icon_folder.gif   Conference Reports for NATAP  
  13th CROI
Conference on Retroviruses and Opportunistic Infections
Denver, Colorado
Feb 5- 8, 2006
Back grey_arrow_rt.gif
The Take Home from CROI on Body Shape
  By David Alain Wohl, MD - The University of North Carolina AIDS Clinical Research and Treatment Unit
The seeds of body shape studies planted years ago blossomed into full bloom presentations at CROI 2006, proving what every gardener knows, patience pays. Since the initiation of studies like AIDS Clinical Trials Group (ACTG) studies 5082 and 5079 five years ago, we have seen a number of investigations of interventions to reverse lipoatrophy including a series of antiretroviral switch studies, several small studies of novel agents and a large randomized clinical trial of the insulin sensitizing agent rosiglitazone. During this time, we have also witnessed the approval of poly-lactic acid (Sculptura®) as a reconstructive approach to HIV-associated facial fat wasting - providing almost immediate relief to those able to access this compound.
However, despite these developments many issues remain unresolved. Conflicting or insufficient data on the use of metformin and thiazolidinediones (e.g. rosiglitazone and pioglitazone) have had a paralyzing effect on their clinical application for abnormalities of fat volume. This has led clinicians and patients to rely more on the interventions that have consistently proved effective - switching from stavudine (and to a lesser extent zidovudine) and reconstructive surgical procedures. Testosterone, one of the most poorly understood agents being prescribed to HIV-infected patients leading to overuse and misuse of this hormone, sometimes with serious consequences, has been used to reverse fat accumulation but there are limited controlled data on its effectiveness and safety.
The body shape data presented in Denver are not definitive but in the case of the thiazolidinediones add weight to prior results that tip the needle back in favor of their use in lipoatrophy. In the cases of metformin and testosterone, the results demonstrate what these drugs are capable and not capable of doing. All of the presentations discussed in this report are available for webcast and podcast. These can be downloaded by going to www.retroconference.org and clicking on 'Webcasts and Podcasts'. Find the Wednesday sessions and scroll to 'Oral Abstracts: Complications of Antiretroviral Therapy and HIV'. Then, plug in - it's the next best thing to being there.
Four separate studies looked at interventions to reverse body fat changes. Collectively, these studies suggest the following:
1. Rosiglitazone and pioglitazone increase limb fat volume both in patients with and without insulin resistance or marked glucose intolerance
2. Metformin does not reduce abdominal fat in those without insulin resistance or marked glucose intolerance
3. Metformin reduces limb fat
4. Testosterone reduces limb fat and abdominal subcutaneous but not visceral fat
TZDs worked to increase limb fat whether you had insulin resistance (A5082), glucose intolerance (A5082) or neither (ANRS study - although these conditions were not excluded).
1. Rosiglitazone and pioglitazone increase limb fat volume both in patients with and without insulin resistance or marked glucose intolerance
The thiazolidinediones rosiglitazone and pioglitazone were found to increase limb adipose tissue volume in two separate studies. In ACTG 5082, subjects with increased abdominal girth and evidence of problems handling glucose (i.e. insulin resistant or glucose intolerant), were randomized to one of four study arms (1):
- Rosiglitazone 4 mg daily plus Placebo
- Metformin 1500 mg daily plus Placebo
- Rosiglitazone 4 mg daily plus Metformin 1500 mg daily
- Placebo plus Placebo
Approximately 25 participants entered the study and were evaluated over 16 weeks with DEXA scans and glucose, insulin and lipid laboratories. Those getting rosiglitazone experienced a significant increase (- 200 g) in leg fat compared to the control subjects. This increase was small but was statistically significant (recall that the time course of the study was a mere 16 weeks). There was an increase in LDL cholesterol and decrease in HDL cholesterol in the rosiglitazone subjects.
A second study of pioglitazone 30 mg daily for 48 weeks also found an increase in limb fat compared to placebo (2). This study conducted by the French Agence Nationale de Recherches sur le Sida (ANRS) enrolled 110 patients with clinically evident lipoatropy. There were no insulin or glucose-related inclusion criteria. While those participating in ACTG 5082 were found to have some evidence of baseline peripheral lipoatrophy, the ANRS subjects had profound limb fat wasting at study entry. The pioglitazone assigned patients experienced a median 380 g increase in limb fat. However, there was a kicker: the - 25% of those on pioglitazone continuing on stavudine did not experience a significant increase in limb fat. As opposed to ACTG 5082, no significant changes in lipids were seen with this thiazolidinedione.
So, the informed reader may ask, 'Why did these two studies find an increase in limb fat while the ROSEY Study, a randomized double blind placebo controlled study of rosiglitazone at double the daily dose of that used in ACTG 5082, find no significant change in limb fat compared to placebo (3)?' Are not randomized controlled trials the gold standard? To that reader I answer, 'Good point.' But, I would ask back why no difference between study groups was seen in the ROSEY Study?
