Relationship between antiretroviral plasma concentration and emergence of HIV-1 resistance mutations at treatment failure

PURPOSE: The relationship between antiretroviral pharmacokinetic exposure and acquisition of human immunodeficency virus-1 (HIV-1) drug resistance mutations (DRM) is not fully understood. The aim of this study was to investigate whether antiretroviral plasma concentration could predict the emergence of DRM at treatment failure. METHODS: The study cohort comprised retrospectively selected patients with failing antiretroviral regimens for whom a protease inhibitor (PI) or non-nucleoside reverse transcriptase inhibitor (NNRTI) trough concentration measurement (TDM) had been obtained before failure, a genotypic resistance test (GRT1) had been performed before the TDM, and a genotypic resistance test (GRT2) had been performed at therapeutic failure. Drug levels were classified as undetectable/detectable or subtherapeutic/therapeutic according to limits of quantification of a high-performance liquid chromatography-ultraviolet assay or pre-defined efficacy thresholds, respectively. The number of DRM acquired at treatment failure was evaluated by comparing the results of the GRT2 and GRT1. RESULTS: A total of ten and 57 failure episodes occurred among our patients on NNRTI-based and PI-based regimens, respectively, and included in the evaluation. PI concentration was subtherapeutic in 28.1% of patients, among which the levels were undetectable in 21.1%. Twenty-five (43.9%) patients acquired at least one new PI-DRM according to the GRT2. Patients with undetectable PI levels showed a lower emergence of PI-DRM (minor + major) than those with detectable levels (8.3 vs. 53.3%, p = 0.007). Multivariate analysis confirmed that undetectable PI levels were independent negative predictors of DRM selection. NNRTI measurements were subtherapeutic in 2/10 (20%) patients. NNRTI-DRM were acquired by all patients regardless of NNRTI levels. CONCLUSIONS: A PI measurement showing undetectable drug levels prior to treatment failure predicted the lack of emergence of PI-DRM at failure. These results suggest that PI levels can help clinicians interpret the reasons for treatment failure and guide the type of interventions needed.