The role of efavirenz compared with protease inhibitors in the body fat changes associated with highly active antiretroviral therapy

Transcription

1 Journal of Antimicrobial Chemotherapy (2008) 62, doi: /jac/dkn191 Advance Access publication 14 May 2008 The role of efavirenz compared with protease inhibitors in the body fat changes associated with highly active antiretroviral therapy José A. Pérez-Molina 1 *, Pere Domingo 2, Esteban Martínez 3 and Santiago Moreno 4 1 Tropical Medicine and Clinical Parasitology Unit, Infectious Diseases Department, Hospital Universitario Ramón y Cajal, Madrid, Spain; 2 Infectious Diseases Unit, Hospital de la Santa Creu i San Pau, Universitat Autònoma de Barcelona, Barcelona, Spain; 3 Infectious Diseases Unit, Hospital Clínic, University of Barcelona- IDIBAPS, Barcelona, Spain; 4 Infectious Diseases Department, Hospital Universitario Ramón y Cajal, Madrid, Spain Highly active antiretroviral therapy plays a central role in the development of lipodystrophy syndrome, which may affect up to 50% of patients depending on the diagnostic criteria used. Most protease inhibitors and nucleoside reverse transcriptase inhibitors (NRTIs) are involved in body fat changes and associated metabolic disturbances. In contrast, non-nrtis have not been directly related to the onset of this syndrome. One of the most widely used methods to evaluate body fat changes is dual-energy X-ray absorptiometry (DEXA), which can detect differences in the distribution of body fat in patients with and without lipodystrophy. New information from a randomized open-label clinical trial suggests that efavirenz could have greater potential for causing lipoatrophy than lopinavir1ritonavir. This paper examines the impact of efavirenz on adipose tissue and body fat composition in order to evaluate whether this drug plays a role in the development of lipodystrophy. We have focused on the evidence obtained from comparative randomized clinical trials that use an objective measurement of fat distribution, such as DEXA. We analysed available in vitro data and evidence from non-comparative clinical trials. Keywords: lipodystrophy, lopinavir, DEXA scan, antiretroviral drugs, HAART Introduction Lipodystrophy arising from the treatment of HIV-1 infection and its associated metabolic alterations (e.g. dyslipidaemia and insulin resistance) is a multifactorial syndrome. Although highly active antiretroviral therapy (HAART) plays a central role in the development of this syndrome, it is also known that adipose tissue disorders are already present in naive HIV-1-infected patients. The mechanisms underlying the onset of this syndrome may represent a response to independent or interdependent factors, such as long-lasting viral infection, host-related factors, specific drugs or a specific combination of drugs. 1 9 The first drugs to be associated with lipodystrophy were the protease inhibitors (PIs), 1,2,6,10 and this relationship was later confirmed by in vitro studies The duration of the therapy with nucleoside reverse transcriptase inhibitors (NRTIs), especially thymidine nucleotides, was subsequently observed to be associated with the onset of lipoatrophy and lactic acidosis, 3,7,18,19 mainly as a consequence of the inhibition of DNA polymerase g in the mitochondria. 20,21 In contrast, non-nrtis (NNRTIs) have not been directly related to the onset of lipodystrophy. In fact, these were first used in studies on switching therapy from PIs as a means of reversing or stabilizing PI-induced lipodystrophy In those studies, the replacement of PIs by NNRTIs, generally after 1 year of follow-up, did not lead to any significant change in body fat (i.e. no improvement or worsening). In clinical terms, lipodystrophy becomes evident by a loss of fat in the face and limbs that may be accompanied by an accumulation of fat on the trunk and the back of the neck. Incidence varies depending on the diagnostic criteria used, but up to 50% of the patients on HAART may experience this condition. 2 4,6,7,27,28 Alterations in the distribution of body fat are a serious aesthetic problem that can stigmatize patients. Even if patients have controlled HIV infection, adhere to therapy and lead a normal life, lipodystrophy can make them look ill and potentially identify them as AIDS patients. It is no surprise, therefore, that in patients suffering from morphological changes, adherence to therapy may be compromised, thus leading to treatment failure. 29 Furthermore, metabolic alterations such as dyslipidaemia and insulin resistance can increase the risk of cardiovascular diseases in the long term.... *Corresponding author. Tel: þ ; Fax: þ ; japerezm.hrc@salud.madrid.org # The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org

2 Unfortunately, there are no objective criteria or case definitions for lipodystrophy. The evaluation of altered fat distribution by patients or physicians is easy; however, it does not allow for comparisons within the same patient or between different studies, because it is not objective. Yet, clinical research requires an evaluation that is as objective as possible in order to reliably determine the incidence of this condition, identify associated risk factors and evaluate the efficacy of preventive and therapeutic measures. In this sense, several different methods have been used to evaluate patients with lipodystrophy: anthropometric measurements, computed tomography, magnetic resonance imaging, sonography and dual-energy X-ray absorptiometry (DEXA). 2 4,7,35 38 This technique analyses regional and whole-body composition and is the most widely used one. It has proved useful in the diagnosis and evaluation of lipodystrophy and can detect differences in the distribution of body fat in patients with and without lipodystrophy. 38 Nevertheless, the results vary according to the manufacturer, site, operator and hardware and software versions used Therefore, the results of DEXA, especially in multicentre studies involving several operators, must be interpreted with caution. The ideal solution would be a central analysis site with as few operators as possible so that variability in the measurements could be reduced. 