Table of Contents

• Introduction
• Body Shape
• Switching Antiretroviral Therapy to Reverse Fat Redistribution
• Lipids and Cardiovascular Disease
• Conclusions
• References

Introduction

For an HIV clinician/clinical researcher with more than a passing interest in the metabolic complications associated with HIV, the 12th Conference on Retroviruses and Opportunistic Infections (CROI) in Boston, Mass., was like dying and going to lipodystrophy heaven (where, it is rumored, Krispy Kreme doughnuts are plentiful). Often pushed aside by sessions dedicated to large clinical trials on antiretroviral therapy, new antiretroviral agents and even epidemiology, metabolic complications have of late become a sidebar topic at HIV conferences. This has much to do with a frustrating lack of major therapeutic developments in the complications field as well as to the slow pace of our understanding of the pathogenesis of these adverse events. In Boston, however, the tables were turned. Results from big antiretroviral therapy trials were almost nonexistent and the strongest competition for attention came from one man in New York with bad luck and a nasty virus. That left plenty of room for metabolic complications. (It should be mentioned that there were also several interesting presentations on the prevention of maternal-to-child transmission of HIV, see Margaret Hoffman-Terry's review.)

Studies of the 2 major complications of HIV therapy, dyslipidemia and body shape, appropriately made up the lion's share of the data presented. The most significant results related to these disorders were presented during a single session on Wednesday morning.

Body Shape

Data emerging from 4 clinical trials with metabolic outcomes were presented during the session. These included 3 AIDS Clinical Trials Group (ACTG) studies -- each conceived several years ago when stavudine (d4T, Zerit) use was rampant and the distress of those with body shape changes had reached a crescendo.

Characterization of Morphologic Changes During HIV Therapy

Kathy Mulligan of the University of California-San Francisco presented data on the body shape changes that developed during the ACTG 5005s study, a metabolic substudy of ACTG 384.¹ The parent study ACTG 384 is a large, 6-arm clinical trial comparing the nucleoside analog combinations of stavudine + didanosine (ddI, Videx) versus zidovudine (AZT, Retrovir) + lamivudine (3TC, Epivir) when combined with either efavirenz (EFV, Sustiva, Stocrin) or nelfinavir (NFV, Viracept) or efavirenz + nelfinavir. Of the 980 patients enrolled in the parent study, 334 participated in the substudy.

Anthropometric measurements of the circumferences of the waist, hip, thigh and arm were performed over 64 weeks in most substudy patients, as was whole body dual energy X-ray absorptiometry (DEXA) scanning in 102 subjects. The results were divided into 2 sections: the first reported the concordance of anthropometrics with DEXA scanning and the second detailed aggregate data on the trend in body shape among patients who were evaluated.

Anthropometric evaluations of body shape have the advantage of being both inexpensive and convenient, and require little more than a non-stretch tape measure and basic training regarding technique. However, there are limited data on the accuracy of these measures in assessing morphologic changes, such as fat redistribution, when compared to automated radiographic techniques, such as DEXA scanning. Mulligan reported that waist circumference and DEXA-assessed truncal fat correlated moderately well, as did limb fat and hip, arm and thigh circumferences. These data indicate that anthropometrics are a reasonable method for the longitudinal evaluation of body shape change -- a finding that supports the use of such techniques, particularly within international antiretroviral therapy trials conducted in locales where DEXA and computerized tomography (CT) scanning are unavailable.

During the study, the overall picture that emerged was one of increased central fat accumulation and peripheral fat loss (Figure 1). However, on closer examination, a more complex pattern emerges. Looking at the hundred or so patients who had DEXA scan results at week 64, for two thirds of the patients, central and peripheral fat correlated positively, that is, when one went up so did the other and vice versa (Figure 2). Among the remainder of the patients in the study, most demonstrated a mixed pattern of increased central fat and decreased limb fat that has become the sine qua non of the lipodystrophy syndrome (Figure 3). A break down of the morphologic changes evolving in each study arm has been previously reported.²

Figure 1. Median Changes in Trunk and Limb Fat by DEXA (All Treatment Groups Combined)

(Slide by Kathy Mulligan.)

Figure 2. Changes in 66% of Subjects Were Concordant

(Slide by Kathy Mulligan.)

Figure 3. Discordant Changes in 32% of Subjects

(Slide by Kathy Mulligan.)

