Effectiveness of an oral cholera vaccine in Zanzibar: findings from a mass vaccination campaign and observational cohort study.

Background Zanzibar, in east Africa, has been severely and repeatedly affected by cholera since 1978. We assessed the effectiveness of oral cholera vaccination in high-risk populations in the archipelago to estimate the indirect (herd) protection conferred by the vaccine and direct vaccine effectiveness. Methods We offered two doses of a killed whole-cell B-subunit cholera vaccine to individuals aged 2 years and older in six rural and urban sites. To estimate vaccine direct protection, we compared the incidence of cholera between recipients and non-recipients using generalised estimating equations with the log link function while controlling for potential confounding variables. To estimate indirect effects, we used a geographic information systems approach and assessed the association between neighbourhood-level vaccine coverage and the risk for cholera in the non-vaccinated residents of that neighbourhood, after controlling for potential confounding variables. This study is registered with ClinicalTrials.gov, number NCT00709410. Findings Of 48 178 individuals eligible to receive the vaccine, 23 921 (50%) received two doses. Between February, 2009, and May, 2010, there was an outbreak of cholera, enabling us to assess vaccine effectiveness. The vaccine conferred 79% (95% CI 47-92) direct protection against cholera in participants who received two doses. Indirect (herd) protection was shown by a decrease in the risk for cholera of non-vaccinated residents within a household's neighbourhood as the vaccine coverage in that neighbourhood increased. Interpretation Our findings suggest that the oral cholera vaccine offers both direct and indirect (herd) protection in a sub-Saharan African setting. Mass oral cholera immunisation campaigns have the potential to provide not only protection for vaccinated individuals but also for the unvaccinated members of the community and should be strongly considered for wider use. Because this is an internationally-licensed vaccine, we could not undertake a randomised placebo-controlled trial, but the absence of vaccine effectiveness against non-cholera diarrhoea indicates that the noted protection against cholera could not be explained by bias.