42 The recent results of a randomized open-label clinical trial suggest that efavirenz could have greater potential for causing lipoatrophy than lopinavirþritonavir. 44 This outcome is surprising, given that NNRTIs were not traditionally considered to be related to the development of lipodystrophy or at least were perceived as lower risk drugs than PIs. This paper aims to analyse the impact of efavirenz on adipose tissue and body fat composition and to evaluate whether this drug plays a role in the development of lipodystrophy. We searched MEDLINE and Embase for randomized clinical trials including efavirenz as part of the antiretroviral treatment schedule and using DEXA to measure fat distribution, as well as the most recent conference abstracts on HIV infection. We also searched for basic science studies on adipocytes in which efavirenz was used. Both clinical and adipocyte studies were included. In vitro data Adipocyte differentiation and metabolism seem to be very sensitive to HIV-1 infection, antiretroviral drugs or both. Historically, PIs were the first drugs associated with lipodystrophy and related metabolic disorders, whereas nucleoside analogues (mainly thymidine analogues) were later linked to the development of lipoatrophy. Several mechanisms have been postulated to explain the effects of PIs. Given that it is difficult to determine the adverse effects of a single drug on adipose tissue in patients taking combined antiretroviral treatment, in vitro models (mainly with mouse cells) have been used to examine the effects of these drugs on adipocyte metabolism and differentiation. The potential mechanisms of the deleterious effects of PIs on fatty tissue include a negative effect on pre-adipocyte differentiation, impaired adipogenesis, promotion of adipocyte death, increased lipolysis and impairment of insulin signalling events and glucose transport. The inhibitory effects of PIs on adipogenesis have been related to the decreased expression of peroxisome proliferator-activated receptors or CCAAT enhancer-binding proteins (C/EBPa) or to altered expression or intranuclear localization of sterol regulatory element-binding protein (SREBP-1c) ,45 54 Nevertheless, the effects of PIs on the physiology of human adipocytes are not the same for all the drugs in this group. In this way, atazanavir does not seem to negatively affect adipocyte lipid synthesis or glucose transport, unlike ritonavir, lopinavir, saquinavir and nelfinavir. 55 Information on the effect of nevirapine and efavirenz on adipocyte metabolism is very scarce. In a study evaluating the effects of nevirapine on brown adipocytes, this drug did not significantly impair cell biology and to some extent promoted adipocyte differentiation. 56 In another study evaluating the effect of switching from indinavir to nevirapine on insulin metabolism and adipocyte apoptosis, only an improvement in the glucose/ insulin ratio was demonstrated. Subcutaneous adipocyte apoptosis continued to occur in lipoatrophic areas, suggesting that this strategy cannot reverse lipoatrophy. 57 To our knowledge, there is only one study on the role of efavirenz in this situation. El Hadri et al. 58 attempted to analyse the effects of efavirenz on adipocyte differentiation and metabolism using an in vitro assay with murine (3T3-F442A) cells and human adipocytes exposed to efavirenz at different concentrations (0 50 mm). When incubated with pre-adipocytes during the differentiation process, efavirenz prevented these cells from accumulating lipids, and at high concentrations (40 50 mm), it also altered the magnitude of adipocyte differentiation. The latter effect was detectable even at low drug concentrations and did not seem to be related to activation of lipolysis, alteration in cell viability (this phenomenon was reversible upon drug withdrawal) or reduced baseline or insulin-stimulated glucose transport that would result in a decreased substrate availability for fatty acid biosynthesis. Nevertheless, the antilipogenic effect of efavirenz is the result of a dramatic down-regulation in the SREBP-1c expression through decreased mrna levels and, consequently, the expression of the related mature 68 kda protein. Because the mature form of SREBP-1c promotes the lipogenic gene expression, the significant decrease in the levels of the 68 kda SREBP-1c protein probably contributes to impaired lipogenesis in efavirenz-treated cells. In this experiment, the inhibitory effect of efavirenz on cell triglyceride accumulation could be reversed by providing the cells with free fatty acids and by overexpression of a dominant positive form of SREBP-1c. As the authors suggested, the effect of efavirenz on adipose tissue could be counteracted by de novo hepatic lipid synthesis. Moreover, although efavirenz can accumulate in the adipose tissue 59 and intracellular concentrations seem to be higher than plasma concentrations, 60 the concentrations used in this experiment were higher than those expected at the therapeutic levels. The average plasma C max in clinical studies ranges from 9.1 to 12.9 mm levels over 12.7 mm are considered toxic and 3.1 mm is considered the minimum efficacious concentration. Further clinical investigation is warranted in this field to ascertain the medical relevance of these observations and whether efavirenz could also alter lipogenesis in hepatocytes. Recently, Hammond et al. 66 presented data regarding the effects of efavirenz and lopinavir on adipose tissue in HIV-infected patients initiating the first-line therapy. Biopsy samples were assessed for mitochondrial DNA and tissue histology, as well as for cytokine expression. There was no evidence that the use of efavirenz or lopinavir affected mitochondrial DNA depletion or cytokine expression in the adipose tissue, whereas thymidine analogues were strongly associated 235

