Associations between self-reported PrEP use and short and long-term pharmacologic metrics of PrEP adherence among Kenyan women

Background: Biomarkers provide reliable information on oral PrEP pill-taking. It remains unclear whether self-reported PrEP adherence has value when used in combination with pharmacologic measures. We examined associations between self-reported and biomarker-based PrEP adherence metrics to improve identification of those with low adherence in resource-constrained settings. Methods: HIV-uninfected women using oral tenofovir (TFV)-based PrEP were enrolled (N=100) in the Point-of-Care Urine Monitoring for Adherence trial in Kenya from 2021–2022. Participants were randomized 1:1 to standard-of-care or urine-based-point-of-care assay-informed drug-level feedback. Adherence was measured at enrollment and quarterly through 12 months via urine, hair and self-report. A positive urine assay (TFV≥1500 ng/mL) indicated short-term adherence. TFV hair concentration≥0.023 ng/mg indicated long-term adherence (4-times-weekly dosing). Self-reported adherence was calculated as the mean score (range=0-100) on a 2-item scale. Generalized estimating equations estimated longitudinal associations between self-reported adherence and biomarkers. Receiver-operating-characteristic curves assessed self-report as a predictor of biomarker metrics. Results: At enrollment, 68.7% had a positive urine assay, 83.7% had detectable hair TFV, and median self-reported adherence was 100 (IQR=90-100). At Month 12, participants with higher adherence scores were more likely to have hair TFV concentrations≥0.023 ng/mg (aRR=1.44; 95% CI=1.20-1.72; p-value<0.001) and positive urine results (aRR=1.56; CI=1.43-1.69; p-value=<0.001). Areas-under-the-curve for self-report as a predictor of hair levels were 0.64 in a model with self-report and 0.74 in a model with both self-report and urine testing. Conclusions: Associations between self-reported and biomarker-verified PrEP adherence demonstrated accurate reporting. Incorporating biomarkers in studies may result in improved adherence and/or improved accuracy of self-report.