The 7th Conference on Retroviruses and Opportunistic Infections (CROI) included close to sixty reports and a symposium devoted to looking at the changes in metabolism and body composition that people on long-term antiretroviral therapy have been experiencing in recent years. These changes -- increased triglycerides, increased cholesterol, insulin resistance, glucose intolerance, fat accumulation in the abdomen, breasts and upper back, fat loss in the face, buttocks and limbs (lipoatrophy) -- had previously been lumped together under the term lipodystrophy. Osteoporosis and cardiovascular disease have recently been added to the list.

It is becoming increasingly clear that no single factor can explain such varied and confounding conditions. Protease inhibitors were initially thought to be the culprits because many of these conditions first appeared (or were paid attention to) in 1996, when the protease inhibitors came into use -- thus the terms "Crix belly" and "protease paunch." But 1996 was also the beginning of widespread use of long-term combination therapy. Protease inhibitors clearly play a major role, but mounting evidence shows that nucleoside analogs and HIV itself contribute to at least some of these conditions as well.

Many people are working hard to develop a common definition of this syndrome or, more likely, syndromes. Without a systematic definition in use, researchers and physicians employ various measurements to report data. Depending on the methods of measurement, studies of people on protease inhibitors, for example, have reported lipodystrophy prevalence rates ranging from 2% to 83%. These discrepancies are partly due to the lack of a common terminology, and they make it difficult to understand the causes and clinical implications of the conditions. They also make figuring out the risk/ benefit ratio of long-term antiviral treatment a far more complicated process for everyone.

Who's at Risk for What?

A number of studies at the conference reported on the prevalence of lipodystrophy syndrome in large groups of people. The CPCRA (Community Programs for Clinical Research on AIDS) evaluated 1,370 trial participants for fat accumulation or loss, diabetes and coronary artery disease. Almost 25% of the participants were antiviral-naive, 14% were women, 34% African-American and 12% Latino. Overall, the incidence of diabetes and coronary artery disease was almost identical in both the antiviral-naive and antiviral-experienced groups, although certain factors significantly increased the risks.

Diabetes rates were significantly higher in non-whites, older people, and those with higher CD4 and viral load, while older age was associated significantly with a greater risk of coronary artery disease, as might be expected. Overall, fat accumulation or loss appeared more frequently in older people, whites, and those who were antiviral-experienced (5.4% vs. none in antiviral-naive). This comparatively low rate suggests that African-Americans and Hispanics may be at lower risk for changes in body composition than whites. There was a clear correlation between length of time on treatment and an increased risk of body shape changes, regardless of protease inhibitor use: 14.3% for a subgroup with more than eight years of treatment compared to 2.7% for people with less than two years of treatment.

A study from Australia looked at body fat changes in 1,350 people surveyed between November 1998 and June 1999. Fat accumulation and/or loss was reported in 51% of all participants. Within this group, most of whom had extensive antiviral experience, body fat changes occurred in 81% who were protease inhibitor-experienced, 33% who were protease inhibitor-naive, and 5% who were completely antiviral-naive. As in many other studies, there was a clear correlation between length of treatment, particularly with a protease inhibitor, and the degree of fat redistribution; the longer you've been on treatment, the more likely you are to develop fat redistribution.

In real life, we know that women often experience different patterns of fat redistribution than men, but most studies have so many more male participants that it has been difficult to sift through the data by sex. An update of the SALSA study (Self-ascertained Lipodystrophy Syndrome Assessment) described metabolic and body shape changes in 395 HIV-negative participants (324 men, 71 women). Almost all are on antiviral treatment and most have taken protease-containing regimens (81% of the men and 63% of the women). Fat accumulation occurred in 84% of the men compared to 97% of the women, fat loss occurred in 77% of the men compared to 61% of the women, and a mix of accumulation and loss occurred in 65% of the men compared to 58% of the women. Significantly higher levels of triglycerides and cholesterol were reported in the men than the women, while elevated glucose levels didn't significantly differ by sex. These data seem to verify the experiences of many people on HAART -- women often have less fat loss and more fat accumulation than men do.

Longer duration of HIV infection also often correlates with higher rates of fat redistribution, but a report from Spain put a dent in that assumption. The abstract described a surprisingly high incidence of lipodystrophy in seventeen people who began therapy (d4T+3TC+indinavir) during or within 90 days of primary infection. Five of the seventeen switched from indinavir (Crixivan) to nelfinavir (Viracept) or nevirapine (Viramune) due to virologic failure or kidney stones. Within six months to two years of starting therapy, 35% of these individuals (four men, two women) developed lipodystrophy, defined as fat loss in the face, buttocks or limbs and/or fat gain in the abdomen, neck or chest. Incidence of lipodystrophy increased the longer people were on treatment. The overall incidence of elevated triglycerides and cholesterol was also high, 53% and 33% respectively.

