Background: In 2 recent clinical trials (M03-613 and ACTG 5142) an arbitrary cut-off of at least 20% of limb fat loss was used to define lipoatrophy. However, it is unknown how much fat loss is needed for lipoatrophy to become clinically evident. Thus, the objective of this study was to assess the extent of limb fat loss evaluated by an objective measurement (DEXA scan) that becomes clinically evident lipoatrophy when evaluated by subjective assessment (physician/patient observation).

Methods: In a subgroup of patients enrolled in the ABCDE study (stavudine [d4T]/lamivudine [3TC]/efavirenz [EFV] vs abacavir [ABC]/3TC/EFV > lipoatrophy in naive patients ), DEXA scans were performed to evaluate body fat changes at baseline, 48, and 96 weeks. Extent of limb fat loss in several predefined anatomical sites was categorized as no, mild, moderate, or severe changes with a standardized questionnaire. For the purpose of this study, clinically evident lipoatrophy was defined as moderate/severe changes (group 1) while no or mild changes were considered no clinically evident lipoatrophy (group 2).

Results: In 54 patients evaluated (13 group 1; 41 group 2), DEXA scans showed a significant difference between groups in fat change from baseline (–45% [–2300 g] vs. +5% [+400 g], p <0.001). Limb fat loss at 96 weeks was >20% in 85% (11 of 13) in group 1 vs 37% (15 of 41) in group 2 (p = 0.003), and >35% in 69% (9 of 13) in group 1 vs 24% (10 of 41) in group 2 (p = 0.003). Considering exclusively patients with clinically evident lipoatrophy (group 1), the absolute amount of limb fat loss after 96 weeks was greater in those with baseline limb fat in DEXA >15 kg (the median value of the cohort) vs those with baseline limb fat <15 kg (3.9 vs 2.0 kg, p = 0.028) and in those with baseline weight >65 kg (the median value of the cohort) vs those with baseline weight <65 kg (3.9 vs 2.3 kg, p = 0.30). However, the relative limb fat loss was similar in all the previous four groups: –48%, –43%, 40%, and –47%, respectively.

Conclusions: Percentage (not absolute) changes in fat correlate best with clinical evidence of lipoatrophy. More than one third of our patients without clinical lipoatrophy would be diagnosed with lipoatrophy if the cut-off of 20% is used. Most patients with clinically apparent lipoatrophy had a fat loss of at least 35%, irrespective of baseline limb fat content or weight.