Switching Effective Antiretroviral Therapy: A Review

Transcription

1 INVITED ARTICLE HIV/AIDS Kenneth H. Mayer, Section Editor Switching Effective Antiretroviral Therapy: A Review Henning Drechsler and William G. Powderly Division of Infectious Diseases, Washington University School of Medicine, St. Louis, Missouri One approach to target the long-term metabolic toxicity and disfiguring body-shape changes associated with antiretroviral therapy is to switch one component of a regimen to an alternative drug, usually from a different class of antiretrovirals. Most commonly, substitutions have involved protease inhibitors, but the thymidine analogue nucleosides, especially stavudine, have been investigated more recently. Certain trends from these studies have emerged. First, if the patient has had sustained viral suppression, switching therapy is generally virologically safe. Second, metabolic disturbances, such as insulin resistance and dyslipidemia, appear to be at least partially reversible. Substitution of other agents for protease inhibitors has not been associated with reversal or improvement in fat redistribution. Studies in which thymidine analogue reverse-transcriptase inhibitors have been switched have reported modest improvements in peripheral lipoatrophy. Larger, controlled, long-term studies and a more standardized approach to definition of metabolic and morphological abnormalities are needed. With the clear success of antiretroviral therapy (ART) has come increasing concern about long-term adverse events. In particular, serious concerns have been raised about the metabolic side-effects and body-shape changes that have been increasingly recognized among many patients receiving ART. The metabolic effects, especially hyperlipidemia and insulin resistance (or frank diabetes mellitus), may increase the risk of cardiovascular events. The body-shape changes, both peripheral fat loss and fat gain, cause considerable distress for patients. The same holds true for the inconvenient dosing schedules and the number of pills of many antiretroviral regimens. All of these factors may lead to decreased adherence to therapy. A possible strategy to counteract this is to switch treatment for patients receiving a successful ART regimen to a simpler regimen for example, exchanging the protease inhibitor (PI) for a drug from a different class. This review will concentrate on the virological and metabolic effects of switching a component of a successful potent antiretroviral regimen. Received 15 March 2002; accepted 17 June 2002; electronically published 21 October Financial support: National Institutes of Health (grants AI and DK-59532). Reprints or correspondence: Dr. William G. Powderly, Div. of Infectious Diseases, Washington University School of Medicine, St. Louis, MO (Wpowderl@imgate.wustl.edu). Clinical Infectious Diseases 2002;35: by the Infectious Diseases Society of America. All rights reserved /2002/ $15.00 REVIEWED STUDIES In recent years, there has been a steady increase in clinical studies of patients taking effective ART who exchanged one component of an antiretroviral regimen for a different drug. These studies are referred to here as switch studies. Prospective studies with a duration of follow-up of 6 months and a sample size of 10 patients were included. Only a few of these studies have been published in peer-reviewed journals to date; most have been presented only in abstract form. The majority of the studies were not randomized. However, where conclusions were made, greater weight was given to studies in which randomization occurred, allowing a comparison with subjects who continued to receive the original therapy. Most studies investigated a switch from PI-based regimens to nevirapine-, efavirenz-, or abacavir-based regimens. We included a total of 35 PI-switch studies in our review. Metabolic substudies were included as separate studies. All PI-switch studies are summarized in tables 1 4. More recently, there has been increasing interest in studying the effects of switching thymidine analogues (particularly stavudine) on metabolism and body shape. Early data from these studies are included in table 5. VIROLOGICAL SAFETY All of the studies reported to date have examined the effect of switching therapy for patients in whom viral replication is suc- HIV/AIDS CID 2002:35 (15 November) 1219

2 Table 1. Switch studies in which the protease inhibitor (PI) was changed to abacavir (ABC). Reference Description Design No. of subjects (no. in switch arm/ no. in control arm) Median duration of follow-up, months HIV load rebound, % of switch/ control subjects Lipodystrophy at baseline, % of patients d4t, Change in body shape, % a % of patients Total Mean D, mg/dl HDL LDL Triglycerides Insulin resistance Comment [1] ABC/AZT/3TC vs. previous PI regimen RCT 209 (106/103) 12 22/22 (ITT) Clinical sign of lipodystrophy: ABC, 40r28; PI, 50r42 f29 Fasting, f9 [2] ABC/AZT/3TC vs. previous PI regimen RCT 163 (84/79) 19 14/3 (OTA; P p.01) 0 No change in weight or WHR f35 None f43 [3] RCT 87 (58/29) 6 0/15 (ITT) Adherence improved in both treatment arms [4] RCT 211 (105/106) 12 12/23 (P p.03) 35 f19 NS f Self-reported adherence improved with ABC, worsened with PI (NS) [5] Substudy from [4] 31 (16/15) 12 f50 Fasting, f39 f [6] Substudy from [4] 34 (17/17) AM: no change (WHR, skin folds); 1 clinical sign of lipodystrophy: ABC, 65r35 (77% subjectively better); PI: 71r67 (41% subjectively better) f29 at 6 months, NS D at 12 months Fasting, NS f NS f NOTE. Italic reference number indicates that the study is a substudy. AM, anthropomorphic measurements; AZT, zidovudine; d4t, stavudine; HDL, high-density lipoprotein; ITT, intention-to-treat analysis; LDL, low-density lipoprotein; NS, not significant; OTA, on-treatment analysis; RCT, randomized controlled trial; 3TC, lamivudine; WHR, waist-to-hip ratio; f, lower(ed) or decrease(d); F, higher or increase(d); r, changed to. a Percentage of patients in switch group taking d4t after the switch. 1220

