Background: HIV, the immune system, and antiretroviral therapy play unclear roles in body composition alterations.

Methods: From the Nutrition For Healthy Living Study, we evaluated the effect of CD4, viral load, and HAART on trunk and appendicular (APP) body composition in 121 men and 45 women with $2 study visits (211 follow-up intervals). Visits included were 11 months apart, with blood sampling, health/medication questionnaires, and DXA scans. Intervals including nutrition or body composition research interventions were excluded. HAART users (3 months use prior and continuous use during interval) were compared to HAART-naive subjects. General linear models were adjusted for HAART use, age, and gender.                   

Results: Subject age was 42.7 ±7.2 yrs, time from HIV diagnosis: 7.6 ±3.5 yrs; median CD4 count: 401 (IQR: 258–590) and log₁₀ viral load: 2.3 (IQR: 2.3–3.8); 59% were gay men, 27% female, 10% had used IV drugs, and 36% were of minority ethnicity. Median interval length was 12.6 months (IQR: 11.4–17.7), with 173 HAART-use (75% PI-based) and 38 HAART-naive intervals. In all subjects, higher baseline CD4 count predicted increased trunk fat (1.7% ±0.8 per 100 cells; p = 0.03) at interval end. HAART use independently predicted additional trunk fat (11.4% ±4.7; p = 0.02) and trunk bone mineral content (BMC) (3.3% ±1.1; p = 0.003), and less trunk lean mass (LM: -1.1% ±0.6; p = 0.05).

Independent of HAART use, improved CD4 status during the interval (per 100 cells) predicted increased APP lean (0.7% ±0.2; p = 0.004) and a trend toward decreased APP fat (-1.8% ±0.9; p = 0.06). In HAART users, CD4 count improvement predicted increased trunk BMC (4.5% ±1.1; p < 0.001), decreased trunk LM (-1.1% ±0.6; p = 0.05), and perhaps increased trunk fat (8.9% ±5.2; p = 0.09).

Higher baseline VL (log₁₀) predicted decreased trunk fat (-5.2% ±2.3, p = 0.03) and APP LM (-1.1 ±0.4%, p = 0.01) at the interval end. In HAART users, higher baseline VL predicted increased trunk BMC (4.0 ±1.3%, p = 0.003) and a trend toward trunk LM loss (-1.1 ±0.6%, p = 0.07); increased VL during the interval predicted increased trunk BMC (3.8 ±1.1%; p < 0.001), and a trend toward increased trunk fat (10.3 ±5.6%; p = 0.07).

Conclusions: Baseline CD4 count and viral load and their changes during follow-up intervals, independent of HAART use, predict alterations in trunk fat, trunk bone, and appendicular LM. Among HAART users, immunologic responders experience trunk fat accumulation, appendicular LM loss, and trunk bone improvements.