Managing the metabolic side effects of antiretroviral therapy

10 March 2002. Related: Side effects, Lipodystrophy and metabolic complications.

Mark Mascolini, for Treatment Directory at amfAR.org

Introduction

Quick approval of antiretrovirals in the late 1990s saved lives, but at a price. The speedy development and testing of these potent drugs, in relatively small studies, gave researchers little opportunity to size up side effects. Obvious short-term problems — nausea with ritonavir and rash with nevirapine, for example — seemed a fair trade-off for the rapidly recognised clinical benefits. Only when thousands of people began taking antiretrovirals for more than a typical study’s 48 weeks did today’s familiar, portentous, and sometimes devastating side effects become apparent.

The emergence of lipodystrophy, hyperlipidemia, and insulin resistance, coupled with closer scrutiny of liver toxicity, high lactates, and other threats, fostered the dramatic rethinking of antiretroviral tactics that continues today. These tactics come in many guises — delayed treatment of drug-naive people, swapping one drug for another, treatment interruptions, and pulsed therapy. But all these approaches share a common feature: avoiding antiretrovirals, whether that means one drug, one class, or all of them.

Shunning anti-HIV drugs has its risks, and its benefits remain largely unproven. But so far clinical research has produced only a handful of other tools to counter side effects. Often these other tools are other medicines with their own toxicities, plus interactions with protease inhibitors (PIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), and nucleoside reverse transcriptase inhibitors (NRTIs or nucleoside analogues). This article surveys recent findings on managing hyperlipidemia, insulin resistance, and lipodystrophy.

Elevated blood lipids

Advanced HIV infection itself can boost triglycerides in the blood and lower levels of high-density lipoprotein cholesterol (HDL-C, or “good” cholesterol). Protease inhibitors compound these problems, hiking blood serum triglycerides, total cholesterol, and low-density lipoprotein cholesterol (LDL-C, or “bad” cholesterol).

Ritonavir raises all these lipids in HIV-negative people after only two weeks. As noted in US Department of Health and Human Services (HHS) antiretroviral guidelines, this evidence moves some experts to recommend measuring cholesterol and triglycerides before starting PIs and every three to four months afterward. People who have elevated triglyceride levels before beginning PIs, the HHS panel adds, should probably have their blood lipids measured one to two months after starting treatment. Classic heart disease risk factors should also be assessed.

What should be done with the results of these tests? Different authorities have different suggestions. The US National Cholesterol Education Program (NCEP) recommends diet and exercise for people with:

LDL-C at or above 100 mg/dL (if they already have heart disease and a 10-year risk above 20%)
LDL-C at or above 130 mg/dL (if they do not already have two or more other coronary risk factors)
LDL-C at or above 160 mg/dL (if they do not already have heart disease and have fewer than two other coronary risk factors)

For these three respective groups, the NCEP says “consider drug therapy” with:

LDL-C at or above 130 mg/dL
LDL-C at or above 160 mg/dL (if the 10-year heart disease risk is below 10%) or at or above 130 mg/dL (if the ten-year risk is 10% to 20%)
LDL-C at or above 190 mg/dL

The NCEP did not have people with HIV infection in mind when it formulated those recommendations. In its 2001 guidelines, the British HIV Association (BHIVA) issued much more specific advice for HIV-infected people:

Cholesterol repeatedly above 250 mg/dL (with LDL:HDL ratio above 4:1):

Switch PI to PI-sparing regimen (in people taking their first regimen)
Recommend dietary advice, exercise, stopping smoking, and blood pressure control
Pravastatin (40 mg) each night, or consider atorvastatin (10 mg) each night
LDL-C:HDL-C ratio above 4:1: as above

Fasting triglycerides repeatedly above 310 mg/dL:

Switch PI to PI-sparing regimen (in people taking their first regimen)
Recommend dietary advice, exercise, stopping smoking, and blood pressure control
Fenofibrate (67 to 267 mg once daily), or gemfibrozil (300 to 600 mg once daily)

Unfasted triglycerides repeatedly above 390 mg/dL: as above.

Diet and exercise

The AIDS Clinical Trials Group (ACTG) Cardiovascular Disease Focus Group recommends dietary advice and regular aerobic exercise for any HIV-infected person with a fasting triglyceride tally above 200 mg/dL. But little work has weighed the effects of diet and exercise on high lipids or on the abdominal weight gain of lipodystrophy.

