Lipodystrophies, by contrast, are a relatively new body-composition abnormality -- first reported in significant numbers following the widespread adoption of multidrug antiretroviral therapies that contained at least one protease inhibitor. This temporal coincidence initially led a number of researchers to implicate this potent class of anti-HIV drugs as the source of these unusual and often unsightly fatty deposits, but it now seems clear that lipodystrophies occur in patients who have received only protease-inhibitor-sparing antiretroviral regimens.
From a clinical standpoint, the news regarding body-composition abnormalities is decidedly mixed: The good news is that several efficacious treatments are now available to increase lean body mass and improve physical performance in patients who develop wasting syndrome. The bad news is that there is still very little that we can do to reverse lipodystrophy when it occurs, and there is much that we still do not understand about this phenomenon.
The loss of more than 5% of total body weight is associated with worse prognosis in recent as well as older studies in HIV-infected individuals (4-6). Both lean tissue and fat stores are depleted when seropositive patients lose weight, but prognosis and function correlate with lean-tissue losses, not with the loss of body fat or of weight per se (5). Although it has proven considerably more difficult to restore lean tissue than body fat (7-11), rebuilding lean body mass is what matters: treatments that successfully increase lean body mass also increase objective measures of strength or physical performance (12-14), whereas the mere accrual of body fat is not associated with improved performance or survival (9, 10).
The primary therapeutic objective of treating patients for wasting is, therefore, the accrual of functional lean tissue, with attendant improvement in performance. A number of therapeutic options have been tested with this clinical objective in mind, among them appetite stimulants.
Megestrol acetate is an extremely effective appetite stimulant in people with advanced HIV infection and weight loss (9, 10). Food intake generally increases by 20% to 25% (i.e. 500-600 kilocalories per day), and weight gain averages about a pound per week, when patients are given high doses of megestrol acetate. Between 70% and 100% of the weight gained is body fat, however, and no improvements in function or survival have been demonstrated in large multicenter trials. These fat gains generally occur on the abdomen, trunk, back, and face -- a pattern that tends to exaggerate the changes in body-fat deposition associated with protease-inhibitor-containing antiretroviral regimens (1, 3).
Megestrol acetate therapy also has some adverse effects. It can lead to diabetes or insulin resistance in both sexes, and in men it results in a high incidence of hypogonadism, which is marked by lower serum testosterone concentrations, fatigue, and loss of libido (15). These adverse effects may exacerbate problems that are already present due to HIV disease itself or to the effects of antiretroviral therapy.
The reasons that megestrol acetate therapy primarily increases body fat and fails to increase lean tissue remain uncertain. Secondary hypogonadism may be a contributory factor in men, although a study that combined megestrol with testosterone replacement did not improve lean-tissue response (15). This may be due to the fact that megestrol has intrinsic glucocorticoid-agonist actions, and these may favor body-fat anabolism and lean-tissue catabolism. Alternatively, disturbances of macronutrient metabolism due to HIV infection itself may intrinsically limit lean-tissue anabolism (7, 11, 13, 17-19).
The usefulness of megestrol acetate in HIV-related wasting is limited, given these disappointing results. For people who simply want to gain weight or get more enjoyment out of meals -- or who want more body fat, to counteract facial wasting, for example -- this is an effective agent. Optimal dosing is 400-800 mg (10-20 mL of the oral suspension), taken once a day in the morning. Higher doses lead to more rapid weight gain, but they may increase the tendency for fat accrual to predominate. At San Francisco General Hospital we routinely initiate testosterone replacement therapy prophylactically, to prevent secondary hypogonadism, whenever we prescribe megestrol acetate (16).
Marinol, an orally available cannabinoid derivative, has not been proven to increase body weight or even food intake in HIV-associated wasting. Subjects have reported increased appetite, but little or no weight gain has been observed (20). Sedation, impaired concentration, and other adverse psychoactive effects are common with marinol therapy -- which makes administration to patients with a history of substance abuse problematic. If this appetite stimulant is used, side effects can be minimized by administering the agent as a single dose at bedtime. Studies with inhaled marijuana are ongoing, but results have not yet been reported.
