, after adjusting for test performance, sampling quality, and case definition (Table E in S1 Appendix).
Discussion
Here, we estimated that, on average, half of medically attended suspected cholera cases represent true V. cholerae O1/O139 infections. We found that V. cholerae positivity was higher when a minimum age was set in case definitions and when surveillance was initiated in response to an outbreak. Additionally, we found substantial heterogeneity in V. cholerae positivity between studies, so that simply multiplying the number of suspected cholera case counts by this global proportion positive to estimate the true number of cases will not be appropriate in most settings. To our knowledge, this is the first study to systematically synthesize data globally to estimate overall V. cholerae positivity and examine factors that contribute to variation in positivity.
A remaining question is why only about half of medically attended suspected cholera cases represent true infections. It is possible that we overestimated test sensitivity and have not fully accounted for false negatives; unfortunately, this is difficult to evaluate without a gold standard diagnostic test. A portion of the remaining suspected cases could also be infections with other enteric pathogens, especially those with similar transmission modes as cholera that may have outbreaks or high levels of endemic transmission concurrently. For example, in Uvira, Democratic Republic of the Congo, 36% of suspected cholera cases were positive for Enterotoxigenic Escherichia coli and 28% for Cryptosporidium [10]. In rural Bangladesh, the majority of acute watery diarrhea in children under 18 months was attributable to rotavirus, while older children were more often infected with V. cholerae [12]. In Haiti, 64% of acute watery diarrhea cases tested positive for V. cholerae O1, 4% for rotavirus, and <1% for Shigella and Salmonella, though rotavirus positivity was higher among children under 5 [11]. Thus, the relative contribution of non-cholera watery diarrhea varies with age distribution and other location-specific drivers of enteric infections.
One of the limitations of this study was that we could not account for all potential drivers of V. cholerae positivity, which contributed to the large heterogeneity we found between studies. In addition, V. cholerae positivity may be highest in the early stages of an outbreak [7,9,131], but we could not account for this, given the temporal resolution of our dataset. However, a strength of our approach is that we pooled estimates from studies across diverse geographies, time periods, and epidemiological contexts. A further potential limitation is that, without a gold standard diagnostic test, sensitivity and specificity estimates may be biased if the tests are less sensitive and/or specific for shared reasons. The hierarchical conditional dependence model we used accounted for this pairwise dependence and increased uncertainty around our estimates accordingly. This approach also allowed us to pool test performance estimates across studies from 4 countries. Thus, to our knowledge, we adjusted our estimates for test sensitivity and specificity using the best generic estimates available. Still, we likely overestimated sensitivity of culture for settings where samples had to be sent to a reference lab. Variation in the timing of tests in relation to when sample was taken could mean that one sensitivity and specificity estimate per diagnostic method is not appropriate. For example, a 2023 study in Haiti found that stool culture had a sensitivity of 33% during the waning phase of the 2018 to 2019 cholera outbreak [142], which is much lower than previous estimates. Overall, we have high confidence in our average estimates of V. cholerae positivity, despite the difficulty of accurately estimating positivity in a new location/time/setting without confirmation tests.
These findings have several implications for cholera surveillance policy. The GTFCC defines suspected cholera in areas where an outbreak has not yet been reported as acute watery diarrhea and severe dehydration or death in individuals 2 years and older [17]. Our finding that setting any minimum age increases specificity for identifying a true V. cholerae infection in suspected cases supports using an age restriction in this case definition. The February 2023 interim guidance from the GTFCC on cholera surveillance provides concrete recommendations for systematic and frequent testing of suspected cholera cases at the health facility or surveillance unit scale [17]. Our finding of high variability in positivity across settings and times lends support to these recommendations of systematically generating local data that can be used to scale suspected to true cholera. Our finding that high-quality sampling also increases specificity for V. cholerae suggests that systematically selecting cases to test is important for accurately evaluating endemic cholera. Finally, that V. cholerae positivity was lower during non-outbreak surveillance suggests that systematic confirmation testing is additionally important for understanding cholera burden and epidemiology in endemic, non-outbreak settings where cocirculation of other enteric pathogens is common.
These estimates of V. cholerae positivity address one part of the challenge in establishing the true burden of cholera: cases that are overcounted due to nonspecific suspected case definitions. A crucial next step will be to estimate missed cases due to care seeking and poor clinical surveillance. This could be done in part through systematically synthesizing data from studies of care seeking behavior for diarrheal symptoms (e.g., [143,144]), including where potential cholera cases seek care (e.g., at pharmacies, traditional healers, or hospitals). This could additionally be done through population representative surveys and active case finding, similar to studies conducted in Haiti [145] and Tanzania [146], respectively, which demonstrated higher mortality rates associated with cholera than had been reported through passive surveillance. Together, these studies will help to understand whether and to what degree missed cholera cases compensate for the biases described here in overcounting.
Ultimately, a better understanding of V. cholerae positivity will help us move toward estimates of true cholera incidence and mortality. Given the large heterogeneity between studies, it will be important to do this in a way that accounts for variation in V. cholerae positivity between sites. Moreover, the proportion of suspected cholera cases missed because of milder symptoms or barriers to healthcare seeking needs to be estimated and accounted for. Such estimates will provide crucial information to guide the allocation of limited resources such as vaccines in a way that most effectively supports cholera prevention and control.
Supporting information
S1 Checklist. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 Checklist.
https://doi.org/10.1371/journal.pmed.1004286.s001
(PDF)
S1 Appendix. Supporting information.
Detailed methods, including systematic review search terms and the full statistical model, as well as additional figures and tables.
https://doi.org/10.1371/journal.pmed.1004286.s002
(PDF)
S1 Data. Full dataset.