3 Table 2. Switch studies in which the protease inhibitor was changed to efavirenz. Reference Design No. of subjects (no. in switch arm/ no. in control arm) Median duration of follow-up, months HIV load rebound, % of switch/ control subjects Lipodystrophy at baseline, % of patients d4t, Change in body shape, % a % of patients Total Mean D, mg/dl HDL LDL Triglycerides Insulin resistance Comment [8] RCT 346 (226/120) 12 7/15 (P p.024) 38 None (baseline level was only mildly elevated) F None F For both groups (NS; baseline only mildly elevated) Improved adherence [9] RCT 134 (69/65) 12 4/10 (NS) 42 None (baseline level was only mildly elevated) None None None (baseline only mildly elevated) [10] RCT 93 (46/47) 12 17/33 (NS) 100 AM: WHR, BMI; sonographic fat pad: face, intra-abdominal NS change F6 NS f Fasting, NS change (both ) f [11] Cohort ( Fat redistribution ) 88 CT: VAT (NS); body weight F39 F165 3/5 OGTT results normalized [12] Cohort ( Fat redistribution ) 90 AM: no change (WHR); better (self-report) None F7 None Fasting, none [13] Cohort (Lipoatrophy) AM: no change (WHR); BMI: no change None None None Fasting, none None [14] Cohort CT: VAT (NS); better, 50 NS f NS f [15] Cohort Patients better, 9 patients worse F4 f27 [16] Cohort DEXA: no sign. Change. None None None Fasting, none [17] Cohort ( Clinical or laboratory abnormality AM: no change None None None (OGGT) NOTE. AM, anthropomorphic measurements; BMI, body mass index; d4t, stavudine; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NS, not significant; OGTT, oral glucose tolerance test; RCT, randomized controlled trial; VAT, visceral adipose tissue; WHR, waist-to-hip ratio; f, lower(ed) or decrease(d); F, higher or increase(d). a Percentage of patients in switch group taking d4t after the switch. 1221

4 Table 3. Switch studies in which the protease inhibitor was changed to nevirapine (NVP). Reference(s) Description Design No. of subjects (no. in switch arm/ no. in control arm) Median duration of follow-up, months HIV load rebound, % of switch/ control subjects Lipodystrophy at baseline, % of patients d4t, Change in body shape, % a % of patients Total HDL Mean D, mg/dl LDL Triglycerides Insulin resistance Comment [18] New regimen: NVP/d4T/ddI RCT b 106 (52/54) 12 26/28 (NS) DEXA: f trunk fat (NS); AM: (f abdominal skin fold) f21 b Fasting, f53 b Improved qualityof-life scores [19] RCT 142 (104/34) 6 4/18 (ITT; P p.015) Self-report: 50% better f For both groups b Fasting, f (both groups b ) [20] Controlled trial 34 (16/18) 6 No data 100 f17 b (Baseline mod. elevated) F9 b (Baseline low); HDL particle size f20 b (Baseline mod. elevated) f b VLDL triglycerides [21, 22] Cohort AM: WHR; subjective improvement (predominantly lipoatrophy), 91 f71 f Fasting, f200 f [23] Cohort None NS trend to f NOTE. AM, anthropomorphic measurements; didanosine; DEXA, dual-energy radiographic absorptiometry; d4t, stavudine; ddi, HDL, high-density lipoprotein; ITT, intention-to-treat analysis; LDL, low-densitylipoprotein; mod, moderately; NS, not significant; RCT, randomized controlled trial; VLDL, very low-density lipoprotein; WHR, waist-to-hip ratio; f, lower or decrease; F, higher or increase. a Percentage of patients in switch group taking d4t after the switch. b Compared with baseline value, not with control group. 1222

5 Table 4. Switch studies in which 11 drug was used to replace the protease inhibitor (PI). Ref. Description Design No. of subjects (no. in switch arm/ no. in control arm) Median duration of follow-up, months HIV load rebound, % of switch/ control subjects Lipodystrophy at baseline, % of patients d4t, % a Change in body shape, % of patients Total Mean D, mg/dl HDL LDL Triglycerides Insulin resistance Comment [7] ABC or EFV Case-control cohort 334 (81 ABC/86 EFV/167 PI) 12/16 15% ABC vs. 4.7% (EFV), 5.4% (PI) (P!.02) 27% (EFV group) None (EFV group, n p 51) f338 (BL elevated; EFV group, n p 51) [24] NVP or EFV vs. PI RCT 77 (NVP, 26; EFV, 25; PI, 26) 12 NVP, 4;EFV,8;PI, 8 (NS) No improvement by DEXA NVP, f23 b ; EFV, none f NVP, f45 b ; EFV, NS F Improved QOL scores b ; good adherence all3arms [25] ABC/NVP/ADV/ HU vs. PI RCT 81 (49/32) 12 14/10 (NS) 100 (peripheral lipoatrophy) 80 DEXA: f abdominal and peripheral fat; CT: no change in VAT f50 None Fasting, f195 None [26] NVP vs. EFV RCT (no PI control group) 76 (NVP, 39; EFV, 37) 6 2.5/0 None None [27] ABC vs. EFV vs. NVP RCT (no PI control group) 460 (NVP, 155; EFV, 156; ABC, 149) 6 ITT/OTA: NVP, 22/6; EFV, 26/6; ABC, 23/ mg/dl: NVP, 24%; EFV, 22%; ABC, 9% (P!.05) [28] Substudy from [27] 81 (NVP, 28; EFV, 26; ABC, 27) 6 f For total/ HDL (all, NVP, EFV) F (All, NVP, EFV); f (ABC) f (All together) Fasting, f82 (NVP only) f Insulin and insulin resistance (all) [29] NVP or EFV Cohort 39 (NVP, 32; EFV, 7) DEXA: increased lean body mass; MRI: no change (VAT, SAT) None (baseline normal) F8 (baseline low) None Fasting, f80 b NS f fasting glucose No change in bone mineral density [30] ABC/EFV Cohort CT: SAT, TAT; BIA: fat loss 1.6 kg; better (self-report): most; weight: f4 kg 6 Months f; 1 year, none Initial f; 1 year, none Initial f; 1 year, none Fasting, initial f; 1 year, none 2/4 Cases of diabetes resolved [31] NVP or EFV Cohort 100 (NVP, 68; EFV, 32); some results pooled 12 NVP, 24; EFV, Subjective improvement, waist and breast f35 f92 Improved adherence [32] NVP or EFV Cohort 73 (NVP, 63; EFV, 10) AM: better, 41; 3/73 patients had new lipodystrophy NS f f77 [33] NVP, EFV, or ABC Cohort 109 (NVP, 38; EFV, 58; ABC, 13) 9 29 [34] NNRTI Cohort 59 6 VL 1500: 31r 34 NS F NS F [35] ddi/hu only vs. PI-containing regimen RCT 223 (116/107) 12 55/52 (ITT) % Partial reversion of lipodystrophy 1200 mg/dl: 71%r38% (P!.05) 1200 mg/dl: 44%r28% (P!.05) NOTE. Italic reference number indicates that the study is a substudy. ABC, abacavir; ADV, adefovir; AM, anthropomorphic measurements; AZT, zidovudine; BIA, bioelectrical impedance; ddi, didanosine; DEXA, dual-energy radiographic absorptiometry; d4t, stavudine; EFV, efavirenz; HDL, high-density lipoprotein; HU, hydroxyurea; ITT, intention-to-treat analysis; LDL, low-density lipoprotein; NS, not significant; NVP, nevirapine; OTA, on-treatment analysis; QOL, quality-of -life; RCT, randomized controlled trial; ref., reference; SAT, subcutaneous adipose tissue; TAT, total adipose tissue; VAT, visceral adipose tissue; VL, virus load; f, lower(ed) or decrease(d); F, higher or increase(d); r, changed to. a Percentage of patients in switch group taking d4t after the switch. b Compared with baseline value, not with control group. 1223