One problem with prescribing diets for people taking antiretrovirals is that “the need for lipid lowering and weight gain may coexist in patients who often experience prominent gastrointestinal symptoms.” The ACTG also notes that people on low-fat diets may add carbohydrates, which could raise triglycerides and lower HDL-C.

Complicating matters further, the best diet for someone with severely elevated triglycerides differs from the best diet for someone with high cholesterol. Clinicians not well versed in these matters clearly need a dietitian’s help.

In one study that followed the NCEP guidelines, diet and exercise trimmed cholesterol levels by 29% (a significant drop) in eight HIV-positive people with lipid abnormalities. But diet and exercise did little to help 12 others who tried it. Dietary advice had little impact on cholesterol in a 24-week comparison of advice versus advice plus pravastatin (see “Statins, fibrates, fish oil” below).

Eighteen men who followed a program of 64 resistance-training sessions exhibited increased lean muscle mass while lowering fasting serum triglycerides from 281 to 204 mg/dL, a significant change. The authors suggested that building muscle through exercise “may promote triglyceride clearance from the circulation.” A study of six men with lipodystrophy who had a ten-week course of aerobic and resistance training confirmed the significant drop in triglycerides and also logged a significant 19% decrease in total cholesterol.

Trading a PI for nevirapine or abacavir

A handful of studies, most of them not randomised, found that lipids elevated during protease inhibitor therapy fell after a switch to nevirapine or abacavir. Substituting efavirenz for a PI does not consistently improve lipid levels. Such results inspired BHIVA’s advice to try a PI-sparing combination when someone with high cholesterol or triglycerides is taking a first-line PI. Some work found a greater incidence of treatment failure when switching from a first PI to nevirapine or abacavir in people with prior NRTI therapy.

Those with NRTI exposure before taking a protease inhibitor may have HIV that has developed some resistance to those drugs. The nevirapine or triple NRTI route becomes riskier for them because of the relative ease with which further drug resistance could evolve.

Nevirapine may be the best switch option, if other research confirms the results from the Atlantic study. That trial randomised treatment-naive people to take ddI and d4T with nevirapine, indinavir, or 3TC. Participants in the 98-person lipid metabolism substudy started with cholesterol and triglyceride levels within the normal range, except that HDL-C was low. After 96 weeks, HDL-C was 40% above baseline in the nevirapine group, compared with 6% in the indinavir group and 20% in the 3TC group.

Like the other regimens, the nevirapine-containing one exhibited an increase in LDL-C levels. This somewhat mitigated nevirapine’s protective effect. The total cholesterol:HDL-C ratio, considered by many to be the best predictor of heart disease, fell 6% in the nevirapine group (total cholesterol up 22%) while it rose 25% in the indinavir group (total cholesterol up 22%) and just 2% above baseline in the 3TC group (total cholesterol up 14%). Except for the HDL-C rise, the differences between nevirapine and 3TC were not statistically significant.

Statins, fibrates, fish oil

Like the BHIVA experts, the ACTG heart disease panel calls for statins in PI-treated people with high cholesterol, but not just any statins. Some of these lipid-lowering agents interact with PIs more than others due to common metabolic pathways in the liver’s cytochrome P450 system. The ACTG gives the nod to pravastatin (at 20 mg daily) or atorvastatin (at 10 mg daily), with fluvastatin as an “acceptable alternative agent.” (Cerivastatin, once recommended by ACTG as another alternative statin, has since been recalled by the manufacturer after reports of sometimes fatal rhabdomyolysis.) Lovastatin and simvastatin, they advise, should not be given with PIs. BHIVA also recommends pravastatin and atorvastatin as the agents of choice, but they set the pravastatin dose at 40 mg daily.

In the study that followed NCEP intervention rules, atorvastatin started at 10 mg daily cut cholesterol by 19% and triglycerides by 21% in the 10 people who took only that antilipid medication. A more recent study randomised 31 PI-treated men with cholesterol levels above 6.5 mmol/L (250 mg/dL) to get dietary advice or advice plus 40 mg of pravastatin daily for 24 weeks. Total cholesterol dwindled 17.3% in the advice-plus-pravastatin group, a significant fall, versus 4% in men who received only dietary advice. Drops in dangerous LDL-C entirely accounted for the total cholesterol decrease.

