HIV-Associated Adipose Redistribution Syndrome (HARS) and Other Metabolic Disorders

The 39th ICAAC devoted significant attention to the HIV-associated adipose redistribution syndrome (HARS), the new umbrella term for the various body fat composition changes that have become a grave concern to many patients on HAART. The disorder has, accordingly, become the center of attention of antiretroviral drug developers, clinicians, marketing specialists and stock investors. Many have committed themselves to finding strategies to reverse HARS or a preventive strategy to avoid it altogether.

The natural history of HARS is still unclear, but various prospective epidemiological studies seem to provide a better idea of its incidence and prevalence, as well as potential causal associations. The largest study, called the HIV Outpatient Study (HOPS), was conducted by Ward et al.(1) and included 1077 individuals (15% women and 27% non-Caucasian) who were recruited from eight large HIV clinics from seven cities in the US. Patients were evaluated with face-to-face interviews and exams for evidence of abnormal fat distribution. The researchers devised a new classification system that took into account the severity of the abnormality as well as the number of body parts affected. Results indicated a significant association between fat maldistribution and days on d4T, days on indinavir, increasing age, time since HIV diagnosis, the minimum CD4 percentage and a greater CD4 percentage rebound since HIV diagnosis. High levels of cholesterol and triglycerides were also correlated with the severity of fat redistribution. While the rate of various conditions varied, the most common were central obesity (62%), fat loss in extremities (54%), fat loss in hips and buttocks (47%), sunken cheeks (38%) and buffalo hump (21%).

A second prospective study by Romeu and others determined the risk of developing aspects of lipodystrophy in a cohort of over 450 HIV-positive subjects.(2) They found that both the incidence of body composition and metabolic changes increased according to time on protease inhibitors with 26% developing hypercholesterolemia (high levels of cholesterol in blood) after one year, 51% after two years and 83% after three years.

The third prospective study(3), conducted by Muurahainen and colleagues, gave preliminary results of the first 210 patients enrolled (160 men and 50 women). For this study, both the patient and caregiver had to agree on the diagnosis of lipodystrophy. Results showed that women had a higher mean body mass index (BMI) of 29 compared with men (25) and were more likely to report body habitus changes than men. Men, on the other hand, more often reported fat depletion, especially facial fat wasting.

There are a number of different theories about how HARS might evolve. Several researchers investigated the hypothesis that protease inhibitors affect retinoic acid metabolism which regulates fat cell (adipocyte) differentiation and survival. It has been postulated that protease inhibitors bind to cis-retinoid acid binding protein type 1 (CRABP-1), in a manner similar to how they bind to the HIV protease enzyme, but now x-ray crystallography of the implicated proteins note clear differences in their three-dimensional structures. Thus, protease inhibitors are not binding to CRABP-1 -- at least not directly. In addition, the conversion of trans- to cis-retinoic acid -- theorized to be blocked by protease inhibitors, thus affecting adipocyte maturation and fat storage -- was not affected by cytochrome p450 inhibition. This had been the proposed reason that ritonavir was most closely associated with the hyperlipidemia (high blood lipid levels).

Another hypothesis about the pathogenesis of HARS is that it is partly the result of the mitochondrial toxicity caused by nucleoside analogs. One particularly severe expression of this toxicity is lactic acidosis, a rare but deadly metabolic disorder. Lactic acidosis can present as weight loss, fatigue, malaise, nausea, vomiting, abdominal pain, shortness of breath and low serum bicarbonate levels. The incidence of lactic acidosis were presented by a group from the FDA(4), who described 108 cases. Fifty-six were associated with a single NRTI and 60 with dual NRTI therapy, the most common being d4T/3TC (36 cases), followed by ddI/d4T (nine), AZT/3TC (seven), AZT/ddI (seven) and AZT/d4T (one). Liver biopsies revealed hepatic fatty infiltration in 69%, and 22% of the patients had pancreatitis. Half of these cases were female and the associated mortality rate was 56%. The trend implicating d4T/3TC may be just reflecting a greater utilization of this dual combination. Another report found an incidence of lactic acidosis of 0.84% per year (11 patients out of 867) in patients who had received NRTI therapy for an average duration of 27 months. Ten patients survived due to early diagnosis and discontinuation of NRTI therapy.