3 with mitochondrial DNA depletion and pathological changes characteristic of clinical lipoatrophy. In summary, in vitro studies show that both PIs and NNRTIs may have an impact on adipocyte metabolism. Although the evidence is greater for PIs, information on the potential role of NNRTIs remains scarce and needs further evaluation. Clinical data Lipoatrophy has mainly been associated with NRTIs, especially thymidine analogues and, more specifically, stavudine. 3,7,18,19,67,68 Therefore, several studies have explored the possibility of reversing lipoatrophy by switching from thymidine analogues to thymidine-sparing combinations. Nevertheless, once lipoatrophy is established, the results up to 1 2 years have not been promising and recovery of limb fat has been limited On the contrary, the good tolerability and ease of use of NNRTIs meant that they were initially explored for reducing the toxicity and complexity of PI-based therapy. 22,77 80 This, together with their antiviral efficacy, has made the combination of efavirenzþ2 NRTIs one of the preferred combinations for naive patients. 81,82 Unlike NRTIs or PIs, NNRTIs have not been directly related to lipodystrophy (specifically lipoatrophy) in cohort studies or clinical trials. To assess the potential role of efavirenz compared with PIs in the emergence of lipodystrophy, we reviewed comparative and non-comparative published clinical studies in which an objective measure of fat distribution, such as DEXA, was performed. Non-comparative studies Body fat has been measured by DEXA in five non-comparative trials in which efavirenz was the anchor drug: ABCDE 83 (stavudineþlamivudineþefavirenz versus abacavirþlamuvudine þefavirenz), Gilead Study (tenofovirþemtricitabineþ efavirenz versus zidovudineþlamivudineþefavirenz), ACTG 5125s 85 (simplification to 2 NRTIsþefavirenz versus lopinavirþ efavirenz), Gilead (tenofovirþlamivudineþefavirenz versus stavudineþlamivudineþefavirenz) and Sweet 87 (switch from zidovudine/lamivudineþefavirenz to tenofovir/emtricitab ineþefavirenz versus zidovudine/lamivudineþefavirenz) (Table 1). These studies show the expected impact of body fat on the population treated with efavirenz and NRTIs Three involved naive patients and two involved a simplification strategy. In two of the studies (Gilead 903 and Gilead 934), DEXA was not used at the baseline visit, although body fat changes were measured on two occasions during therapy with an interval of 48 weeks between measurements. The follow-ups were long and, except for one case, there were evaluations at 96 weeks. In two of the studies, a clinical evaluation of lipoatrophy was also carried out. In general, thymidine-analogue-sparing regimens showed a continuous weight gain up to week 144. In the ABCDE study, the weight gain was 913 g in the abacavir (ABC) arm compared with a loss of 1578 g in the stavudine arm. In ACTG 5125s, the gain was 782 g in the NRTI-sparing arm (a 10% increase on baseline) compared with a loss of 850 g (15%) in the arm including NRTIs with efavirenz. In ACTG 5142s, patients simplified therapy after at least 18 months on successful HAART (median time on antiretroviral therapy in the efavirenzþ2 NRTIs arm was 127 weeks) and the NRTI pairs used in 25 of the 31 patients included thymidine analogues (stavudine 6; zidovudine 19). In Gilead 903 and 934, baseline DEXA was not performed because the design of the body composition substudy was established after the initiation of the main study. However, the results of these studies do not show whether body fat remained stable or not with prolonged antiretroviral therapy. Once again, the tenofovir-containing arms, unlike those containing stavudine or zidovudine, continued to show a significant gain in weight (tenofovir 300 g versus zidovudine 700 g in Gilead 934 and tenofovir 2900 g versus stavudine 600 g in Gilead 903). Furthermore, significant differences were also observed in total limb fat between the groups including stavudine or zidovudine compared with tenofovir at week 96: tenofovir 7.7 kg (n ¼ 144) versus zidovudine 5.5 kg (n ¼ 136) (P, 0.001). There were no significant differences in trunk fat between the groups. Diagnosis of lipodystrophy was based on the physician s evaluation. In Gilead 903, clinical lipodystrophy was included among the adverse effects. In both studies, the incidence of lipodystrophy in the thymidine-analogue-sparing groups was very low: tenofovir 3% versus stavudine 19% at 144 weeks in Gilead 903 (P, 0.001) and abacavir 4.8% versus stavudine 38.3% at 96 weeks in the ABCDE study (P, 0.001). Data from studies in which efavirenz is used as an anchor drug consistently show that when the backbone is composed of non-thymidine analogues, the long-term incidence of lipodystrophy is very low, with a significant gain in weight that continues at least until week 144. Even in those patients treated with zidovudine/lamivudine for a median of 2.9 years, the change to tenofovir/emtricitabine improves the limb fat by an average of 261 g. The development of clinical lipoatrophy in patients treated with efavirenz, after 2 3 years of therapy, has ranged from 15% to 38% when a thymidine analogue is used and from 3% to 4.8% when abacavir or tenofovir is used. Similarly, limb fat changes have ranged from a loss of 1578 g to a gain of 600 g in thymidine analogue-treated patients compared with a net gain of g in patients not treated with thymidine analogues. Comparative studies Four prospective randomized studies have compared alterations in body fat distribution using DEXA: ACTG 5005s 88 (efavirenz versus nelfinavir versus efavirenzþnelfinavir), BMS (efavirenz versus non-boosted atazanavir), M (simplification to lopinavir monotherapy after an induction period of weeks of zidovudine/lamivudineþlopinavir versus zidovudine/ lamivudineþefavirenz) and ACTG (efavirenz versus lopinavir versus efavirenzþlopinavir). In all four, the backbone included thymidine analogues, although in ACTG 5142, 34% of the patients received tenofovir instead of stavudine or zidovudine (in ACTG 5142, allocation of NRTIs was not randomized), and in M03-613, patients assigned to lopinavir monotherapy discontinued zidovudineþlamivudine. All patients underwent a DEXA scan at the baseline visit. This was compared with the follow-up scans at week 48, 96 or 144 (Table 1). The patients were mainly males aged In ACTG 5142, the patients had a lower CD4 count, were older and the number of non-white patients was greater than in the other two studies (64% in ACTG 5142 versus 55% in BMS-034 and 49% in ACTG 5005s). Limb 236