In The ROSEY trial, the rosiglitazone group experienced a 140 g increase in limb fat compared to a 180 g increase in the placebo group. The gain in limb fat in the placebo group is unexplained. Further, 53% of those in the rosiglitazone group versus 29% in the placebo group were on stavudine - which the ANRS study found blunted the limb fat increase of pioglitazone. Interestingly, in the same sort of subgroup analysis done by the ANRS investigators, the ROSEY investigators found that those subjects not on stavudine or zidovudine who received rosiglitazone had a 340 g increase in limb fat (almost identical to that seen in the ANRS study) but the control group not on these NRTIs had a 390 g (!) increase in limb fat. Again, this increase in peripheral fat among the placebo-receiving subjects remains unexplained. The high use of abacavir in both study arms (- a third of subjects) raises the question as to whether these subjects switched from stavudine to abacavir well prior to study entry leading to slow accumulation of limb fat during the trial.
Another important point is that the magnitude of the changes in limb fat both in A5082 and the ANRS study were small and on surveys poorly recognized by the subjects. However, further gains may be seen with longer term treatment and any improvement, even if mild, means that there is no worsening.
Metformin does not reduce abdominal fat in those without insulin resistance or marked glucose intolerance
In addition to the interesting finding of the effect of rosiglitazone on limb fat, ACTG 5082 found that over 16 weeks there was no significant effect of rosiglitazone or metformin or the combination on abdominal fat. Neither visceral (deep) fat nor subcutaneous (surface) fat changed to a statistically significant extent with in any of the treatment arms compared to those only receiving placebo. Further, metformin was poorly-tolerated in this study with 8 of 26 (31%) of subjects assigned metformin alone developing treatment-limiting toxicity. Therefore, as only a few patients were able to stay on metformin the ability to truly evaluate this agent was limited. This may explain why metformin had no observed effect in reducing visceral fat in ACTG 5082 while the opposite result was found in other studies of similar HIV-infected patients (4). In the context of the patient with insulin resistance or marked glucose intolerance the effect of metformin on visceral fat remains unclear.
Another study of metformin conducted by investigators at Tufts University and presented at CROI enrolled 48 patients who also had increased abdominal girth but, in contrast to A5082, were required not to have insulin resistance (5). Again, 1500 mg of metformin daily was compared to placebo. At the end of the day, metformin did not reduce visceral fat volume beyond that seen with placebo even when the analysis adjusted for potentially confounding factors such as age, height and baseline visceral adipose tissue volume.
These data support restricting the use of metformin for increased abdominal girth to those with clear insulin resistance and marked glucose intolerance.
Metformin reduces limb fat
Both ACTG 5082 and the study from Tufts suggest that metformin reduces limb fat. In the Tufts study, there was a statistically significant drop in limb adipose mass in the treated group compared to the controls (a 686 g drop in metformin group versus a 161 g increase in the control subjects, p = 0.03). In ACTG 5082 the metformin group did not have a net decrease in limb fat relative to the double placebo group; however, the limb fat gain experienced with rosiglitazone in this study appeared to be negated when both rosiglitazone and metformin were combined - suggesting a potential effect of metformin on fat wasting that was not able to be observed in the metformin alone group due to a high drop out rate in that study arm.
These results coupled with the toxicity of metformin in ACTG 5082 indicate that this drug should be used very cautiously and that clinicians and patients must be aware that wasting of limb fat may be a consequence of this therapy.
Testosterone reduces limb fat and abdominal subcutaneous but not visceral fat
Testosterone treatment of fat accumulation makes sense. Like all androgens, testosterone lyses fat cells. The ability to topically administer testosterone as a gel and its frequent use in the setting of HIV-infection for hypogonadism make it an ideal candidate to study for increased abdominal girth associated with HIV therapy. ACTG 5079 enrolled 88 HIV-infected, HAART-treated men with abdominal obesity and mild to moderately reduced testosterone levels (total testosterone level 124-400 ng/dL or, if > 400 ng/dL a level of bioavailable testosterone that was below normal) (6). Half the participants received 10 g of testosterone gel to apply topically or placebo.
After 24 weeks there was a statistically significant decrease in abdominal subcutaneous fat (about 7%) compared to placebo (who experienced an 8% gain) but no significant change in visceral fat - where there were small increases seen in both study groups. Further, limb fat decreased significantly with testosterone, as expected, with a median drop in limb fat mass of almost 500 g with testosterone gel compared to a slight increase in the control subjects.
Testosterone seems to help reduce abdominal fat by reducing subcutaneous adipose tissue. This lipolytic effect on subcutaneous fat is not restricted to the abdomen and is accompanied by the loss of limb fat. As in the case of metformin, use of testosterone for any reason must take into account untoward effects of the drug on peripheral fat.
1. Mulligan K, et al. Abstract 147, CROI 2006
2. Slama L, et al. Abstract 151b, CROI 2006
3. Carr A, Workman C, Carey D, et al. No effect of rosiglitazone for treatment of HIV-1 lipoatrophy: randomised, double-blind, placebo-controlled trial. Lancet. 2004 Feb 7;363(9407):429-38
4. Hadigan C, Corcoran C, Basgoz N, et al. Metformin in the treatment of HIV lipodystrophy syndrome: a randomized controlled trial. JAMA 2000;284:472-477
5. Kohli R, et al. Abstract 148, CROI 2006
6. Shikuma C, et al. Abstract 149, CROI 2006