In Reply Dr Kreimer and colleagues point out that there appears to be little difference in the risk of condyloma with different numbers of doses of HPV vaccine if a 12-month buffer period is used. We appreciate the thoroughness with which they studied the supplementary material and regret that space limitations did not allow a more detailed explanation of the sensitivity analysis in question.
In short, the buffer period and delayed case counting were introduced to account for prevalent cases diagnosed shortly after vaccination, which would otherwise bias the estimated risk of condyloma upward. In Figure 1 in the article, which shows cumulative incidence of condyloma after the first dose compared with incidence in an unvaccinated cohort, we found a divergence of the cumulative incidence curves between 90 and 150 days (3-5 months), suggesting that the bias introduced by prevalent infections waned at this point.
Ploner A, Herweijer E, Arnheim-Dahlström L. Number of Human Papillomavirus Vaccine Doses and Condyloma—Reply. JAMA. 2014;311(23):2439-2440. doi:10.1001/jama.2014.5012