In a chart review of 77 HIV-infected people who took statins, gemfibrozil, or both for hyperlipidemia, median total cholesterol fell 13% in the statin-only group, and fasting triglycerides dropped only 5%. Both changes lacked statistical significance. A little more than half of the 29 people taking only a statin used atorvastatin, while 27% took pravastatin and 18% simvastatin. The researchers suspected poor statin adherence because lipid levels actually rose in some individuals. When they dropped these nonresponders from the analysis, triglycerides did fall significantly, but cholesterol still did not. In 55 people taking only gemfibrozil, median triglycerides fell by half, a highly significant change, and total cholesterol sank 13%. Among 17 people taking a statin plus gemfibrozil, triglycerides dropped by 36% and cholesterol by 22%.

Gemfibrozil falls into the antilipid class called fibrates, which the ACTG panel lists as “viable alternative agents” when people have high cholesterol and high triglycerides. Fibrates are the prime choice, the ACTG says, for isolated hypertriglyceridemia, advice echoed by BHIVA. The ACTG experts suggest fenofibrate may have an edge over gemfibrozil because it is easier to take and does a better job lowering LDL-C. But without further evidence, they see “no compelling reason” to pick fenofibrate over gemfibrozil.

In the NCEP-guided study, only six of 25 people reduced cholesterol and triglyceride levels with gemfibrozil alone. In the 19 people who had to add atorvastatin to gemfibrozil, average cholesterol sank by 30% and triglycerides by 60%. Combining fibrates and statins raises a risk of muscle toxicity, but these clinicians saw no myopathy (muscle wasting) in the 19 people taking the two drugs for an average 6.5 months.

A randomised, double-blind study of a low-fat diet plus either gemfibrozil or placebo in 36 men taking PIs recorded lower triglycerides in the gemfibrozil group and triglyceride gains in the diet-only group after 20 weeks. But the triglyceride drop in men taking gemfibrozil lacked statistical significance. Low-fat diet alone may have failed to lower triglycerides because half the enrolees were already following such a diet when the study began.

Fish oils, or omega-3 fatty acids, can pare high triglycerides, and the ACTG panel suggests that they “may be tried” in people with triglycerides above 1,000 mg/dL. The panel notes, though, that fish oils sometimes paradoxically boost triglycerides, that they have not been studied in people with PI-induced hypertriglyceridemia, and that high doses may abet insulin resistance in people with diabetes. Moderate doses of 1.7 g daily apparently do not promote insulin resistance.

High-dose niacin increases HDL-C to a much greater extent than the statins, which mainly act to reduce LDL-C. It also has a number of side effects. The ACTG panel recommends staying away from niacin because it causes insulin resistance even in people without diabetes. Bile-sequestering resins like cholestyramine (Questran) and colestipol (Colestid) bind to liver bile in the intestines, causing the liver to absorb more cholesterol in order to increase bile production. The ACTG panel also advised against these agents because they can inflate triglyceride levels in the blood.

High blood glucose and insulin resistance

Insulin regulates glucose by inhibiting its production in the liver and aiding its uptake by muscle, fat, and other cells. When these mechanisms falter — a problem called insulin resistance — glucose levels climb. For people without diabetes, HHS guideline writers note that “some experts” call for fasting blood glucose measures every three to four months during the first year of PI therapy. They do not recommend routine glucose tolerance tests. People starting a PI, they add, should know the warning signs of high glucose: excessive thirst, hunger, or urination.

PI-sparing regimens and diet

Several studies implicate PIs in insulin resistance, even in people without lipodystrophy or without HIV infection. David Nolan and Simon Mallal (Royal Perth Hospital) cite studies like these in suggesting “that insulin resistance precedes, and contributes to, visceral adiposity.”