4 Table 1. Study characteristics Study (ref.) Design Intervention n Measurement BL characteristics Outcome ABCDE 83 open-label trial naive pts starting HAART: d4tþ3tcþefv versus ABCþ3TCþEFV DEXA substudy d4t: 25 ABC: 32 DEXA at BL, and 48 and 96 weeks; performed and read centrally in a blinded fashion median CD4 cells/mm 3 median VL (log) weight (kg) BMI (kg/m 2 ) ABC d4t net limb fat gain at week 96: ABC 913 g versus d4t g (P, 0.001) fat loss in one or more location at week 96: ABC 4.8% versus d4t 38.3% (P, 0.001) fat accumulation in one or more location at week 96: ABC 10.7% versus d4t 24.7% (P ¼ 0.023) GILEAD open-label trial naive pts starting HAART: TDFþFTCþEFV versus ZDVþ3TCþEFV DEXA substudy a DEXA at 48 and 96 weeks (not at BL); read centrally in a blinded fashion TDF: 49 ZDV: 44 median CD4 cells/mm 3 median VL (log) TDF ZDV difference in limb fat between week 96 and week 48: ZDV: 20.7 kg (P ¼ 0.001); TDF: 0.3 kg (P ¼ 0.01) absolute change in limb fat: ZDV: 6.0 kg at 48 weeks; 5.5 kg at 96 weeks (P ¼ 0.001) TDF: 7.4 kg at 48 weeks; 8.1 kg at 96 weeks (P ¼ 0.01) 237 ACTG 5125s 85 GILEAD open-label trial double-blind trial pts on their first HAART regimen with VL,200 copies/ml for at least 18 months who simplified to EFVþ2 NRTIs b, EFVþLPV/r naive pts starting HAART: TDF/d4T-Pboþ3TCþEFV, d4t/tdf-pboþ3tcþefv NRTIs: 31 LPV/r: 31 TDF: 128 D4T: 134 DEXA at BL, week 48 and the last visit of follow-up (median 102 weeks); results analysed centrally lipodystrophy (investigator-defined) was assessed in all participating pts; DEXA was performed and read locally in a subset of pts at 96 and 144 weeks median CD4 cells/mm 3 VL (,200 copies/ml) weight (kg) limb fat (kg) mean age (years) mean CD4 cells/mm 3 mean VL (log) weight (kg) NRTIs TDF LPV/r D4T median change in limb fat (week 48): NRTIs: 2242 g (24%); LPV/r: 562 g (þ6%) (P ¼ 0.086) median change in limb fat (at the time of the last DEXA): NRTIs: 2850 g (215%); LPV/r: 782 g (þ10%) (P ¼ 0.002) trunk fat remained stable in both groups investigator-defined lipodystrophy at week 144: d4t: 19% (n ¼ 301); TDF: 3% (n ¼ 299) (P, 0.001) limb fat at week 96: d4t: 5.0 kg; TDF: 7.9 kg (P, 0.001) limb fat at week 144: d4t: 4.5 kg; TDF: 8.6 kg (P, 0.001) mean fat gain through week 144: d4t: 0.6 kg; TDF: 2.9 kg (P ¼ 0.001) Review Sweet 87 open-label trial subjects stable on CBVþEFV (HIV RNA,50 copies/ml for 6 months) switched to: TVDþEFV or continued CBVþEFV CBV: 38 TRV: 36 DEXA at BL and week 48 median CD4 cells/mm 3 VL (,50 copies/ml) (%) median BMI (kg/m 2 ) median limb fat (kg) TVD CBV median change in total limb fat at week 48: TVD: þ261 g; CBV: 2187 g (P ¼ 0.02) whole body fat at week 48 increased equivalently in both arms ACTG 5005s 88 double-blind trial naive pts starting HAART: EFV/NFV-Pbo or NFV/ EFV-Pbo or NFVþEFV plus ZDVþ3TC or ddiþd4t EFV: 51 NFV: 47 NFVþEFV: 59 DEXA at BL and every 16 weeks up to week 144; analysed centrally in blinded fashion median CD cells/mm 3 median VL (log) BMI (kg/m 2 ) total body mass (kg) median limb fat median trunk fat EFVþNFV EFV NFV median percentage and absolute change from BL at week 144: limb fat changes (P values for comparisons from BL) EFV: 2.4% (0.2 kg) P ¼ 0.90 NFV: 223.8% (20.8 kg) P ¼ 0.69 NFVþEFV: 221.4% (22.0 kg) P ¼ 0.05 combined NFV/NFVþEFV: 223.8% (21.9 kg) P ¼ 0.05 Trunk fat changes (P values for comparisons from BL) EFV: 32% (3.5 kg) P ¼ 0.01 NFV: 2.7% (0.3 kg) P ¼ 0.4 Continued

5 Table 1. Continued Study (ref.) Design Intervention n Measurement BL characteristics Outcome 238 BMS randomized, multicentre, double-blind trial naive pts starting HAART: ZDVþ3TC þ ATV (400 mg)/ EFV-Pbo or EFV/ATV-Pbo ATV: 77 DEXA at BL and week 48, analysed centrally; NFVþEFV: 6.9% (0.6 kg) P ¼ 0.5 combined NFV/NFVþEFV: 6.4% (0.3 kg) P ¼ 0.3 measurements: VAT/SAT/ TAT (VAT, SAT and TAT) ATV ATV BL week 48 EFV: 63 median age (years) VAT (mean kg) SAT (mean kg) median CD4 cells/mm TAT (mean kg) median VL (log) EFV BL week 48 mean weight (kg) 68 67s VAT (mean kg) BMI (kg/m 2 ) SAT (mean kg) TAT (mean kg) there were no significant differences between the ATV and EFV arms M open-label trial naive patients starting HAART: ZDVþ3TCþLPV/r weeks (induction) followed by LPV/r monotherapy. ZDVþ3TCþEFV throughout all study period LPV/r: 74 EFV: 32 DEXA at BL and every 24 weeks until week 96; lipoatrophy was defined as.20% loss of limb fat lipohypertrophy was defined as.20% increase in trunk fat mass median CD cells/mm 3 mean VL (log) mean weight (kg) mean trunk fat (kg) mean limb fat LPV/r EFV EFV median change in limb fat at 96 weeks: LPV/r: 18%; EFV: 29% (P, 0.001) median change in trunk fat at 96 weeks: LPV/r: 14%; EFV: 15% (P. 0.05) lipoatrophy (.20% loss extremity fat) at 96 weeks: LPV/r: 5%; EFV: 34% (P, 0.001) lipohypertrophy (.20% increase in trunk fat) at 96 weeks: LPV/r: 45%; EFV: 44% (P. 0.99) lipoatrophy and lipohypertrophy at 96 weeks: LPV/r: 0%; EFV: 16% (P ¼ 0.002) Review ACTG open-label trial naive pts starting HAART: LPV/rþEFV, EFVþ2NRTIs c, LPV/ rþ2nrtis c LPV/rþEFV: 234 EFV: 225 LPV/r: 234 DEXA at BL, and 48 and 96 weeks analysed centrally; lipoatrophy was defined as.20% loss of limb fat median CD cells/mm 3 VL.10 5 (%) trunk fat (kg) median limb fat EFVþLPV/r EFV LPV/r median change from BL to 96 weeks limb fat LPV/rþEFV 18%; EFV 1.4%; LPV/r 9.8%, P0.01 for all pairwise comparisons between groups trunk fat LPV/rþEFV 17%; EFV 12%; LPV/r 19%; there were no significant differences between groups lipoatrophy (.20% loss extremity fat) at week 96 LPV/rþEFV 9%; EFV 32%; LPV/r 17%, P0.013 for all pairwise comparisons between groups ABC, abacavir; ATV, atazanavir; BL, baseline; BMI, body mass index; CBV, zidovudine/lamivudine; ddi-ec, didanosine enteric coated; d4t, stavudine; EFV, efavirenz; FTC, emtricitabine; HAART, highly active antiretroviral therapy; LPV, lopinavir; NFV, nelfinavir; NRTI, nucleoside reverse transcriptase inhibitor; Pbo, placebo; pts, patients; SAT, subcutaneous adipose tissue; TAT, total adipose tissue; TDF, tenofovir; TVD, tenofovir/emtricitabine; VAT, visceral adipose tissue; VL, viral load; ZDV, zidovudine; 3TC, lamivudine. a Those patients with DEXA scan at 48 and 96 weeks. b NRTI pairs allowed were ddi-ecþ3tc, ddi-ecþzdv or ddi-ecþd4t, ZDVþ3TC and d4tþ3tc. c NRTI selected before randomization. All pairs consisted of 3TC plus one of the following: ZDV or d4t-xr or TDF.

