February 11, 2004
There was not much news this year in the arena of metabolic complications. In fact, the one big news item during this meeting did not even take place in the conference (the Abbott price increase and the resulting activism by doctors and patients). Other big news at this meeting included the failure of the first large vaccine trials run in Thailand. Although most everyone had heard the news before the conference, this did not make the presentation of the study more palatable. Everybody has a different take on it, but, in the end, everyone seems to be asking for more money -- obviously the vaccine researchers are asking for more money for more trials, basic scientists are asking for more money to continue HIV basic science research and microbicide researchers are asking for more money to try to get a microbicide to market. Everybody is right in a different way.
The other important (and extremely sad) news was the data presented regarding the price that African women and women from other developing countries are paying when they use single-dose nevirapine (NVP, Viramune) as a way to prevent maternal-fetal transmission of HIV infection. The response rate of subsequent NNRTI-containing regimens for these women is at least 20% less potent. All this made the general tone of the meeting pretty somber.
Regarding metabolic studies, although some interesting data were presented, unfortunately, none of it was encouraging. From a clinical perspective, the metabolic presentations that I think were the most important included the following:
Michael Dubé presented data from ACTG 5005,1 the metabolic substudy of ACTG 384. In this study, patients were randomized to receive zidovudine (ZDV, Retrovir) plus lamivudine (3TC, Epivir) or stavudine (d4T, Zerit) plus didanosine (ddI, Videx) with efavirenz (EFV, Sustiva), nelfinavir (NFV, Viracept) or both. The original study was published in the New England Journal of Medicine.2
The hypothesis of ACTG 5005 was that regimens containing the protease inhibitor (PI) nelfinavir would cause hyperlipidemia and insulin resistance, as compared to regimens containing efavirenz. This did not happen, and it is this which makes the results interesting.
The primary analysis was performed after 16 and 32 weeks of treatment, but data up to 64 weeks was presented. A summary of the conclusions discussed lipid metabolism: All regimens were associated with increases in fasting lipids. Regimens containing efavirenz or nelfinavir were associated with similar increases. Changes in high-density lipoprotein cholesterol (HDL-C) and total HDL-C ratio were more favorable with efavirenz, but small in magnitude. Zidovudine/lamivudine tended to have more favorable changes in total-C and non-HDL-C than did didanosine/stavudine; didanosine/stavudine had a more favorable effect on HDL-C.
With regards to glucose metabolism, insulin resistance, as estimated by homeostasis model assessment (HOMA-IR), increased modestly in the group as a whole; regimens containing efavirenz and nelfinavir affected HOMA-IR in a similar manner, with no early increases with either agent. The lack of an early increase in HOMA-IR with nelfinavir suggests that acute insulin resistance is not a PI-drug-class effect (which is something that had been long assumed).
Indinavir (IDV, Crixivan) is the PI that produces the most insulin resistance, and it does so really quickly. Zidovudine/lamivudine and didanosine/stavudine-containing regimens affect insulin resistance in a similar manner.
In summary, this study challenges the common belief that all PIs -- that is until atazanavir (ATV, Reyataz) was approved last year -- have similar metabolic effects. It also challenges the idea that these metabolic side effects are mainly mediated through increasing insulin resistance and that NNRTIs are metabolically neutral. None of these assumptions turn out to be completely true.
Lipoatrophy is the most difficult to treat of the metabolic complications associated with HIV and HIV treatment. Nothing seems to work well except switching off some of the antiretrovirals most commonly associated with its development -- stavudine, and, to a lesser degree, zidovudine. Unfortunately not all patients can switch antiretrovirals easily. The other alternative is an expensive surgical procedure which can produce an aesthetically acceptable result in some patients, but ignores the root of the problem.
Rosiglitazone (Avandia), a drug used for the treatment of type II diabetes, has been suggested as a possible therapeutic agent to treat lipoatrophy in HIV-infected individuals. Rosiglitazone belongs to a family of drugs called glitazones. The glitazones are insulin-sensitizing drugs, which tend to increase the storage of fatty acids in peripheral fat tissue and to stimulate a receptor called
Two small studies on rosiglitazone use have been presented so far:
As usually happens in medicine, a positive study gets more publicity than a negative one. Both studies were too small and underpowered to provide any definitive answer regarding how useful this drug is for the treatment of lipoatrophy.
During this conference, Andrew Carr from Australia presented the results of the ROSEY study.5 This study evaluated the use of rosiglitazone for the treatment of lipoatrophy. One hundred eight HIV-1-infected, lipoatrophic adults on antiretroviral therapy were randomized to rosiglitazone 4 mg twice daily (n=53) or matching placebo (n=55) for 48 weeks.
Limb fat increased by 0.14 kg in the rosiglitazone group and 0.18 kg in the placebo group. Rosiglitazone had no significant benefit on any other measure of lipodystrophy.
So, clearly, 48 weeks of rosiglitazone use did not improve lipoatrophy in HIV-1-infected adults receiving antiretroviral therapy.
However, one issue with this study is that patients were allowed to switch off stavudine and PIs, which the same authors have already demonstrated increases peripheral fat in patients with lipoatrophy. That could have significantly contaminated the results of this trial, although I do not think it would have affected the main results.
Thus the bad news regarding lipoatrophy treatment continues with yet another possible intervention that does not work very well, so lipoatrophy remains a critical concern for patients. The study is important because it dramatically demonstrates that the best way to treat lipoatrophy now is by preventing it through prescribing HIV medications that do not produce it.