6 Table 5. Switch studies in which the thymidine analogue was changed. Reference Description a Design No. of subjects (no. in switch arm/no. in control arm) Median duration of follow-up, months HIV load rebound, % of switch/ control subjects Lipodystrophy at baseline, % of patients Change in body shape, % of patients Total HDL Mean D, mg/dl LDL Triglycerides Insulin resistance Other changes [36] d4t (84%) or AZT (16%) r ABC vs. continued regimen (54% PI use) RCT 111 (55/56) (lipoatrophy) DEXA, CT: F limb fat, F sc thigh fat, F sc abdominal fat at 6 months; no subjective improvement None None None f Lactate; no change in quality of life [37] d4t/3tc/idv, d4t/3tc/nfv, or AZT/3TC/IDV r AZT/3TC/ABC vs. continued regimen RCT 39 (22/17) 12 DEXA: F limb fat (arms 1 legs); no change in VAT None None f When discontinued IDV None No change in lactate [38] NVP/ABC/3TC/d4T r NVP/ ABC/3TC RCT 19 (9/10) (lipoatrophy) DEXA: F limb fat at 6 months; no change in weight, central fat [39] d4t r AZT/3TC or ABC Cohort 118 (ABC, 86; AZT/3TC, 32) 6 6, no change 100 (lipoatrophy or F lactate) DEXA, CT: arm fat, F25%; leg fat, F6%; subjective improvement: 27% Trend towards f lactate [40] d4t/nrti r NNRTI/AZT/3TC (50%); PI/d4T/NRTI r PI/ABC/ NRTI (50%) Cohort 36 9 No change in mean load 100 (lipoatrophy) CT, AM: F sc fat; BIA: total body fat, F1.7 kg; subjective improvement: 31% major, 58% partial None f C-peptide in PI group f Lactate, pyruvate [41] d4t r other NRTI Cohort (facial lipoatrophy) Facial lipoatrophy, 91% improved (self report); weight: F5.4 kg None [42] NNRTI/d4T/3TC or ddi r NNRTI/ABC/3TC or ddi Cohort (lipoatrophy) Subjective improvement in 25% f NS change f Lactate (NS) NOTE. ABC, abacavir; AM, anthropomorphic measurements; AZT, zidovudine; BIA, bioelectrical impedance; d4t, stavudine; ddi, didanosine; DEXA, dual-energy radiographic absorptiometry; HDL, high-density lipoprotein; IDV, indinavir; LDL, low-density lipoprotein; NFV, nelfinavir; NNRTI, nonnucleoside reverse transcriptase inhibitor; NRTI, nucleoside reverse transcriptase inhibitor; NS, not significant; NVP, nevirapine; PI, protease inhibitor; RCT, randomized controlled trial; VAT, visceral adipose tissue; VL, virus load; f, lower(ed) or decrease(d); F, higher or increase(d); r, changed to. a Percentages are % of patients.

7 cessfully suppressed (usually for 6 months). Given that, even in patients with undetectable plasma viremia, suppression of HIV replication is incomplete, it is important to examine only randomized, controlled trials to assess virological safety, because a comparison can be made between patients who switched therapy and those who continued to receive PI-based therapy. Three randomized, controlled trials in which a PI was switched to efavirenz have been reported. In the largest of these studies, 226 patients switched their PI to efavirenz, whereas 120 control subjects continued to receive PI-based regimens. After 48 weeks, 7% of the patients receiving efavirenz had detectable plasma HIV RNA, compared with 15% of those who continued to receive PI-based therapy [8]. This difference was statistically significant. A smaller study found a trend toward more-sustained virus suppression in the efavirenz treatment arm [9], whereas another smaller study showed no difference [10]. The studies involving nevirapine have been smaller, with one study showing more virus suppression in patients whose therapy was switched [19] and the other showing equal results [18]. Studies involving abacavir have had more mixed results. The largest of these studies found that patients who switched to abacavir were less likely to experience viral failure at 48 weeks (12% vs. 23% for those who continued to receive the original regimen) [4]. This result was confirmed in another randomized, controlled trial that involved 81 patients [3]. However, in another randomized, controlled trial from Switzerland, patients who switched to abacavir were less likely to continue to have undetectable HIV RNA [2]. A possible explanation for the difference in the Swiss study is that it involved patients who had received nucleoside therapy before they started receiving a PIbased suppressive regimen. It is possible that, for such patients, the abacavir-based regimen was not sufficiently potent to fully suppress a virus that had developed some degree of nucleoside resistance. In a recently reported Spanish study [27], 460 patients were randomized to switch to efavirenz, nevirapine, or abacavir. The intent-to-treat analysis did not show any differences (treatment failure occurred in 22% 26% of patients). However, the duringtreatment analysis showed more virologic failures in the abacavir arm (13% vs. 6% in the nonnucleoside reverse-transcriptase inhibitor [NNRTI] arms). More patients in the NNRTI arms discontinued treatment because of difficulties in tolerating the regimen. The majority of patients who had virological failure had received prior nucleoside analogue therapy. This study suggests that the overall effectiveness of switching from a PI to any of these drugs is similar. However, there is a higher risk of virological failure associated with switching to abacavir, especially for patients who have received prior nucleoside monoor dual therapy. ADHERENCE TO THERAPY AND QUALITY OF LIFE AFTER DISCONTINUING THE PI Some of the large PI-switch studies have documented that improved adherence with simplified dosing schedules is one of the most striking features of switching ART regimens. The aforementioned large efavirenz study that involved 226 patients showed significantly improved adherence 1 year after the switch [8]. The same holds true for a trial in which 105 patients switched from PI to abacavir [4]. Another abacavir-switch study did not observe a difference in adherence (improved selfreported adherence was seen in both treatment arms) but nevertheless had a better virological outcome [3]. One year after switching from a PI to nevirapine, Spanish patients in 2 studies reported a significantly better quality of life, as assessed by a standardized questionnaire, compared with the baseline score [24] or with the quality of life of a control group that continued to receive PI-based therapy [18]. ADVERSE EFFECTS ASSOCIATED WITH DRUGS SUBSTITUTED FOR PIS The observed adverse effects associated with use of abacavir, nevirapine, and efavirenz in switch studies corresponded with the known figures from previous clinical studies. Sixteen of 35 PI-switch studies reported frequencies of treatment discontinuation or treatment change because of adverse drug effects. Hypersensitivity was the predominant side-effect noted for abacavir, and it resulted in the discontinuation of abacavir use for 7% of patients. For the switch studies that examined NNRTIs, adverse effects that led to discontinuation or change of the drug were slightly more common for nevirapine than for efavirenz (14% vs. 10%). The aforementioned large, randomized, controlled trial from Spain that compared abacavir, efavirenz, and nevirapine found significantly more treatment discontinuations for nevirapine or efavirenz (16% and 17%) than for abacavir (6%) [27]. The frequency of nevirapine-associated rash ranged from 4% to 19% in the individual studies and was greater than the rate of nevirapine-associated hepatitis at a ratio of 2:1 as a reason for study drug discontinuation. Nevirapine-associated hepatitis was predominantly observed in hepatitis C antibody positive patients. Rash was rare among patients who received efavirenz and did not lead to treatment discontinuations. All treatment discontinuations in the PI-efavirenz switch studies (4.6%) were because of CNS side-effects. INSULIN SENSITIVITY AFTER SUBSTITUTING THE PI As many as 40% of patients undergoing PI-based ART develop increased insulin resistance. This can lead to diabetes mellitus HIV/AIDS CID 2002:35 (15 November) 1225

