Prevalence of Lipoatrophy and Mitochondrial DNA Content of Blood and Subcutaneous Fat in HIV-1-Infected Patients Randomly Allocated to Zidovudine- or Stavudine-Based Therapy

INTRODUCTION: Mitochondrial toxicity resulting from mitochondrial DNA (mtDNA) depletion is suggested to be involved in the pathogenesis of lipodystrophy. METHODS: We cross-sectionally assessed lipodystrophy both clinically and radiographically in patients who, 4 years before, had been enrolled in a randomized comparative trial of stavudine- or zidovudine-based therapy. mtDNA content was measured in peripheral blood mononuclear cells (PBMCs) and subcutaneous adipose tissue from the thigh and back. RESULTS: Twenty-eight of the 45 patients enrolled in the original trial were included. Despite comparable exposure to stavudine or zidovudine (51 and 50 months, respectively), lipoatrophy prevalence by intent-to-treat analysis was significantly greater in stavudine recipients (82 vs 9%, P=0.0001). Likewise, those allocated to stavudine had significantly less peripheral fat. In an analysis restricted to patients who had remained on randomly allocated nucleoside reverse transcriptase inhibitors (NRTIs), mtDNA in PBMCs decreased after the start of treatment in both groups (P<0.0001) (-73% for stavudine and -67% for zidovudine, P=0.11), resulting in significantly lower levels in patients with lipoatrophy (P=0.007). The mtDNA content in subcutaneous adipose tissue from the thigh, but not from the back, was significantly lower in patients allocated to stavudine compared to zidovudine (P=0.01). mtDNA in adipose tissue from either location did not differ significantly between those with or without lipoatrophy. DISCUSSION: This study objectively confirms that regimens containing stavudine are associated with a greater risk of lipoatrophy than those containing zidovudine. mtDNA in PBMCs markedly declined with both treatments and was lowest in patients with lipoatrophy. The lack of difference in mtDNA in adipose tissue from patients with as opposed to without lipoatrophy may have been masked by a relative preponderance of stromal and vascular tissue in the subcutaneous tissue samples from these patients, combined with compensatory mitochondrial proliferation in remaining adipocytes. However, our findings may also suggest that the different risk of lipoatrophy observed between NRTIs cannot solely be explained by differences in mtDNA depletion directly at the level of peripheral adipose tissue.