HIV Articles  
Efavirenz & AZT/3TC Did Not Cause Limb Fat Loss in ACTG 364  
  "Glucose metabolism, lipid, and body fat changes in antiretroviral-naive subjects randomized to nelfinavir or efavirenz plus dual nucleosides"
AIDS: Volume 19(16) 4 November 2005
Dub, Michael Pa; Parker, Robert Ab; Tebas, Pablof; Grinspoon, Steven Kc; Zackin, Robert Ab,*; Robbins, Gregory Kd; Roubenoff, Ronenne; Shafer, Robert Wg; Wininger, David Ah; Meyer, William A IIIi; Snyder, Sally Wj; Mulligan, Kathleenk
ACTG 364 has been the subject of a number of conference reporting & publications in journals. I served as the original patient community representative on this study, when it was being organized. Patients were randomized to receive nelfinavir or efavirenz and ddI/d4T or AZT/3TC. To keep this brief, the most interesting finding reported in this published article is that limb fat (arm plus leg) was reduced (as measured by DEXA) by 13% after week 64 for patients receiving nelfinavir and limb fat increased by 1.8% (virtually no change) for patients receiving efavirenz, suggesting that efavirenz does not reduce limb fat, at least in this study. As well, AZT/3TC use in this study did not result in limb fat loss after 64 weeks therapy.
The other significant findings in this study are well known. Author summary: "Greater changes in percentage changes in limb fat occurred with didanosine/stavudine (-16.8%) than with zidovudine/lamivudine (+4.0%; P < 0.001 for overall change from baseline) and with nelfinavir (-13.1%) compared with efavirenz (+1.8%; P = 0.003).... Over 64 weeks, all regimens were associated with increases in lipids but insulin resistance did not differ between groups. Regimens containing didanosine/stavudine and regimens containing nelfinavir were associated with greater loss of limb fat."
Objective: To determine if particular components of antiretroviral drug regimens are associated with greater insulin resistance, dyslipidemia, and peripheral lipoatrophy.
Methods: Metabolic and body composition variables were measured prospectively over 64 weeks in 334 antiretroviral-naive, HIV-infected subjects who were randomized to receive nelfinavir, efavirenz, or both, combined with zidovudine/lamivudine or didanosine/stavudine in a factorial design, multicenter trial. Subjects assigned to efavirenz (n = 110) were compared with those assigned to nelfinavir (n = 99); subjects assigned to zidovudine/lamivudine (n = 154) were compared with those assigned to didanosine/stavudine (n = 180). A subset of 157 subjects had serial dual-energy X-ray absorptiometry (DEXA) scans.
Results: Lipid measures increased in all groups. Greater increases in high density lipoprotein (HDL) cholesterol occurred with efavirenz than with nelfinavir. Greater increases in total cholesterol, non-HDL cholesterol and HDL cholesterol occurred with stavudine and didanosine than with zidovudine/lamivudine. There were no differences in insulin resistance in the comparisons. After initial increases in the first 16 weeks, median limb fat decreased. Greater changes in percentage changes in limb fat occurred with didanosine/stavudine (-16.8%) than with zidovudine/lamivudine (+4.0%; P < 0.001 for overall change from baseline) and with nelfinavir (-13.1%) compared with efavirenz (+1.8%; P = 0.003).
Conclusions: Over 64 weeks, all regimens were associated with increases in lipids but insulin resistance did not differ between groups. Regimens containing didanosine/stavudine and regimens containing nelfinavir were associated with greater loss of limb fat.
This prospective metabolic and body composition study nested in a large randomized trial of potent antiretroviral regimens demonstrates clear differences between the tendency of particular antiretroviral drug-containing regimens to induce limb fat loss compared with baseline measurements. When combined with a protease inhibitor (nelfinavir), an NNRTI (efavirenz), or both, assignment to the combination of ddI/d4T was associated with greater loss of limb fat than was the combination of ZDV/3TC. When combined with dual NRTI, assignment to the protease inhibitor nelfinavir was associated with greater loss of limb fat than was the NNRTI efavirenz, although the magnitude of the difference was smaller than that with the NRTI drugs and the on-treatment analysis did not reach statistical significance. Contrary to our expectations, use of nelfinavir in combination with a pair of NRTI did not induce greater insulin resistance or dyslipidemia than use of efavirenz.
The use of antiretroviral therapy that includes protease inhibitors has been associated with new-onset diabetes mellitus [1,2,20] and insulin resistance [3-5,7,21,22]. However, indinavir is the only protease inhibitor clearly associated with early (within 8 weeks or less) reduction of insulin sensitivity [21-23], likely in part via inhibition of the function of the insulin-sensitive glucose transporter GLUT4 [24,25]. There is human clinical evidence that amprenavir [26] and atazanavir [27,28] do not cause early insulin resistance, while conflicting data exist for lopinavir-ritonavir [28,29]. There was no evidence of an effect on insulin sensitivity in this study with nelfinavir at 8 weeks, suggesting that an early, direct effect on insulin resistance also does not occur with this particular protease inhibitor.