This rigorous longitudinal study provides important information and expands on the conclusions of the FRAM (Fat Redistribution and Metabolic Change in HIV Infection) study -- a cross-sectional evaluation of body shape and other metabolic parameters in HIV-infected patients (for which, it should be disclosed, this author serves as an investigator).³ FRAM famously reported that when comparing HIV-infected patients with HIV-uninfected controls, lipoatrophy was the dominant body shape abnormality and was more often than not accompanied by central fat loss rather than gain. A5005s has the advantage of being able to examine changes in shape over time and reveals that there may be no one dominant feature of antiretroviral-associated body shape change, with some patients experiencing fat wasting of both the extremities and trunk, others experiencing fat gain in both and a proportion of patients noting a mixture of fat wasting and fat gain. What influences which pattern emerges is not clear. Leading suspects include genetics, antiretroviral therapy composition and pre-antiretroviral therapy body shape.

Switching Antiretroviral Therapy to Reverse Fat Redistribution

Two complimentary ACTG trials examined the effect of different strategies of antiretroviral therapy modification to reverse fat redistribution in antiretroviral-treated patients. A5110, chaired by Pablo Tebas from the University of Pennsylvania, enrolled patients with self-reported lipoatrophy of the extremities confirmed on physical examination that developed during thymidine analog (stavudine or zidovudine)-containing antiretroviral therapy.⁴

Participants were required to have a CD4+ cell count equal to or greater than 100 cells/µL and a plasma HIV-RNA level of less than 500 copies/mL. A total of 101 patients were recruited to participate in the trial and were randomized to either switch their thymidine analog to abacavir (ABC, Ziagen) while maintaining their pre-entry protease inhibitor (PI) or non-nucleoside reverse transcriptase inhibitor (NNRTI) or initiate a completely new regimen of lopinavir/ritonavir (LPV/r, Kaletra) + nevirapine (NVP, Viramune).

Approximately three quarters of the patients were receiving stavudine prior to study entry and there did not appear to be any differences in the proportion in each arm on this nucleoside reverse transcriptase inhibitor (NRTI).

At 24 weeks after the antiretroviral therapy switch, there was a significant median gain in subcutaneous fat of the leg, as assessed by CT scan in the lopinavir/ritonavir + nevirapine group (P = .02). Little change, however, was noted among the abacavir group (the study was not specifically powered to detect differences between the study arms).

Patients in both the thymidine analog-sparing and the NRTI-sparing arms experienced an increase in subcutaneous trunk fat (P<.01), but visceral fat decreased significantly only in the abacavir arm (P<.01). The visceral adipose tissue (VAT) to total adipose tissue (TAT) ratio decreased significantly in both study arms (P<.01). Six patients experienced viral load rebounds during the study; 3 in each arm of the study and 6 patients in each group experienced treatment-limiting drug toxicity. Analysis of the lipid costs of the individual switch strategies as well as the effect of the antiretroviral therapy modification on other metabolic parameters and quality of life are in progress.

A sister study to the A5110 Tebas study just discussed was A5125s.⁵ Like A5110, this trial examined the impact of 2 distinct antiretroviral therapy switch strategies on body shape; however, this study did not require patients to have lipoatrophy in order to enroll. Further, the main endpoint of this trial was change in limb fat; changes in central fat were not reported. Patients in this study had a viral load of less than 200 copies/mL and were receiving either

1. 2 NRTIs + indinavir (IDV, Crixivan),
2. 2 NRTIs + indinavir + efavirenz, or
3. 2 NRTIs + indinavir + nelfinavir.

All were randomized to either continue their NRTIs but change their PI and/or NNRTI to efavirenz versus change their regimen to lopinavir/ritonavir + efavirenz. Therefore, as opposed to A5110, in this study thymidine analogs were included in one of the switch arms -- recall that this study was designed before the results of the MITOX and TARHEEL studies on the substitution of thymidine analogs with abacavir were reported.^6,7 These 2 studies demonstrated that the switch from thymidine analogs (predominantly stavudine) to abacavir led to modest improvements in peripheral fat among lipoatrophic patients.

A total of 62 patients were enrolled in A5125s and there were no significant differences between the study arms at baseline. Slightly less than a quarter of the patients were receiving stavudine at study entry and although it was not reported, it is assumed, given the design, that the majority of the remainder were receiving zidovudine.

