Simulation links Shigella trial design to underestimated vaccine efficacy

A simulation-based analysis found that vaccine efficacy estimates in planned Shigella trials may be biased downward when naturally acquired immunity develops rapidly in high-burden communities. Researchers assessed six trial and analysis approaches across low-, medium- and high-burden settings, finding that only a design combining active surveillance for Shigella infection with a single-outcome model was consistently unbiased under all scenarios examined. The other approaches underestimated efficacy against both severe, Shigella-attributable disease and any Shigella-attributable diarrhea by as much as 20 percentage points.

The simulations examined how the timing and extent of infection-acquired immunity could affect the apparent performance of Shigella vaccines. In high-incidence settings, repeated exposure may allow unvaccinated infants to accumulate natural immunity, potentially bringing their immune status closer to that of vaccinated infants. A trial conducted after this convergence could therefore record a lower vaccine efficacy than the vaccine would achieve in a fully immunologically naïve population. The magnitude of the effect depended on the trial design and on assumptions about the natural history of Shigella, including how quickly immunity develops and how long it persists.

Researchers modeled trials in settings with low, medium and high disease burdens. For each setting, they evaluated six combinations of surveillance and statistical analysis: active surveillance for Shigella infection or symptom-based reporting, each paired with a single-outcome model, a stratified recurrent-outcome model or a crude recurrent-outcome Cox proportional hazards model. The analysis measured bias in vaccine-efficacy estimates, the coverage of nominal 95% confidence intervals and the false-negative rate. Active infection surveillance with a single-outcome model was the only approach that remained unbiased throughout the scenarios tested. Among the other five approaches, the single-outcome model paired with symptom-based reporting performed least poorly.

Bias generally increased in higher-burden settings, where the accumulation of natural immunity was expected to be more pronounced. The results for realistically sized trials were also less favorable at higher-burden sites: confidence-interval coverage was worse, and the false-negative rate was higher. These findings identify a specific measurement challenge for late-stage Shigella vaccine trials rather than evidence that the vaccines themselves are ineffective. They also do not establish the size of the discrepancy in an actual trial, because the estimates are conditional on simulated inputs and limited published information about Shigella immunity and natural history. The study’s conclusion is that observed efficacy is likely to understate true efficacy in some circumstances, while careful design is needed to reduce that bias.

Subject of Research: Simulation analysis of how natural immunity and trial design affect estimates of Shigella vaccine efficacy

Article Title: Analysis of the impact of naturally-acquired immunity on Shigella vaccine performance to inform clinical trials: A simulation-based modeling study

Article References: Amin, A. B., Garcia Quesada, M., Trudell, D., Dean, N. E., Lopman, B. A., Platts-Mills, J. A., & Rogawski McQuade, E. T. (2026). Analysis of the impact of naturally-acquired immunity on Shigella vaccine performance to inform clinical trials: A simulation-based modeling study. PLOS Medicine, 23(10), e1004878. https://doi.org/10.1371/journal.pmed.1004878

Image Credits: AI Generated

DOI: 10.1371/journal.pmed.1004878

Keywords: Shigella, vaccine efficacy, natural immunity, clinical trials, epidemiology, simulation modeling, active surveillance, diarrheal disease, statistical bias, Analysis, impact, naturally-acquired