Abstract
BACKGROUND: Demonstrating the efficacy of new Vi-conjugate typhoid vaccines is challenging, due to the cost of field trials requiring tens of thousands of participants. New trial designs that use serologically defined typhoid infections (seroefficacy trials) rather than blood culture positivity as a study endpoint may be useful to assess efficacy using small trials. METHODS: We developed a model for Vi-immunoglobin G antibody responses to a Vi-vaccine, incorporating decay over time and natural boosting due to endemic exposures. From this, we simulated clinical trials in which 2 blood samples were taken during follow-up and the relative risk of a serologically defined typhoid infection (seroefficacy) was computed. We aimed to determine (1) whether seroefficacy trial designs could substantially reduce sample sizes, compared with trials using blood culture-confirmed cases; (3) whether the rate of case detection was higher in seroefficacy trials; and (3) the optimal timing of sample collection. RESULTS: The majority (>90%) of blood culture-positive typhoid cases remain unobserved in surveillance studies. In contrast, under-detection in simulated seroefficacy trials of equivalent vaccines was as little as 26%, and estimates of the relative risk of typhoid infection were unbiased. For simulated trials of non-equivalent vaccines, relative risks were slightly inflated by at least 5%, depending on the sample collection times. Seroefficacy trials required as few as 460 participants per arm, compared with 10 000 per arm for trials using blood culture-confirmed cases. CONCLUSIONS: Seroefficacy trials can establish the efficacy of new conjugate vaccines using small trials that enroll hundreds rather than thousands of participants, and without the need for resource-intensive typhoid fever surveillance programs.