Given the results of A5110 and the earlier switch studies, can you guess what happened to limb fat following the switch? Smart money would hazard that the NRTI + efavirenz arm would not see a gain in peripheral fat, and, in fact, this arm actually saw a continued and significant decline in extremity adipose tissue. The median loss of limb fat was approximately 1 kg after a median of 106 weeks.

In contrast, the lopinavir/ritonavir + efavirenz group experienced an increase in fat of over 750 g, leading to a significant difference between the 2 study arms (P = .002). Not unexpectedly, the lopinavir/ritonavir + efavirenz arm experienced increases in triglycerides and low-density lipoprotein (LDL) cholesterol, while there was little change in these lipid parameters in the NRTI + NNRTI arm. Bone mineral density (BMD) and serum bone markers did not change from entry in either group.

The overall effect of the switch strategies studied in the 2 ACTG studies is summarized in Table 1.

Table 1. Effect of Antiretroviral Therapy Switch Strategies on Fat Compartments

The pair of studies provide interesting data regarding the effects of different antiretrovirals on fat redistribution. Elimination of NRTIs had the most profound effect on the recovery of limb fat. For reasons that are unclear, the switch from thymidine analogs to abacavir in A5125s did not result in the limb fat increases seen in other studies. Extremity fat was assessed using CT scans in this study, rather than DEXA, and these can be a bit tricky to use to quantify fat changes. Otherwise, it is hard to reckon this discrepancy compared to the MITOX and TARHEEL studies. Changing to abacavir, however, did decrease visceral fat.

The increase in lipids seen with lopinavir/ritonavir + efavirenz suggests that the switch in therapy may also mean a swap of toxicities -- not exactly an appealing proposition. Lastly, the combined data suggest that not all fat loss can be blamed solely on stavudine. A quarter of the lipoatrophic patients enrolled in A5110 were not on stavudine and, one presumes, were instead on zidovudine. Similarly, although A5125s participants were not required to have lipoatrophy, they did experience an increase in peripheral fat when their mostly non-stavudine NRTIs were discontinued.

The substitution of abacavir for stavudine, as described above, has been found to improve lipoatrophy. In vitro assays of mitochondrial function suggest tenofovir (TDF, Viread), like abacavir, is relatively nontoxic to mitochondrial DNA polymerases.⁸

Further, data from Gilead 903, a head-to-head trial comparing stavudine and tenofovir when combined with lamivudine and efavirenz, demonstrated significant metabolic toxicity from stavudine but no such toxicity from tenofovir.⁹ Therefore, a switch from stavudine to tenofovir in the face of metabolic complications is logical and occurs quite commonly in clinical practice today.

To compare such a swap to the switch to abacavir, Graeme Moyle and colleagues in London randomized 105 patients with moderate to severe lipoatrophy on thymidine analog-containing antiretroviral therapy and a viral load of less than 50 copies/mL to trade their thymidine analog to either tenofovir or abacavir while maintaining the other components of their pre-entry regimen.

The study arms were well matched, with the exception of stavudine exposure -- 77% of the tenofovir- versus 59% of the abacavir-assigned patients had been receiving stavudine at entry.¹⁰ Over 48 weeks, both groups experienced increases (300-400 g) in DEXA-assessed limb fat, with no significant differences observed between arms (Figure 4). Likewise, trunk fat also increased in both groups similarly. Tenofovir led to significantly lower total and LDL cholesterol as well as triglyceride levels; 1 tenofovir-assigned patient initiated lipid-lowering therapy compared to 8 abacavir patients. There were no differences between study arms in change from baseline in DEXA derived T-scores, although data on differences in the change in BMD itself was not reported. Only a single patient experienced a confirmed increase in viral load of more than 200 copies/mL. Therefore, tenofovir appeared to be as effective as abacavir in reversing antiretroviral-associated fat wasting but was associated with a better lipid profile; however, the imbalance between arms in stavudine use at baseline adds a confounding element.

Figure 4. RAVE: Median Changes at Week 48 in Regional Fat by DEXA

(Slide by Graeme Moyle; reprinted with permission.)

A related study, which was conducted in Spain, randomized 58 patients on a virologically suppressive regimen that included stavudine to continue their stavudine, switch the stavudine to tenofovir or reduce the dose of the stavudine to 30 mg twice a day.¹¹ Again, the switch to tenofovir produced increases in peripheral fat by DEXA. Surprisingly, the stavudine-dose reduction led to similar increases in extremity fat, but was not as likely as tenofovir to lead to reductions in LDL cholesterol or triglycerides. Curiously, for unexplained reasons, lean body mass decreased in the tenofovir arm, a finding that had not been previously described. Data on BMD were not reported.