Total parenteral nutrition is rarely indicated for treatment of HIV-associated wasting. TPN tends to increase body fat rather than lean tissue in these patients, although a subset of patients with relatively pure "starvation" (i.e., deficient nutrient intake in the absence of infection or inflammation) did demonstrate increases in lean-tissue mass (8). Preferential accrual of fat over lean has similarly been observed when the cachexia associated with end-stage cancer is treated with TPN (21, 22). In cancer wasting, meta-analyses suggest that TPN reduces rather than increases survival, due to infectious complications. Long-term, prospective, controlled outcome studies of TPN in HIV infection have not yet been reported.
Intractable cryptosporidial watery diarrhea, which was a major indication for the use of TPN in previous years, has virtually disappeared since the advent of combination antiretroviral therapy -- and so, as a result, has the need for total parenteral nutrition. Because administration of TPN is expensive, associated with infection and other risks, inconvenient, and of unproven efficacy for increasing lean tissue or overall function, we do not recommend its use except in special circumstances.
This recommendation does not apply to enteral nutrition, in the form of oral supplements -- which can be an extremely useful adjunctive therapy in HIV-associated wasting. Patients with clear histories of nutrient deficiency (23) or with objective biochemical measures indicating "starvation" rather than inflammation (11) may accrue lean-tissue mass when given oral supplements. Many forms of enteral nutrition are currently available, and most are well tolerated by patients with AIDS-wasting; experienced dietitians can generally find dietary supplements that are acceptable even to patients with diarrhea, altered taste, or oropharyngeal pain. As a rule, however, oral supplementation does not significantly increase lean body mass (11).
Over the past few years there have been major advances in the therapeutic use of androgens to increase lean body mass. From a clinical perspective, it is useful to separate discussions of androgen use into three categories: in states of low androgen level (hypogonadism), in eugonadal states, and in combination with exercise.
It is well established that administration of male hormones to any person who does not have the testosterone levels of a normal adult male -- that is, to women, pre-pubertal boys, or hypogonadal men -- results in substantial nitrogen retention, an increase in lean body mass that includes muscle, and improvements in strength. This is, of course, why use of androgens is banned among female competitive athletes.
This pattern of response to testosterone supplementation is highly relevant to the treatment of HIV-associated wasting, because hypogonadism, defined as a significant drop in serum testosterone concentrations, is the most common endocrine abnormality in HIV-infected men. Close to 50% of men with HIV infection exhibit a persistent lowering of serum total testosterone in the course of their illness (17). The typical reduction, however, is not into the frankly hypogonadal range (< 200 ng/dL) but into the lowest quartile or tertile of serum values (200-500 ng/dL) (13, 17). In a study that my colleagues and I conducted, we found that the bell-shaped population distribution observed in healthy controls is not present in Bay Area men with HIV-associated weight loss (13). The most common value for serum total testosterone was 200-450 ng/dL, which was present in close to half the subjects, rather than the 450-700 ng/dL range that was most common in healthy controls. This borderline hypogonadism may represent the "hypothalamic hypogonadism" seen in other settings of stress, infection, or weight loss (24, 25).
These findings suggested that there were two important clinical questions that needed to be addressed regarding androgen replacement therapy in HIV- associated wasting: First, is this treatment effective in frank hypogonadism (testosterone < 200 ng/dL), or is there resistance, in HIV disease, to the usual anabolic effects of androgens? And second, do patients with borderline hypogonadism also benefit? In other words, is this commonly observed condition a functionally hypogonadal state? Both of these questions have now been answered in the affirmative.
Grinspoon et al. gave testosterone injections to men with serum testosterone levels in the frankly hypogonadal range and observed significant increases in both fat-free mass and overall body weight (15). My colleagues and I have compared intramuscular nandrolone decanoate to placebo injections in men with weight loss and total testosterone levels between 200 and 500 ng/dL (13). The nandrolone arm showed a significant positive nitrogen balance (equivalent to about 0.5 kg of muscle gained per week) over the first two weeks of treatment when compared with the placebo arm.
