Relationship between untimed plasma lopinavir concentrations and virological outcome on second-line antiretroviral therapy

Background: Resource constraints in low and middle-income countries necessitate practical approaches to optimizing antiretroviral therapy outcomes. We hypothesised that an untimed plasma lopinavir concentration (UPLC) at week 12 would predict loss of virological response in those taking lopinavir as part of a second-line antiretroviral regimen. Methods: We measured plasma lopinavir concentration at week 12 on stored samples from the SECOND-LINE study. We characterized UPLC as: detectable and optimal (≥1000-μg/l); detectable but suboptimal (≥25 to <-1000-μg/l); and undetectable (<25-μg/l). We used Cox regression to explore the relationship between UPLC and loss of virological response over 48 weeks and backwards stepwise logistic regression to explore the relationship between UPLC and other predictors of virological failure at week 48. Results: At week 48, we observed virological failure in 15/32 (47%) and 53/485 (11%) of patients with undetectable and detectable UPLC, respectively, P-<-0.001. Both suboptimal [adjusted hazard ratio (HR) 2.94; 95% confidence interval (CI) 1.54-5.62; P-=-0.001], and undetectable (adjusted HR 3.55; 95% CI 1.89-6.64; P-<-0.001) UPLC were associated with higher rates of loss of virological response over 48 weeks. In multivariate analysis, an independent association with virological failure at week 48 and undetectable UPLC was observed after adjustment (odds ratio 5.48; 95% CI 2.23-13.42; P-<-0.01). Conclusion: In low and middle-income countries implementing a public health approach to antiretroviral therapy treatment, an untimed plasma drug concentration may provide a practical method for early identification of patients with inadequate medication adherence and facilitate timely corrective interventions to prevent virological failure.