Background: Adherence to highly active antiretroviral therapy (HAART) for HIV-1 infection is essential for plasma HIV-1 RNA suppression. Self-report is the most frequently used measure of adherence to HAART, but its validity is controversial. Studies on the relation between self-reported adherence and virologic treatment response have shown inconsistent results. We investigated whether this variability between studies about the effect of self-reported adherence on virologic treatment response could be attributed to study design features.
Methods: We searched for studies reporting on adult nonpregnant patients prescribed antiretroviral therapy for chronic HIV-1 infection using a self-reported adherence measure and providing information about the relation between adherence and plasma HIV-1 RNA concentrations. Meta-analysis with random effects modeling was used to pool data and to investigate sources of heterogeneity.
Results: Sixty-five studies fulfilled inclusion criteria, containing data from 15,351 patients. The pooled odds ratio (95% confidence interval) of detectable plasma viral load in nonadherent patients was 2.31 (1.99-2.68). There was significant heterogeneity among studies (P < 0.001). Not ascertaining confidentiality of responses, use of actual viral load measurements, an adherence threshold lower than 95%, higher percentages of patients on their initial antiretroviral regimen, and higher percentages of patients with a history of intravenous drug use within a study were associated with higher point estimates.
Conclusions: Overall, we observed that self-reported adherence measures can distinguish between clinically meaningful patterns of medication-taking behavior. Distinct study characteristics were significantly associated with the relation between adherence and virologic response. These characteristics should be taken into consideration when interpreting results from studies on self-reported adherence.