Background:  Progression of HIV infection results in fat loss and wasting. However, leptin levels have been reported to be normal in advanced HIV infection in the absence of secondary infections, and repeated ART interruptions may increase leptin levels. Overall, the direct effects of HIV replication on fat metabolism, particularly in women, remain unclear.

Methods:  In Johannesburg, South Africa, we studied 85 HIV-1-infected, ART-naïve women (median CD4, 247 cells/µL; median log₁₀ viral load, 4.78 copies/mL, no prior opportunistic infection); 50 healthy female volunteers served as controls. Informed consent was obtained from all subjects. Fresh blood samples were stained on pre-mixed monoclonal antibody lyophilized on 96-well plates, and analyzed by flow cytometry. We collected 54 metabolic variables, including:  total, LDL, and HDL cholesterol, triglycerides, leptin, free fatty acids (FFA), insulin, and c-peptide. Fasting glucose was assessed on fresh or cryopreserved serum. Body fat distribution was assessed by dual-energy X-ray absorptiometry (DEXA) scan (with bone density), magnetic resonance imaging (MRI) (L4–L5 cross section), and anthropometric measurements; body mass index was calculated. Statistical analyses included between-group differences and correlations (Spearman test, Kruskall-Wallis ANOVA); principal component on variance analysis; standard least-squares analysis.

Results:  As expected, viral load was negatively correlated to CD4 counts and positively correlated to CD95/HLA-DR expression on T cells. Plasmacytoid dendritic cells (but not myeloid dendritic cells (DC) or natural killer (NK) subsets) were lower in women whose viral load was above the median. Both leptin and all body fat measurements, but not bone density, were negatively correlated to viral load. Accordingly, a significant difference in leptin and fat accumulation was observed in women whose viral load was above the median, as compared to controls or HIV-infected women whose viral load was low. No difference was observed in FFA or other serum factors, indicating the absence of differential lipolysis. Leptin levels were positively correlated to fat levels in all subjects. In HIV-infected women the leptin/body mass index ratio remained negatively correlated to viral load, suggesting an independent effect of viral replication on leptin levels. We further explored this eventuality using a standard least square model using log₁₀ viral load and total fat (assessed by MRI) as predictors; the results indicate that viral load is an independent predictor of serum leptin.

Conclusions:  HIV-1 replication—associated with fat loss, immune activation, and depletion of adaptive and innate immunity effectors—is an independent predictor of serum leptin levels in women.