[Long-term efficacy of second-line treatment of HIV infection after class change following virological failure on protease inhibitor-based therapy]

Abstract

Background and objective: While there are evidence-based recommendations for the initial combination antiretroviral treatment (cART) of HIV infection, there are no comparative studies on long-term efficacy of different second-line strategies after initial virological failure. The aim of this study was to compare different second-line strategies after virological failure of an initial protease inhibitor (PI) based regimen, specifically the comparison between change to a different PI and class change to a non-nucleoside reverse transcriptase inhibitor (NNRTI).

Patients and methods: This cohort study retrospectively analyzed patient data documented for the Clinical Surveillance of HIV Disease project (ClinSurv) between 1999 and 2008, run by the Robert Koch Institute in Berlin, Germany. Follow-up data for at least three months of a treatment switch after virological failure of the first-line regimen were available for 157 patients out of the 14,377 patients in the ClinSurv cohort. Eighty-four (54%) of these had a PI-based first-line regimen and were therefore included into the analysis. Fifty-one (61%) of the 84 patients were switched to a different PI (group 1), 33 (39%) to a NNRTI (group 2). Primary end points were the probability of virological failure of the second-line regimen, the duration of a successful second-line regimen and the time to suppression of viral load below the level of detectability.

Results: There was no significant difference in the median time to virological suppression with 88 days in group 1 and 57 days in group 2 (p = 0.16). After > 3,000 days more than 50% of patients in group 2 (class switch to NNRTI) were still on an effective regimen, their risk of virologic failure thus was significantly lower than in group 1 (switch to a different PI), where the median duration of second-line therapy was only 581 days. Multivariate Cox regression analysis did not identify any of the available covariates as significant predictors of duration of the second-line treatment or as confounders. For group 1, with patients switching within the PI class, there was a more than two-fold risk of virological failure during the time of observation (HR = 2.3; 95%CI 1.1 - 4.9; p = 0.03).

Conclusion: Class switch to a NNRTI as opposed to changing to a different PI following virological failure of a PI-based first-line regimen is associated with significantly better durability of the second-line regimen.