When examined together, these studies provide some handy data on switch options for lipoatrophic patients. But, they also offer us a glimpse of what life would be like without thymidine analogs. On the plus side, it looks like fat wasting could become eligible to be placed on the "endangered metabolic toxicity list." However, many switch strategies arrive with their own toxicity baggage -- particularly the NRTI-sparing regimens, which may entail the use of agents more likely to increase lipid parameters. The data support the use of tenofovir in place of thymidine analogs when NRTI-related metabolic toxicity develops. Now that stavudine use has all but dried up (domestically, that is), the question is whether clinically evident lipoatrophy will continue to be seen, albeit at lower rates, among zidovudine users or among patients on quadruple-NRTI therapy. Further monitoring of body shape changes, even in the post-stavudine era, is warranted.

Lipids and Cardiovascular Disease

D:A:D Study

The D:A:D (Data Collection on Adverse Events of Anti-HIV Drugs) Study is one of those well-designed and ambitious efforts that enters the scene with impressive results, is warmly embraced and then continues to deliver tasty follow-ups as the study matures -- another fine example of such a study is the Multicenter AIDS Cohort Study (MACS). D:A:D is a prospective meta-cohort study tracking over 23,000 HIV-infected patients in Europe and North America and collecting data on the incidence and risk factors for myocardial infarction.

At CROI, El-Sadr, for the D:A:D study investigators, supplied us with an updated analysis of the cohort.¹² Since the last report, the number of myocardial infarctions has increased to 277: 59% confirmed, 24% possible and the remainder unclassifiable. About three quarters of the cohort are receiving combination antiretroviral therapy, with a mix of PI- and NNRTI-based regimens, 42% had dyslipidemia and an astounding 47% were current smokers.

Exposure to combination antiretroviral therapy continued to be associated with an increased risk of myocardial infarction on the order of a 17% increased risk per year on therapy (e.g., if your risk was 1% before combination antiretroviral therapy, with therapy, your risk increases to 1.17%). Older age, male gender, previous cardiovascular disease, smoking and family history of cardiovascular disease also continued to be predictive in multivariable analysis for myocardial infarction. The study now includes a sufficient number of women to permit analysis of myocardial infarction risk by gender. Women were also found to have a similar increased myocardial infarction risk with combination antiretroviral therapy exposure (relative risk = 1.23 [1.01-1.49]) compared to men. Atherogenic lipid profiles enhanced the chance of myocardial infarction, but did not fully account for the effect of HIV therapy.

The study continues to support the role of antiretroviral therapy in cardiovascular disease, and as longer-term data and endpoints accumulate, this conclusion becomes stronger. The study also makes clear that mutable risk factors for cardiovascular disease, such as smoking and dyslipidemia are highly prevalent among HIV-infected patients and should be targeted for interventions. Indeed, one wonders how many fewer myocardial infarctions there would have been in this cohort if more participants were ex- rather than current smokers. That said, it is important to put the risk of cardiovascular disease among these cohorts into perspective. A separate D:A:D presentation on mortality among participants found the leading cause of death was AIDS (30%), followed by liver disease (14%, of which 79% was associated with viral hepatitis) and malignancy (8%). Heart disease accounted for 9% of the D:A:D deaths.¹³

In the same vein, Klein and colleagues at California's Kaiser Permanente health maintenance organization (HMO) updated previously reported data on the risk of cardiovascular disease among the program's male HIV-infected members aged 35-64 years.¹⁴ The rates of cardiovascular disease among the HIV-infected patients from 1996 onward were compared to rates from a random sample of patients in the HMO without known HIV infection and of similar age to the HIV-infected group.

Over 5,160 HIV-infected patients were included. There were 131 cardiovascular disease events among these patients, including 81 myocardial infarctions. Both cardiovascular disease and myocardial infarction rates were significantly higher among the HIV-infected patients. However, PI exposure was not significantly associated with cardiovascular disease -- suggesting that any lipid-raising effect of this class of antiretroviral therapy was not associated with an increase risk of cardiovascular disease over the time period studied. Risk factor information was sparse and it is unclear whether HIV-infected patients simply have more cardiovascular disease risk factors than HIV-uninfected patients. This study also provides insights into the power that the potential for metabolic complications has on clinicians' prescribing behaviors. For example, during the study period stavudine use fell from 48% in 2001 to 17% at end of 2004, atazanavir (ATV, Reyataz) use is currently at 27% and lipid-lowering therapy among HIV-infected patients tellingly increased from 1% in 1997 to 22% in 2004.