8 and may contribute to an increased risk of cardiovascular disease. Insulin resistance may also be a factor in the development of other adverse events, especially visceral fat gain [43]. Although the mechanism of insulin resistance is unclear, it has been closely linked, in vitro and in vivo, to the use of PIs. Short-term administration of indinavir to healthy HIV-negative volunteers has been shown to be associated with the rapid onset of decreases in insulin sensitivity [44]. Thus, it might be expected that discontinuing use of a PI might improve insulin sensitivity. Eleven of 35 PI-switch studies have examined glucose metabolism. The results have not been clear-cut, in part because of design or methodological problems. A cohort study that involved 39 patients found nonsignificant decreases in fasting glucose levels and improvements in insulin resistance (measured by an intravenous insulin tolerance test) 6 and 12 months after switching from a PI- to a nevirapine-based regimen. Glucagon levels had decreased significantly, and the oral glucose tolerance test results for 3 patients with baseline hyperglycemia became normal [29]. A Spanish trial that involved 23 patients with self-reported lipodystrophy who had opted to switch their PI-based therapy to a nevirapine-based regimen found 15% lower serum glucose values and an 80% improved fasting insulin resistance index [45] 12 months after the switch [21, 22]. Similar results were reported by the same group for patients with elevated serum glucose levels whose therapy was switched to efavirenz. For these patients, the fasting insulin resistance index was significantly decreased after 6 months [10]. In a study of 26 patients with clinical lipodystrophy, a 21% decrease in insulin levels and a normalization of previously abnormal glucose tolerance test findings in 3 out of 5 patients were reported 1 year after switching from PI to efavirenz [11]. A metabolic substudy of 90 patients from the aforementioned large Spanish switch trial of 460 patients [27] showed significantly decreased insulin levels and improved insulin sensitivity for all 3 treatment arms [28]. In contrast, another Spanish study of 41 patients with markedly increased fasting insulin levels did not find any improvements in glucose metabolism 1 year after the PI component was switched to efavirenz [13]. DYSLIPIDEMIA AFTER DISCONTINUING THE PI Lipid disorders in patients with HIV infection are complicated. HIV infection itself has been associated with an increase in the serum triglyceride level, a lowering of the total level, and a lower high-density lipoprotein (HDL) level [46, 47]. Treatment with current potent regimens results in variable changes. PIs, especially ritonavir, cause a characteristic dyslipidemia pattern with sometimes dramatic increases in triglyceride-rich lipoproteins. Total and low-density lipoprotein (LDL) levels also increase, whereas HDL levels rarely change [48]. Initiation of therapy with the NNRTIs, in contrast, is associated with more-modest increases in total levels and minimal changes in fasting triglyceride levels [49]. Indeed, the HDL level may increase with the initiation of the NNRTIs; this has been best documented with nevirapine [50]. Approximately one-half of all included studies did not measure fasting lipid values or did not specify how their lipid parameters were measured. This may be a confounding factor, especially for the outcomes for triglyceride and LDL values. Abacavir. Three randomized, controlled trials evaluated dyslipidemia after abacavir was substituted for a PI. All showed significantly lower total levels after a median followup period of 12 months, compared with levels for subjects in the PI treatment arm (mean difference, 24 mg/dl) [1, 2, 4]. Two of the 3 trials also showed significant decreases in triglyceride levels [1, 2]. Two abacavir-switch trials have measured HDL levels: a randomized, controlled Swiss trial of 81 patients found no significant differences between treatment arms after 19 months [2], whereas a smaller Spanish study of 27 patients showed significantly lower values after 6 months [28]. Efavirenz. Evolving dyslipidemia after discontinuing efavirenz was studied in 752 patients in 14 clinical trials. Onethird of these studies were randomized (with a total of 366 switched patients). Ten efavirenz trials also included measurements of subclasses. The only discernible trend after a median duration of follow-up of 12 months was elevated HDL level compared with the baseline value. This holds true both for the smaller cohorts that focused on patients with clinical lipodystrophy and for the large randomized trials. The aforementioned large multicenter study of 346 patients showed a small increase in HDL levels in the efavirenz arm and a small decrease after 12 months in the PI arm; the difference was statistically significant [8]. Eight studies (3 of which were randomized, controlled trials) measured HDL levels; 5 of these studies showed a significant increase in the HDL level [8, 10, 12, 15, 28], and 3 did not report significant changes [9, 13, 16]. Total and LDL values were seemingly unaffected by the switch to efavirenz: none of the 13 trials showed any significant changes. The outcome for triglyceride levels was heterogeneous, with 11 of the 14 studies not reporting any significant changes. Nevirapine. The lipid results for nevirapine switch trials stand somewhat in the middle between those for the other 2 switch drugs. Most trials were performed with patients who were reported to have clinical lipodystrophy. Although there is a clear trend toward improvement in triglyceride levels, data on total are conflicting. Three of 9 studies included HDL and LDL values, and all showed significantly increased HDL values after the therapy switch [20, 28, 29]. Two randomized, controlled studies from Spain have compared cho CID 2002:35 (15 November) HIV/AIDS

