Background:  The utility of plasma ART concentrations for therapeutic drug monitoring in HIV is limited by intra-individual variability in drug levels. Adherence patterns prior to clinic visits may be atypical, and fluctuations in pharmacokinetic parameters influence single measurements. Hair levels of ART provide an average level of exposure over weeks to months and may enrich predictive models of treatment outcome.

Methods:  Our accompanying abstract describes methods for analyzing lopinavir/ritonavir (LPV/RTV) levels in small samples of human hair (10 to 20 strands or ~2 mg). We analyzed hair levels of drug in 72 women in the Women’s Interagency HIV Study approximately 6 months (range 2 to 12 months) after initiating LPV/RTV-based combination regimens. Multivariate logistic regression estimated the effect of hair drug levels and a number of other variables (including self-reported adherence to ART over the past 30 days being ≥95% vs <95%) on a dichotomous measure of virologic success (achieving viral load undetectability or at least a 10-fold drop by 6 months). Because hair concentrations of the 2 drugs were substantially collinear, models including levels of both are not presented.

Results:  Hair concentrations of either LPV or RTV showed strong associations with virologic success after 6 months on therapy in multivariate models. The odds ratio (OR) for achieving success was 2.0 (95%CI 1.25 to 3.3, p = 0.004) for every 2-fold increase of LPV level in hair and the OR for success per doubling in RTV level was 2.3 (95%CI 1.29 to 3.9, p = 0.004) when controlling for age, race, pre-treatment viral load, total time on drug, self-reported adherence, and prior ART experience. Having a high self-reported adherence level over the preceding 30 days was moderately associated with the outcome (OR for achieving virologic success 1.90 (95%CI 0.55 to 6.6, p = 0.31) with ≥95% adherence vs <95% adherence in models with LPV hair levels; OR by adherence level was 2.05 (95%CI 0.58 to 7.2, p = 0.26) in models with RTV).

Conclusions:  In models examining predictors for virologic success after initiating LPV/RTV-based regimens, hair levels of LPV or RTV at 6 months were each strong, independent predictors of virologic success, even when controlled for self-reported adherence, pre-treatment viral load, and prior ART experience. Levels of ART in small samples of hair estimate long-term exposure to a drug and may serve as noninvasive monitoring tools for treatment outcomes.