NON-NUCLEOSIDE REVERSE TRANSCRIPTASE INHIBITOR USE IS NOT ASSOCIATED WITH INCREASED MORTALITY AMONG PERSONs INITIATING ANTIRETROVIRAL THERAPY IN A POPULATION-BASED SETTING

Robert S Hogg, Katherine Heath, Benita Yip, Julio SG Montaner

Objective: To characterize the impact of non-nucleoside reverse transcriptase inhibitor (NNRTI) versus protease inhibitor (PI) based triple drug antiretroviral therapy on survival.
Methods: Population-based analysis of 1282 antiretroviral therapy naïve HIV-positive individuals aged 18 years and older in British Columbia who initiated started triple combination therapy between August 1996 and December 1999. Cumulative mortality rates were estimated using Kaplan-Meier methods. Event-free subjects were right censored as of September 30, 2000. Cumulative mortality rates from the initiation of triple-drug antiretroviral therapy to September 30, 2000 were determined using various CD4 and plasma HIV RNA thresholds.
Results: Our analysis was based on 1282 persons, of whom 346 (27%) received NNRTIs [324 (94%), on nevirapine, 11 (3%) on efavirenz and 11 (3%) on delavirdine] and 936 (73%) received PIs [692 (74%) on inidinavir, 128 (14%) on nelfinavir, 79 (8%) on saquinavir and 37 (4%) on ritonavir]. As of September 30, 2000, a total of 106 subjects had died. Cumulative mortality was 3.9% (±0.5%) at 12 months. The inclusion of NNRTIs in the initial regimen (OR=0.47; 95% CI: 0.26, 0.86), a prior diagnosis of AIDS, HIV-1 RNA levels, CD4 cell count, physician experience, and intermittent use of therapy were found to be prognostic predictors of survival in the univariate analysis. In a multivariate model, after controlling for other variables that were significant in the univariate analyses, NNRTIs use in the initial regimen (AOR=0.66; 95% CI: 0.35, 1.22) was not associated with mortality.
Conclusion: Our study demonstrates that initiating therapy with NNRTI versus PI based triple drug therapy was associated with similar outcomes. These results are consistent with those of recent clinical trials, but contrast with some cohort analyses. This difference could be at least partially attributed to the population-based nature of our database which protects against some of the potential subtle confounding biases that can emerge in cohort based studies.