9 lesterol and triglyceride levels [18, 19]. There were no significant differences between treatment arms (partially because of improving lipid values in the PI treatment arms). Three of 5 nevirapine trials, however, have shown improvement when and/or triglyceride values after the switch were compared with baseline values [19, 21, 24]. Two other studies observed only trends toward improvement [18, 23]. BODY HABITUS CHANGES AFTER DISCONTINUING THE PI Alterations in fat distribution as a consequence of treatment for HIV infection have been estimated to occur in as many as 50% 60% of patients who are receiving long-term, successful ART [43, 51]. At least 2 phenotypes have been described: fat accumulation (mainly visceral fat gain, although sometimes the fat is gained at other body sites) and peripheral fat loss (especially in the face and limbs). The phenotype is often mixed, although 1 component (loss or gain) tends to predominate in an individual patient. Studies of this syndrome (or syndromes) are hampered by the lack of a case definition and by the lack of uniformity measured across studies [43]. These factors lead to great difficulties in interpreting the effects of discontinuation of the PI component of therapy on body composition. Twenty-two of the 35 referenced PI-switch studies commented on the frequency of clinical lipodystrophy (either at baseline or during the course of the study), but most studies did not define the criteria on which this diagnosis was made. Body habitus changes after treatment switches were evaluated in 18 of 35 PI-switch studies. The most common diagnostic tools were self-report and physician s assessment, which are also the least reliable and most subject to patient and physician bias. The outcome was almost uniformly in favor of the switch regimen, with reported subjective improvement of the lipodystrophy syndrome. Sixteen of the 22 PI-switch studies that included data on body-shape changes included 1 objective test. The most commonly used objective methods were anthropomorphic measurements, including body-part circumferences, skin-fold thickness, and waist-to-hip ratio. These methods are subject to considerable observer variation because of variability in technique. Dual-energy radiographic absorptiometry was used in 5 studies [16, 18, 24, 25, 29]. This technique is a reasonable measure of peripheral fat loss and total fat, but it cannot distinguish between visceral and subcutaneous fat and, thus, cannot detect changes in visceral fat accumulation. Only 5 studies with a total of 155 patients used the most reliable techniques for measurement of visceral fat, CT or MRI of the abdomen [11, 14, 25, 29, 30]. Given the wide variability in the definitions and methods of assessment, it is difficult to make definitive statements about the effect on fat redistribution of switching from a PI. Overall, 6 of 16 studies showed, by objective measurements, body-shape changes that were determined to be significant [10, 18, 21, 25, 29, 30]. These changes could be referred to as moderate loss of fat, as documented by dual-energy radiographic absorptiometry; in some cases, this is associated with weight loss, and, in 1 case, it was associated with a gain of lean body mass. However, as noted above, the low accuracy of these tests (or, in the case of dual-energy radiographic absorptiometry scans, the ability to distinguish visceral from total abdominal fat) makes interpretation of these data difficult. Three of the 5 PIswitch studies that used quantitative methods that directly assess visceral fat accumulation (abdominal CT or MRI) showed trends toward reduction of visceral adipose tissue [11, 14, 30], whereas 2 failed to document any changes [25, 29]. Of importance, no study showed any improvement in fat loss after discontinuation of the PI component of therapy. NUCLEOSIDE-SWITCH TRIALS Cohort studies have suggested that thymidine analogue nucleosides (particularly stavudine) might be important contributors to the development of lipoatrophy [52 55]. Nucleoside-switch trials were first initiated in All enrolled patients with lipoatrophy, many of whom were already receiving PI-free regimens. To date, 7 nucleoside-switch studies with a total of 350 patients (94% of whom discontinued use of stavudine) have been reported; 3 of the studies were randomized, controlled trials. The most recent and largest nucleoside-switch trial randomized 111 Australian abacavir-naive patients with lipoatrophy to either switching the thymidine analogue component of therapy to abacavir or to continuing the previous regimen [36]. The majority of patients were taking stavudine (77%); 16% were taking zidovudine at the time of the switch. Fifty-two percent of the patients were receiving PI-containing combination therapy, which was not changed. Peripheral wasting, as measured by dual-energy radiographic absorptiometry and CT, improved 6 months after the switch. The mean change in the amount of limb fat was modest (an 0.39-kg increase from baseline). Of interest, there was no subjective improvement and there were no correlations between subjective severity scores and objective data. There was a trend toward decreased lactate levels, but no changes in lipid or glucose metabolism were seen. Four of these 7 nucleoside-switch studies used dual-energy radiographic absorptiometry [36 39], which is a sensitive diagnostic tool for peripheral lipoatrophy. All of these studies confirmed that there was a modest but statistically significant gain in peripheral fat after the switch from or discontinuation of stavudine. Subjective improvement of lipoatrophy symptoms, however, was reported only for a minority of patients (mean, 37%) [39 42]. Four of 7 studies thus far have also HIV/AIDS CID 2002:35 (15 November) 1227

10 shown decreased lactate levels [36, 39, 40, 42]. The impact on other metabolic parameters was less clear. SUMMARY The management and understanding of ART-associated lipodystrophy and metabolic disturbances has become a major clinical concern, because as many as 60% of the patients after 11 year of therapy might be affected. Since 1998, many clinical studies have investigated the evolution of ART-associated sideeffects after removing the presumed causative agent, the PI. In the absence of effective therapeutic options for the disfiguring body habitus changes and dyslipidemia that many patients experience, some of the switch trials were at the same time attempts at therapeutic intervention. These switch trials have contributed significantly to a better understanding of ARTrelated side-effects. Before analyzing the complex phenomena of lipid metabolism and body fat distribution, one should conclude that discontinuing the PI often results in a much more convenient regimen and is overall