Drugs, not host or immune system, create conditions for fat loss

Antiretrovirals, not features of the host or the immune response to HIV, are overwhelmingly responsible for the development of lipoatrophy, according to studies presented on Monday at the Seventh International Workshop on Adverse Drug Reactions and Lipodystrophy in HIV, in Dublin, Ireland.

Controversy in the HIV field has raged for over eight years regarding the causes of fat loss in people receiving antiretroviral treatment, with some continuing to argue that a low baseline CD4 cell count, lower body mass index and race all predict the risk of developing lipoatrophy just as strongly as the use of d4T and AZT (the two drugs thought to be responsible for the condition).

However a new analysis of the first study to unequivocally demonstrate the effect of specific antiretroviral combinations on the development of lipoatrophy failed to show any effect of race on the risk of developing lipoatrophy. The study also showed that patients with higher baseline CD4 counts (above 200 cells/mm³) had a higher risk of developing lipoatrophy, as did patients with higher baseline triglyceride or total cholesterol levels.

Professor William Powderley of University College Dublin, a co-chair of the Workshop, said: "Large numbers of people are being exposed to an avoidable toxicity. The presentations at this meeting show the overwhelming influence of drug choice on the development of lipoatrophy."

The analysis, of participants in the A5005s sub-study of ACTG 384, looked at 102 patients for whom DEXA scan data were available for baseline and week 64.

ACTG 384 was a treatment strategy study looking at the respective merits of starting treatment with different drug combinations and the effect of the first line regimen on second line treatment response. Participants were randomised to receive a nucleoside backbone of either AZT/3TC or d4T/ddI and either efavirenz or nelfinavir. After virologic failure participants switched to the opposite combination of drugs, and the primary endpoint of the study was the time to failure of the second regimen.

At week 64, 44 of those originally assigned to d4T/ddI had developed lipoatrophy (defined as a 20% reduction in arm and leg fat), compared to 17 in the AZT/3TC group (p=0.006).

Patients assigned to d4T/ddI were three times as likely to develop lipoatrophy. Multivariate analysis showed an odds ratio for developing lipoatrophy of 3.19 for d4T/ddI.

A 100mg/dl increase in triglyceride levels at week 8 was also highly predictive (OR 3.27) of developing lipoatrophy regardless of whether participants received d4T or not (d4T treatment is also associated with triglyceride increases). However participants who developed lipoatrophy on d4T had a median triglyceride increase of +103mg/dl at week 8, compared with an increase of +24mg/dl in participants who developed lipoatrophy on AZT/3TC.

In contrast race was not associated with lipoatrophy, and nor was a low baseline CD4 cell count. Indeed participants with higher CD4 cell counts (above 200 cells) were at greater risk of developing lipoatrophy in ACTG 384 (p=0.001), in contrast to findings from a number of cohort analyses.

An analysis of subcutaneous fat biopsies from 32 male patients from the Western Australia Cohort also supported the view that the choice of drug in the initial regimen was the main factor associated with lipoatrophy, rather than any host or viral factors.

Paired biopsy samples taken at baseline and after a median of seven months on treatment from 20 individuals showed no evidence of unusual changes in the structure or appearance of subcutaneous fat cells in treatment-naïve individuals, and no changes were observed in seven patients after exposure to abacavir.

However patients taking thymidine analogues (d4T or AZT, n=13) showed evidence of mitochondrial DNA depletion, increased proinflammatory cytokine expression and increased presence of macrophages in subcutaneous fat.

During discussion Jacqueline Capeau of the Universite Pierre et Marie Curie, Paris, argued that the presence of macrophages reflected the recruitment of macrophages into subcutaneous fat as a result of damage caused by antiretrovirals, rather than an effect of HIV disease.

Samples from a small number of patients who switched from thymidine analogues to abacavir n=5) were also analysed, and showed modest improvements in tissue morphology and mitochondrial DNA content, but changes in cytokine expression were less pronounced, with the only significant improvement occurring in IL-8 secretion. A lower CD4 cell count at the time of the switch was associated with less reduction in macrophage content and cytokine expression, suggesting that future analyses of switch studies should look at immunologic and host correlates of fat restoration in more detail to determine how early lipoatrophy becomes irreversible as a result of drug switches, and whether there is any correlation between immunologic and host factors and the speed and extent of fat restoration after switching to a non-thymidine-containing regimen.