Excel sheet with the complete data extracted from all 131 studies that met the inclusion criteria (tab 1) as well as all variable descriptions (tab 2). Data extracted from the 119 nonoverlapping studies included in the main analysis dataset can be found by filtering for the values “1” in the column “Primary dataset.”
https://doi.org/10.1371/journal.pmed.1004286.s003
(XLSX)
Acknowledgments
We thank Morgane Dominguez for feedback on this manuscript, Lori Rosman for assistance developing the literature search strategy, and Javier Perez-Saez for feedback on the analytical methods.
References
- 1. Ali M, Nelson AR, Lopez AL, Sack DA. Updated global burden of cholera in endemic countries. PLoS Negl Trop Dis. 2015;9:e0003832. pmid:26043000
- 2. Lessler J, Moore SM, Luquero FJ, McKay HS, Grais R, Henkens M, et al. Mapping the burden of cholera in sub-Saharan Africa and implications for control: an analysis of data across geographical scales. Lancet. 2018;391:1908–1915. pmid:29502905
- 3. Azman AS, Lauer SA, Bhuiyan TR, Luquero FJ, Leung DT, Hegde ST, et al. Vibrio cholerae O1 transmission in Bangladesh: insights from a nationally representative serosurvey. Lancet Microbe. 2020;1:e336–e343. pmid:33345203
- 4. Mosley WH, Benenson AS, Barui R. A serological survey for cholera antibodies in rural East Pakistan. Bull World Health Organ. 1968;38:327–334.
- 5. Weil AA, Begum Y, Chowdhury F, Khan AI, Leung DT, LaRocque RC, et al. Bacterial shedding in household contacts of cholera patients in Dhaka. Bangladesh Am J Trop Med Hyg. 2014;91:738–742. pmid:25114012
- 6. Khan AI, Rashid MM, Islam MT, Afrad MH, Salimuzzaman M, Hegde ST, et al. Epidemiology of cholera in Bangladesh: findings from nationwide hospital-based surveillance, 2014–2018. Clin Infect Dis. 2020;71:1635–1642. pmid:31891368
- 7. Camacho A, Bouhenia M, Alyusfi R, Alkohlani A, Naji MAM, Radiguès X de, et al. Cholera epidemic in Yemen, 2016–18: an analysis of surveillance data. Lancet Glob Health. 2018;6: e680–e690. pmid:29731398
- 8. Sauvageot D, Njanpop-Lafourcade B-M, Akilimali L, Anne J-C, Bidjada P, Bompangue D, et al. Cholera incidence and mortality in sub-Saharan African sites during multi-country surveillance. PLoS Negl Trop Dis. 2016;10:e0004679. pmid:27186885
- 9. Jameel SK, Shafek MA, Abdulmohsen AM, Mohamed NS, Naji SR, Mohammed TT. The isolation of Vibrio cholera and other enteric bacteria with molecular characterization of Vibrio cholera during the outbreak of Baghdad/Iraq in 2015. Adv Microbiol. 2016;6:699–715.
- 10. Williams C, Cumming O, Grignard L, Rumedeka BB, Saidi JM, Grint D, et al. Prevalence and diversity of enteric pathogens among cholera treatment centre patients with acute diarrhea in Uvira, Democratic Republic of Congo. BMC Infect Dis. 2020;20:741. pmid:33036564
- 11. Steenland MW, Joseph GA, Lucien MAB, Freeman N, Hast M, Nygren BL, et al. Laboratory-confirmed cholera and rotavirus among patients with acute diarrhea in four hospitals in Haiti, 2012–2013. Am J Trop Med Hyg. 2013;89:641–646. pmid:24106190
- 12. Siddique AK, Ahmed S, Iqbal A, Sobhan A, Poddar G, Azim T, et al. Epidemiology of rotavirus and cholera in children aged less than five years in rural Bangladesh. J Health Popul Nutr. 2011;29:1–8. pmid:21528784
- 13. Nadri J, Sauvageot D, Njanpop-Lafourcade B-M, Baltazar CS, Banla Kere A, Bwire G, et al. Sensitivity, specificity, and public-health utility of clinical case definitions based on the signs and symptoms of cholera in Africa. Am J Trop Med Hyg. 2018;98:1021–1030. pmid:29488455
- 14. Lucien MAB, Schaad N, Steenland MW, Mintz ED, Emmanuel R, Freeman N, et al. Identifying the most sensitive and specific sign and symptom combinations for cholera: results from an analysis of laboratory-based surveillance data from Haiti, 2012–2013. Am J Trop Med Hyg. 2015;92:758–764. pmid:25732682
- 15. Muzembo BA, Kitahara K, Debnath A, Okamoto K, Miyoshi S-I. Accuracy of cholera rapid diagnostic tests: a systematic review and meta-analysis. Clin Microbiol Infect. 2022;28(2):155–162. pmid:34506946
- 16. Ontweka LN, Deng LO, Rauzier J, Debes AK, Tadesse F, Parker LA, et al. Cholera rapid test with enrichment step has diagnostic performance equivalent to culture. PLoS ONE. 2016;11:e0168257. pmid:27992488
- 17.
Global Task Force on Cholera Control (GTFCC) Surveillance Working Group. Public health surveillance for cholera interim guidance. 2023. Available from: https://www.gtfcc.org/wp-content/uploads/2023/02/gtfcc-public-health-surveillance-for-cholera-interim-guidance.pdf
- 18.