Despite this evidence that PIs provoke insulin resistance, the HHS panel observes that “the reversibility [of insulin resistance and hyperglycaemia] is currently unknown due to limited data.” Studies substituting an NNRTI for a PI to ease insulin resistance, the panel maintains, are “inconclusive.” BHIVA experts disagree. They recommend switching to a PI-sparing regimen when symptoms of glucose irregularities appear, but only if the PI is part of a person’s first antiretroviral combination. In some studies, switching from a PI to nevirapine significantly lowered glucose and insulin resistance. Insulin resistance also improved in a trial of 15 people who substituted abacavir for a PI. In 20 people who traded a PI for efavirenz, insulin resistance improved within six months of the switch. But another study in 41 people found no change in insulin resistance one year after efavirenz replaced a PI.

BHIVA also recommends dietary advice and exercise for people with glucose intolerance, but data on these interventions remain slim. A study of 62 men and 23 women with HIV infection and fat redistribution linked only three factors to high insulin levels: older age, higher polyunsaturated-to-saturated fat ratio, and longer PI use. More dietary fibre correlated with lower insulin levels. The authors suggest that polyunsaturated fats and fibre “may be important targets for dietary modification” in patients like these.

Metformin and glitazones

Metformin (Glucophage) and the glitazones (eg, Avandia) are both therapies used to reverse conventional diabetes. They both increase cellular sensitivity to insulin. High blood levels of glucose and insulin decrease as a result. The latest BHIVA guidelines offer specific treatment advice for glucose intolerance and high insulin:

All symptoms of glucose intolerance:

Dietary advice and exercise
Switch PI to PI-sparing regimen (in people taking their first regimen)

Glucose intolerance (fasting glucose 101 to 115 mg/dL or two-hour glucose tolerance test 117 to 185 mg/dL) with body mass index above 25 mg/kg2, insulin above 17 mU/L, and haemoglobin A1C above 6.5 mU/L):

Consider metformin (500 mg twice daily)
Switch off PI

Diabetes (fasting glucose above 117 mg/dL random value or two-hour glucose tolerance test above 185 mg/dL with the same additional parameters listed above):

Metformin (500 mg twice daily)
Review after three months

Body mass index 18 to 25 mg/kg2 with insulin above 17 mU/L and haemoglobin A1C above 6.5 mU/L:

Sulphonylurea therapy
Consider metformin (500 mg twice daily)

Body mass index below 18 mg/kg2 with lipoatrophy, insulin above 17 mU/L, and haemoglobin A1C above 6.5 mU/L:

Consider rosiglitazone (2 to 5 mg daily) or pioglitazone
Seek clinical opinion

Evidence supporting this advice is not overwhelming, but there is some. A randomised study of metformin versus no treatment in 27 PI-treated people with central weight gain found significant drops in fasting glucose, insulin, and triglycerides in the metformin group after two months. Visceral fat also fell. But the high dose, 850 mg three times daily, probably contributed to two study dropouts caused by severe diarrhoea and abdominal cramps. Metformin also cut subcutaneous fat in this study, so it is not the best option for people with lipoatrophy. (Note that BHIVA opts for a glitazone in people with lipoatrophy and high insulin.) Another problem with metformin in people taking antiretrovirals is that it can cause lactic acidosis, which is also a rare but dangerous side effect of NRTIs.

Other researchers studied a lower metformin dose in 26 people with lipodystrophy, 500 mg twice daily, and BHIVA adopted that dose. This three-month placebo-controlled trial charted significant drops in insulin, weight, and diastolic blood pressure in the metformin group. Those taking metformin also lost more visceral fat than those on placebo, but not significantly more. As in the high-dose metformin study, subcutaneous fat also decreased. These investigators stress that the study excluded people with a history of liver trouble, kidney failure, diabetes, or substance abuse—problems in many people with HIV infection.

Whereas metformin lowers subcutaneous fat, glitazones add fat cells. An ongoing trial, ACTG 5082, may show whether glitazones offset the subcutaneous fat wasting caused by metformin. This trial randomised people with insulin resistance and excess abdominal fat to take metformin, rosiglitazone (Avandia), or both. In a pilot study of troglitazone in six antiretroviral-treated men with diabetes, the drug improved insulin sensitivity in four of them. It also increased lean body mass and subcutaneous fat while decreasing visceral fat. A larger study of troglitazone in HIV-negative people with congenital lipodystrophy had similar results. Troglitazone (Rezulin) has been removed from the market because of liver toxicity, so researchers now focus on rosiglitazone and pioglitazone.

