October 29, 2001
While we are waiting for more definitive data from an on-going "2NN" comparative study between efavirenz (EFV) and nevirapine (NVP), the two main NNRTIs, cohort studies are endeavoring to compare these agents. Additionally, as the 2NN study may not continue long enough to evaluate the effects of a drug on the development of lipodystrophy, cohort analyses provide a potentially unique opportunity to examine this issue. A previously published study of a cohort of 138 patients on first-line therapy with either EFV or NVP found no difference between them with regard to lipid disturbances as assessed by total cholesterol and triglycerides [Matthews GV, et al. "Absence of Association Between Individual Thymidine Analogues or Non-Nucleoside Analogues and Lipid Abnormalities in HIV-1 Infected Persons on Initial Therapy." JAIDS 2000; 24:310-5]. However, a cross-study comparison of these agents in switch studies suggests EFV benefits total cholesterol and triglycerides levels less than NVP. Large cohort studies assessing lipodystrophy (however that may be defined) have suggested that lipodystrophy may be more common when PIs are combined with nucleoside analogs relative to nucleosides alone or PIs alone. However, no differences have been reported between the (low) rate of lipodystrophy with nucleosides alone relative to nucleosides with NNRTIs.
The Dagna-LipART study is a collaborative study involving 79 centers across Germany. It reports on patients who started therapy as antiretroviral-naive individuals in 1996. The total cohort consists of 221 individuals, 77 of whom are on NNRTI-based regimens with either NVP (n=39, mean 9.7 months) or EFV (n=38, mean 5.8 months). Patients were assessed for the presence of lipodystrophy after approximately three years of therapy, with 34% being found to have the syndrome. This may represent one of the best available estimates of the cumulative prevalence of lipodystrophy over 3 years. A CD4 nadir (low point) of <200cells/mm3, and d4T use for >12 months were associated with an increased relative risk for the syndrome. Use of NNRTIs for >12 months was associated with a reduced risk of lipodystrophy relative to other (essentially PI-containing) regimens. The association with CD4 nadir has been observed in several previous studies. The association of d4T with lipodystrophy remains controversial, as two studies comparing individuals who have only ever received d4T or AZT did not find differences between the drugs [Bogner, JAIDS 2001; Bonjoch, IAS meeting 2001). However, some cross-sectional studies have indicated an association. The reduced risk with NNRTIs may simply represent the other side of the increased risk with PIs story.
The analysis of this cohort presented at Athens suggested that the reduction in risk of lipodystrophy was something which accumulated over time. Across all NNRTIs, the risk of lipodystrophy was reduced modestly (and not statistically significant) by around 25% relative to other regimens. In those who were on NNRTIs for 12 months or more, the reduction, about 70%, was significant. This effect was largely driven by NVP therapy. Although it was not statistically significant when NVP alone was tested, the same trends exist in the data. The chances of demonstrating a benefit with EFV were limited by the short exposure in the cohort to this drug (5.8 months average).
Regarding lipids, NNRTI use was also associated with a relative increase in the risk of a high cholesterol but no effects regarding triglycerides. However, this effect was exclusively related to EFV use where the risk of high cholesterol was 2.3-fold higher relative to other regimens. With NVP, the risk was unchanged (relative risk 1.0). Risk of a high triglyceride was not significantly higher with efavirenz (1.13-fold higher) and not significantly (45%) lower with NVP. The reasons for these differences are unclear. While they may be explained by intrinsic differences among the drugs, they may also represent a persistence of the effects of prior PI therapy for a period of time after switching to the NNRTI. Randomized comparative data will ultimately be required to answer this question.