Fish Oil for Antiretroviral-Associated Hypertriglyceridemia

There is an accumulated body of data supporting the benefits of consuming omega-3-fatty acids from marine sources (fish oil) in order to prevent cardiovascular disease. How fish oil helps to reduce the risk of cardiovascular disease is not completely understood given the many actions that have been attributed to fish oil, including triglyceride reduction, anti-inflammatory and inhibitor of platelet aggregation. The triglyceride-lowering effect of fish oil in HIV-uninfected cohorts is well described and, despite the absence of data in the setting of HIV infection, products containing fish oil are increasingly being prescribed to HIV-infected patients with elevated triglyceride levels.

To examine fish oil in such patients, Pierre De Truchis and other investigators from France conducted a well-designed, randomized, placebo-controlled trial of fish oil in patients with high triglyceride levels.¹⁵ The study enrolled 146 patients with fasting triglycerides of more than 300 mg/dL. All subjects received dietary counseling for the first 4 weeks of the study and those with triglyceride levels remaining above 200 mg/dL entered the randomized phase. A total of 122 participants were ultimately randomized to 2 g of fish oil 3 times a day or a placebo made from paraffin oil for 8 weeks. After week 8, all subjects received open-label fish oil and were followed an additional 8 weeks.

The baseline triglyceride levels were around 450 mg/dL in each arm and all patients were receiving combination HIV therapy. At week 8 there was a mean percentage reduction from baseline triglycerides of -25% in the fish oil arm compared to a 1% increase in the control group (P = .003). Further, 22% of the fish oil patients achieved a triglyceride level that was less than 200 mg/dL versus only 6.5% of the placebo-assigned patients (P = .012). Following the open-label phase, the placebo group experienced a mean 21% reduction in triglycerides. Total cholesterol decreased by 8.5% in the fish oil arm, but did not change in the control arm.

Figure 5. Maxepa Study: Global Evolution of Triglycerides (Mean) During the Study

(Slide by Pierre De Truchis.)

There were 16 adverse events considered to be possibly related to study medication (7 in the fish oil arm and 9 in the control). Gastrointestinal (GI) complaints were the most common toxicity seen in both arms. Whether paraffin oil itself can cause GI upset was asked during the question and answer session, but the investigator thought this unlikely.

This nice study demonstrates the activity of fish oil in the treatment of HIV-associated hypertriglyceridemia. Certainly, whether a 25% (or even a 50%) reduction in triglycerides is sufficient to negate antiretroviral-associated lipid effects is difficult to discern. However, cardiovascular disease risk increases along the continuum of atherogenic lipid elevation (i.e., the higher the level, the greater the risk). Therefore, a sizable reduction should lead to a sizable lowering of risk.

Some may complain that the study did not look at the effect of fish oil on the pharmacokinetics of HIV drugs and this is an important point; however, this trial provides the rationale for further and more extensive study of fish oil. In fact, the ACTG is in the midst of a study comparing fish oil and fenofibrate (Tricor); it includes pharmacokinetics and immunologic evaluations. Meanwhile, those of you who are either prescribing or taking fish oil may have a bit more faith in this alternative approach to lipid-reducing therapy.

Atazanavir Is Nice to Lipids

A couple of posters told us what we basically already knew, which was that a switch from an older PI-based therapy to atazanavir -- even when boosted with ritonavir (RTV, Norvir) -- leads to reductions in LDL cholesterol and triglycerides. Martinez et al in Barcelona analyzed lipid, immunologic and virologic data among the 255 patients who were enrolled in the Spanish atazanavir expanded access program.

All the patients had their atazanavir boosted with ritonavir and had fasting triglyceride levels greater than 500 mg/dL or total cholesterol greater than 200 mg/dL or LDL cholesterol greater than 130 mg/dL at baseline.¹⁶ Over 40% of the patients were switched from lopinavir/ritonavir and the remainder from other boosted PIs, nelfinavir or NNRTIs. Of course, when starting the boosted atazanavir, other antiretrovirals, such as the NRTIs, were also changed and such changes could also impact lipids. As expected, triglyceride and total and LDL cholesterol levels dropped after the switch. At a year, triglycerides were 43% lower and LDL cholesterol was reduced by 18%; lipid-lowering therapy usage dropped by a third. At the time of switch, only 50% of the patients had an HIV-RNA level of less than 500 copies/mL, but after one year 75% of the patients had a viral load below this level. Only 8 patients experienced treatment limiting toxicity, including 3 with icterus.