Fat accumulation

Making sense of studies that attempt to reverse therapy-associated increases in fat tissue (lipohypertrophy) remains a challenge for many reasons. Probably most important is the lack of a case definition for lipodystrophy, which may include various types of fat accumulation and/or fat loss (lipoatrophy). Lipohypertrophy occurs around the abdominal organs or viscera (causing truncal enlargement or abdominal paunch), between the shoulder blades (causing “buffalo hump”) and in women’s breasts. Lipoatrophy affects the subcutaneous fat layer, especially in the limbs and cheeks. Most studies lack control arms; they use different (often subjective) measures to gauge fat changes; and follow-up has generally been short. As a result, neither the HHS nor BHIVA ventures any advice on managing lipodystrophy, besides viewing with scepticism the value of switching from PIs.

Switching to a non-PI regimen

Switch studies are particularly hard to evaluate, for the reasons just mentioned. But a review of 10 switch studies at one meeting last year noted only one (involving 17 people switching to abacavir) that claimed substantial improvement in lipohypertrophy. A similar review of eight switch studies at another meeting in 2000 counted only one in which fat abnormalities improved — an update of the 17-person abacavir study.

These dismal results are not surprising. William Powderly (Washington University, St. Louis) noted that several factors may explain the failure of switching to reverse any suspected drug toxicity: the agent switched from is uninvolved or it is required for the establishment, but not the maintenance, of the abnormality; the change is irreversible; the toxicity is multifactorial; or follow-up is too short.

The HIV Outpatient Study (HOPS) made plain the multifactorial nature of lipodystrophy. Statistical analysis of factors contributing to fat gains in 104 members of a 1,077-person cohort found that treatment duration was only one of six variables that independently predicted lipohypertrophy. Among the others were age, higher body mass index (ratio of weight to height), lower pretreatment viral load, HIV viral load below the level of quantification (especially for more than two years), and haemophilia.

The Atlantic Study’s fat substudy mentioned above underscored the danger in blaming one class of antiretrovirals for fat accumulation. (Atlantic randomised treatment-naive people to take indinavir, nevirapine, or 3TC with ddI and d4T.) After a median 96 weeks of treatment, 13% taking indinavir, 10% taking nevirapine, and 23% taking 3TC had lipohypertrophy without lipoatrophy. Both hypertrophy and atrophy affected 15% on indinavir, 15% on nevirapine, and 16% on 3TC. None of the differences was statistically significant.

Such results bolster BHIVA’s advice that “individuals switching must consider that they may risk their long-term HIV management in exchange for an uncertain outcome with regard to their lipodystrophy.”

Exercise

Three published studies evaluated the effect of exercise on weight and fat in HIV-infected people without wasting. The first tried resistance training — weight lifting — four times a week for 16 weeks in 18 men. All were taking PIs, and 14 had excess central fat. Lean body mass increased significantly and the men got stronger, but they did not lose fat in the trunk, arms, or legs.

The second study combined resistance training with aerobic exercise (treadmill or stationary bike) in 10 men with abdominal weight gain, nine of whom were taking PIs. After they exercised three times weekly for 16 weeks, total body fat dropped significantly by about 2%, mostly from the trunk. Weight and lean mass did not change significantly.

A third study evaluated only aerobic exercise (on treadmills and other machines) in 54 men and eight women, many of them above ideal body weight. Half exercised three times a week for 12 weeks and the other half did not. Weight, subcutaneous fat (by skinfold thickness), central fat, and waist-to-hip ratio fell significantly among exercisers compared with controls. Waist size, which the authors call “the most robust anthropometric predictor of visceral adipose tissue,” also fell significantly in the exercisers.

These findings cannot be readily applied to people with antiretroviral-related fat gains because only 14 of the 62 enrolees were taking HAART. The last person to complete the exercise program did so in 1998, and the researchers did not report any baseline abnormalities that looked like lipodystrophy. They note the need “to assess whether central fat can be preferentially reduced [by aerobic exercise]…without exaggerating the loss of peripheral fat.” A possible confounding variable is a significant on-study reduction in dietary fat among the exercisers.

A smaller randomised study of 12 men with lipodystrophy compared no exercise with aerobic plus resistance training. As in the earlier aerobic/anaerobic exercise study, fat decreased significantly among the exercisers. Much of it was apparently from the midriff, since the exercisers’ waist-to-hip ratio improved significantly. The authors did not directly report changes in peripheral fat, but arm and leg circumference both grew significantly in the exercise group.