6 fat at baseline was slightly lower in the ACTG 5005s patients, whereas in the other three studies, it was almost normal. 1,91 In ACTG 5005s, for the group as a whole, limb fat increased during the first 32 weeks and subsequently declined. Patients who received efavirenz did not show significant changes in limb fat at 144 weeks (an increase of 200 g, 2.4% from baseline; P ¼ 0.90). However, in those patients who received nelfinavircontaining regimens, a weight loss of 1900 g corresponding to 23.8% from baseline (P ¼ 0.05) was observed. After week 32, there was an additional decrease in limb fat of 28.7% per year in the combined nelfinavir and nelfinavirþefavirenz group when compared with efavirenz alone, after adjusting for the nucleoside backbone (P ¼ 0.03). If those patients who received didanosineþstavudine are excluded and only those who were treated with zidovudineþlamivudine are considered, after week 32, limb fat changed by 2.7% per year with efavirenz and by 27.9% per year for the combined nelfinavir and nelfinavirþ efavirenz group (P ¼ 0.03). Trunk fat increased during the first 32 weeks and was maintained over time. The only group that presented a significant increase in body fat was the efavirenz group: 3500 g (32%) (P ¼ 0.01) versus nelfinavir 300 g (2.7%) (P ¼ 0.4) versus nelfinavirþefavirenz 600 g (6.9%) (P ¼ 0.5). If we focus only on patients receiving zidovudineþlamivudine, after week 32, trunk fat changed by 18.9% per year with efavirenz and by 23.5% per year for the combined nelfinavir and nelfinavirþefavirenz arm (P, ). In BMS-034, visceral, subcutaneous (appendicular) and total adipose tissue (VAT, SAT and TAT, respectively) did not change significantly within the efavirenz and atazanavir groups between the baseline and week 48 visit, and there were no differences between the efavirenz and atazanavir arms. As for the SAT to TAT ratio and the VAT to TAT ratio, no significant changes from baseline were observed in either treatment group. Although computed tomography results were comparable for the two treatment arms at baseline, there was a modest but statistically significant increase for the atazanavir group compared with the efavirenz group in VAT (33% versus 23%) and TAT (21% versus 10%). Nevertheless, in this study, the follow-up was only made up to 48 weeks, thus making it impossible to draw longterm conclusions. In M03-613, a significant improvement in limb fat was observed at 96 weeks in the lopinavir monotherapy arm (18% versus 29%), whereas it was not observed for the trunk fat. Nevertheless, differences were more evident from 24 and 48 weeks onwards, maybe reflecting the effect of zidovudine discontinuation at that point. ACTG 5142 compared three class-sparing regimens for naive subjects: lopinavir (soft-gel)þefavirenz, lopinavirþ2 NRTIs and efavirenzþ2 NRTIs. NRTIs were selected before randomization and the choice was zidovudine 42%, stavudine-xr 24% or tenofovir 34%. The study differs from previous trials in several ways, for example, the definition of lipoatrophy (not used previously in any clinical trial) or the non-randomized choice of NRTI. The median change from baseline through week 96 in limb fat was significantly different between the groups: lopinavirþefavirenz 18%, lopinavir 9.8% and efavirenz 1.4% (P 0.01 for all pairwise comparisons). However, the median change in the trunk fat did not disclose significant differences in any group: lopinavir 19%, lopinavirþefavirenz 17% and efavirenz 12%. The incidence of lipoatrophy (defined as.20% loss in limb fat) at week 96 was significantly different between the groups: efavirenz 32%, lopinavir 17% and lopinavirþefavirenz 9% (P values for pairwise comparisons: lopinavirþefavirenz versus lopinavir, P ¼ 0.013; lopinavirþefavirenz versus efavirenz, P, 0.001; and lopinavir versus efavirenz, P ¼ 0.007). If we look at the differences in the incidence of lipoatrophy at 96 weeks with regard to the NRTIs used, the highest percentage was observed with the thymidine analogues stavudine (42%) and zidovudine (27%), and the lowest percentage was observed with tenofovir (9%). All pairwise comparisons of NRTIs were significant: zidovudine versus tenofovir, P, 0.001; stavudine versus tenofovir, P, 0.001; and stavudine versus zidovudine, P ¼ Similarly, the incidence of lipoatrophy was greater in the efavirenz and lopinavir groups when these were associated with thymidine analogues (stavudine. zidovudine. tenofovir) (Table 2). A logistic regression model including a randomized arm and NRTIs was applied to the week 96 lipoatrophy data for the NRTI-containing regimens. For lipoatrophy, this analysis showed an odds ratio (OR) of 2.7 (95% CI: ) for efavirenz versus lopinavir (P, 0.001), 1.9 (95% CI: ) for stavudine versus zidovudine (P ¼ 0.029) and 0.24 (95% CI: ) for tenofovir versus zidovudine (P, 0.001). There were no interactions analyses to examine whether the effect of each NRTI was significantly different for efavirenz or lopinavir. The most surprising aspect of these results is the analysis according to the number of subjects who presented a reduction of 20% or more in limb fat (taking this cut-off point as that which defines lipoatrophy). Although there is no objective definition of lipoatrophy, an incidence of 32% at 96 weeks seems to be well above that reported for patients treated with efavirenz, even more if we bear in mind that one-third of these patients received tenofovir. Furthermore, the number of patients with a limb fat loss 20% depended to a large extent on the NRTIs that formed the backbone: 51% with stavudine-xr, 49% with zidovudine and 12% with tenofovir. This behaviour was also observed with lopinavir, although less markedly: 33% for stavudine-xr, 16% for zidovudine and 6% for tenofovir. In this sense, the only regimen with no NRTI presented a significantly lower incidence (only 9%). In ACTG 5142, higher baseline CD4 cell count and lower gain in the body and trunk weight were associated with lipoatrophy and smaller gain in limb fat, independently of the antiretroviral regimen used. Analysing fat change as a continuous variable also found male sex, non-aids patient, lower baseline extremity fat and smaller increases in cholesterol and LDL cholesterol to be significantly associated with a smaller gain in limb fat. It seems that several factors other than antiretroviral therapy may be important in the pathogenesis of fat loss. 92 How should these results be interpreted? First, we still need information on the baseline characteristics of the patients (body Table 2. Patients with.20% fat loss in efavirenz and lopinavir/ ritonavir arms in ACTGs 5142 d4t ZDV TDF Total EFV (n) 51% (44) 40% (63) 12% (67) 32% (171) LPV/r (n) 33% (43) 16% (73) 6% (50) 17% (166) d4t, stavudine; EFV, efavirenz; LPV/r, lopinavir/ritonavir; TDF, tenofovir; ZDV, zidovudine. 239