In addition to lactic acidosis, various studies implicated the nucleoside analogs d4T, ddI and 3TC in the pathogenesis of HARS. Polo and others noted a marked difference in the rate of lipodystrophy among patients on regimens containing d4T/ddI (89%) and d4T/3TC (52%) as compared to those on AZT/3TC (4%).(5) Most subjects in this study received indinavir. Women were more likely to have fat accumulation than men, yet men were more likely to have fat depletion and hyperlipidemia.

Treatment of HARS

Various strategies to ameliorate the signs and symptoms of HARS were presented at ICAAC. Several groups felt that the same rules for treatment of hyperlipedemia promulgated by the National Cholesterol Education Program (NCEP) for HIV-negative persons should be observed in HIV-positive patients because the latter often have other risk factors for coronary artery disease, such as smoking, diabetes, a family history, age, hypertension and obesity.

There are a large number of pharmaceutical treatments for high cholesterol and other lipid disorders. However, there is concern about using some of these drugs in people on HAART because there is a potential for drug interactions since many share the same metabolic pathway (CYP3A4) as many antiretrovirals. Pravastatin and fluvastatin, though, do not, so they are favored for treatment of hypercholesterolemia. Similarly, because the most common agents to treat hypertriglyceridemia -- gemfibrozil and troglitazone -- inhibit and induce (respectively) that metabolic pathway, more recently approved drugs that do not, rosiglitatzone and pioglitazone, are preferred. In addition, a separate study of patients with HARS receiving fenofibrate showed that 72% had a reduction in triglyceride levels in three months.(6) Another drug, metformin, used in adult onset diabetes, was evaluated in a small study of HIV-positive patients with lipodystrophy and increased abdominal fat (assessed by abdominal CT scan). However, no improvements in fat redistribution or lipodystrophy were noted.

Several open-label clinical studies have demonstrated a significant effect of recombinant human growth hormone (rHGH) on some of the manifestations of HARS. Gabriel Torres, MD, (author of this article) and his colleagues from New York presented the first report at the 6th National Conference on Human Retroviruses and Opportunistic Infections in 1998. He described a middle aged woman with wasting syndrome who had developed truncal adiposity, a buffalo hump, peripheral lipodystrophy and hyperlipidemia. Upon receiving rHGH, a marked reduction in the adipose tissue collections and improvement in fat-free mass were noted, despite no improvement in lipodystrophy and hyperlipidemia. Torres subsequently published a report(7) of ten patients treated with routine doses of rHGH (4-6 mg/day, sq) who have had responses of varying degrees to the hormone. Some patients had complete resolution of the truncal adiposity and buffalo humps while other patients with uncontrolled hyperglycemia and advanced HIV disease saw only minimal effects. Side effects included myalgia (muscular pain), flu-like symptoms, carpal tunnel syndrome, hyperglycemia and joint and facial swelling.

Additionally, Wanke and others from Harvard reported that another group of ten patients with HARS treated with subcutaneous injections of rHGH had reductions in waist-to-hip ratios and mid-thigh circumference measurements, as well as increases in body mass indices.(8)

Serono Labs, maker of rHGH, is conducting a large survey of HIV-positive patients with and without manifestations of HARS in order to establish its overall prevalence and incidence. Serono is planning a randomized, placebo-controlled study of rHGH in patients with HARS that will use DEXA (Dual Energy X-Ray Absorbtiometry) to determine changes in body composition and abdominal CT scans to evaluate visceral adiposity. Until the FDA approves rHGH for the indication of lipodystrophy, insurance companies can reject claims for this treatment. Nevertheless, most are aware of the psychological distress caused by the syndrome and the potential negative consequences of nonadherence to HAART, so they have sometimes covered three months of rHGH therapy.