virologically safe if the virus load has been undetectable for the past 6 months. If a switch in therapy is indicated or contemplated, patients and physicians should not be concerned about the risk of breakthrough viremia. This is particularly true for patients for whom the PI-based regimen is their first antiviral therapy. Although it is unlikely that the new regimens are more potent than PI-based treatments, they are more simple. This and the improved quality of life may contribute to the observed better long-term adherence and, thus, sustained virus suppression. In addition, discontinuing the PI component also preserves this class of agents for future use. PI-free antiretroviral regimens have the advantage of a lower potential for medication interactions. In conclusion, our analysis suggests that switching from a PI to either an NNRTI or abacavir is generally virologically safe and often results in an improved metabolic profile. It is therefore a reasonable approach to offer to many patients. In contrast, the virological safety of switching components (either PIs or nucleoside analogues) for nucleoside reverse-transcriptase inhibitor experienced patients or for patients whose virus is not fully suppressed remains an open question. Full interpretation of the metabolic effects of switching from a PI is difficult for many reasons. Baseline values for parameters of lipid and glucose metabolism before starting the PI were often not provided. Many patients did not have significant metabolic abnormalities while taking the PI-containing regimen. Thus, there would be little opportunity to see benefit after switching. The NNRTI class of antiretrovirals, which are commonly used as PI substitutes, may have a significant impact on lipid metabolism on their own [49, 50]. Nevertheless, there is a clear trend toward an improved metabolic profile when patients discontinue use of the PI. This holds especially true for insulin resistance and triglyceride levels and, to a lesser degree, for total and HDL levels (especially when therapy is switched to include an NNRTI). It should be kept in mind that there is still no generally accepted case definition for ART-associated lipodystrophy, and little is known about the natural course of this syndrome. A plateau effect has been observed, both on an individual basis [56] and on a population basis (stable prevalence after 2 years of receiving ART) [47]. Given the wide variability in methods of assessment, there is a clear bias toward overrating the impact of PIs on body habitus changes. If we focus only on studies that used the most reliable measures, there might be a trend toward reduced visceral fat. None of the PI-switch studies showed any objective evidence for reversibility of peripheral wasting. Thus, peripheral wasting might be more closely related to the nucleoside analogue components of ART, as suggested by the recent Australian stavudine/zidovudine switch trial [36]. It is noteworthy, though, that almost one-half of the patients in this study were not taking PIs at the time of the switch. On the other hand, most PI-switch studies that have involved patients with clinical lipodystrophy maintained the original nucleoside regimen. Therefore, it is reasonable to propose that the development of ART-associated lipodystrophy and metabolic disturbances could be triggered by a combination of PIs and thymidine analogues. Although the overall improvement in limb fat levels seen in the nucleoside-switch studies was small, these trials are important proof of concept and strongly suggest a role for thymidine analogues (especially stavudine) in the pathogenesis of lipoatrophy. The virological safety of discontinuing use of nucleoside analogues is not yet well documented. Many more questions remain with regard to switching strategies for patients with lipodystrophy. Will metabolic abnormalities and symptoms of lipodystrophy return to baseline levels if the PI component of therapy and stavudine are discontinued at the same time? Is there a best time for switching ART? A small randomized switch trial comparing the effects of discontinuing PI versus stavudine versus both drugs is currently underway [57]. Other studies in the future might evaluate completely nucleoside reverse-transcriptase inhibitor free regimens. All future trials will, it is hoped, include baseline metabolic and body composition data and will routinely assess adherence to ART and quality of life. It is preferable that they will concentrate on patients with new metabolic disturbances seen after starting ART to determine the effectiveness of switching as a therapeutic option. References 1. Lafeuillade A, Mamet JP, Chéret A, et al. Comparison of the metabolic disorders and clinical lipodystrophy 48 weeks after switching from 1228 CID 2002:35 (15 November) HIV/AIDS

11 highly active antiretroviral therapy to Trizizvir versus continued highly active antiretroviral therapy [abstract 28]. In: Programs and abstracts of the 3rd International Workshop on Adverse Drug Reactions and Lipodystrophy (Athens). Antiviral Therapy 2001; 6(Suppl 4). 2. Opravil M, Hirschel B, Lazzarin A, et al. A randomized trial of simplified maintenance therapy with abacavir, lamivudine, and zidovudine in human immunodeficiency virus infection. J Infect Dis 2002; 185: Pulvirenti J, Godwin D, Slater L, et al. Simplification of protease inhibitor containing HAART regimen with abacavir maintains viral suppression and favorable adherence in HIV-1 infected adults [abstract 689]. Clin Infect Dis 2001; 33: Clumeck N, Goebel F, Rozenbaum W, et al. Simplification with abacavir-based triple nucleoside therapy versus continued protease inhibitor based highly active antiretroviral therapy in HIV-1 infected patients with undetectable plasma HIV-1 RNA. AIDS 2001; 15: Walli R, Michl G, Bogner JR, Goebel FD. Improvement of HAARTassociated insulin resistance and dysplipidemia after replacement of protease inhibitors with abacavir. Eur J Med Res 2001; 6: Rozenbaum W, Delfraissy JF, Bentata M, et al. Metabolic and clinical evaluation of lipodystrophy 48 weeks after switching from two nucleoside reverse transcriptase inhibitors/protease inhibitor to two nucleoside reverse transcriptase inhibitors/abacavir (French Substudy, CNA30017) [abstract P89]. In: Programs and abstracts of the 2nd International Workshop on Adverse Drug Reactions and Lipodystrophy (Toronto). Antiviral Therapy 2000; 5(Suppl 5). 