CROI: Disappointment for glitazones, metformin and testosterone for body fat changes

Four studies examining interventions to treat lipoatrophy or central fat gain reported this week at the Thirteenth Conference on Retroviruses and Opportunistic Infections in Denver were largely disappointing from a patient's point of view, despite the positive spin the investigators put on their findings.

Metformin, with and without rosiglitazone, fails to affect central fat accumulation

Two different studies from the United States failed to show that metformin - alone or with rosiglitazone - reduces lipodystrophy-associated central fat gain in HIV-positive individuals with or without insulin resistance.

Investigators from Tufts University School of Medicine in Boston, randomised 48 HIV-positive participants with normal glucose tolerance (56% male; median age 42 years; 69% Caucasian) but with a self-reported increase in abdominal girth and waist-hip ratio in a double-blind fashion to either 1500mg metformin daily, or placebo. The primary endpoint was change in visceral fat as measured by DEXA after 24 weeks.

Rakhi Kohli, presenting, reported no significant change in visceral fat after 24 weeks of metformin versus placebo (-22.2 vs -3.85 cm², p = 0.17). No changes were seen in any metabolic or lipid parameters, either.

However, although the investigators found that metformin use was significantly associated with a decrease in limb fat mass compared to placebo (-686g vs 161.0g, p = 0.03), this reduction in limb fat became less significant after adjusting for age, height, and baseline limb fat mass (-614.0 vs 95.3g, p = 0.12). Still, the participants on metformin lost weight during the study, reducing body mass (BMI) significantly from a mean of 27.3 to 25.7.

Dr Kohli concluded by reiterating that their study failed to show a benefit of metformin on fat redistribution or dyslipidaemia in HIV-positive individuals with lipodystrophy and normal glucose tolerance, and warned that "metformin should be used with caution in HIV-associated lipodystrophy, and if used, should be reserved for persons with impaired glucose tolerance and adequate subcutaneous fat."

A second study, ACTG 5082, in 105 individuals with some evidence of insulin resistance at baseline, also found that metformin alone, or in combination with rosiglitazone, failed to have any impact upon abdominal visceral fat. This study used higher doses of metformin (1000mg twice daily) and twelve of the 26 individuals randomised to the metformin alone arm of the study left due to side-effects (primarily diarrhoea) before the study's end at 16 weeks.

Kathleen Mulligan of the University of California, San Francisco, presenting, admitted that significant dose reductions and discontinuations in the metformin arm may have affected their results. However, only four of the 25 individuals on both metformin and rosiglitazone discontinued early.

The study, which used lower doses of rosiglitazone (4mg once daily) than in previous studies, concluded that neither metformin alone, nor in combination with rosiglitazone, had any effect on either abdominal visceral or subcutaneous fat over 16 weeks.

Interestingly, concluded Dr Mulligan, a secondary endpoint of the study did find that rosiglitazone increased leg fat compared to placebo, "providing additional evidence that rosiglitazone may increase subcutaneous fat in some individuals."

Is pioglitazone better than rosiglitazone at restoring lost fat?

Although the exact causes of lipoatrophy (fat loss) in HIV-positive individuals on antiretrovirals are still being investigated, there is increasing agreement that the thymidine analogues, d4T (stavudine, Zerit) and AZT (zidovudine, Retrovir) are major contributors.

Fat restoration as a result of switching from d4T or AZT to abacavir (Ziagen) or tenofovir (Viread) is slow and often imperceptible, and experience with rosiglitazone has been mixed, with a major double-blinded, placebo-controlled study finding no impact on fat loss, although a more recent, smaller, non double-blinded, non-placebo-controlled study suggested some effect.

However, rosiglitazone adversely affects lipid levels, and investigators from the French Agency for AIDS Research (ANRS) reported on their double-blind, placebo-controlled study examining pioglitazone, which has a different lipid profile, in a group of highly treatment-experienced HIV-positive individuals (82% male; median 7.7 years of prior antiretroviral therapy; 89% exposed to d4T for a median of 3.2 years; 28% still using d4T at baseline) with self-reported lipoatrophy confirmed by physical exam. All had been on the same antiretroviral regimen for at least six months, had plasma RNA levels below 400 copies/mL and CD4 cells above 200 cells/mm³.

A total of 130 participants were randomised to receive either 30mg pioglitazone once daily or placebo for 48 weeks, although only 119 participants had validated DEXA scans at baseline and at week 48. The study had a 80% power to detect its primary endpoint - a 325g change in limb fat between the two groups at week 48 by intent-to-treat analysis (a fairly modest degree of fat restoration).

Willy Rozenbaum from Tenon and Saint Louis Hospitals in Paris, presenting, noted that at baseline the median limb fat mass of the study's' participants was a very lean 2.9kg. Limb fat in the general population is around 8kg. After 48 weeks there was a significant difference of +330g of limb fat in the pioglitazone group compared with the placebo group (p = 0.051).

When the results were analysed by current use of d4T, pioglitazone significantly increased limb fat in the 94 individuals not taking d4T (a difference between the groups of +400g; p = 0.013), but did not significantly increase limb fat in the 36 individuals who were currently taking d4T (+110g; p = 0.404).

Significant improvements in thigh circumference (+1.4 cm vs. 0.2; p = 0.017) and skin-fold thickness at the triceps (+0.9 mm vs. 0.4; p = 0.047) were also observed in the pioglitazone group. The only effect seen on lipids was an improvement in HDL ("good") cholesterol which increased in the pioglitazone group (+0.08 mmol/L vs -0.08; p = 0.005) compared with the placebo group.

Dr Rozenbaum noted that although pioglitazone was successful in meeting the study's primary endpoint, and was well-tolerated, "no improvement was perceived by the patients." He concluded that his study's results support the use of pioglitazone for HIV-related lipoatrophy, but suggested that since "no effect was seen in the people exposed to d4T during the study" it should be combined with a switch away from thymidine analogues. He added that further studies with higher doses and longer follow-up times might yield results that were perceptible to patients.

Testosterone replacement for central fat accumulation?

A third study from the United States, ACTG 5079, examined the effects of a 1% testosterone gel rubbed onto the skin and providing 10g of testosterone daily in moderately hypogonadal men. This randomised, double-blind study's primary endpoint was determining whether testosterone replacement would reduce visceral abdominal fat as measured by CT scan.

A total of 88 men on stable antiretroviral therapy with HIV RNA

Cecilia Shikuma of the University of Hawaii, presenting, noted that the gel was "very well tolerated", but that by 24 weeks the median percentage change in visceral fat did not differ significantly between groups (testosterone 0.3%, placebo 3.1%; p = 0.76).

She noted, however, that the men on testosterone lost significant amounts of subcutaneous (-7.2% vs. 8.1%, p

As expected, lean body mass, as measured by DEXA, increased in testosterone-treated men compared to placebo (1.3% vs. -0.3, p = 0.02).

During questioning, Dr Shikuma conceded that the loss of subcutaneous fat, particularly in the limbs, may mean that testosterone replacement is problematic for individuals with both lipoatrophy and central fat accumulation. However, she said that this needed to be balanced with the improvements seen in lean body mass and in quality of life in general.