Global Task Force on Cholera Control. Identification of priority areas for multisectoral interventions (PAMIs) for cholera control. [cited 2023 May 16]. Available from: https://www.gtfcc.org/resources/identification-of-priority-areas-for-multisectoral-interventions-pamis-for-cholera-control/
- 19. Weill F-X, Domman D, Njamkepo E, Tarr C, Rauzier J, Fawal N, et al. Genomic history of the seventh pandemic of cholera in Africa. Science. 2017;358:785–789. pmid:29123067
- 20. Herve M. RVAideMemoire: testing and plotting procedures for biostatistics. 2023. Available from: https://cran.r-project.org/web/packages/RVAideMemoire/index.html
- 21. Wickham H, Chang W, Henry L, Pedersen TL, Takahashi K, Wilke C, et al. ggplot2: create elegant data visualisations using the grammar of graphics. 2023. Available from: https://cran.r-project.org/web/packages/ggplot2/index.html
- 22. Wang C, Lin X, Nelson KP. Bayesian hierarchical latent class models for estimating diagnostic accuracy. Stat Methods Med Res. 2020;29:1112–1128. pmid:31146651
- 23. Plummer M, Stukalov A, Denwood M. rjags: Bayesian graphical models using MCMC. 2022. Available from: https://CRAN.R-project.org/package=rjags
- 24.
R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. http://wwwR-project.org/. 2016 [cited 2022 Aug 17]. Available from: https://cir.nii.ac.jp/crid/1574231874043578752
- 25. Debes AK, Murt KN, Waswa E, Githinji G, Umuro M, Mbogori C, et al. Laboratory and field evaluation of the Crystal VC-O1 cholera rapid diagnostic test. Am J Trop Med Hyg. 2021;104:2017–2023. pmid:33819171
- 26. Mwaba J, Ferreras E, Chizema-Kawesa E, Mwimbe D, Tafirenyika F, Rauzier J, et al. Evaluation of the SD bioline cholera rapid diagnostic test during the 2016 cholera outbreak in Lusaka, Zambia. Trop Med Int Health. 2018;23:834–840. pmid:29851181
- 27. Sayeed MA, Islam K, Hossain M, Akter NJ, Alam MN, Sultana N, et al. Development of a new dipstick (Cholkit) for rapid detection of Vibrio cholerae O1 in acute watery diarrheal stools. PLoS Negl Trop Dis. 2018;12:e0006286. pmid:29538377
- 28. Gabry J, Češnovar R, Bales B, Morris M, Popov M, Lawrence M, et al. R interface to CmdStan. 2022 [cited 2022 Aug 17]. Available from: https://mc-stan.org/cmdstanr/
- 29. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–1558. pmid:12111919
- 30. Viechtbauer W. I2 for multilevel and multivariate models. 2022 [cited 2022 Aug 26]. Available from: https://www.metafor-project.org/doku.php/tips:i2_multilevel_multivariate
- 31. Abdullahi KN, Mutindin D, Kabugi W, Mowlid S. Epidemiological description of a protracted cholera outbreak in Hagadera refugee camp and the surrounding host community within Fafi Sub County and Garissa County in Kenya during march-September 2019. Epidemiol Open J. 2019;4:31–35.
- 32. Ahmed S, Afzal RK, Mian UA. A localized outbreak of cholera due to Vibrio cholerae 01, Ogawa resistant to tetracyclines. Pak Armed Forces Med J. 2015;65:595–599.
- 33. Alajo SO, Nakavuma J, Erume J. Cholera in endemic districts in Uganda during El Niño rains: 2002–2003. Afr Health Sci. 2006;6:93–97. pmid:16916299
- 34. Alkassoum S, Djibo I, Amadou H, Bohari A, Issoufou H, Aka J, et al. The global burden of cholera outbreaks in Niger: an analysis of the national surveillance data, 2003–2015. Trans R Soc Trop Med Hyg. 2019;113:273–280. pmid:30715523
- 35. Amadu DO, Abdullahi IN, Seibu E, Fadeyi A, Kamaldeen K, Akanbi AA, et al. Retrospective analysis of the serovars and antibiogram of Vibrio cholerae isolates of the 2017 Ilorin Cholera Outbreak. Nigeria Infect Chemother. 2021;53:368–373. pmid:34216130
- 36. Anh DD, Lopez AL, Thiem VD, Grahek SL, Duong TN, Park JK, et al. Use of oral cholera vaccines in an outbreak in Vietnam: a case control study. PLoS Negl Trop Dis. 2011;5:e1006. pmid:21283616
- 37. Baltazar CS, Langa JP, Baloi LD, Wood R, Ouedraogo I, Njanpop-Lafourcade B-M, et al. Multi-site cholera surveillance within the African Cholera Surveillance Network shows endemicity in Mozambique, 2011–2015. PLoS Negl Trop Dis. 2017;11:e0005941. pmid:28991895
- 38. Bhattacharya MK, Dutta D, Ramamurthy T, Sarkar D, Singharoy A, Bhattacharya SK. Azithromycin in the treatment of cholera in children. Acta Paediatr. 2003;92:676–678. pmid:12856976
- 39. Bhuiyan NA, Qadri F, Faruque ASG, Malek MA, Salam MA, Nato F, et al. Use of dipsticks for rapid diagnosis of cholera caused by Vibrio cholerae O1 and O139 from rectal swabs. J Clin Microbiol. 2003;41:3939–3941. pmid:12904424
- 40. Bin-Hameed EA, Joban HA. Cholera outbreak in Hadhramout, Yemen: the epidemiological weeks 2019. Int J Epidemiol Res. 2021;8:40–46.
- 41. Brazilay E, Schaad N, Magloire R. Cholera surveillance during the Haiti epidemic-the first two years. N Engl J Med. 2013;368:599–609.
- 42. Bukar AM, Goni HB, Bwala AB, Kolo FB, Isa A, Ibrahim A, et al. Determination of cholera outbreak among internally displaced persons (IDPs) in complex emergency settings within Maiduguri, Borno State-Nigeria. Int J Pure Appl Sci Res. 2020;12(2):28–36.