The strategy to reverse HARS that received the most attention at the conference was switching from a protease inhibitor-based regimen to one based on an NNRTI. A recent report(9) described 23 patients with HARS who switched from a protease inhibitor to nevirapine due to psychological distress associated with body changes. After a median follow-up of eight months, significant improvements were noted in serum cholesterol and triglyceride levels, blood glucose and waist-to-hip ratios. All but two patients reported an improvement in peripheral fat wasting though none felt that they had regained their original body composition.

The UK's Graeme Moyle reported 48 week follow-up data of nine patients and 24 week follow-up data for another 20 patients whose fat redistribution, after switching from protease inhibitor to efavirenz therapy, was evaluated by self-report, DEXA and CT scans.(10) Most gained weight and lean body mass and lost visceral fat. At 48 weeks, there were improvements in visceral abdominal tissue and improved glucose tolerance. Most cases of hyperlipidemia did not improve, but there was a trend towards normalization of subcutaneous fat. In addition, the Maintavir study(11) described 53 patients who changed from protease inhibitors to nevirapine or efavirenz. After six months of follow-up, all but seven patients had maintained viral suppression with an average CD4 cell increase of 83 cells. Eight of twelve patients with HARS had subjective improvements in body habitus changes and seven had objective improvements, but there was no reversal of abnormal fat changes in lipid or glucose levels.

Another group from Strasbourg(12) followed 51 patients who switched from a protease inhibitor to efavirenz and had sustained viral suppression of under 20 copies/ml after six months of follow-up (the two patients who did not maintain suppression had stopped all antiretrovirals). Both triglyceride and cholesterol levels fell slightly and no changes in body habitus were reported. Finally, in DPC 006, the pivotal study comparing efavirenz to indinavir, both plus AZT/3TC(13), four cases of lipodystrophy occurred among those in the efavirenz arm, versus nine in the indinavir arm and twelve among those taking efavirenz and indinavir. The effects of protease inhibitors may take a long time to reverse and continued nucleoside therapy may also be contributing to the fat redistribution.

It is also possible that not all protease inhibitors will be as frequently associated with HARS. Structurally, amprenavir, being non-peptidic, is significantly different from the other protease inhibitors; and over the course of its clinical development, there have been few reported cases of fat redistribution. At the recent Lisbon meeting(14), there was a report on a randomized trial comparing amprenavir to indinavir in a group of 504 heavily NRTI-experienced, protease-inhibitor-naive patients who were experiencing virologic or treatment failure. In the 48-week, intent-to-treat analysis, 30% of the amprenavir group and 46% of the indinavir group attained <400 copies RNA/ml. But the difference in response on this analysis was caused by a high drop out rate due to adverse effects, clinical progressions or other reasons -- the difference in the incidence of virologic failures was less (26% and 18% respectively). Notably, only 3% of the amprenavir group developed fat redistribution as compared to 12% of those receiving indinavir -- a statistically significant difference. There were minimal changes in cholesterol and triglyceride levels reported in this study, but grade III and IV laboratory abnormalities were more frequent in the indinavir group (32%) than in the amprenavir group (20%).

These results support the in vitro work reported last year in which amprenavir had little effect on the metabolism of fat in cultured cells as compared to other protease inhibitors.(15)


1. Abstracts 1298 and 1299. 2. Abstract 1293. 3. Abstract 1301. 4. Abstract 1284. 5. Abstract 1306. 6. Abstract 2063. 7. AIDS, 13: in press. 8. Wanke et al. AIDS, 13: 2099-2103. 9. AIDS 1999, 13:805-810. 10. Abstract 2064. 11. Abstract 2198 12. Abstract 2196. 13. Abstract 1304. 14. Abstract 813. 15. Lenhard et al. 6th Conference on Retroviruses and Opportunistic Infections, 1999, Chicago , Illinois, Abstracts 665 and 666.

Back to the GMHC Treatment Issues September/October 1999 contents page.