After 12 weeks of nandrolone decanoate therapy (100 mg/wk), participants gained, on average, about 3 kg of lean body mass and 4.5 kg of total weight. These values for lean-tissue accrual with nandrolone are similar to results achieved with recombinant growth hormone, but at much lower cost. Some measures of exercise performance also improved with nandrolone therapy. In our view, it is therefore reasonable to expand the criteria for androgen replacement therapy to include not only frank hypogonadism but also borderline hypogonadal patients -- that is, those in the lowest quartile or tertile (values < 500 ng/dL) who have weight loss or other signs or symptoms of functional hypogonadism.
Administration of replacement doses of androgens to healthy, eugonadal men has little effect on body composition. This might be expected, since endogenous production of testosterone is suppressed by exogenous testosterone. Administration of supraphysiologic doses of androgens to healthy men, in the absence of exercise, typically has modest effects on body composition, with little or no increase in strength (26). Similar results have been reported in eugonadal men with HIV-associated weight loss (27).
The exception to the rule just noted for eugonadal men is when high-dose androgens are combined with resistance exercise such as weight-lifting. Bhasin et al. performed an important placebo-controlled study of this therapeutic combination in healthy, eugonadal men. In this cohort, very high doses of testosterone (600 mg IM weekly, or six times the usual replacement dose) were given alone and in combination with supervised resistance exercise (26). Substantially greater gains in muscle size and strength were observed in the combined treatment group, compared with either treatment alone. The subjects in this study gained, on average, 6 kg of lean tissue with the combination of high-dose testosterone plus exercise, compared to 2 kg with exercise alone; strength also increased more in the testosterone-plus-exercise group. We recently performed an analogous study in eugonadal men with HIV infection and an average 9% weight loss, asking three questions:
Our study compared the addition of oxandrolone (20 mg/day) or placebo with a regimen of supervised resistance exercise plus replacement doses of testosterone (100 mg IM/week) for 8 weeks. The replacement dose of testosterone was given for several reasons: to prevent suppression of endogenous testosterone production by exogenous oxandrolone; to overcome any borderline hypogonadism that might have been present; and to prevent exercise-induced hypogonadism, which has been described in other settings (28).
Oxandrolone was given at 20 mg/day because this is the approved dose and has been shown to be safe over the longterm in other clinical settings, such as alcoholic liver disease (29). We observed a significant improvement in all measures of body composition and strength in the oxandrolone/testosterone/exercise group (14). Subjects gained an average of 7 kg of lean body mass with oxandrolone/exercise, compared with about 4 kg in the placebo/exercise group. Nitrogen balance was also more positive in the oxandrolone group. Muscle strength (determined by the maximum amount of weight lifted or the maximum power exerted on a Cybex machine) improved in the placebo/exercise group but was considerably greater with the addition of oxandrolone.
During a subsequent three-month open-label phase of this study, the subjects in the oxandrolone/testosterone/exercise arm continued to increase their lean body mass. Significantly, the use of protease inhibitors did not affect gains in lean tissue or strength. The exercise regimen and oxandrolone were well tolerated, and there were no effects on CD4 counts or viral load. One subject out of the 12 in the oxandrolone group developed moderate elevations in liver function tests and had to be withdrawn from the study. One potentially important adverse effect was that HDL cholesterol levels fell significantly (by 9 mg/dL) in the oxandrolone group. No adverse behavioral effects were observed.
We conclude that combining a moderately supraphysiologic androgen regimen that contains an anabolic steroid, oxandrolone, with supervised resistance exercise has extraordinarily beneficial effects on lean body mass and muscle strength, is well tolerated, and is not affected by protease-inhibitor therapy. The average lean-tissue gain of 15 pounds over eight weeks in the exercise/oxandrolone group is considerably greater than has previously been reported for any anabolic intervention in HIV infection, cancer, or other causes of wasting disorders (22).