In another study, patients with an LDL cholesterol of greater than 130 mg/dL, and who were virologically suppressed on a PI (ritonavir boosted or not) regimen, were randomized to either switch the PI(s) immediately to unboosted atazanavir (400 mg daily) or delay the switch for 24 weeks.¹⁷ NRTI modification was permitted, but tenofovir prohibited due to the drug-drug interaction of tenofovir and unboosted atazanavir. A total of 246 patients were enrolled. Most switched from indinavir or nelfinavir. By week 12, in the treated arm, fasting LDL cholesterol dropped by 15%, triglycerides by 35% and cardiovascular disease risk marker LP(a) by 21% -- all statistically significant improvements.

These data support all previous reports of atazanavir's lipid friendliness. However, we have yet to see a study evaluating lipids after treatment-naive patients have initiated a ritonavir-boosted atazanavir regimen. It is unlikely boosted atazanavir is completely lipid "neutral." I suspect that such a study would demonstrate a rise in triglycerides and LDL cholesterol, albeit certainly to a lesser degree than that seen with drugs like lopinavir/ritonavir and nelfinavir. It will take an actual study, however, to tell if I am correct.

Conclusions

Body shape and lipid-related abnormalities are the most common clinically encountered metabolic disorders. At CROI, the spotlight appropriately returned to these and to other metabolic complications of HIV and its therapies. The attention is welcome, but its glare reveals exactly how much and how little we know about these conditions. Several excellent presentations reviewing the likely pathogenesis of lipid and body shape disorders displayed an improved understanding of the underlying mechanism of these problems and their inter-relatedness.

However, on the therapeutic end, there seemed to be less to be impressed with and one could leave with the impression that after all these years, what we have to show for our efforts are a few, somewhat obsolete, switch studies reinforcing the well appreciated message that stavudine should be avoided. However, these studies raise important issues. First, stavudine is hardly dead. Its worldwide use is increasing and the metabolic consequences of this reality will come back to haunt us. During a plenary session dedicated to lipodystrophy, Peter Reiss¹⁸ of the Netherlands made a strong appeal to examine and remedy the widespread use of stavudine in resource poor nations. Increased availability of alternative agents would reduce dependence on stavudine. Second, the switch studies really brought home the message that PIs and NRTIs act differently when it comes to fat redistribution. Lastly, the data presented suggests zidovudine may also play a role in fat wasting, although almost certainly not nearly as quickly or completely as stavudine.

The data on cardiovascular disease risk among those receiving antiretroviral therapy are mixed. Alarmingly, HIV therapy is associated with enhanced risk of myocardial infarction, but it must be appreciated that the rates are low. Further, a significant proportion of HIV-infected patients have risk factors for cardiovascular disease that can be modified, most notably smoking tobacco. HIV clinicians have spent considerable time learning about the use of lipid-lowering therapy in the setting of HIV infection. Now, we should also become well acquainted with techniques for smoking cessation. For as antiretroviral therapy becomes kinder and gentler, vis-à-vis lipids and cardiovascular disease, the same cannot be said of the Marlboro Man.

References

1. Mulligan K, Parker R, Komarow L, et al, and the ACTG 5005s and 384 Study Teams. Mixed patterns of changes in central and peripheral fat following initiation of ART. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 38.

2. Dubé MP, Zackin R, Yang Y, et al. Prospective study of regional body composition in antiretroviral-naive subjects randomized to receive zidovudine+lamivudine or didanosine+stavudine combined with nelfinavir, efavirenz, or both: A5005s, a substudy of ACTG 384. In: Program and abstracts of the 4th International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV; September 22-25, 2002; San Diego, Calif. Abstract 27.

3. Saag M, Tien PC, Gripshover B, Osmond D, Bacchetti P, Grunfeld C, for the Investigators of the Fat Redistribution and Metabolic Change in HIV Infection (FRAM) Study. Body composition in HIV+ men with and without peripheral lipoatrophy is different than controls. In: Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Mass. Abstract 733.