7. Maggiolo F, Miglorino M, Maserati R, et al. Simplified therapeutic strategies in PI-experienced patients successfully treated with HAART [abstract 1916]. In: Programs and abstracts of the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy (Chicago). Washington, DC: American Society for Microbiology Press, Becker S, Rachlis A, Gill J, et al. Successful substitution of protease inhibitors with SUSTIVA (efavirenz) in patients with undetectable plasma HIV-1 RNA levels: results of a prospective, randomized, multicenter, open label study (DMP ) [abstract 20]. In: Programs and abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, Katlama C, Staszewski S, Cluck N, et al. Successful substitution of protease inhibitors with Sustiva (efavirenz) in patients with undetectable plasma HIV-1 RNA: results of a prospective, randomized, multicenter, open label study (DMP ) [abstract LbPeB7044]. In: Programs and abstracts of the 13th International Conference on AIDS (Durban, South Africa). Rome: International AIDS Society, 2000: Martinez E, Romeu J, Garcia-Viejo A, et al. An open randomized study on the replacement of HIV-1 protease inhibitors by efavirenz in chronically suppressed HIV-1 infected patients with lipodystrophy [abstract 668]. In: Programs and abstracts of the 8th Conference on Retroviruses and Opportunistic Infections (Chicago). Alexandria, VA: Foundation for Retroviruses and Human Health, Moyle GJ, Baldwin C, Mandalia S, et al. Changes in metabolic parameters and body shape after replacement of protease inhibitor with efavirenz in virologically controlled HIV-1 positive persons: single arm observational cohort [letter]. J Acquir Immune Defic Syndr 2001; 28: Viciana P, Alarcon D, Martin P, et al. Partial improvement of lipodystrophy after switching from HIV-1 protease inhibitors (PI) to efavirenz [abstract 48]. In: Programs and abstracts of the 7th conference on Retroviruses and Opportunistic Infections (San Francisco). Alexandria, VA: Foundation for Retroviruses and Human Health, Estrada V, de Villar N, Larrad MT, Lopez AG, Fernandez C, Serrano- Rios M. Long-term metabolic consequences of switiching from protease inhibitors to efavirenz in therapy for human immunodeficiency virus infected patients with lipoatrophy. Clin Infect Dis 2002; 35: Lafon E, Bani Sadr F, Chandemerle C, et al. LIPSTOP study: evolution of clinical lipodystrophy, blood lipids, visceral and subcutaneous adipose tissue after switching from protease inhibitor to efavirenz in HIV- 1 infected patients [abstract 1535]. In: Programs and abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, Knechten H, Stürner KH, Höhn C, Braun P. Switch to efavirenz in a a protease inhibitor containing regimen. HIV Clin Trials 2001; 2: Bonnet EP, Yovanovitch JD, Mularczyk MB, et al. Switching from protease inhibitors to efavirenz does not improve clinical lipodystrophy and hyperlipidemia in HIV-infected patients [abstract 1538]. In: Programs and abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, Gharakhanian S, Salhi Y, Adda N, et al. Identification of fat redistribution/metabolic abnormalities in a cohort treated by 2 NRTIs 1 PI and absence of significant modification following PI substitution with efavirenz [abstract 46]. In: Programs and abstracts of the 7th Conference on Retroviruses and Opportunistic Infections (San Francisco). Alexandria, VA: Foundation for Retroviruses and Human Health, 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: Barreiro P, Soriano V, Blanco F, et al. Risks and benefits of replacing protease inhibitors by nevirapine in HIV-infected subjects under longterm successful triple combination therapy. AIDS 2000; 14: Negredo E, Masana L, Ribalta J, et al. Nuclear magnetic resonance (NMR) assessed changes of lipid metabolism in HIV-1 infected patients with lipodystrophy after switching the protease inhibitor by nevirapine [abstract 485]. In: Programs and abstracts of the 1st International AIDS Society Conference on HIV Pathogenesis and Treatment (Buenos Aires). Stockholm: International AIDS Society, Gatell JM, Conget I, Lozano R, et al. One year effects of switching from HIV-1 protease inhibitors to nevirapine on metabolic abnormalities [abstract 040]. In: Programs and abstracts of the 1st International Workshop on Adverse Drug Reactions and Lipodystrophy (San Diego). Antiviral Therapy 1999; 4(Suppl 2). 22. Martinez E, Conget I, Lozano R, et al. Reversion of metabolic abnormalities after switching from HIV-1 protease inhibitors to nevirapine. AIDS 1999; 13: Buisson M, Grappin M, Piroth L, et al. Simplified maintenance therapy with NNRTI (nevirapine) in patients with long-term suppression of HIV-1 RNA: first results of a cohort study [abstract 1541]. In: Programs and abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, Negredo E, Cruz L, Paredes R, et al. Virological, immunological, and clinical impact of switching from protease inhibitors (PI) to nevirapine (NVP) or efavirenz (EFV) in patients with HIV infection and longlasting viral suppression. Clin Infect Dis 2002; 34: Carr A, Hudson J, Chuah J, et al. HIV protease inhibitor substitution in patients with lipodystrophy: a randomized, controlled, open-label multicentre study. AIDS 2001; 15: Patterson P, Krolewiecki A, Ochoa C, et al. A randomized study of treatment simplification with nevirapine or efavirenz in patients responding to a protease inhibitor based combination [abstract 58]. In: Programs and abstracts of the 1st International AIDS Society Conference on HIV Pathogenesis and Treatment (Buenos Aires). Stockholm: International AIDS Society, Martinez E, Podzamczer D, Ribera E, et al. Switching protease inhibitors to nevirapine, efavirenz or abacavir: a randomized, multicenter, open-label simplification trial [abstract LB17]. In: Programs and abstracts of the 9th conference on Retroviruses and Opportunistic Infections (Seattle). Alexandria, VA: Foundation for Retroviruses and Human Health, Fisac C, Fumero E, Crespo M, et al. A randomized trial of metabolic HIV/AIDS CID 2002:35 (15 November) 1229