- 43. Bwire G, Malimbo M, Maskery B, Kim YE, Mogasale V, Levin A. The burden of cholera in Uganda. PLoS Negl Trop Dis. 2013;7:e2545. pmid:24340106
- 44. Bwire G, Waniaye JB, Otim JS, Matseketse D, Kagirita A, Orach CG. Cholera risk in cities in Uganda: understanding cases and contacts centered strategy (3CS) for rapid cholera outbreak control. Pan Afr Med J. 2021;39:193. pmid:34603574
- 45. Chibwe I, Kasambara W, Kagoli M, Milala H, Gondwe C, Azman AS. Field evaluation of Cholkit rapid diagnostic test for Vibrio cholerae O1 during a cholera outbreak in Malawi, 2018. Open Forum. Infect Dis Ther. 2020:7. pmid:33241067
- 46. Chirambo R, Mufunda J, Songolo P, Kachimba J, Vwalika B. Epidemiology of the 2016 cholera outbreak of Chibombo district, central Zambia. Med J Zambia. 2016;43:61–63.
- 47. Chowdhury G, Senapati T, Das B, Kamath A, Pal D, Bose P, et al. Laboratory evaluation of the rapid diagnostic tests for the detection of Vibrio cholerae O1 using diarrheal samples. PLoS Negl Trop Dis. 2021;15:e0009521. pmid:34129602
- 48. Das S, Gupta S. Diversity of Vibrio cholerae strains isolated in Delhi, India, during 1992–2000. J Health Popul Nutr. 2005;23:44–51. pmid:15884751
- 49. De Guzman A, de los Reyes VC, Sucaldito MN, Tayag E. Availability of safe drinking-water: the answer to cholera outbreak? Nabua, Camarines Sur, Philippines, 2012. Western Pac Surveill Response J. 2015;6:12–16. pmid:26668761
- 50. Debes AK, Ateudjieu J, Guenou E, Ebile W, Sonkoua IT, Njimbia AC, et al. Clinical and environmental surveillance for Vibrio cholerae in resource constrained areas: application during a 1-year surveillance in the Far North Region of Cameroon. Am J Trop Med Hyg. 2016;94:537–543. pmid:26755564
- 51. Dengo-Baloi LC, Semá-Baltazar CA, Manhique LV, Chitio JE, Inguane DL, Langa JP. Antibiotics resistance in El Tor Vibrio cholerae 01 isolated during cholera outbreaks in Mozambique from 2012 to 2015. PLoS ONE. 2017;12:e0181496. pmid:28792540
- 52. Djomassi L, Gessner B, Andze G, Mballa G. Cholera epidemiology in Cameroon based on national surveillance data. J Infect Dis. 2013;208:S92–S97.
- 53. Dutta BP, Kumar N, Meshram KC, Yadav R, Sodha SV, Gupta S. Cholera outbreak associated with contaminated water sources in paddy fields, Mandla District, Madhya Pradesh. India Indian J Public Health. 2021;65:S46–s50. pmid:33753592
- 54. Dzotsi EK, Dongdem AZ, Boateng G, Antwi L, Owusu-Okyere G, Nartey DB, et al. Surveillance of bacterial pathogens of diarrhoea in two selected sub metros within the Accra metropolis. Ghana Med J. 2015;49:65–71. pmid:26339088
- 55. Eurien D, Mirembe BB, Musewa A, Kisaakye E, Kwesiga B, Ogole F, et al. Cholera Outbreak Caused by Drinking Unprotected Well Water Contaminated with Feces from an Open Storm Water Drainage—Kampala City, Uganda. January 2019. 2020.
- 56. Fouda AAB, Kollo B. Epidémie de choléra à Douala en 2011 épidémiologie, clinique et bactériologie Cholera outbreak in Douala in 2011 epidemiology, clinic and bacteriology.
- 57. Franke MF, Jerome JG, Matias WR, Ternier R, Hilaire IJ, Harris JB, et al. Comparison of two control groups for estimation of oral cholera vaccine effectiveness using a case-control study design. Vaccine. 2017;35:5819–5827. pmid:28916247
- 58. Fredrick T, Ponnaiah M, Murhekar MV, Jayaraman Y, David JK, Vadivoo S, et al. Cholera outbreak linked with lack of safe water supply following a tropical cyclone in Pondicherry, India, 2012. J Health Popul Nutr. 2015;33:31–38. pmid:25995719
- 59. George CM, Rashid MU, Sack DA, Bradley Sack R, Saif-Ur-Rahman KM, Azman AS, et al. Evaluation of enrichment method for the detection of Vibrio cholerae O1 using a rapid dipstick test in Bangladesh. Tropical Med Int Health. 2014;19:301–307. pmid:24401137
- 60. Grandesso F, Kasambara W, Page AL, Debes AK, et al. Effectiveness of oral cholera vaccine in preventing cholera among fishermen in Lake Chilwa, Malawi: a case-control study. Vaccine. 2019;37:3668–3676. pmid:31133470
- 61. Guévart E, Noeske J, Sollé J, Mouangue A, Bikoti JM. Large-scale selective antibiotic prophylaxis during the 2004 cholera outbreak in Douala (Cameroon). Sante. 2007;17:63–68.
- 62. Gupta PK, Pant ND, Bhandari R, Shrestha P. Cholera outbreak caused by drug resistant Vibrio cholerae serogroup O1 biotype ElTor serotype Ogawa in Nepal; a cross-sectional study. Antimicrob Resist Infect Control. 2016;5:23. pmid:27274815
- 63. Gupta S, Jhamb U, Uppal B, Chakraverti A, Mittal SK. Diagnosing cholera in the young: a review of W.H.O. criteria. JK Sci. 2007;9:137–139.
- 64. Haque F. Cholera outbreak in Netrokona Municipality, 2013. Health Sci Bull. 2014:12.