Administration of recombinant growth hormone (0.1 mg/kg S.C. daily) produces a positive nitrogen balance and leads to a gain of 3 kg lean tissue over 12 weeks time in HIV-associated wasting (30). Modest improvement in treadmill exercise endurance has also been reported (12). Unfortunately, the evidence suggests that patients are likely to lose all of the weight and lean tissue they have gained, if their rGH therapy is discontinued, so treatment may have to be lifelong. Side effects such as hyperlipidemia, diabetes, headaches, and carpal tunnel syndrome, although uncommon, can also be problematic. Clinical use of rGH has been relatively infrequent to date, primarily because of the extremely high cost, which can reach $1,000 a week.
Anticytokine therapies, including thalidomide treatment and n-3 fatty-acid supplementation, have also been evaluated (19, 31). Although the cytokine hypothesis of wasting remains attractive, studies to date with potent anticytokine agents have shown only a modest improvement in lean-tissue mass in patients with AIDS-related cachexia, with no documented effect on function or performance (31, 32). Patients coinfected with tuberculosis and HIV who have constitutional symptoms may exhibit a better response to thalidomide (31) than patients with HIV infection alone (32). The toxicities of thalidomide are substantial and include sedation, rash, and fever in addition to the agent's notorious teratogenicity.
A number of syndromes involving altered content or distribution of body fat have recently been described in HIV-infected patients on combination antiretroviral therapy. These alterations include abdominal obesity ("protease paunch"), peripheral lipodystrophy (often combined with atrophy, resulting in "AIDS face"), and the build-up of interscapular fat ("buffalo hump"). Almost all of the important questions about these syndromes remain unresolved at this point, and clinical trials evaluating potential therapies have yet to be published. Nevertheless, we are closer than we were two years ago, when the first of these cases were reported, to understanding these anomalous changes in body composition.
Estimates of the prevalence of these body-fat alterations have ranged from less than 10% to upward of 80% in patients on combination antiretroviral therapy (1-3, 33-35). Some of these differences may be due to different stages of illness or duration of multidrug therapy among study populations. Differences in inclusion criteria may also explain different estimates, and whether the alterations in corpus habitus were observed by care providers or were self-reported by patients may have some bearing as well. For the present, an increase in the waist:hip ratio may be the most reproducible objective measure for clinical studies.
It does seem clear, at this point, that weight is gained by patients after they start a protease-inhibitor-containing multidrug regimen, and that most of the weight that is gained is body fat. In a study of almost 1,000 patients taking ritonavir, the average weight gain was about 2 kg (36). The weight gains seen in patients who start therapy with a regimen that includes indinavir are similar, and older studies of zidovudine in nucleoside-naïve patients also noted weight gains of about 2 kg. When patients are followed longitudinally, almost all the weight they gain on protease-inhibitor-containing therapies appears to be body fat (37). After longer periods on protease-inhibitor-containing therapies, patients tend to lose much of this added weight, often from subcutaneous fat stores.
Are these antiretroviral-therapy-associated changes in body composition due to protease inhibitors per se or to suppression of viral replication? This essential question remains unanswered -- and highly controversial. Several authors have now described body-fat changes associated with the use of classes of anti-HIV drugs other than protease inhibitors (34, 38). Others have presented preliminary evidence of both improvement (39, 40) and no improvement (41) in abnormal body-fat distribution in patients who were switched to regimens that did not contain a protease inhibitor. It may be that antiretroviral therapy itself is responsible for some of these metabolic changes, or it may be that protease inhibitors have effects that exacerbate underlying metabolic disorders. At present, therefore, it is premature to recommend changing from, or avoiding, protease-inhibitor-containing regimens solely to reduce body-composition abnormalities.
