Correlation between pill counts and biologic effects in an HIV-1 prevention clinical trial: implications for measuring adherence.

Clinic-based pill counts of unused study medication are frequently used to measure adherence in HIV-1 prevention trials. Monthly pill count adherence data from the Partners in Prevention HSV/HIV Transmission Study, a double-blind, placebo controlled trial of twice-daily acyclovir suppression of herpes simplex virus type 2 (HSV-2) in HIV-1 infected persons was used to compare changes between 3,381 placebo and active arm participants in two objective biologic measures of acyclovir's drug activity: reduction in plasma HIV-1 RNA and HSV-2 genital ulcer disease (GUD). Higher acyclovir pill count adherence was associated with greater reductions in plasma HIV-1 RNA and GUD, indicating pill count data is strongly correlated with biological effects of adherence. However, when calculated adherence exceeded 102 % (i.e., fewer pills returned than expected) and when pill counts were missing because bottles were not returned, plasma HIV-1 RNA and GUD effects were diminished, likely indicating periods of non-adherence.