7 mass index, weight or dispersion of body mass values) to help us interpret correctly the variations observed. A gain in fat in patients who start with normal values may not be desirable or clinically irrelevant in this sense, the patients in ACTG 5142 were within the normal range for limb fat values (7.1 kg). 28,91,93, Furthermore, even if the definition of lipoatrophy is objective because it is based on DEXA, it is not validated as a measure of the presence of clinically significant lipoatrophy or even as a predictor of the development of lipoatrophy. In fact, patients with evident lipoatrophy have limb fat levels that are 40 50% lower than those in healthy patients; 71,73,75,85,86 those with clinically apparent lipoatrophy have at least 30% limb fat loss regardless of baseline total fat or weight 94 and anthropomorphic changes correlate modestly with DEXA measurements. 89 Even in ACTG 5142, only 30% of the subjects with lipoatrophy as defined by DEXA reported fat loss. 92 Therefore, a cut-off at 20% could be too sensitive in clinical terms. 69 In addition, measurement by DEXA does not distinguish between subcutaneous and intra-abdominal fat. 73,95 As far as thymidine analogues are concerned, it seems obvious that there is a clear relationship between the inclusion of these drugs in the antiretroviral regimen and a greater incidence of limb fat loss 20%. In fact, this was very low (9%) at 96 weeks in the NRTI-sparing group. Similarly, in ACTG 5125s, those patients whose regimens were simplified to lopinavirþefavirenz showed an average gain of 10% (782 g) in limb fat, despite the fact that they had been on thymidineanalogue-containing HAART for 18 months. What ACTG 5142 did reveal was a greater effect of thymidine analogues in the efavirenz group than in the lopinavir group. The OR of efavirenz versus lopinavir for developing lipoatrophy at 96 weeks was 2.7 (95% CI: ). Nevertheless, no interaction test was carried out to determine whether the effect of the NRTIs (mainly thymidine analogues) was significantly greater in one group than in the other. What seems obvious is that if efavirenz alone had a potentially important role in the development of lipoatrophy, an impact on body fat would also have been observed in the lopinavirþefavirenz group. The behaviour of efavirenz may have been more neutral with respect to limb fat, in that lopinavir was more likely to be associated with fat gain and NRTIs with fat loss. This effect of greater fat gain in the limbs in the PI-containing regimens than in the PI-sparing regimens has also been observed in other studies. 28 In summary, the results of comparative, randomized clinical trials seem to indicate that therapy with efavirenz leads to a gain in the limb fat in all cases. This gain, although modest (1.4% 9.3%), is higher than that produced by nelfinavir, similar to that of atazanavir and lower than that observed with lopinavir. Trunk fat also increases (3.2 32%), although with no significant differences with the PI comparator. Surprisingly, the incidence of lipodystrophy (defined as.20% loss in limb fat) in the ACTG 5142 trial was significantly higher than that of lopinavir (32% versus 17%) and was severely affected by the accompanying NRTI. Data for lipid outcome In general, it seems that the lipid profile worsens when the nucleoside backbone includes a thymidine analogue (more evident with stavudine than with zidovudine) than when it includes tenofovir or abacavir (Table 3). With tenofovir, in particular, plasma triglyceride levels remained unchanged or experienced minor increases. Cholesterol increased moderately with higher increases in HDL-C than in total cholesterol (TC), suggesting an improvement in the atherogenic profile. When efavirenz was combined with lopinavir or NRTI as the backbone in patients switching therapy, the lipid profile was significantly worse in the lopinavir group. 85 The fact that all patients switched from PI-containing regimens probably contributes to this effect. Increases in triglycerides of over mg/dl were more common with thymidine analogues (14 16% versus 3 4%) or lopinavir (8% versus 4%) than with the comparator, and the use of lipid-lowering agents was also more frequent with NRTIs (16 17% versus 4 11%). As for comparative trials between efavirenz and PIs, comparison with nelfinavir disclosed a slightly better lipid profile for efavirenz with a higher increase in HDL-C leading to a lower TC/HDL-C ratio. When compared with atazanavir (nonboosted), efavirenz showed a neutral lipid profile, with a slight increase in cholesterol levels and a stable atherogenic index (TC/HDL-C), whereas atazanavir improved its lipid figures in comparison with baseline. The figures for lipid outcome in ACTG 5142 were equivalent for efavirenz and lopinavir with regard to TC, HDL-C and non-hdl-c, whereas increases in triglycerides were significantly higher in the lopinavir arm. The worst performing arm was the NRTI-sparing regimen, which showed significant elevations in all the parameters tested. Efavirenz generally behaves neutrally with respect to lipids (depending to a large extent on the accompanying NRTI), with increases in cholesterol that are usually accompanied by increases in HDL-C; therefore, the atherogenic index remains stable or even improves slightly. Therapy involving lopinavir usually leads to an increase in triglyceride values that is not always accompanied by an increase in cholesterol, whereas NRTI-sparing PI-containing regimens typically produce a much higher deterioration in lipids than that observed with 44,83 86,89, lopinavirþ2 NRTIs. In this sense, the progression of lipids in ACTG 5142 did in fact agree with the information from previous studies, and the greatest increases were detected with the NRTI-sparing regimens, whereas the lopinavir group had a significant increase in triglycerides in comparison with the efavirenz group. Nevertheless, we do not have the necessary number of patients with grades 3 and 4 cholesterol or increased triglyceride levels to enable us to provide a more accurate evaluation of these increases. In summary, the efavirenz lipid profile seems to be neutral and is substantially affected by the NRTI backbone, the most favourable values being observed with tenofovir and abacavir. When compared with other PIs, efavirenz showed a profile that was very similar to that of nelfinavir (worse for cholesterol) or lopinavir (worse for triglycerides) and slightly worse than that of non-boosted atazanavir. Conclusions and clinical implications Lipodystrophy remains a major problem for HIV-infected patients. Aside from the metabolic complications that potentially contribute to increased cardiovascular risk, the change in the body appearance can be very stigmatizing. Therefore, any new data in this field are extremely relevant, as shown by the metabolic substudy of ACTG