Mitochondrial Damage

Some of the long-term side effects that people experience are certainly due to mitochondrial damage caused by nucleoside analogs, although it's unclear how much this damage contributes to fat accumulation or loss. Mitochondria, found inside all human cells, use oxygen, fat and sugar to produce energy for the cells. A single human cell can have thousands of mitochondria, depending on how much energy is required for the cell to function properly. Mitochondria use the enzyme polymerase gamma to reproduce. The nucleoside analogs (AZT, ddI, ddC, d4T, 3TC, abacavir) inhibit reverse transcriptase, an enzyme that HIV uses to work its RNA into human DNA. They can also inhibit polymerase gamma. The result is the production of fewer mitochondria and a greater likelihood of mutations in the mitochondria that are produced.

"Depending on the methods of measurement, studies of people on protease inhibitors have reported lipodystrophy prevalence rates ranging from 2% to 83%."

Mitochondrial damage caused by nucleoside analogs has been recognized since the introduction of AZT in 1987. Some of the resulting symptoms may have been under-diagnosed in the past, but as people have been on these drugs for longer periods of time, increasing attention is being paid to their role in metabolic and morphologic changes. The degree to which each of the nucleosides contributes to mitochondrial damage is unclear. d4T (Zerit) often gets the lion's share of the blame, but combining two or more nucleosides may inhibit polymerase gamma in such a way as to exponentially increase the risk. Mitochondrial damage may be responsible for many of the common side effects of the nucleoside analogs: myopathy (inflammation of muscle tissue), peripheral neuropathy (nerve damage in the feet and hands), pancreatitis, and low levels of red blood cells (anemia), neutrophils (neutropenia), or platelets (thrombocytopenia).

Lactic Acidosis

Two particularly serious conditions can also result from mitochondrial damage -- lactic acidosis and hepatic steatosis, or fatty liver. All of us are familiar with the aching muscles that often follow a physical workout. That soreness is caused by a buildup of lactate. Our bodies usually clear excess lactate, but mitochondrial damage can create very high levels of lactate in the blood, sometimes leading to lactic acidosis, a rare but potentially fatal condition. Symptoms of lactic acidosis are difficult to discern. They can include shortness of breath, abdominal pain, nausea, vomiting, fatigue and weight loss, subtle symptoms that can easily be ignored or mistaken for something else. If you experience these symptoms while on nucleoside analogs, see your doctor right away.

There are no simple blood tests to check lactate levels. However, serum bicarbonate levels are measured as part of routine blood work, and low levels are a sign that some kind of excess acid production is occurring. If you're taking a nucleoside and your serum bicarbonate levels are low, lactic acidosis should be suspected. Although riboflavin and coenzyme Q10 are sometimes used to treat lactic acidosis, there is no evidence yet to support the value of either. Usually, the only recourse is to stop nucleoside analog therapy or, if appropriate, reduce the dose.

At CROI, a poster from the Netherlands described four cases of fatal lactic acidosis. The four people had been on nucleoside-containing combinations (all with d4T) for six to twenty months, and all had previously experienced at least one nucleoside-related side effect. They entered the hospital with gastrointestinal and respiratory problems and died within three weeks.

A team from Johns Hopkins University looked at lactate levels in 509 individuals who had been on combinations that included two nucleosides and a protease inhibitor for varying lengths of time. Although these data are only suggestive, people on combinations that included d4T+3TC had significantly higher lactate level than those on AZT+3TC, d4T+ddI or AZT+ddI. The potential for these people to develop lactic acidosis is unclear.

Another poster discussed abnormally high lactate levels in twenty patients on nucleoside analog-containing regimens (again, all included d4T) at the University of California Medical Center in San Diego from July 1998 to September 1999. The problems were identified early enough that no deaths resulted, and all twenty had normal lactate levels within 7 to 176 days of stopping antiviral therapy. Three of the twenty resumed antiviral therapy (without d4T), and still had normal lactate levels three months later.