- 65. Haque F, Hossain MJ, Kundu SK, Naser AM, Rahman M, Luby SP. Cholera outbreaks in Urban Bangladesh in 2011. Epidemiology (Sunnyvale). 2013:3. pmid:26702366
- 66. Harris JR, Cavallaro EC, De Nóbrega AA, Dos S Barrado JC, Bopp C, Parsons MB, et al. Field evaluation of Crystal VC Rapid Dipstick test for cholera during a cholera outbreak in Guinea-Bissau. Trop Med Int Health. 2009;14: 1117–1121. pmid:19624473
- 67. Im J, Islam MT, Ahmmed F, Kim DR, Chon Y, Zaman K, et al. Use of oral cholera vaccine as a vaccine probe to determine the burden of culture-negative cholera. PLoS Negl Trop Dis. 2019;13:e0007179. pmid:30870416
- 68. Ingelbeen B, Hendrickx D, Miwanda B, van der Sande MAB, Mossoko M, Vochten H, et al. Recurrent cholera outbreaks, Democratic Republic of the Congo, 2008–2017. Emerg Infect Dis. 2019;25:856–864. pmid:31002075
- 69. Islam MT, Khan AI, Sayeed MA, Amin J, Islam K, Alam N, et al. Field evaluation of a locally produced rapid diagnostic test for early detection of cholera in Bangladesh. PLoS Negl Trop Dis. 2019;13:e0007124. pmid:30703097
- 70. Issahaku GR, Asiedu-Bekoe F, Kwashie S, Broni F, Boateng P, Alomatu H, et al. Protracted cholera outbreak in the Central Region, Ghana, 2016. Ghana Med J. 2020;54:45–52. pmid:33536668
- 71. Jain A, Choudhary S, Saroha E, Bhatnagar P, Harvey P. Cholera outbreak in an informal settlement at Shahpur huts, Panchkula District, Haryana State, India, 2019. Indian J Public Health. 2021;65:S51–s54. pmid:33753593
- 72. Jeandron A, Cumming O, Rumedeka BB, Saidi JM, Cousens S. Confirmation of cholera by rapid diagnostic test amongst patients admitted to the cholera treatment centre in Uvira, Democratic Republic of the Congo. PLoS ONE. 2018;13:e0201306. pmid:30067824
- 73. Jones FK, Wamala JF, Rumunu J, Mawien PN, Kol MT, Wohl S, et al. Successive epidemic waves of cholera in South Sudan between 2014 and 2017: a descriptive epidemiological study. Lancet Planet Health. 2020;4:e577–e587. pmid:33278375
- 74. Khatib A, Ali M, von Seidlein L, Kim D, Hashim R, Reyburn R. Direct and indirect effectiveness of an oral cholera vaccine in Zanzibar, East Africa: findings from a large mass vaccination campaign followed by an observational cohort study. Lancet Infect Dis. 2012;12:837–844.
- 75. Khazaei HA, Rezaei N, Bagheri GR, Moin AA. A six-year study on Vibrio cholerae in southeastern Iran. Jpn J Infect Dis. 2005;58:8–10. pmid:15728982
- 76. Kisera N, Luxemburger C, Tornieporth N, Otieno G, Inda J. A descriptive cross-sectional study of cholera at Kakuma and Kalobeyei refugee camps, Kenya in 2018. Pan Afr Med J. 2020;37:197. pmid:33505566
- 77. Koley H, Ray N, Chowdhury G, Barman S, Mitra S, Ramamurthy T, et al. Outbreak of cholera caused by Vibrio cholerae O1 El Tor variant strain in Bihar, India. Jpn J Infect Dis. 2014;67:221–226. pmid:24858614
- 78. Kulkarni S, Chillarge C. Antibiotic susceptibility pattern of Vibrio cholerae causing diarrohea outbreaks in Bidar, North Karnataka, India. Int J Curr Microbiol App Sci. 2015;4:957–961.
- 79. Kuttiat VS, Lodha R, Das B, Kohli U. Prevalence of cholera in pediatric patients with acute dehydrating diarrhea. Indian J Pediatr. 2010;77:67–71. pmid:20135270
- 80. Kwesiga B, Pande G, Ario AR, Tumwesigye NM, Matovu JK, Zhu B-P. A prolonged, community-wide cholera outbreak associated with drinking water contaminated by sewage in Kasese District, western Uganda. BMC Public Health. 2018;18:1–8.
- 81. Landoh DE, Gessner BD, Badziklou K, Tamekloe T, Nassoury DI, Dagnra A, et al. National surveillance data on the epidemiology of cholera in Togo. J Infect Dis. 2013;208(Suppl 1):S115–S119. pmid:24101639
- 82. Lenglet A, Khamphaphongphane B, Thebvongsa P, Vongprachanh P, Sithivong N, Chantavisouk C, et al. A cholera epidemic in Sekong Province, Lao People’s Democratic Republic, December 2007-January 2008. Jpn J Infect Dis. 2010;63:204–207. pmid:20495276
- 83. Ley B, Khatib AM, Thriemer K, von Seidlein L, Deen J, Mukhopadyay A, et al. Evaluation of a rapid dipstick (Crystal VC) for the diagnosis of cholera in Zanzibar and a comparison with previous studies. PLoS ONE. 2012;7:e36930. pmid:22662131
- 84. Llanes R, Lazo A, Somarriba L, Mas P. Sentinel surveillance detects low circulation of Vibrio cholerae serotype Inaba in Haiti, 2011–2012. MEDICC Rev. 2015;17:43–46. pmid:26947158
- 85. Luquero FJ, Grout L, Ciglenecki I, Sakoba K, Traore B, Heile M, et al. Use of Vibrio cholerae vaccine in an outbreak in Guinea. N Engl J Med. 2014;370:2111–2120. pmid:24869721
- 86. Mahamud AS, Ahmed JA, Nyoka R, Auko E, Kahi V, Ndirangu J, et al. Epidemic cholera in Kakuma Refugee Camp, Kenya, 2009: the importance of sanitation and soap. J Infect Dev Ctries. 2012;6:234–241. pmid:22421604
- 87. Matias WR, Cademil A, Julceus FE, Mayo-Smith LM, Franke MF, Harris JB, et al. Laboratory evaluation of immunochromatographic rapid diagnostic tests for cholera in Haiti. Am J Trop Med Hyg. 2015;93:569–569.