8 Table 3. Lipid outcome in clinical studies Study (ref.) Lipid parameters outcome Comments ABCDE 83 at 96 weeks, both treatments were associated with significant (P, 0.05) mean changes from BL in: in comparison with Tx, ABC pts showed a lower increase in TG greater increase in HDL-C (P ¼ 0.001) and a greater reduction in TC/HDL-C (P ¼ 0.03), a ratio (P ¼ 0.005); use of lipid-lowering agents was d4t ABC significantly higher in d4t than the ABC group: 17% versus 4% TC (mg/dl) LDL-C (mg/dl) HDL-C (mg/dl) TG (mg/dl) TC/HDL-C ratio GILEAD at week 96, pts in the TDF group had a significantly (P, 0.05) lower mean increase in TC and HDL-C: there were no differences in treatment emergent TG 400 md/dl: TDF 15% versus ZDV 16% TDF ZDV TC (mg/dl) LDL-C (mg/dl) HDL-C (mg/dl) TG (mg/dl) 0 30 ACTG 5125s 85 at week 48, pts in LPV/r group had a significantly (P, 0.05) greater median increase in TG, TC and non-hdl-c: LPV/r NRTIs TC (mg/dl) non-hdl-c (mg/dl) TG (mg/dl) GILEAD at week 144, pts in TDF group had a significantly (P, 0.05) lower mean increase in TG, TC and LDL-C and higher increase in HDL-C: ACTG 5005s 96 BMS TDF d4t TC (mg/dl) LDL-C (mg/dl) HDL-C (mg/dl) 9 6 TG (mg/dl) 1 34 at week 64, median increase in HDL-C was significantly (P ¼ 0.005) higher for EFV versus NFV group (resulting in a lower TC/HDL-C ratio for EFV), whereas median increases in TC, TG and non-hdl-c were similar for both groups at week 48, ATV was not associated with statistically significant increases in LDL-C or TC (not stated for EFV) mean change from BL (%): ATV EFV TC 1 20 LDL-C 0 17 HDL-C 11 1 TG 266 M both LPV/r and EFV therapy increased TC, LDL-C and HDL-C; these increases generally occurred early in therapy (by week 24) and remained stable thereafter; TG seem to increase more with LPV/r than with EFV therapy (P ¼ 0.06), Apo C3 with LPV/r (P ¼ 0.004) and hs-crp with EFV (P ¼ 0.009) median change from BL (mg/dl): changes in HDL-C and free fatty acids were comparable between groups; percentage of pts with TG.500 mg/dl at week 48 was 8% and 4% for the LPV/r and NRTI groups, respectively increases in TGgrade 3 were significantly higher in d4t compared with ABC group: 3% versus 14%; use of lipid-lowering agents was significantly higher in d4t than the TDF group: 16% versus 11% there were no between-arm differences in the percentage of pts with TC.200 mg/dl, LDL-C.130 mg/dl or TG.200 mg/dl; 1 pt in the NFV group and 4 pts in the EFV group started lipid-lowering treatment pts who took lipid-lowering agents were censored from this analysis; ratio of TC/HDL-C remained stable in EFV pts during the study period (4.3 to 4.3) and decreased in ATV pts (4.4 to 4.0); significance in differences between groups not stated discontinuation of CBV did not affect lipid changes in LPV/r treated subjects LPV/r EFV TC TG LDL-C HDL-C LDL:HDL ratio Apo A Apo B Apo C Hs-CRP Continued 241

9 Table 3. Continued Study (ref.) Lipid parameters outcome Comments ACTG there was no significant difference between EFV and LPV/r in the change from BL to week 96 for total, HDL and non-hdl cholesterol levels, whereas there was a significant difference for TG (P ¼ 0.006); a significantly higher proportion of lipid elevations was observed in the LPV/rþEFV arm median change at 96 weeks (mg/dl): EFV/LPV/r LPV/r EFV TC HDL-C non-hdl-c TG by week 96, 10%, 12% and 26% of EFV, LPV and LPV/EFV subjects had used a lipid-lowering agent ABC, abacavir; BL, baseline; CBV, zidovudine/lamivudine; d4t, stavudine; EFV, efavirenz; HDL-C, HDL cholesterol; hs-crp, high-sensitivity C-reactive protein; LDL-C: LDL cholesterol; LPV, lopinavir; pts, patients; TC, total cholesterol; TDF, tenofovir; TG, triglycerides; Tx, treatment; ZDV, zidovudine. ACTG 5142 compared the efficacy of two of the preferred regimens for naive patients, efavirenzþ2 NRTIs and lopinavirþ2 NRTIs, with the NRTI-sparing regimen efavirenzþ lopinavir. This study showed the greater efficacy of efavirenzþ2 NRTIs over lopinavirþ2 NRTIs ( proportion without virological failure at 96 weeks: 76% versus 67%) with no significant differences in the comparison with efavirenzþlopinavir (73%). CD4 lymphocyte recovery was significantly better with lopinavir than with efavirenzþ2 NRTIs (285 versus 241 cells/mm 3 ) and similar to that of lopinavirþefavirenz (268 cells/mm 3 ). In the metabolic substudy, the incidence of lipoatrophy, defined as 20% loss of limb fat, was significantly higher in the efavirenzþ2 NRTIs groups than in the lopinavirþ2 NRTIs and efavirenzþlopinavir groups (32% versus 17% versus 9%). To date, several studies that measure fat variation using DEXA have revealed that the potential role of efavirenz in the development of lipodystrophy is minimal or non-existent and that it depends on the NRTIs that form the backbone. The same cannot be said for lopinavir because before ACTG 5142 the effect of standard regimens containing NRTIs plus lopinavir on body fat was unknown. Direct comparisons of efavirenz with other third-arm drugs have been made in clinical trials with PIs. Patients treated with efavirenz in these studies gained weight and did not show significant losses of limb fat. The same occurred when the comparator was atazanavir, and only when patients received nelfinavir concomitantly was a significant reduction in limb fat observed. The results of ACTG 5142 are consistent with this: a 1.4% gain in limb fat was detected at week 96 for patients receiving efavirenz. Even though the incidence of lipodystrophy has been significantly reduced by the avoidance of thymidine analogues and the availability of new antiretrovirals, it continues to be a serious problem for those who suffer from it. Therefore, it is essential to have long-term data on the potential for developing lipodystrophy of the most commonly used combinations and of those that include new drugs. To do this, we must have standardized criteria to objectively evaluate lipodystrophy. These criteria would enable us to compare the results of different studies using the same conditions for all, evaluate the long-term outcome of therapeutic and preventive measures, and even design studies to predict the appearance of morphological alterations associated with HAART. Transparency declarations J. A. P.-M. has received grant support and has served as a consultant for Bristol-Myers Squibb, Boehringer-Ingelheim, Gilead and GSK. P. D. has received grant support and has served as a consultant/speaker for Bristol-Myers Squibb, Glaxo Smith Kline, Boehringer Ingelheim, Roche, Abbott, Janssen Cilag, Gilead Sciences, Pfizer and Merck Sharp and Dohme. E. M. has participated on advisory boards and received research grants, travel grants, or honoraria from Abbott, BMS, Boehringer- Ingelheim, Gilead, GSK and Janssen. S. M. has received grant support and has served as a consultant for Abbott Laboratories, Boehringer-Ingelheim, Bristol-Myers Squibb, Gilead, GSK, MSD, Pfizer and Roche. References 1. Carr A, Samaras K, Chisholm DJ et al. Pathogenesis of HIV-1-protease inhibitor-associated peripheral lipodystrophy, hyperlipidaemia, and insulin resistance. Lancet 1998; 351: Carr A, Samaras K, Thorisdottir A et al. Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor-associated lipodystrophy, hyperlipidaemia, and diabetes mellitus: a cohort study. Lancet 1999; 353: Saint-Marc T, Partisani M, Poizot-Martin I et al. A syndrome of peripheral fat wasting (lipodystrophy) in patients receiving long-term nucleoside analogue therapy. AIDS 1999; 13: Lichtenstein KA, Ward DJ, Moorman AC et al. Clinical assessment of HIV-associated lipodystrophy in an ambulatory population. AIDS 2001; 15: Lichtenstein KA. Redefining lipodystrophy syndrome: risks and impact on clinical decision making. J Acquir Immune Defic Syndr 2005; 39: Martinez E, Mocroft A, Garcia-Viejo MA et al. Risk of lipodystrophy in HIV-1-infected patients treated with protease inhibitors: a prospective cohort study. Lancet 2001; 357: Carr A, Miller J, Law M et al. A syndrome of lipoatrophy, lactic acidaemia and liver dysfunction associated with HIV nucleoside analogue therapy: contribution to protease inhibitor-related lipodystrophy syndrome. AIDS 2000; 14: F

