Concordance between daily diary reported pre-exposure prophylaxis intake and intraerythrocytic tenofovir diphosphate in the Amsterdam Pre-exposure Prophylaxis demonstration project

Abstract

Objective: We assessed the association and concordance between self-reported oral pre-exposure prophylaxis (PrEP) intake in a diary app and intraerythrocytic drug metabolite concentrations.

Design: AMPrEP was a prospective demonstration study providing daily and event-driven PrEP to MSM in Amsterdam, the Netherlands (2015-2020).

Methods: Participants could record their PrEP intake in a diary app. Dried blood spots (DBS) were taken at 6, 12, 24, and 48 months and analysed for tenofovir diphosphate (TFV-DP) and emtricitabine triphosphate (FTC-TP) concentrations. We included TFV-DP measurements preceded by diary completion on at least 90% of days in the 6 weeks prior. We examined the association between self-reported PrEP intake (i.e. number of pills) and TFV-DP concentrations using tobit regression with a random intercept per participant. We also calculated concordance between categorized PrEP intake (i.e. <2, 2-3, 4-6 or 7 pills per week) and categorized TFV-DP concentrations (i.e. <350, 350-699,700-1249 or ≥1250 fmol/punch) using weighted Cohen's kappa. Last, we calculated concordance between self-reported recent PrEP intake (yes/no, in past 2 days) and quantifiability of FTC-TP (yes/no) using Cohen's kappa.

Results: Seven hundred and fifty-nine DBS measurements from 282 MSM were included. Self-reported PrEP intake was strongly and positively associated with TFV-DP concentration ( β = 0.77, 95% CI = 0.70-0.84, P < 0.0001). Concordance between categorized PrEP intake and TFV-DP concentration was moderate ( κ = 0.44, 95% CI = 0.39-0.50). Concordance between self-reported recent PrEP intake and FTC-TP quantifiability was perfect ( κ = 0.83, 95% CI 0.76-0.90).

Conclusion: Self-reported PrEP intake in a diary app is strongly correlated with actual use, and therefore reliable for comparing PrEP adherence between groups. Still, suboptimal criterion validity according to clinically relevant categories warrants caution when assessing 6-week reported adherence for individuals.