- 88. Mbala-Kingebeni P, Vogt F, Miwanda B, Sundika T, Mbula N, Pankwa I, et al. Sachet water consumption as a risk factor for cholera in urban settings: findings and implications from a case control study in Kinshasa, Democratic Republic of the Congo during the 2017–2018 outbreak. PLoS Negl Trop Dis. 2021;15:e0009477. pmid:34237058
- 89. Michel E, Gaudart J, Beaulieu S, Bulit G, Piarroux M, Boncy J, et al. Estimating effectiveness of case-area targeted response interventions against cholera in Haiti. elife. 2019:8. pmid:31886768
- 90. Mishra A, Taneja N, Sharma M. Environmental and epidemiological surveillance of Vibrio cholerae in a cholera-endemic region in India with freshwater environs. J Appl Microbiol. 2012;112:225–237. pmid:22054549
- 91. Monje F, Ario AR, Musewa A, Bainomugisha K, Mirembe BB, Aliddeki DM, et al. A prolonged cholera outbreak caused by drinking contaminated stream water, Kyangwali refugee settlement, Hoima District, Western Uganda: 2018. Infect Dis Poverty. 2020;9:154. pmid:33148338
- 92. Mugoya I, Kariuki S, Galgalo T, Njuguna C, Omollo J, Njoroge J, et al. Rapid spread of Vibrio cholerae O1 throughout Kenya, 2005. Am J Trop Med Hyg. 2008;78:527–533. pmid:18337355
- 93. Mukherjee P, Ghosh S, Ramamurthy T, Bhattacharya MK, Nandy RK, Takeda Y, et al. Evaluation of a rapid immunochromatographic dipstick kit for diagnosis of cholera emphasizes its outbreak utility. Jpn J Infect Dis. 2010;63:234–238. pmid:20657061
- 94. Mwenda V, Niyomwungere A, Oyugi E, Githuku J, Obonyo M, Gura Z. Factors associated with cholera outbreaks, Nairobi County, July 2017: a case control study. bioRxiv. 2019;719641.
- 95. Ndugwa Kabwama S, Riolexus Ario A, Guwatudde D. Cholera outbreak caused by drinking lakeshore water contaminated by feces washed down from a hill-side residential area: Kaiso Village, Uganda. Pan Afr Med J Conference Proceedings. 2017.
- 96. Noora CL, Issah K, Kenu E, Bachan EG, Nuoh RD, Nyarko KM, et al. Large cholera outbreak in Brong Ahafo Region. Ghana BMC Res Notes. 2017;10:389. pmid:28797285
- 97. Nsubuga F, Garang SC, Tut M, Oguttu D, Lubajo R, Lodiongo D, et al. Epidemiological description of a protracted cholera outbreak in Tonj East and Tonj North counties, former Warrap State, South Sudan, May-Oct 2017. BMC Infect Dis. 2019;19:4. pmid:30606126
- 98. Okello PE, Bulage L, Riolexus AA, Kadobera D, Kwesiga B, Kajumbula H, et al. A cholera outbreak caused by drinking contaminated river water, Bulambuli District, Eastern Uganda, March 2016. BMC Infect Dis. 2019;19:516. pmid:31185939
- 99. Page AL, Alberti KP, Mondonge V, Rauzier J, Quilici ML, Guerin PJ. Evaluation of a rapid test for the diagnosis of cholera in the absence of a gold standard. PLoS ONE. 2012;7:e37360. pmid:22666350
- 100. Pal BB, Khuntia HK, Samal SK, Kerketta AS, Kar SK, Karmakar M, et al. Large outbreak of cholera caused by El Tor variant Vibrio cholerae O1 in the eastern coast of Odisha, India during 2009. Epidemiol Infect. 2013;141:2560–2567. pmid:23461927
- 101. Pal BB, Khuntia HK, Samal SK, Das SS, Chhotray GP. Emergence of Vibrio cholerae O1 biotype E1 Tor serotype Inaba causing outbreaks of cholera in Orissa, India. Jpn J Infect Dis. 2006;59:266.
- 102. Pande G, Kwesiga B, Bwire G, Kalyebi P, Riolexus A, Matovu JKB, et al. Cholera outbreak caused by drinking contaminated water from a lakeshore water-collection site, Kasese District, south-western Uganda, June-July 2015. PLoS ONE. 2018;13:e0198431. pmid:29949592
- 103. Phukan AC, Borah PK, Biswas D, Mahanta J. A cholera epidemic in a rural area of northeast India. Trans R Soc Trop Med Hyg. 2004;98:563–566. pmid:15251406
- 104. Ramazanzadeh R, Rouhi S, Shakib P, Shahbazi B, Bidarpour F, Karimi M. Molecular characterization of Vibrio cholerae isolated from clinical samples in Kurdistan Province, Iran. Jundishapur J Microbiol. 2015;8:e18119. pmid:26060565
- 105. Rosewell A, Addy B, Komnapi L, Makanda F, Ropa B, Posanai E, et al. Cholera risk factors, Papua New Guinea, 2010. BMC Infect Dis. 2012;12:287. pmid:23126504
- 106. Roskosky M, Acharya B, Shakya G, Karki K, Sekine K, Bajracharya D, et al. Feasibility of a comprehensive targeted cholera intervention in the Kathmandu Valley. Nepal Am J Trop Med Hyg. 2019;100:1088–1097. pmid:30887946
- 107. Roy S, Parande MV, Mantur BG, Bhat S, Shinde R, Parande AM, et al. Multidrug-resistant Vibrio cholerae O1 in Belgaum, south India. J Med Microbiol. 2012;61:1574–1579. pmid:22878247
- 108. Sack RB, Siddique AK, Longini IM Jr, Nizam A, Yunus M, Islam MS, et al. A 4-year study of the epidemiology of Vibrio cholerae in four rural areas of Bangladesh. J Infect Dis. 2003;187:96–101. pmid:12508151
- 109. Saha R, Das S, Waghmare M, Ramachandran VG. Paradoxical reduction in prevalence of vibrio cholerae in its niche environment. Int J Pharm Bio Sci. 2013;4:B1099–B1107.
