Background: In previous calculations of qHPV vaccine efficacy, a conservative approach was taken, whereby observed incident cases were attributed to any vaccine type detected in a lesion, including those with multiple HPV types present. Here we apply an alternate method to evaluation of vaccine efficacy in lesions with multiple HPV types present.

Methods: 598 MSM aged 16-26 years with <=5 lifetime sex partners were randomized to vaccine or placebo at Day 0 and Months 2 and 6. AIN cases in which >1 HPV type was found were assessed for evidence of infection with a non-vaccine HR HPV type by evaluation of HPV types in swabs from the 2 immediately preceding visits. If an HPV type was detected on at least one of the two visits as well as in the lesion, that type was considered to be causal. If not, the lesion was attributed to the vaccine HPV type present.

Results: Of the 5 and 24 vaccine or placebo group subjects with 6/11/16/18-related AIN cases, 3 and 7, respectively, had lesions with multiple HPV types. Of these, case assignment methodology led to the re-assignment of three subjects in the qHPV group and none in the placebo group, resulting in a 2/24 vaccine/placebo case split and efficacy of 91.1% (95% CI: 64.2%, 99.0%), compared to 77.5% efficacy in the original analysis. The vaccine/placebo case split for HPV 6/11/16/18-related high-grade AIN is now 1/13 (compared to 3/13 in the original analysis), resulting in efficacy of 91.7% (95% CI: 44.6, 99.8%).

Conclusions: Using a case assignment methodology suggests that efficacy against vaccine-type related AIN may be higher than indicated by the protocol-specified analyses.