Background
Females may require higher oral Pre-exposure Prophylaxis (PrEP) adherence to achieve equivalent therapeutic drug concentrations compared to males, suggesting the need for sex-specific cutoffs for rapid urine tenofovir (TFV) assays. We developed a next-gen urine TFV Lateral Flow Assay (LFA) without a predefined threshold to (1) determine if the quantified LFA readouts differ between females and males with the same adherence levels, and (2) to evaluate its diagnostic performance for detecting non-adherence across sexes.
Methods
Using a next-gen urine TFV LFA, we tested urine samples from HIV-negative adults randomized to 2 or 4 or 7 doses/week of Emtricitabine/Tenofovir Disoproxil Fumarate for directly-observed oral PrEP. We obtained the averaged visual scores of the LFA test line from two operators using a graded color intensity reference card. Optical readings of the test and control line were objectively measured by an LFA reader. We assessed the effect of sex on LFA readouts among adherent samples with a dose taken in the prior day using Generalized Estimating Equations. We calculated the overall and sex-specific sensitivities/specificities for recent non-adherence defined as no dosing in the prior day.
Results
A total of 586 urine specimens were evaluated from 28 participants. Among 139 urine samples with a dose taken in the prior day, males had 33% lower (p <0.01) optical readings, and a 1.82 unit lower (p <0.01) averaged visual score of the LFA test line, when compared to females. With the same cutoffs (>705) for the test line optical reading, we found similar sensitivities (86% vs 86%) and specificity (100% vs 95%) among males versus females. With similar specificities (97% vs 92%), the sensitivity among females (82%, 95% confidence interval (CI): 77-88%) was slightly higher than that among males (76%, 95% CI: 71-81%) when using the same cutoff (>1.25) for the averaged visual score. The ratio of the LFA test/control line optical readings (>0.095) was highly sensitive (93%, 95% CI: 91-95%) for detecting non-adherence as early as missing one dose, while maintaining a high specificity (96%, 95% CI: 93-99%) in the total population.
Conclusions
While females showed higher LFA readouts (indicating lower urine TFV levels) than males, sex-specific cutoffs for a next-gen TFV LFA may not be necessary to measure antiretroviral adherence. The next-gen LFA demonstrated enhanced diagnostic accuracy in identifying early non-adherence.
Table or Figure