- 110. Sévère K, Rouzier V, Anglade SB, Bertil C, Joseph P, Deroncelay A, et al. Effectiveness of oral cholera vaccine in Haiti: 37-month follow-up. Am J Trop Med Hyg. 2016;94:1136–1142. pmid:26928838
- 111. Shah WA, Shahina M, Ali N. First report of Vibrio cholerae infection from Andaman and Nicobar, India. J Commun Dis. 2002;34:270–275. pmid:14710858
- 112. Sharma A, Dutta BS, Rasul ES, Barkataki D, Saikia A, Hazarika NK. Prevalence of Vibrio cholerae O1 serogroup in Assam, India: A hospital-based study. Indian J Med Res. 2017;146:401–408. pmid:29355149
- 113. Shikanga OT, Mutonga D, Abade M, Amwayi S, Ope M, Limo H, et al. High mortality in a cholera outbreak in western Kenya after post-election violence in 2008. Am J Trop Med Hyg. 2009;81:1085–1090. pmid:19996441
- 114. Siddiqui FJ, Bhutto NS, von Seidlein L, Khurram I, Rasool S, Ali M, et al. Consecutive outbreaks of Vibrio cholerae O139 and V. cholerae O1 cholera in a fishing village near Karachi, Pakistan. Trans R Soc Trop Med Hyg. 2006;100:476–482. pmid:16443247
- 115. Sinha A, Sengupta S, Ghosh S, Basu S, Sur D, Kanungo S, et al. Evaluation of a rapid dipstick test for identifying cholera cases during the outbreak. Indian J Med Res. 2012;135:523–528. pmid:22664501
- 116. Sreedhara H, Mohan N. Molecular epidemiology of vibrio cholerae causing outbreaks and sporadic cholera in and around Hassan district and its antibiotic susceptibility pattern. IP Int J Med Microbiol Trop Dis. 2019;5:41–46.
- 117. Sugunan AP, Ghosh AR, Roy S, Gupte MD, Sehgal SC. A cholera epidemic among the Nicobarese tribe of Nancowry, Andaman, and Nicobar, India. Am J Trop Med Hyg. 2004;71:822–827. pmid:15642977
- 118. Sur D, Deen JL, Manna B, Niyogi SK, Deb AK, Kanungo S, et al. The burden of cholera in the slums of Kolkata, India: data from a prospective, community based study. Arch Dis Child. 2005;90:1175–1181. pmid:15964861
- 119. Sur D, Sarkar BL, Manna B, Deen J, Datta S, Niyogi SK, et al. Epidemiological, microbiological & electron microscopic study of a cholera outbreak in a Kolkata slum community. Indian J Med Res. 2006;123:31–36.
- 120. Tamang M, Sharma N, Makaju R, Sarma A, Koju R, Nepali N, et al. An outbreak of El Tor cholera in Kavre district, Nepal. Kathmandu Univ Med J (KUMJ). 2005;3:138–142. pmid:16415609
- 121. Taneja N, Kaur J, Sharma K, Singh M, Kalra JK, Sharma NM, et al. A recent outbreak of cholera due to Vibrio cholerae O1 Ogawa in & around Chandigarh, North India. Indian J Med Res. 2003;117:243–246.
- 122. Thiem VD, Deen JL, von Seidlein L, Canh DG, Anh DD, Park JK, et al. Long-term effectiveness against cholera of oral killed whole-cell vaccine produced in Vietnam. Vaccine. 2006;24:4297–4303. pmid:16580760
- 123. Torane V, Kuyare S, Nataraj G, Mehta P, Dutta S, Sarkar B. Phenotypic and antibiogram pattern of V. cholerae isolates from a tertiary care hospital in Mumbai during 2004–2013: a retrospective cross-sectional study. BMJ Open. 2016;6:e012638. pmid:27888174
- 124. Tripurari K, Deepak B, Kaur TA, Pushpa VV, Aakash S, Prakash NJ, et al. Vibrio cholerae outbreak in Batala town, Punjab, India 2012. J Commun Dis. 2017;49:35–40.
