Prediction of pertussis vaccine efficacy using a correlates of protection model

Abstract

A previously developed statistical model relates vaccine immune responses to protection against pertussis disease in a household contact setting. Before this model can be used to predict the risk of disease based on immune responses, it must be validated to demonstrate reliable predictions. The model is shown here to be validated in terms of statistical criteria (Prentice surrogacy measures) as well as predictive capability in an independent efficacy trial (meta-analysis). Additionally, the model is used to predict efficacy from two recent immunogenicity trials comparing a 5-component acellular pertussis pentavalent combination vaccine with separate administration of its vaccine components. The model predicted similar protective efficacy rates: 82% after three doses and 83% after four doses. Follow-up of these subjects the model also predicted sustained and comparable efficacy for the combination and separate vaccines (75%) up to the pre-school booster age.