Protection against cholera from killed whole-cell oral cholera vaccines: a systematic review and meta-analysis

Abstract

Background: Killed whole-cell oral cholera vaccines (kOCVs) are becoming a standard cholera control and prevention tool. However, vaccine efficacy and direct effectiveness estimates have varied, with differences in study design, location, follow-up duration, and vaccine composition posing challenges for public health decision making. We did a systematic review and meta-analysis to generate average estimates of kOCV efficacy and direct effectiveness from the available literature.

Methods: For this systematic review and meta-analysis, we searched PubMed, Embase, Scopus, and the Cochrane Review Library on July 9, 2016, and ISI Web of Science on July 11, 2016, for randomised controlled trials and observational studies that reported estimates of direct protection against medically attended confirmed cholera conferred by kOCVs. We included studies published on any date in English, Spanish, French, or Chinese. We extracted from the published reports the primary efficacy and effectiveness estimates from each study and also estimates according to number of vaccine doses, duration, and age group. The main study outcome was average efficacy and direct effectiveness of two kOCV doses, which we estimated with random-effect models. This study is registered with PROSPERO, number CRD42016048232.

Findings: Seven trials (with 695 patients with cholera) and six observational studies (217 patients with cholera) met the inclusion criteria, with an average two-dose efficacy of 58% (95% CI 42-69, I2=58%) and effectiveness of 76% (62-85, I2=0). Average two-dose efficacy in children younger than 5 years (30% [95% CI 15-42], I2=0%) was lower than in those 5 years or older (64% [58-70], I2=0%; p<0·0001). Two-dose efficacy estimates of kOCV were similar during the first 2 years after vaccination, with estimates of 56% (95% CI 42-66, I2=45%) in the first year and 59% (49-67, I2=0) in the second year. The efficacy reduced to 39% (13 to 57, I2=48%) in the third year, and 26% (-46 to 63, I2=74%) in the fourth year.

Interpretation: Two kOCV doses provide protection against cholera for at least 3 years. One kOCV dose provides at least short-term protection, which has important implications for outbreak management. kOCVs are effective tools for cholera control.

Funding: The Bill & Melinda Gates Foundation.

Copyright This is an Open Access article published under the CC BY 3.0 IGO license which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. In any use of this article, there should be no suggestion that WHO endorses any specific organisation, products, or services. The use of the WHO logo is not permitted. This notice should be preserved along with the article's original URL.

Figures

Figure 1

PRISMA flow chart highlighting details of the systematic review and data abstraction process Of the four studies that did not meet the study design inclusion criteria, one was excluded because it included medically attended cases and those detected from active surveillance in a clinical trial, one was excluded because of a non-standard study design, the others, only used suspected, not confirmed cholera, as the study endpoint. The 359 other studies related to cholera vaccine that did not match eligibility criteria included commentaries, reviews, policy pieces, computational modelling, and studies with non-cholera outcomes related to any cholera vaccine.

Figure 2

Effectiveness and efficacy main pooled analyses for two-dose killed whole-cell oral cholera vaccine 95% CIs in the figure and used for meta-analyses are not necessarily the same as those in the original study because they were based on a reconstruction of a two-sided 95% CI from estimates of the standard error of the estimate from each study. All estimates (except for those from Clemens and colleagues, 1990a and 1990b) use the main vaccine dose used in the study. Qadri and colleagues' estimate is an estimate of total protection including both direct and indirect effects. Observational study effectiveness estimates had an 18-month weighted mean duration, and the trial efficacy estimates had a 28-month weighted mean duration.

Figure 3

Efficacy and effectiveness by time since vaccination and dose Follow-up durations are shown here as the midpoint of the time window during which the estimate was measured. (A) Two-dose efficacy estimates at 0–12 months (from left to right, data from , , , , [WC-BS group], [whole-cell group], and 27), 12–24 months (, [WC-BS], [whole-cell], and 27), 24–36 months (, [WC-BS], and [whole-cell]), 36–48 months (, [WC-BS], [whole-cell]), and 48–60 months (reference 6). The study by van Loon and colleagues has two estimates at each relevant timepoint, because results were obtained for a whole-cell vaccine group and a whole-cell with B-subunit group. (B) One-dose efficacy estimate at 0–12 months. (C) Two-dose effectiveness estimates at 0–12 months (from left to right, data from 7, 34, 35, 8), 12–24 months, and 24–36 months. (D) One-dose effectiveness estimates at 0–12 months (from left to right, data from 7, 37, 34, 8) and 24–36 months. Estimates are grouped by timeframe of analysis with zero representing the day of last dose of vaccine (dose dependent). Grey bars and squares show 95% CIs and point estimates of efficacy or effectiveness from the literature. Black diamonds show the average efficacy or effectiveness and 95% CI. Studies were grouped by time period and not all studies cover the entire 12-month period. kOCV=killed whole-cell oral cholera vaccine. WC-BS=whole-cell with B subunit.

Figure 4

Relationship between protection by vaccine, duration of follow-up, and median age of patients with cholera (A) Relationship between protection and months of follow-up. Shading shows median age of patients with cholera. (B) Relationship between protection and the median age of patients with cholera measured in years. Shading shows duration of follow-up. Lines in black were fit with a polynomial spline with three degrees of freedom. These plots only include a subset of data in which the age distributions of patients were available (all observational studies and two trials)., , , , , , Error bars show 95% CI.