- 125. Uthappa CK, Allam RR, Nalini C, Gunti D, Udaragudi PR, Tadi GP, et al. An outbreak of cholera in Medipally village, Andhra Pradesh, India, 2013. J Health Popul Nutr. 2015;33:7. pmid:26825056
- 126. Von Nguyen D, Sreenivasan N, Lam E, Ayers T, Kargbo D, Dafae F, et al. Cholera epidemic associated with consumption of unsafe drinking water and street-vended water-Eastern Freetown, Sierra Leone, 2012. Am J Trop Med Hyg. 2014;90:518–523. pmid:24470563
- 127. Wang XY, Ansaruzzaman M, Vaz R, Mondlane C, Lucas ME, von Seidlein L, et al. Field evaluation of a rapid immunochromatographic dipstick test for the diagnosis of cholera in a high-risk population. BMC Infect Dis. 2006;6:17. pmid:16451731
- 128. Wierzba TF, Kar SK, Mogasale VV, Kerketta AS, You YA, Baral P, et al. Effectiveness of an oral cholera vaccine campaign to prevent clinically-significant cholera in Odisha State, India. Vaccine. 2015;33:2463–2469. pmid:25850019
- 129. Zachariah R, Harries AD, Arendt V, Nchingula D, Chimtulo F, Courteille O, et al. Characteristics of a cholera outbreak, patterns of Vibrio cholerae and antibiotic susceptibility testing in rural Malawi. Trans R Soc Trop Med Hyg. 2002;96:39–40. pmid:11925988
- 130. Zereen F, Akter S, Sobur MA, Hossain MT, Rahman MT. Molecular detection of Vibrio cholerae from human stool collected from SK Hospital, Mymensingh, and their antibiogram. J Adv Vet Anim Res. 2019;6:451–455. pmid:31819871
- 131. Zgheir SM, Mustafa NM, Ali AA, Al-Diwan J. Cholera outbreak in Iraq, 2017. Indian J Public Health Res Dev. 2019;10:686.
- 132. Azman AS, Parker LA, Rumunu J, Tadesse F, Grandesso F, Deng LL, et al. Effectiveness of one dose of oral cholera vaccine in response to an outbreak: a case-cohort study. Lancet Glob Health. 2016;4:e856–e863. pmid:27765293
- 133. Blake A, Keita VS, Sauvageot D, Saliou M, Njanpop BM, Sory F, et al. Temporo-spatial dynamics and behavioural patterns of 2012 cholera epidemic in the African mega-city of Conakry, Guinea. Infect Dis Poverty. 2018;7:13. pmid:29448965
- 134. Boncy J, Rossignol E, Dahourou G, Hast M, Buteau J, Stanislas M, et al. Performance and utility of a rapid diagnostic test for cholera: notes from Haiti. Diagn Microbiol Infect Dis. 2013;76:521–523. pmid:23886437
- 135. Bwire G, Orach CG, Abdallah D, Debes AK, Kagirita A, Ram M, et al. Alkaline peptone water enrichment with a dipstick test to quickly detect and monitor cholera outbreaks. BMC Infect Dis. 2017;17:726. pmid:29157211
- 136. Ferreras E, Blake A, Chewe O, Mwaba J, Zulu G, Poncin M, et al. Alternative observational designs to estimate the effectiveness of one dose of oral cholera vaccine in Lusaka, Zambia. Epidemiol Infect. 2020;148:e78. pmid:32167038
- 137. Franke MF, Ternier R, Jerome JG, Matias WR, Harris JB, Ivers LC. Long-term effectiveness of one and two doses of a killed, bivalent, whole-cell oral cholera vaccine in Haiti: an extended case-control study. Lancet Glob Health. 2018;6:e1028–e1035. pmid:30103980
- 138. George CM, Monira S, Sack DA, Rashid M, Saif-Ur-Rahman KM, Mahmud T, et al. Randomized controlled trial of hospital-Bbased hygiene and water treatment intervention (CHoBI7) to reduce cholera. Emerg Infect Dis. 2016;22:233–241. pmid:26811968
- 139. Ivers LC, Hilaire IJ, Teng JE, Almazor CP, Jerome JG, Ternier R, et al. Effectiveness of reactive oral cholera vaccination in rural Haiti: a case-control study and bias-indicator analysis. Lancet Glob Health. 2015;3:e162–e168. pmid:25701994
- 140. Lucas MES, Deen JL, von Seidlein L, Wang X-Y, Ampuero J, Puri M, et al. Effectiveness of mass oral cholera vaccination in Beira, Mozambique. N Engl J Med. 2005;352:757–767. pmid:15728808
- 141. Roy S, Dutta B, Ghosh AR, Sugunan AP, Nandy RK, Bhattacharya SK, et al. Molecular tracking of the lineage of strains of Vibrio cholerae O1 biotype El Tor associated with a cholera outbreak in Andaman and Nicobar Islands, India. Trop Med Int Health. 2005;10:604–611. pmid:15941425
- 142. Guillaume Y, Debela M, Slater D, Vissieres K, Ternier R, Franke M, et al. Poor sensitivity of stool culture compared to PCR in surveillance for V. cholerae in Haiti, 2018–2019. Open Forum. Infect Dis Ther. 2023:ofad301. pmid:37383250
- 143. Chowdhury F, Khan IA, Patel S, Siddiq AU, Saha NC, Khan AI, et al. Diarrheal illness and healthcare seeking behavior among a population at high risk for diarrhea in Dhaka, Bangladesh. PLoS ONE. 2015;10:e0130105. pmid:26121650
- 144. Fissehaye T, Damte A, Fantahun A, Gebrekirstos K. Health care seeking behaviour of mothers towards diarrheal disease of children less than 5 years in Mekelle city, North Ethiopia. BMC Res Notes. 2018;11:749. pmid:30348211
- 145. Luquero FJ, Rondy M, Boncy J, Munger A, Mekaoui H, Rymshaw E, et al. Mortality rates during cholera epidemic, Haiti, 2010–2011. Emerg Infect Dis. 2016:22. pmid:26886511
- 146. McCrickard LS, Massay AE, Narra R, Mghamba J, Mohamed AA, Kishimba RS, et al. Cholera mortality during urban epidemic, Dar es Salaam, Tanzania, August 16, 2015–January 16, 2016. Emerg Infect Dis. 2017;23. pmid:29155665
- 147. Watts V. Confidence intervals for a population proportion. 2022 [cited 2023 Jul 10]. Available from: https://ecampusontario.pressbooks.pub/introstats/chapter/7-4-confidence-intervals-for-a-population-proportion/