Determinants of heterogeneous adherence to HIV-antiretroviral therapies in the multicenter AIDS cohort study

Assessment of adherence to HIV antiretroviral therapy (ART) is required for studying therapeutic effectiveness and identifying subgroups needing focused education. The study's goals were to describe the level of ART adherence using self-reported recall over a 4-day period and to characterize determinants of lower adherence. The interaction between adherence and drug holidays on level of HIV RNA also was investigated. Perfect self-reported adherence was defined as taking all doses and numbers of pills as prescribed for current HIV medications. Independent predictors of <100% adherence were determined using multivariate logistic regression. Among 539 men, 419 (77.7%) were 100% adherent by the algorithm using self-reported data. HIV-1 RNA was <50 copies/ml in 48.2% of the adherent group versus 33.7% in the less adherent group (p = .015). This proportion dropped to 28% if a drug holiday was reported in addition to lower adherence. A drug holiday was not virologically detrimental if the participant was otherwise adherent. Determinants of lower adherence included African American race (odds ratio [OR], 2.4; p = .008), income <U.S.$50,000 (OR, 2.2; p = .002), no outpatient visits (OR, 3.6; p = .003) and increasing numbers of ART medications (OR, 4.5; p = .001). These data support the validity of using a questionnaire to assess adherence in observational studies. Identification of individuals with characteristics associated with lower adherence provides the basis for interventions to enhance adherence and optimize effective therapies.