4. Murphy R, Zhang J, Hafner R, et al, and the AACTG 5110 Study Team. Switching to a thymidine analog-sparing or a nucleoside-sparing regimen improves lipoatrophy: 24-week results of a prospective randomized clinical trial, AACTG 5110. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 45LB.

5. Tebas P, Zhang J, Yarasheski K, et al, and the Adult AIDS Clinical Trials Group (AACTG). Switch to a protease inhibitor-containing/nucleoside reverse transcriptase inhibitor-sparing regimen increases appendicular fat and serum lipid levels without affecting glucose metabolism or bone mineral density. The results of a prospective randomized trial, ACTG 5125s. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 40.

6. Carr A, Workman C, Smith DE, et al, for the Mitochondrial Toxicity (MITOX) Study Group. Abacavir substitution for nucleoside analogs in patients with HIV lipoatrophy: a randomized trial. JAMA. July 10, 2002;288(2):207-215.

7. McComsey GA, Ward DJ, Hessenthaler SM, et al, for the Trial to Assess the Regression of Hyperlactatemia and to Evaluate the Regression of Established Lipodystrophy in HIV-1-Positive Subjects (TARHEEL; ESS40010) Study Team. Improvement in lipoatrophy associated with highly active antiretroviral therapy in human immunodeficiency virus-infected patients switched from stavudine to abacavir or zidovudine: the results of the TARHEEL study. Clin Infect Dis. January 15, 2004;38(2):263-270.

8. Brinkman K, Smeitink JA, Romijn JA, Reiss P. Mitochondrial toxicity induced by nucleoside-analogue reverse-transcriptase inhibitors is a key factor in the pathogenesis of antiretroviral-therapy-related lipodystrophy. Lancet. September 25, 1999;354(9184):1112-1115.

9. Gallant J, Staszewski S, Pozniak AL, et al. Long-term efficacy and safety of tenofovir disoproxil fumarate (TDF): a 144-week comparison versus stavudine (d4T) in antiretroviral-naive patients. In: Program and abstracts of the XV International AIDS Conference; July 11-16, 2004; Bangkok, Thailand. Abstract TuPeB4538.

10. Moyle G, Sabin C, Cartledge J, et al, and the Rave Study Group. A 48-week, randomized, open-label comparative study of tenofovir DF vs abacavir as substitutes for a thymidine analog in persons with lipoatrophy and sustained virological suppression on HAART. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 44LB.

11. Milinkovic A, Lopez S, Vidal S, et al. A randomized open study comparing the effect of reducing stavudine dose vs switching to tenofovir on mitochondrial function, metabolic parameters, and subcutaneous fat in HIV-infected patients receiving antiretroviral therapy containing stavudine. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 857.

12. El-Sadr W, Reiss P, De Wit S, et al, for the D:A:D Study Group. Relationship between prolonged exposure to combination ART and myocardial infarction: effect of sex, age, and lipid changes. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 42.

13. Weber R, Friis-Møller N, Sabin C, et al, for the D:A:D Study Group. HIV and non-HIV-related deaths and their relationship to immunodeficiency: the D:A:D Study. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 595.

14. Klein D, Hurley L, Quesenberry C, et al. Hospitalizations for CHD and MI among northern California HIV+ and HIV- men: additional follow-up and changes in practice. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 869.

15. De Truchis P, Kirstetter M, Perier A, Meunier C, Gardette J, Melchior JC, and the Maxepa-VIH Study Group. Treatment of hypertriglyceridemia in HIV-infected patients under HAART, by (n-3)polyunsaturated fatty acids: a double-blind randomized prospective trial in 122 patients. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 39.

16. Martinez E, Azuaje C, Antela A, et al, and the BMS study 900 ATV Early Access Program-Spain. Effects of switching to ritonavir-boosted atazanavir on HIV-infected patients receiving antiretroviral therapy with hyperlipidemia. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 850.

17. Sension M, Grinsztejn B, Molina J, et al. AI424067: improvement in lipid profiles after 12 weeks of switching to atazanavir from boosted or unboosted protease inhibitors in patients with no previous PI virologic failure and hyperlipidemia at baseline. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 858.

18. Reiss P. Lipodystrophy: fitting the pieces of the puzzle. In: Program and abstracts of the 12th Conference on Retroviruses and Opportunistic Infections; February 22-25, 2005; Boston, Mass. Abstract 65.