10 8. Munoz-Sanz A, Rodriguez-Vidigal FF, Domingo P. Pathogenesis of lipodystrophy and metabolic syndromes associated with HIV infection. Med Clin (Barc) 2006; 127: Giralt M, Domingo P, Guallar JP et al. HIV-1 infection alters gene expression in adipose tissue, which contributes to HIV-1/ HAART-associated lipodystrophy. Antivir Ther 2006; 11: Miller KD, Jones E, Yanovski JA et al. Visceral abdominal-fat accumulation associated with use of indinavir. Lancet 1998; 351: Zhang B, MacNaul K, Szalkowski D et al. Inhibition of adipocyte differentiation by HIV protease inhibitors. J Clin Endocrinol Metab 1999; 84: Lenhard JM, Weiel JE, Paulik MA et al. Stimulation of vitamin A(1) acid signaling by the HIV protease inhibitor indinavir. Biochem Pharmacol 2000; 59: Caron M, Auclair M, Vigouroux C et al. The HIV protease inhibitor indinavir impairs sterol regulatory element-binding protein-1 intranuclear localization, inhibits preadipocyte differentiation, and induces insulin resistance. Diabetes 2001; 50: Caron M, Auclair M, Sterlingot H et al. Some HIV protease inhibitors alter lamin A/C maturation and stability, SREBP-1 nuclear localization and adipocyte differentiation. AIDS 2003; 17: Domingo P, Matias-Guiu X, Pujol RM et al. Subcutaneous adipocyte apoptosis in HIV-1 protease inhibitor-associated lipodystrophy. AIDS 1999; 13: Vernochet C, Azoulay S, Duval D et al. Differential effect of HIV protease inhibitors on adipogenesis: intracellular ritonavir is not sufficient to inhibit differentiation. AIDS 2003; 17: Wentworth JM, Burris TP, Chatterjee VK. HIV protease inhibitors block human preadipocyte differentiation, but not via the PPARg/ RXR heterodimer. J Endocrinol 2000; 164: R Smith DE, Carr A, Law M et al. Thymidine analogue withdrawal for lipoatrophic patients on protease-sparing therapy improves lipoatrophy but compromises antiviral control: the PIILR extension study. AIDS 2002; 16: Mallal SA, John M, Moore CB et al. Contribution of nucleoside analogue reverse transcriptase inhibitors to subcutaneous fat wasting in patients with HIV infection. AIDS 2000; 14: Brinkman K, Smeitink JA, Romijn JA et al. Mitochondrial toxicity induced by nucleoside-analogue reverse-transcriptase inhibitors is a key factor in the pathogenesis of antiretroviral-therapy-related lipodystrophy. Lancet 1999; 354: Lee H, Hanes J, Johnson KA. Toxicity of nucleoside analogues used to treat AIDS and the selectivity of the mitochondrial DNA polymerase. Biochemistry 2003; 42: Martinez E, Arnaiz JA, Podzamczer D et al. Substitution of nevirapine, efavirenz, or abacavir for protease inhibitors in patients with human immunodeficiency virus infection. N Engl J Med 2003; 349: Ruiz L, Negredo E, Domingo P et al. Antiretroviral treatment simplification with nevirapine in protease inhibitor-experienced patients with HIV-associated lipodystrophy: 1-year prospective follow-up of a multicenter, randomized, controlled study. J Acquir Immune Defic Syndr 2001; 27: Martinez E, Garcia-Viejo MA, Blanco JL et al. Impact of switching from human immunodeficiency virus type 1 protease inhibitors to efavirenz in successfully treated adults with lipodystrophy. Clin Infect Dis 2000; 31: Rey D, Schmitt MP, Partisani M et al. Efavirenz as a substitute for protease inhibitors in HIV-1-infected patients with undetectable plasma viral load on HAART: a median follow-up of 64 weeks. J Acquir Immune Defic Syndr 2001; 27: Hirschel B, Flepp M, Bucher HC et al. Switching from protease inhibitors to efavirenz: differences in efficacy and tolerance among risk groups: a case control study from the Swiss HIV Cohort. AIDS 2002; 16: Bogner JR, Vielhauer V, Beckmann RA et al. Stavudine versus zidovudine and the development of lipodystrophy. J Acquir Immune Defic Syndr 2001; 27: Mulligan K, Anastos K, Justman J et al. Fat distribution in HIV-infected women in the United States: DEXA substudy in the Women s Interagency HIV Study. J Acquir Immune Defic Syndr 2005; 38: Duran S, Saves M, Spire B et al. Failure to maintain long-term adherence to highly active antiretroviral therapy: the role of lipodystrophy. AIDS 2001; 15: Johnsen S, Dolan SE, Fitch KV et al. Carotid intimal medial thickness in human immunodeficiency virus-infected women: effects of protease inhibitor use, cardiac risk factors, and the metabolic syndrome. J Clin Endocrinol Metab 2006; 91: Jacobson DL, Tang AM, Spiegelman D et al. Incidence of metabolic syndrome in a cohort of HIV-infected adults and prevalence relative to the US population (National Health and Nutrition Examination Survey). J Acquir Immune Defic Syndr 2006; 43: Bergersen BM. Cardiovascular risk in patients with HIV infection: impact of antiretroviral therapy. Drugs 2006; 66: Friis-Moller N, Reiss P, Sabin CA et al. Class of antiretroviral drugs and the risk of myocardial infarction. N Engl J Med 2007; 356: Friis-Moller N, Weber R, Reiss P et al. Cardiovascular disease risk factors in HIV patients association with antiretroviral therapy. Results from the DAD study. AIDS 2003; 17: Dinges WL, Chen D, Snell PG et al. Regional body fat distribution in HIV-infected patients with lipodystrophy. J Investig Med 2005; 53: Schwenk A. Methods of assessing body shape and composition in HIV-associated lipodystrophy. Curr Opin Infect Dis 2002; 15: Martinez E, Bianchi L, Garcia-Viejo MA et al. Sonographic assessment of regional fat in HIV-1-infected people. Lancet 2000; 356: Carr A, Emery S, Law M et al. An objective case definition of lipodystrophy in HIV-infected adults: a case control study. Lancet 2003; 361: Kistorp CN, Svendsen OL. Body composition analysis by dual energy X-ray absorptiometry in female diabetics differ between manufacturers. Eur J Clin Nutr 1997; 51: Tothill P, Avenell A, Love J et al. Comparisons between Hologic, Lunar and Norland dual-energy X-ray absorptiometers and other techniques used for whole-body soft tissue measurements. Eur J Clin Nutr 1994; 48: Pritchard JE, Nowson CA, Strauss BJ et al. Evaluation of dual energy X-ray absorptiometry as a method of measurement of body fat. Eur J Clin Nutr 1993; 47: Smith DE, Hudson J, Martin A et al. Centralized assessment of dual-energy X-ray absorptiometry (DEXA) in multicenter studies of HIV-associated lipodystrophy. HIV Clin Trials 2003; 4: Yang Y, Zhu WD, Paton NI. Comparison of dual-energy X-ray absorptiometry machines for measuring fat distribution changes of HIV-associated lipodystrophy. Antivir Ther 2004; 9: Haubrich R, Riddler S, DiRienzo D et al. Metabolic outcomes of ACTG 5142: a prospective, randomized, phase III trial of NRTI-, PI-, and NNRTI-sparing regimens for initial treatment of HIV-1 infection. In: 14th Conference on Retrovirus and Opportunistic Infections, Los Angeles, February Bastard JP, Caron M, Vidal H et al. Association between altered expression of adipogenic factor SREBP1 in lipoatrophic 243