Thus far, no interventions have been proven effective in eliminating any of the body-composition abnormalities associated with multidrug antiretroviral therapy. Anecdotal observations have suggested that exercise plus dietary-fat restriction may be effective in reducing paunch. Dawn Averitt, a member of the editorial advisory board of HIV Newsline and an 11-year veteran of various anti-HIV regimens, has an interesting personal take on this subject (see "Trekking with AIDS, Part 2 -- A Special Report" from AIDS Care in this issue). Ms. Averitt, who has spent the past four-and-a-half months hiking the Appalachian Trail, reports that her strenuous daily routine has not only eliminated her lipodystrophy, it has brought her serum cholesterol down from an all-time high of more than 600 to 225 at last count.
Some investigators have suggested that rGH may also reduce so-called protease paunch (42). It should be recognized, however, that rGH therapy is likely to worsen peripheral lipoatrophy (since rGH increases lipolysis and fat oxidation), and it may lead to hyperlipidemia and insulin resistance as well (30). Off-label use of rGH should therefore be considered very carefully, and treated patients should be followed very closely.
Modestly supraphysiologic doses of androgens, combined with resistance training, tend to encourage the deposition of abdominal fat -- leading to so-called android obesity -- and they suppress HDL cholesterol concentrations. Androgens plus exercise can reduce body fat, but the effects on specific depots have not been determined (15). Therefore, these agents seem to be less than ideal candidates for treatment of this syndrome, although definitive studies have not yet been performed.
Another treatment for lipodystrophy that has been considered is metformin. This is an antihyperglycemic agent that is used in Type 2 diabetes mellitus. Metformin improves insulin sensitivity (largely by reducing hepatic glucose overproduction), reduces serum triglyceride concentrations, and tends to cause weight loss in Type 2 diabetics (43, 44).
HIV-infected patients on combination therapy often exhibit insulin resistance, hypertriglyceridemia, and central obesity (1-3), and when they do, their insulin resistance primarily reflects sharply increased hepatic glucose production (45). Whether metformin will prove beneficial in treating the body-composition abnormalities seen in patients on multidrug antiretroviral therapy remains uncertain, although a preliminary report is encouraging (46).
The recently reported syndrome of lactic acidosis, abdominal pain, and nausea -- often combined with fatigue and weight loss -- that is seen in patients on multidrug therapies that contain a non-nucleoside reverse-transcriptase inhibitor, has no specific treatment (47). Clinicians must learn to identify this syndrome, because the lactic acidosis can be life-threatening. At present the provider's only option is to switch affected patients to a regimen that does not contain an NNRTI.
As with the management of obesity in the general population, patients and providers are faced with a frustrating inability to affect the debilitating changes in body fat that are observed in individuals on combination antiretroviral therapy. Until such time as we do have effective treatments for these lipodystrophies, it is very important to work on those variables that can be controlled. We can -- and should -- encourage our patients to reduce the amount of fat, especially saturated fat, in their diets. At the same time we should urge them to reduce their intake of sugar and alcohol while increasing their level of physical activity, even if that activity is limited to walking.
In addition, we can prescribe drug treatments for the metabolic complications that occur in association with these body-composition changes. These complications include hypertriglyceridemia, which can be extremely severe and can be associated with low HDL cholesterol concentrations, and insulin resistance or frank diabetes.
The emergence of altered body-fat content and distribution as a clinically important syndrome in patients on multidrug antiretroviral therapy has added an extra dimension to the clinical management of body-composition changes in HIV disease. Whereas treatments to manipulate body-fat content or distribution remain uncertain, with no interventions of proven efficacy, the options for efficacious treatment aimed at increasing lean-tissue mass and muscle strength have advanced considerably in recent years. Clinicians can now offer several effective therapies to combat lean-tissue wasting.
It seems reasonable to hope that in the not too distant future we will be able to deal as effectively with alterations in body fat. In the meantime, providers can attend to secondary disorders (such as dyslipidemia and diabetes), for which treatments are available, they can offer reasonable dietary suggestions, and they can keep an eye out for encouraging results of a number of ongoing clinical trials.
Marc K. Hellerstein, M.D., Ph.D., is with the UCSF Medical School, San Francisco General Hospital, San Francisco, CA.