Background:  Lipoatrophy has been associated with hypertriglyceridemia in HIV lipodystrophy, but temporal relationship is not known. We analyzed the relationship between baseline triglyceride levels and the development of lipoatrophy on regimens containing different formulations of stavudine.

Methods: This post-hoc analysis of data from two randomized, multinational, double-blind, placebo-controlled studies (096 and 099 rolled over to study 110) compared stavudine extended-release/prolonged release capsules (XR/PRC) formulation with stavudine immediate-release (IR) formulation, in combination with lamivudine and efavirenz, in antiretroviral naïve patients. For purposes of this analysis, only individuals with available fasting baseline triglyceride levels were analyzed and stratified accordingly (<200 mg/dL and ≥200 mg/dL). Lipoatrophy was investigator defined and included facial or extremity subcutaneous fat loss or wasting. We used multivariate logistic regression modeling to determine predictive variables and developed a scoring system for risk assessment.

Results:  A total of 934 patients who participated in study 110 with median follow-up of 116 weeks for XR/PRC and 114 weeks for IR. A sub-population of 877 (94%) with available fasting triglyceride levels (436/468 for XR/PRC and 441/466 for IR) had similar baseline characteristics (median = 103; range 30 to 797 mg/dL). Total incidence of lipoatrophy was lower in the XR/PRC group compared with the IR group (8% vs 14%). When stratified according to triglyceride levels, lipoatrohy rates were significantly higher in individuals with baseline triglyceride ³200 mg/dL when compared with baseline triglyceride <200 mg/dL on IR but not on XR/PRC formulation:

Multivariate logistic regression modeling identified triglyceride levels <200 mg/dL at baseline (p = 0.030), age <40 years (p = 0.002), and XR/PRC (p = 0.004) as significant predictors for the reduced incidence of any lipoatrophy. In contrast, gender, race, baseline body mass index, fasting glucose, waist circumference, baseline CD4, and viral load were not significant predictors. Utilizing the scoring system (1 point each for age >40 years, baseline triglyceride >200 mg/dL, or IR use), lipoatrophy incidence increased with number of risk points: 0 points‑6.9%, 1 point‑9.7%, 2 points‑17.7%, and 3 points‑34.5%.

Conclusions:  These data suggest that XR/PRC formulation may lessen the occurrence of lipoatrophy. Prudent patient selection may also lessen the incidence of lipoatrophy with stavudine IR containing HAART regimens. Confirmation of these findings is needed.

Keywords: lipodystrophy; hypertriglyceridemia; stavudine