A disturbing poster described the history of a child who developed extreme mitochondrial damage. At three months of age, he started AZT+ddI+nelfinavir, which resulted in a good clinical response -- undetectable viral load and rising CD4s. A year and a half later, however, he had unusual patches on his brain, elevated lactate levels, liver damage, severe atrophy of muscle and nerve fibers, and an astounding 79% depletion of mitochondrial DNA compared to HIV-negative children his age. He was taken off antiviral therapy for three weeks, during which time his viral load rebounded. Then he was started on a combination of ritonavir+nelfinavir+efavirenz (Norvir®+ Viracept®+Sustiva®) and his condition has improved. This is the first reported case of a child experiencing such severe mitochondrial damage, seemingly as a direct result of nucleoside analogs.

Early last year, investigators in France reported on two HIV-negative one-year-olds who died of neurologic disease associated with mitochondrial damage and whose mothers had taken AZT+3TC during pregnancy. As more attention is directed at the potential for nucleoside analogs to cause damage to the mitochondria, an effort is underway in the United States to look at HIV-negative children born to positive women who took nucleosides (primarily AZT) during pregnancy. Using databases from the National Institutes of Health and the Center for Disease Control, this effort has so far focused on 227 HIV-negative children who have died for any reason. Mitochondrial damage has not been found in any of these cases. The next step is to look for possible mitochondrial damage in the thousands of children who are alive.

Nucleosides and Fat Redistribution

Fat tissue cells also contain mitochondria. If nucleosides affect these mitochondria, lipid levels could rise in the blood and excess fat could build up in other body tissues, contributing to abnormal fat distribution. A poster from a French team looked at fat redistribution in 83 people from the ALBI trial, all of whom had been on dual nucleoside regimens for two and a half years. Half of the group had added either a protease or non-nucleoside to their regimen. Although 35% of the entire group had at least one symptom of lipodystrophy at 30 months, the lipodystrophy rate was 37% for the 42 who had remained on only two nucleosides the whole time. Of the dual nucleoside arms (d4T+ddI, AZT+3TC or d4T+ddI followed by AZT+3TC), the d4T+ddI arm showed significantly higher rates of fat redistribution.

Another French team looked at fat redistribution in 149 protease-naive study participants taking combinations of two, three or four nucleosides, with an average time on antiviral therapy of almost four years. Lipodystrophy was diagnosed in close to 40% of the participants -- 29% with fat loss, 39% with fat accumulation, and 32% with a mix of both. There was no significant difference in triglyceride, cholesterol or glucose levels between those with fat redistribution and those without. Of particular interest, lipodystrophy was more frequent in people taking d4T than those taking AZT.

Other studies also implicated d4T as having a particular role in causing both metabolic and body shape changes. One of many reports on switching or stopping therapy as a means of improving these abnormalities showed that stopping d4T significantly normalized triglyceride and lactate levels as well as resulting in major or partial improvement in fat loss. These data are particularly important because of unresolved questions about whether or not fat redistribution is reversible.

Switch Studies

Most of the "switch" studies involved replacing a protease inhibitor with a non-nucleoside (NNRTI), in most cases, Sustiva®. Sustiva® is known to raise lipid levels in some people. The good news in most of the switch studies is that viral loads generally remained stable upon switching. The results concerning metabolic and body shape abnormalities are inconclusive, showing very little change in fat redistribution in most cases, even a year after switching. A Spanish study, for example, reported that switching 31 people with lipodystrophy from d4T+3TC+Crixivan to d4T+ddI+Viramune® resulted in significant improvements in cholesterol and triglyceride levels, but no significant improvement in body shape after nine months.

A French study reported on 32 people, all of whom had been on two nucleosides plus a protease inhibitor. Half of them stayed on their original regimen; the other half substituted abacavir (Ziagen®) for their protease inhibitor. Three months after switching, the protease group continued to experience elevated triglycerides, glucose intolerance and clinical signs of lipodystrophy. Triglyceride levels and glucose metabolism had improved for participants in the three-nuke arm, but there was only minor improvement in body shape abnormalities. In another study, there were significant improvements in cholesterol, triglycerides and insulin sensitivity six months after 106 people substituted their protease with Ziagen®. In a subset, three out of nine people who had lipodystrophy before switching to abacavir reported body shape improvement. The nine who stayed on their protease reported further fat redistribution.

HAART and Heart Disease

There were several reports concerning the possibility that people on protease inhibitors may be at increased risk for cardiovascular disease. A study from the University of Wisconsin found that blood vessel function (endothelial function) was impaired in 21 people on protease inhibitors compared to seven who weren't on a protease inhibitor. Using ultrasound technology, the study measured the width of the participants' brachial arteries, the main artery in the upper arm. Endothelial dysfunction can put you at higher risk for coronary artery disease. Some studies found no similar correlation, while others reported a correlation between coronary heart disease and HIV, but not necessarily protease use. Most of these studies, however, looked at very few people over a relatively short period of time.

