Background:
HIV incidence is high among young men who have sex with men (YMSM) and transgender women (TGW). PrEP is a key strategy to reduce new HIV infections, and monitoring PrEP adherence is essential to guide implementation programs. We aimed to assess the accuracy of indirect PrEP adherence measures with drug concentrations in dried blood spots (DBS) among YMSM and TGW enrolled in the ImPrEP study.
Methods:
ImPrEP was an implementation project offering same-day oral PrEP for 9509 MSM/TGW in Brazil, Mexico, and Peru (Feb/2018-Jun/2021), with follow-up visits scheduled 4 weeks post-enrolment and quarterly thereafter, that included YMSM aged 18-24 years and TGW (all ages) who collected at least one DBS during follow-up. We compared two indirect adherence measures with DBS: medication possession ratio (MPR) (ratio between tablets dispensed in prior visit and days between the two visits) and self-reported information (single-question at each visit; 30-days recall). We used generalized estimating equations and area under the curve (AUC) to assess the accuracy of each indirect measure with protective drug DBS levels (TFV-DP ≥ 550 fmol/punch [week 4] and ≥ 800 fmol/punch [other weeks]), and the DeLong test to compare the curves. We calculated optimal cut-off points for discriminating protective drug levels based on Youden index and their respective sensitivity, specificity, negative (NPV) and positive (PPV) predictive values.
Results:
We included 4274 DBS samples from 2096 participants (week 4: 1905[44.6%], week 28: 1170[27.4%], week 52: 745[17.4%], week 76: 254[5.9%], week 100: 135[3.1%], week 124: 65[1.5%]). Overall, 1692 (80.7%) participants were MSM and 404 (19.3%) TGW; most were aged 18-24 years (1802; 86.0%), non-white (1582; 75.5%), and had ≥12 years of education (1374; 65.5%). Of all DBS samples, 2871(67.2%) had protective drug levels. AUC was 0.75(95%CI:0.74-0.77) for MPR and 0.76(95%CI:0.74–0.78) for self-report adherence (Table), with no difference between adherence assessment methods’ curves (p>0.38). Calculated cut-off points for MPR and self-reported adherence were 97.0% and 93.3%, respectively.
Conclusions:
Self-reported adherence and MPR adequately discriminated protective levels of PrEP among key populations in Latin America at different time points during the study follow-up. These measures are low-cost, easy to implement, and allow for immediate action to support PrEP adherence at health service level and ultimately contribute to monitoring PrEP programs.
