Evolving patterns of antiretroviral drug interactions in people with HIV in British Columbia, Canada

Abstract

Objectives: To characterize the annual prevalence of antiretroviral/nonantiretroviral drug interactions in relation to antiretroviral therapy (ART)-prescribing patterns, and to describe drug interaction-related ART changes.

Design/methods: This cohort study included ART-treated adults in British Columbia, Canada between 01 January 2010 and 31 December 2016. Medication dispensing records were abstracted from a population-based, linked administrative-health dataset and used to identify antiretroviral-comedication drug interactions ('caution'/'avoid' drug interactions in HIV-focused drug interaction checkers). We identified temporal trends in annual drug interaction prevalence and quantified the association between taking higher drug interaction-risk ART and receiving nonrecommended antiretroviral-comedication combinations using Poisson regression models, modified for binary outcomes and correlated data. Clinician-reported, drug interaction-related ART changes and associated adverse events were abstracted from an HIV drug treatment registry and summarized descriptively.

Results: Among 8571 ART-treated adults who received nonantiretroviral comedications, prevalence of having any drug interaction or receiving nonrecommended drug combination(s) significantly declined from 85 to 71% and 5.6 to 3.2%, respectively, between 2010 and 2016 ( P < 0.001). This paralleled a shift from higher drug interaction-risk ART (e.g. ritonavir/cobicistat-boosted protease inhibitors) to lower drug interaction-risk ART (e.g. unboosted integrase inhibitors). Risk of receiving a nonrecommended antiretroviral-comedication combination was greater for persons taking higher vs. lower drug interaction-risk ART [adjusted risk ratio (aRR) 3.12, 95% confidence interval (CI) 2.24-4.35]. Boosted antiretroviral-inhaled corticosteroid drug interactions accounted for the most commonly dispensed, nonrecommended drug combinations, and the most commonly reported drug interaction-related adverse events (adrenal insufficiency).

Conclusion: The prevalence of antiretroviral-comedication drug interactions is declining as ART shifts towards antiretrovirals with lower drug interaction potential but nonrecommended drug combinations remain a concern. Healthcare providers should screen for drug interactions whenever drugs are prescribed or dispensed.