A retrospective study compared hypertension in 42 people with lipodystrophy to 42 HIV-positive people without lipodystrophy and 13 people who were HIV-negative. High blood pressure rates were significantly greater in the group with lipodystrophy; yet blood pressure was also higher in the HIV-negative participants without lipodystrophy compared to the HIV-negative group.

Two studies looked into large existing databases to get an idea of the prevalence of coronary heart disease. The records of 4,526 HIV-negative people who received care through Northern California Kaiser Permanente between January 1996 and June 1999 were looked at. Rates of hospitalization for coronary heart disease in people on protease inhibitors were compared to those of people with HIV not on a protease inhibitor. There was no difference between the two groups. Another poster pooled data from various clinical trials that included protease inhibitor-containing and nucleoside-only arms. Rates of myocardial infarction (cell death in the heart wall) within the arms were compared to each other and to rates from large non-HIV databases. There was no evidence of increases in myocardial infarctions, although the mean follow-up was only one year.

"Mitochondrial damage may be responsible for many of the common side effects of the nucleoside analogs. . . ."

Elevated lipids, often found in people on protease inhibitors, may increase the risk of cardiovascular disease. Therefore, it's difficult to know what to make of the short-term, somewhat contradictory results reported at the conference. A large, international study through EuroSIDA will look at the incidence of cardiovascular events (heart attacks, strokes, myocardial infarctions) in more than 30,000 people with HIV over a period of two years. These data will be enormously helpful in clarifying both prevalence and risk factors. Until they're available, however, monitoring lipid levels and blood pressure remain our best clinical tools.

So Now What?

An overview like this can only touch briefly on some of the recent data concerning the changes in metabolism and body composition that people on HAART are experiencing. Progress is being made toward a better understanding of the causes, clinical implications and potential management of these long-term side effects. But even with so much data on the subject presented at the retrovirus conference, a clear understanding seems far in the future. As it becomes increasingly evident that researchers, physicians and people with HIV need to be talking the same language to accurately describe these conditions, it becomes equally evident that the development of useful treatment strategies is more complicated than ever.

Risk factors for changes in metabolism and body composition appear to be as varied as the individuals who develop them. In addition to protease inhibitors, nucleoside analogs, and at least one non-nucleoside (Sustiva®), factors that contribute to at least some elements of the syndrome include older age, longer period of HIV infection, longer period of time on treatment, and lower CD4 count and percentage before treatment. Sex, race and genetic predisposition also appear to play a role, although perhaps these are more indicative of the type of changes that might occur.

At the conclusion of the symposium at CROI, Dr. William Powderly of Washington University School of Medicine in St. Louis offered a cautionary overview of how little we understand about the causes, treatment, and clinical implications of these metabolic and body shape changes. Without denying the remarkable success of antiretroviral treatment, he urged the need to reassess the current goals of treatment as outlined in the federal treatment guidelines, including the questions of when to start treatment and which drugs to start with. As Powderly pointed out, the goal of therapy is to let your patient "live long and prosper." When you factor in the possible long-term side effects of treatment, figuring out how to best achieve this goal is difficult. Powderly concluded his summary with a fact that may be obvious but is sometimes downplayed: "The success of current therapy has a price."

Protease Inhibitors & Osteoporosis

Two reports at CROI highlighted a possible connection between the use of protease inhibitors and osteoporosis. An Australian group looked at 74 men who had been on protease inhibitor-containing combinations for various lengths of time and were experiencing symptoms of lipodystrophy. 28% had evidence of osteopenia (moderately low bone mineral density), and almost 10% had evidence of osteoporosis (severely low bone mineral density). Although there was a correlation between these conditions and lower weight and lean muscle mass, there was no correlation with the duration of antiviral therapy, baseline CD4, or baseline viral load.

The other report, from Washington University School of Medicine in St. Louis, looked at the rates of osteopenia and osteoporosis in 64 men on a protease inhibitor-containing combination compared to 36 HIV-negative individuals not on a protease and 22 HIV-negative individuals. 50% of those on a protease had osteopenia, and 21% had osteoporosis, clearly significant numbers. In this study, there was no association with weight or muscle mass. So far, there is no understanding of the relationship between these conditions, which can lead to bone fractures, and either HIV itself or protease inhibitor use.

James Learned is the National Technical Assistance Program Director at CRIA and a founding member of Hepatitis C Action & Advocacy Coalition (HAAC).