HAART-Related Fat Redistribution and Metabolic Complications
The phenomenon of fat redistribution (lipodystrophy) and metabolic complications of HIV infection or its therapy have received considerable attention in the past 18 months (see CRIA Update, Summer 1998). Over 30 original research studies on these topics were presented at the recent Sixth Conference on Retroviruses and Opportunistic Infections in Chicago. Although fundamental issues remain unresolved regarding the terminology and classification of the various clinical manifestations, this article will treat fat redistribution, abnormalities of glucose metabolism, lipid abnormalities, and coronary artery disease as separate entities, though they certainly overlap in many people.
The seemingly straightforward question of how frequently fat redistribution occurs has been difficult to answer. The lack of an agreed-upon case definition for diagnosis is a key reason for the discrepancies in reported prevalence of fat redistribution. Many investigators describe fat wasting and fat accumulation as components of fat redistribution which may occur separately or co-exist in some people. Fat wasting (lipodystrophy) generally refers to loss of the fatty layer beneath the skin which may lead to a prominent appearance of veins and a thinning of the arms and legs and hollowed-out appearance of the face. The fat accumulation aspect refers to increased abdominal fat (truncal obesity or adiposity), buffalo hump, and lipomas (fatty tumors beneath the skin).
A large study of the prevalence of body shape changes in persons receiving protease inhibitor therapy was reported at the Retrovirus conference by a group from France. Seventy-eight percent of nearly 500 consecutively evaluated patients had body shape changes, of whom 20% had fat wasting, 16% fat accumulation, and 42% a mix of fat wasting and accumulation. Patients in this study had received protease inhibitor therapy for an average of about 18 months. This high prevalence of fat redistribution was similar to that reported previously by an Australian group led by Dr. Andrew Carr and updated at the conference.
Other researchers described cases of fat redistribution in persons who had never received protease inhibitor therapy, most of whom were receiving d4T (Zerit®) and/or 3TC (Epivir®). These reports, along with others previously published or presented, imply that fat redistribution is not caused exclusively by protease inhibitor therapy, though use of this class of drugs appears to be a significant risk factor.
Several studies focused on potential causal mechanisms of fat redistribution. A leading hypothesis on the mechanism by which protease inhibitors may cause fat redistribution was published last year in the British medical journal The Lancet by Carr and colleagues. They theorized that protease inhibitors bind to human proteins involved in fat metabolism and inhibit their function, leading to inappropriate deposition of fat in some regions of the body and death of fat cells (fat wasting) in other regions. Investigators from Glaxo Wellcome studied some of the key steps in pathways related to fat metabolism in test tube studies with cell culture systems and found that different protease inhibitors either enhanced or inhibited steps in the pathways. While intriguing, these findings did not definitively support Carr's elegant hypothesis. Further work is clearly needed to sort out the mechanism(s) of fat redistribution.
Despite the lack of understanding of the mechanism of fat redistribution, several potential treatments are under study, and preliminary results were reported at the conference. Drs. Gabriel Torres, Kenneth Unger, and colleagues from New York updated a prior presentation of their series of patients with fat redistribution who have been treated with recombinant human growth hormone (Serostim®). They concluded that growth hormone at 5 or 6 mg/day was effective in reducing truncal obesity and buffalo humps but not peripheral fat wasting. Of note are their observations of a relatively rapid re-accumulation of fat after cessation of growth hormone therapy in one patient and the failure of a lower dose of drug (4 mg/day) in another. CRIA's study of growth hormone for truncal obesity, led by Dr. Donald Kotler, is fully enrolled and preliminary results are anticipated in the coming months.
Because insulin resistance is felt by some to contribute to or cause fat redistribution, investigators have begun to study the effects of diabetes drugs which combat insulin resistance as possible treatment. French researchers studied the effects of metformin, a drug which improves sensitivity to insulin, in 16 patients with truncal obesity and insulin resistance. Insulin resistance improved and visceral fat -- that is, fat which accumulates internally as opposed to beneath the skin -- was reduced by 38% at 16 weeks in the group receiving metformin compared with controls who had an average increase in visceral fat of ten percent.
Another European group studied troglitazone, an insulin-sensitizing agent, in a pilot study of six diabetic patients on protease inhibitors. While glucose levels decreased in these patients, effects on insulin resistance and fat distribution were mixed. Although no liver toxicity was seen in this small study, troglitazone has been associated with liver failure in HIV-negative diabetic patients, so potential safety concerns remain.
