Measuring adherence to antiretroviral therapy via hair concentrations in India

Authors : Gandhi M, Devi S, Bacchetti P, Chandy S, Heylen E, Phung N, Kuncze K, Okochi H, Kumar R, Kurpad AV, Ekstrand ML

Publication Year : 2019

Abstract :

BACKGROUND:
Objective adherence measures are of increasing interest in antiretroviral treatment (ART) monitoring. Hair ART levels predict virologic suppression and hair is easy to collect and store. No prior study has examined hair levels in an India-based cohort or laboratory.

METHODS:
Small hair samples were collected from HIV-positive participants on either efavirenz-based or nevirapine-based ART in a South India-based study. Hair samples were split and analyzed for efavirenz or nevirapine in the UCSF-based Hair Analytical Laboratory (HAL) and the Division of Nutrition-analytic lab based at St. John's Research Institute (SJRI), Bangalore, India using liquid chromatography/tandem-mass-spectrometry. Agreement (using Bland-Altman methods) and rank correlation between the two laboratories' hair levels were calculated. Rank correlation between self-reported adherence (SRA) over the prior month using a visual analog scale and hair ART levels was calculated.

RESULTS:
Among 75 participants (38 on nevirapine; 37 on efavirenz), the correlation between nevirapine levels generated by the two laboratories was 0.66 (p less than 0.0001) and between efavirenz levels was 0.87 (p less than 0.0001). Measurements from SJRI were usually within 20% of those from the UCSF HAL. SRA was essentially uncorrelated with hair ARV levels for either drug (all correlations less than 0.04). Hair levels showed variability in adherence although SRA was >85% in all participants.

CONCLUSIONS:
Hair ART levels measured by both an India-based laboratory and the standard U.S.-based laboratory showed generally high agreement and correlation, demonstrating local capacity. As in many other cohorts, hair ART levels and SRA were not well-correlated, likely indicating limitations in self-report and the need for objective adherence monitoring in resource-limited settings.

https://www.ncbi.nlm.nih.gov/pubmed/30865182