Steroids, testosterone, metformin

These exercise studies excluded steroid use, which increases muscle mass. Many clinicians familiar with HIV lipodystrophy are wary of steroid toxicity. BHIVA cautions that “anabolic steroids are not suitable for treatment of lipid abnormalities, due to concerns regarding worsening lipid profiles, fat loss and potential for liver function disturbances.” The effect of steroids on gonadal function remains poorly described, and is a cause for concern since many people with HIV are hypogonadal.

Twelve-week placebo-controlled studies of the steroids oxymetholone (for lipodystrophy and wasting) and nandrolone (for wasting) found that both increased lean muscle, but neither reduced fat mass. Four of 56 people quit the oxymetholone study because of liver toxicity. “Good” HDL-C dropped among people taking nandrolone, as did two hormones that reflect gonadal function.

Testosterone replacement also lowers HDL-C. BHIVA rates it “only suitable for repeatedly hypogonadal and asymptomatic men.” The ACTG has mounted a placebo-controlled trial (A5079) of testosterone skin gel for men with low serum testosterone and abdominal obesity.

A double-blind, placebo-controlled trial of metformin for men with lipodystrophy recorded a significant drop in weight with metformin versus placebo and a substantial, but not statistically significant, decrease in visceral abdominal fat. Subcutaneous abdominal fat also fell in this three-month study, so the visceral-to-subcutaneous fat ratio did not change. The ACTG is studying metformin with or without rosiglitazone in people with central fat gains.

Recombinant human growth hormone

Researchers continue to search for a dose of recombinant human growth hormone (rhGH) that trims visceral fat and still proves tolerable. A meticulous study of HIV-infected men with or without excess visceral fat linked low endogenous growth hormone with visceral adiposity. In a nine-month placebo-controlled trial of rhGH in 30 abdominally obese HIV-negative men, only 1mg of the drug daily lowered visceral fat by 18%. But the lowest dose tried so far in HIV-infected people with lipodystrophy, 4mg every other day, still caused side effects.

This study involved 14 people with “truncal enlargement” who first took 6mg of rhGH daily for 24 weeks. After a 12-week break, they took a 4-mg dose for another 24 weeks. The lower dose significantly reduced visceral fat by 15%, but did not improve skeletal muscle mass or lower cholesterol. Even at this reduced dose recipients complained of pain and stiffness, and triglycerides climbed in some. Growth hormone can also increase resistance to insulin: Diabetes developed in two participants in this study, but the group had high baseline insulin levels.

Another limitation is that even at 6mg daily, growth hormone’s effect on visceral fat is temporary.

Researchers at the University of California, San Francisco, are testing 1mg daily in people with HIV-related lipodystrophy. A placebo-controlled trial is evaluating 4mg every day or every other day. BHIVA advises that “there is not enough evidence for the use of growth hormone [for fat accumulation] outside of clinical trials.”

Surgery

Some people with dorsocervical fat accumulation, so-called buffalo hump, have turned to surgical remedies such as liposuction. Improvements are transient, because surgery does not address potential mechanisms of fat accumulation. Patrick Amard, a Parisian surgeon treating facial atrophy (see below), believes total surgical excision of neck fat may be more effective than liposuction, but long-term follow-up is lacking. Neither liposuction nor surgical excision is practical with visceral fat.

Fat atrophy

The large HIV Outpatient Study mentioned above also looked at the risk factors for fat loss, or atrophy. Fat loss occurred in 171 of the cohort’s 1,077 members. The major identified risk factors were age, time since AIDS diagnosis, low nadir and current CD4 count, white race, and use of either d4T or indinavir.

Antiretroviral-induced fat atrophy has proved resistant to remedies except for cosmetic surgery. BHIVA experts believe “it remains unclear…whether peripheral lipoatrophy is treatable.”

Switching NRTIs

Because research implicates NRTIs in this aspect of lipodystrophy, some investigators have focused on pinpointing a particular nucleoside to justify avoiding or switching from that drug. The leading suspect is d4T. Although several studies have yielded much incriminating evidence, all of it is circumstantial, and other research does not implicate d4T.