12 and body composition changes in patients switching from PI-containing regimens to abacavir, efavirenz, nevirapine [abstract 699]. In: Programs and abstracts of the 9th Conference on Retroviruses and Opportunistic Infections (Seattle). Alexandria, VA: Foundation for Retroviruses and Human Health, Tebas P, Yarasheski K, Henry K, et al. Evolution of multiple metabolic parameters after the switch of protease inhibitors to nevirapine [abstract ThPpB1485]. In: Programs and abstracts of the 13th International Conference on AIDS (Durban, South Africa). Rome: International AIDS Society, 2000: Bickel M, Rickerts V, Jacobi V, et al. Evolution of fasting metabolic parameters and fat distribution after switching protease inhibitor in the PROTRA study [abstract P90]. In: Programs and abstracts of the 2nd International Workshop on Adverse Drug Reactions and Lipodystrophy (Toronto). Antiviral Therapy 2000; 5(Suppl 5). 31. Casado JL, Arrizabalaga J, Antela A, et al. Long-term efficacy and tolerance of switching the protease inhibitor for non-nucleoside reverse transcriptase inhibitors: a 52-week, multicenter prospective study [abstract 673]. In: Programs and abstracts of the 8th Conference on Retroviruses and Opportunistic Infections (Chicago). Alexandria, VA: Foundation for Retroviruses and Human Health, Raffi F, Bonnet B, Ferré V, et al. Substitution of a nonnucleoside reverse transcriptase inhibitor for a protease inhibitor in the treatment of patients with undetectable plasma human immunodeficiency virus type 1 RNA. Clin Infect Dis 2000; 31: Melzi S, Adorni F, Castelnuovo B, et al. Virological and immunological outcome in 109 patients with well-controlled viral replication who switched from a PI-including to a PI-sparing regimen [abstract P78]. In Programs and abstracts of the 8th European Conference on Clinical Aspects and Treatment of HIV Infection (Athens). Paris: European AIDS Clinical Society, Badsi ES, Salord JM. Switching a protease inhibitor for a non-nucleoside reverse transcriptase inhibitor: study of a cohort of 59 HIV patients [abstract 1539]. In: Programs and abstracts of the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy (Toronto). Washington, DC: American Society for Microbiology, Barreiro P, Camino C, de Mendoza V, et al. Simplification to HU-DDI in patients being on long-term success under HAART [abstract 533]. In: Programs and abstracts of the 9th Conference on Retroviruses and Opportunistic Infections (Seattle). Alexandria, VA: Foundation for Retroviruses and Human Health, Carr A, Workman C, Smith D, et al. Abacavir substitution for nucleoside analogs in patients with HIV lipoatrophy: a randomized trial. JAMA 2002; 288: John M, McKinnon E, James I, et al. A randomized, controlled, openlabel study of revision of antiretroviral regimens from stavudine and/ or a protease inhibitor to zidovudine/lamivudine/abacavir to prevent or reverse lipoatrophy: 48-week data [abstract 700]. In: Programs and abstracts of the 9th conference on Retroviruses and Opportunistic Infections (Seattle). Alexandria, VA: Foundation for Retroviruses and Human Health, Smith D, Carr A, Law M, et al. A randomized trial of thymidine analogue withdrawal in lipatrophic HIV patients, virologically controlled on protease-sparing therapy the PIILAR extension study [abstract 96]. In: Programs and abstracts of the 1st International AIDS Society Conference on HIV Pathogenesis and Treatment (Buenos Aires). Stockholm: International AIDS Society, McComsey G, Lonergan T, Fisher R. Improvements in lipoatrophy are observed after 24 weeks when stavudine is replaced by either abacavir or zidovudine [abstract 701]. In: Programs and abstracts of the 9th Conference on Retroviruses and Opportunistic Infections (Seattle). Alexandria, VA: Foundation for Retroviruses and Human Health, Saint-Marc T, Partisani M, Poizot-Martin I, et al. Reversibility of peripheral fat wasting (lipoatrophy) on stopping stavudine therapy [abstract 52]. In: Programs and abstracts of the 7th Conference on Retroviruses and Opportunistic Infections (San Francisco). Alexandria, VA: Foundation for Retroviruses and Human Health, Clark A. Improvement of facial lipoatrophy after substitution of stavudine [abstract P85]. In: Programs and abstracts of the 2nd International Workshop on Adverse Drug Reactions and Lipodystrophy (Toronto). Antiviral Therapy 2000; 5(Suppl 5). 42. García-Benyas T, Blanco F, Gómez JM, et al. Replacement of stavudine by abacavir does not ameliorate lipoatrophy in HAART recipients [abstract 491]. In: Programs and abstracts of the 1st International AIDS Society Conference on HIV Pathogenesis and Treatment (Buenos Aires). Stockholm: International AIDS Society, Safrin S, Grunfeld C. Fat distribution and metabolic changes in patients with HIV infection. AIDS 1999; 13: Noor M, Lo J, Halvorse J, et al. Metabolic effects of indinavir in healthy HIV-seronegative subjects [abstract 010]. In: Programs and abstracts of the 2nd International Workshop on Adverse Drug Reactions and Lipodystrophy (Toronto). Antiviral Therapy 2000; 5(Suppl 5). 45. Duncan MH, Singh BM, Wise PH, et al. A simple measure of insulin resistance. Lancet 1995; 346: Grunfeld C, Pang M, Doerrler W, et al. Lipids, lipoproteins, triglyceride clearance, and cytokines in human immunodeficiency virus infection and the acquired immunodeficiency syndrome. J Clin Endocrinol Metab 1992; 74: Kingsley L, Smit E, Riddler S, et al. Prevalence of lipodystrophy and metabolic abnormalities in the multicenter AIDS cohort study (MACS) [abstract 538]. In: Programs and abstracts of the 8th Conference on Retroviruses and Opportunistic Infections; (Chicago). Alexandria, VA: Foundation for Retroviruses and Human Health, Sullivan AK. Nelson MR. Marked hyperlipidaemia in ritonavir therapy [letter]. AIDS 1997; 11: Tashima K, Stryker R, Skiest D, et al. Lipid profiles and clinical lipodystrophy in study 006 patients [abstract 1304]. In: Programs and abstracts of the 39th Interscience Conference on Antimicrobial Agents and Chemotherapy (San Francisco). Washington, DC: American Society for Microbiology, van der Valk M, Kastelein JJ, Murphy RL, et al. Nevirapine-containing antiretroviral therapy in HIV-1 infected patients results in an antiatherogenic lipid profile. AIDS 2001; 15: Carr A, Samaras K, Thorisdottir A, et al. Diagnosis, prediction, and natural course of HIV-1 protease-inhibitor associated lipodystrophy, hyperlipidemia, and diabetes mellitus: a cohort study. Lancet 1999; 353: Mallal S, 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: 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: 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: Moyle G. Mitochondrial toxicity hypothesis for lipoatrophy: a refutation [letter]. AIDS 2001; 15: Mahajan A, Tashima K, Bausserman L, et al. Plateau in body habitus changes and serum lipid abnormalities in HIV-positive women on highly active antiretroviral therapy: a 3.5-year study. J Acquir Immune Defic Syndr 2001; 28: Moyle G, Baldwin C, Gazzard BG, et al. An open-label randomized trial of substitution of stavudine or protease inhibitors/non-nucleoside reverse transcriptase inhibitors to an abacavir-based therapy for metabolic disturbances in persons on first-line therapy [abstract 97]. In: Programs and abstracts of the 3rd International Workshop on Adverse Drug Reactions and Lipodystrophy (Athens). Antiviral Therapy 2001; 6(Suppl 4) CID 2002:35 (15 November) HIV/AIDS