Lipid measures, including total and non-HDL cholesterol and triglycerides, increased comparably with both nelfinavir- and efavirenz-based treatment. Similar results have been presented in another preliminary report [30]. Of note were the substantial increases in HDL cholesterol with both treatments (median increase of 5-7 mg/dl, or approximately 20-25%). With efavirenz, there was a modest advantage in HDL cholesterol, resulting in a more favorable total cholesterol to HDL cholesterol ratio, suggesting lesser cardiovascular risk [31]. Non-HDL cholesterol, a risk marker that may be more predictive than LDL cholesterol for cardiovascular events [32], increased comparably by approximately 25 mg/dl in both groups. The net effect of these changes on long-term cardiovascular risk in patients with HIV is unknown.
Protease inhibitors may contribute to an increased risk of lipoatrophy when administered with an NRTI. In a descriptive cohort study, dual NRTI therapy plus a protease inhibitor was associated with a more rapid time to fat wasting than was dual NRTI alone [33]. Moreover, as compared with time on nevirapine, cumulative time on protease inhibitors was associated with an increased risk of fat wasting, suggesting that protease inhibitors may accelerate the fat wasting effects of NRTI drugs [33]. When not administered with NRTI agents, the incidence of lipoatrophy with the protease inhibitor combination ritonavir plus saquinavir was low [34]. In the present study, we report evidence from a prospective, randomized trial of an independent effect of assignment to the protease inhibitor nelfinavir on limb fat loss, albeit an effect smaller in magnitude than the effect of assignment to ddI/d4T. Our findings are consistent with in vitro studies in which certain protease inhibitors, including nelfinavir, have been shown to inhibit adipocyte differentiation [35-37]. These results, however, may not be generalizable to all protease inhibitors. The on-treatment analysis was limited by a high frequency of treatment switching, reducing the number of subjects observed. We feel that, because lipoatrophy represents a cumulative toxicity, the on-treatment analysis (which censors subjects after a treatment arm change) is less valuable than our primary, intent-to-treat analysis.
Randomization to ddI/d4T led to a greater loss of limb fat than did ZDV/3TC. A secondary analysis that censored subjects at the time of regimen changes was also statistically significant, and there was no evidence for interaction between treatment group assignments. Many studies have linked d4T use with lipoatrophy [13,33,38,39] and substitution of other NRTI for d4T has improved lipoatrophy [40-42]. A randomized trial that compared d4T or tenofovir in combination with 3TC and efavirenz reported greater limb fat by DEXA among tenofovir recipients at weeks 96 and 144, but did not include baseline DEXA measurements [39] and so could not evaluate the longitudinal patterns of fat gain or loss with the different therapies. The current report from a prospective randomized study among initially antiretroviral-naive subjects using objective measures clearly links use of a particular NRTI combination with loss of limb fat over time compared with baseline values. However, our data cannot address the relative contribution of the individual components (ddI or d4T) to the observed effects. Nonetheless, it is clear that d4T contributes to lipoatrophy risk even when combined with 3TC [39], while the individual contribution of ddI is less certain. Because of the risk of side effects, this drug combination is no longer recommended [43].
A limitation of the body composition analyses in this report is the relatively short follow-up period, 64 weeks. This period was chosen to coincide with the available metabolic data, which were limited to 64 weeks. It is possible that with more extended observation, differences between treatment groups may become even greater, or they may equilibrate over time.
Unlike limb fat, trunk fat tended to remain above baseline values in all groups. Similar increases in DEXA-derived central abdominal fat have been reported in a prospective, non-randomized observational study [44]. It is possible that the early increases in limb fat and the persistent increases in trunk fat represent a generalized 'return-to-health' phenomenon associated with viral suppression and immune reconstitution. This study provides evidence that the subsequent decreases in limb fat are related to use of particular antiretroviral drugs. The treatments that spared limb fat over 64 weeks (ZDV/3TC, efavirenz) were also associated with overall greater increases in trunk fat and total body fat. This may represent a generalized increase in adiposity that preferentially affects trunk fat. If such an increase reflects increased visceral fat, it is a potentially adverse consequence. DEXA scans, however, do not differentiate between visceral and subcutaneous trunk fat.
Our results illustrate the importance of prospective studies to determine the particular roles of individual antiretroviral agents and combinations in the development of metabolic and morphological complications. Extended follow-up will be necessary to define the long-term course of these complications, particularly as individuals are treated with multiple different antiretroviral agents from multiple drug classes. It will be crucial to develop new, less-toxic antiretroviral agents and combination regimens in order to minimize the cardiovascular risks associated with antiretroviral therapy [45,46] and the stigma, non-adherence [47-49], and metabolic disturbances [8] associated with lipodystrophy.
  icon paper stack View Older Articles   Back to Top