Estimating the Effectiveness of Acellular Pertussis Vaccines

TO THE EDITOR—In the 15 June 2012 issue of Clinical Infectious Diseases, Witt et al reported that vaccine effectiveness (VE) of acellular pertussis vaccines was 41% for children aged 2–7 years, 24% for 8to 12-year-olds, and 79% for 13to 18-year-olds [1]. The authors conclude that their data “confirms markedly lower than expected protection afforded by the pre-school series of acellular pertussis vaccinations in the 8–12 year age group” and poor durability of protection from the vaccine. While we agree that protection wanes over time, there are important limitations to the Witt et al analysis. VE estimates are predicated on comparing disease risk in a vaccinated group with disease risk in an unvaccinated group. When calculating VE for multiple dose vaccines, partially vaccinated persons should not be grouped with unvaccinated individuals [2, 3]. Doing so compromises the vaccine-naive comparison group and lowers VE estimates. Although the authors do not specify how undervaccinated individuals were handled in the analyses, the results presented in their Tables 1 and 2 indicate that undervaccinated and unvaccinated individuals were inappropriately categorized together, rather than excluding undervaccinated persons from the analysis. This bias explains the surprisingly low estimates in the 2 younger and combined overall age groups. Second, Witt et al state that VE was estimated using the screening method, often used when precise attack rates cannot be calculated for vaccinated and unvaccinated individuals. However, the authors describe a cohort study and report attack rates, as well as discuss vaccination rates among cases and age-