Comparing a point-of-care urine tenofovir lateral flow assay to self-reported adherence and their associations with viral load suppression among adults on antiretroviral therapy.

Abstract

Point-of-care (POC) lateral flow assays (LFA) to detect tenofovir (TFV) in urine have been developed to measure short-term ART adherence. Limited data exist from people living with HIV in routine care. Adults on TFV-containing regimens, having a routine viral load (VL) at an HIV clinic in Cape Town, South Africa were enrolled in a cross-sectional study. Patients recalled missed ART doses in the past three and 7 days and urine was tested using a POC TFV LFA. VL on the day was abstracted from medical records. Among 314 participants, 293 (93%) had VL <1000 copies/mL, 20 (6%) had no TFV detected and 24 (8%) reported ≥1 missed dose in the past 3 days. Agreement between VL ≥1000 and undetectable TFV was higher compared to 3-day recall of ≥1 missed dose (Kappa 0.504 vs. 0.163, p = 0.015). The AUC to detect VL ≥1000 was 0.747 (95% CI 0.637-0.856) for undetectable TFV. This was statistically significantly better than for 7-day recall (0.571 95% CI 0.476-0.666, p = 0.040) but not for 3-day recall (0.587 95% CI 0.492-0.681, p = 0.071) of ≥1 missed dose. In this largely virally suppressed cohort, TFV in urine had better agreement with VL than self-reported adherence and was a better predictor of viraemia on two of three self-report measures. Used in combination with VL, the POC urine TFV LFA could flag patients with viraemia in the presence of ART. Further research is needed to understand the potential application in different populations on ART, including pregnant women.