Defining long-term drivers of pertussis resurgence, and optimal vaccine control strategies.

Pertussis resurgence has been reported from several developed countries with long-standing immunisation programs. Among these, Australia in 2003 discontinued an 18 months (fourth) booster dose in favour of an adolescent (fifth) dose. We developed a model to evaluate determinants of resurgence in Australia and alternative vaccine strategies for mitigation. Novel characteristics of our model included the use of seroepidemiologic data for calibration, and broad investigation of variables relevant to transmission of, and protection against, pertussis. We simulated multiple parameter combinations, retaining those consistent with observed data for subsequent use in predictive models comparing alternative vaccination schedules. Reproducing the early control of pertussis followed by late resurgence observed in Australia required natural immunity to last decades longer than vaccine-acquired immunity, with mean duration exceeding 50 years in almost 90% of simulations. Replacement of the dose at 18 months with an adolescent dose in 2003 resulted in a 40% increase in infections in the age group 18-47 months by 2013. A six dose strategy (2, 4, 6, 18 months, 4 and 15 years) yielded a reduction in infection incidence (pre-school 43%, infants 8%) greater than any alternative strategies considered for timing of five administered doses. Our finding that natural immunity drives long-term trends in pertussis cycles is relevant to a range of pertussis strategies and provides the necessary context in which to consider maternal vaccination. Comparatively short-lived vaccine-acquired immunity requires multiple boosters over the first two decades of life to maximise reduction in infections.