Interventions for hyperlipidemia do not seem to do much better than those for lipoatrophy. Judy Aberg6 presented the follow-up data for ACTG 5087 (originally presented in Paris) that compared the administration of fenofibrate (Tricor), pravastatin (Pravachol) or both for the treatment of patients with hyperlipidemia. The benefits were as expected. Fenofibrate is good for patients with hypertriglyceridemia, pravastatin better for hypercholesterolemia and the combination a little bit better for the most severe cases.
The problem was that in most of the cases, the patients did not reach National Cholesterol Education Program (NCEP)7 goals. The message here is similar to that of the Andrew Carr paper -- prevention is critical because, when someone develops these antiretroviral therapy-associated problems (hyperlipidemia or lipoatrophy), it is quite difficult to get rid of them with treatment. So far, switching antiretrovirals seems to be the best option, although sometimes this is not possible.
Fish oil supplementation also only had modest effects on high triglyceride levels in a randomized trial by the University of North Carolina. David Wohl presented this randomized trial of patients with high triglycerides (200 to 2,000 mg/dL) on HAART and no history of diabetes mellitus.8 Patients were randomized to receive a nutritionist-administered American Heart Association (AHA)-based diet and exercise counseling (week 0 and 4) alone, versus with 3 g of fish oil daily (1,150 mg docosahexaenoic acid [DHA], 1,750 mg eicosapentaenoic acid [EPA], Coromega Inc) for 16 weeks. Fish oil seemed to have modest effects (17% decrease in triglycerides), which in patients with triglyceride increases of 300 to 400% of the normal values, is probably not clinically useful. Again, prevention and a medication switch have been the most effective approaches to this problem.
Although HIV-associated wasting has become rare in the era of HAART, at least in the developed world, it still affects a significant number of patients. Hengge and colleagues presented a study evaluating oxymetholone (Anadrol) for the treatment of this complication.9 In this trial, 89 patients were randomized to receive two different doses of oxymetholone (50 mg two times a day or 50 mg three times a day) or placebo. Patients in the oxymetholone group gained more weight than the patients who received placebo (a couple of kg or 4 pounds). However, this beneficial effect came at a price: an increased frequency of liver abnormalities and a suppression of pituitary gonadotropins (basically decreasing a patient's endogenous production of testosterone). Both are well-known side effects of anabolic steroids, as many major league baseball players can testify to themselves.
The efficacy of this drug is similar to other known interventions for the treatment of this wasting -- such as growth hormones -- and the cost is probably lower. Anabolic steroids are an alternative for the patient with wasting for whom HAART and nutritional supplementation do not work -- fortunately, a small number of individuals, since these medicines frequently have side effects.
During the last couple of years, osteopenia and osteoporosis have become recognized metabolic complications associated with HIV or its treatment. There were only a few presentations related to this topic during this year's conference. There was a poster10 from an Italian group showing improvements of bone turnover with the use of alendronate (Fosamax). However, surprisingly, the investigators did not see clear improvements in dual-energy x-ray absorptiometry (DEXA) scans of bone mineral density. This study confirms previous results presented last year in Boston11 by the Washington University group and solidifies the evidence that alendronate is a reasonable option for patients with evidence of osteoporosis that warrants treatment.
In a similar way that the development of hyperlipidemia is a risk factor for cardiovascular disease, osteopenia and osteoporosis predispose one to fragility fractures. Although several studies have shown that osteopenia and osteoporosis are frequent among HIV-infected individuals, there have been no studies demonstrating an increased risk of fractures among these patients.
Grace McComsey presented data on the frequency of fragility fractures among HIV-infected patients,12 a potential sign that the elevated prevalence of osteoporosis might have significant clinical implications. Fragility fractures appear to be under-reported. McComsey identified 49 fractures among almost 9,000 patients followed in several clinics in the U.S. There is always the potential for recall bias in this type of study, but it looks like fragility fractures might be a problem.
New data13 was presented regarding the high frequency of obesity in four clinics in Philadelphia, especially among African-American women. Obesity is much more frequent than wasting among HIV-infected individuals in the U.S. Although the prevalence of obesity is lower than in the general population, the numbers are quite concerning -- HIV-infected women and men were equally likely to be overweight (30% versus 31%), however, women were significantly more likely than men to be obese (29% versus 11%). African Americans were more likely to be overweight or obese than non-African Americans.
In a logistic regression model, female sex (RR 2.0), African-American race (RR 1.3), smoking (RR 0.6), and current CD4 (for each 100 cells/?L increment, RR 1.11) were independent predictors of obesity. Obesity is a well-known risk factor for hyperlipidemia, diabetes and the metabolic syndrome. Obesity present at baseline or acquired during treatment of HIV might be a significant contributor to the metabolic abnormalities associated with HIV or its treatment.
The studies presented at CROI 2004 have not brought us any closer to understanding the metabolic side effects of HIV and its treatment. At this point, the best we can do to deal with these problems is to simply avoid the drugs that cause the most metabolic side effects. Clinicians should always individualize HIV treatment for their particular patient. For patients who already have significant cardiovascular risk factors before they begin treatment (such as obesity, smoking or a family history of heart disease), the preferred choice should be NNRTI-based therapies. The drug regimens to be avoided, if at all possible, are those more commonly associated with the development of hyperlipidemia (such as boosted PIs, perhaps with the exception of atazanavir). For patients concerned about the stigmatization of lipoatrophy, stavudine should be avoided.