Several studies explored the effects on body composition of changing from a protease inhibitor-containing regimen to a protease inhibitor-sparing regimen. These studies are summarized in Tim Horn's article on the front page of this issue of CRIA Update.
Prior to the Retrovirus Conference, a number of research groups demonstrated an association between protease inhibitor use and insulin resistance. The body's resistance to the actions of insulin may result in elevated blood glucose levels (hyperglycemia), impaired glucose tolerance (often assessed by the body's ability to metabolize ingested glucose in an oral glucose tolerance test) and, in a small subset of patients, the development of diabetes.
Like the data on fat redistribution, there has been some variability in the reported prevalence of hyperglycemia and impaired glucose tolerance in patients on protease inhibitors. Behrens and colleagues reported that 62% of 98 patients on protease inhibitors had impaired glucose tolerance by an oral glucose tolerance test. Preliminary findings from CRIA's ongoing glucose tolerance study were reported and notable for the finding of glucose intolerance in four of 16 patients who had not yet started protease inhibitor therapy.
The significance of this result is the implication that some HIV-infected patients may be predisposed to altered glucose metabolism in the absence of protease inhibitor therapy and that these drugs may not be the sole cause of insulin resistance and hyperglycemia in the HIV-infected patient population. Walli and colleagues also described insulin resistance in about one-quarter of patients on nucleoside analogs but no protease inhibitors compared with 55% of patients on protease inhibitors.
Abnormalities in lipid metabolism have been well-documented in HIV-infected patients since the early years of the epidemic. Protease inhibitor therapy appears to exacerbate some of these abnormalities, specifically triglyceride elevations, and often results in elevations in total and LDL cholesterol and reductions in HDL cholesterol. These prior-recognized findings were confirmed by several groups of investigators at the Retrovirus conference.
Unfortunately, there was a paucity of data on the management of lipid abnormalities in these patients. Dr. Keith Henry of Minnesota updated his previously reported treatment data which demonstrated limited effects of a dietary and exercise intervention but clinically significant lowering of cholesterol and triglyceride levels with gemfibrozil and atorvastatin. Many of the so-called "statin" drugs have the potential to interact with protease inhibitors since they are metabolized by the same liver enzymes. Hopefully drug interaction studies and further safety data on mixing these classes of drugs will be forthcoming.
Coronary Artery Disease
One of the main concerns about the metabolic complications and fat redistribution phenomena relates to the theoretical increased risk of accelerated atherosclerosis (hardening of the arteries) which can lead to heart attacks and strokes. A handful of cases of premature coronary artery disease has been previously reported in relatively young HIV-infected persons on PI therapy, many of whom have had co-existing risk factors like cigarette smoking.
In an insightful talk, Dr. Carl Grunfeld of San Francisco discussed the theoretical impact of protease inhibitor-induced lipid abnormalities on the rate of heart attacks based on data from the classic Framingham heart study. In the absence of other risk factors, average protease inhibitor-induced elevations of cholesterol had minimal impact on the risk of heart attack for a 44-year-old man (1.4 additional cases per 100 men over ten years). Throwing in other risk factors such as high blood pressure and smoking, however, significantly increased the impact of the cholesterol elevations in the model. Dr. Grunfeld emphasized the importance of low HDL cholesterol as a risk factor for coronary artery disease.
Three studies examined the rates of heart attacks in patients on protease inhibitors. In one study, the rate of heart attacks was increased by five-fold in the era of protease inhibitors compared to the era prior to their availability; this difference, however, was not statistically significant. A larger study from Kaiser Permanente's database found no increased rate of heart attacks attributable to protease inhibitor use at one year of follow-up. Similarly, an analysis by Merck of their data from clinical trials of Crixivan® (indinavir) found no increased risk of heart attacks in patients receiving this PI compared with controls who did not receive protease inhibitors. Since the rate of heart attacks in these patient populations is relatively low, very large studies with longer follow-up might be needed to demonstrate or refute a higher risk attributable to protease inhibitors. Nonetheless, the data available to date seem to rule out a significant effect of protease inhibitors on risk of coronary artery disease in the relatively short term.
Studies of fat redistribution and metabolic complications presented at the Retrovirus conference have added some pieces to the puzzle but much remains to be solved. vast ongoing interest in this field of investigation is evidenced by the announcement at the conference of the First International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV, to be held in San Diego in June, 1999. Hopefully, our limited knowledge base in this area will have greatly expanded by next year's Retrovirus conference.
Marshall Glesby is director of Chelsea's Cornell Clinical Trials Unit, and asst. professor of medicine at Cornell University Medical College.