Most studies finding fault with d4T are cohort analyses, which can suggest, but not prove, cause. Many people who began to suffer lipoatrophy while taking d4T may have started with other NRTIs that failed or proved intolerable, and all of the nucleosides taken could have contributed to the problem. In four cohort studies of first-line d4T, two found a higher risk of lipoatrophy with d4T and two did not.

In another study of 39 people starting antiretrovirals with a d4T regimen and 76 starting with ZDV, rates of lipoatrophy did not differ between the groups after nearly two years. In that time, study participants did not change NRTIs.

One published trial of 29 men switching from d4T to ZDV or abacavir did chart improvements in subcutaneous abdominal and thigh fat but not in facial fat. The study was not randomised, and several people taking two NRTIs added nevirapine when they switched from d4T. Perhaps the more potent triple regimen controlled their HIV better, some suggest, so they gained weight, including peripheral fat.

A more recent study randomised 105 people with lipoatrophy to continue taking regimens containing ZDV or d4T, or to substitute abacavir for those NRTIs; 80% were taking d4T. After 24 weeks, investigators measured significant gains in subcutaneous fat in the switch group. The gains were so small, though, that neither study participants nor clinicians even noticed them. On the basis of these improvements, the researchers estimated that it would take four to five years to reverse lipoatrophy.

Reviewing some of the literature in a recent article, Judith Currier (UCLA) noted potential biases in studies pointing to d4T. She added that recent results “support earlier studies suggesting that [d4T] may be particularly implicated” in lipoatrophy.

Even if lipoatrophy occurs more often in patients taking d4T, it still occurs in patients not taking d4T. One newly published study prospectively followed 86 men and 29 women, all treatment-naive patients who started ZDV- or d4T-containing regimens and did not switch treatments. After a median 101 weeks of observation, the incidence of lipoatrophy (and any lipodystrophy) was the same with either drug.

Surgery

The BHIVA guidelines are adamant on the question of changing therapy to reverse lipoatrophy: “Switching between drugs or drug classes has not led to resolution of lipoatrophy and cannot be routinely recommended. Improvement of significant facial or peripheral limb lipoatrophy does not occur over periods of greater than six months off therapy.”

Switching from d4T may owe some of its popularity to the near-total lack of other remedies for lipoatrophy. In fact, some interventions for other metabolic complications, such as metformin and aerobic exercise, can worsen subcutaneous fat loss. So people with facial atrophy have turned to a tactic that does show some promise:

Cosmetic surgery

One technique involves facial implants of fat harvested from another part of that person’s body. A noncomparative, six-month study of this technique in 12 men and three women recorded a symmetric tripling of facial fat measured by MRI scans. Four of the 15 people rated the improvements “very good,” nine “good,” and two “light.” The durability of this approach remains uncertain, and people with severe atrophy may not have enough fat for the implants.

Another technique pulls Gore-Tex strips into atrophic areas through surgical incisions. The surgery does leave scars and “some lumpiness typically is evident in the implanted areas.” When the lumpiness does not fade, silicone can be injected to smooth the appearance.

The procedure that has stirred the greatest interest is subcutaneous injection of polylactic acid (PLA), marketed in Europe as New-Fill. Parisian surgeon Patrick Amard has used this technique in about 150 HIV-infected people with facial atrophy. PLA works by stimulating collagen production under the skin, but the skin surface remains soft and natural. In 33 men treated for six to 32 months, dermal thickness measured by ultrasound increased significantly. Amard typically sees the best results after four sessions, and the only side effects he has noted are temporary swelling and bruising.

Clinician-surgeon teams in Manchester and London are studying the PLA technique in people with facial atrophy. At least six sites in the US and Mexico have offered PLA injections at prices ranging from $500 to $650 per session. In August 2001, the FDA blocked PLA imports into the US on the grounds that its sponsors had not obtained premarketing clearance to use an unapproved “device.”

Anyone contemplating these procedures should consider that cosmetic surgery is a highly specialized discipline. A series of PLA shots under the skin may sound simple, but it takes an expert to get consistently good results. An HIV clinician who has watched the procedure performed warns people to “be wary of medical doctors willing to apply the treatment without the proper surgical training and skill.”

This article, with all its references, is at:
http://www.amfar.org/cgi-bin/iowa/td/feature/record.html?record=55