BACKGROUND: Little is known about sex-specific effects of HIV infection on energy expenditure. OBJECTIVE: We investigated the determinants of energy expenditure in HIV-infected women. DESIGN: Resting energy expenditure (REE), body composition, and hormonal and nutritional indexes were compared in 33 ambulatory, premenopausal HIV-infected women and 26 weight-matched, healthy premenopausal control subjects. REE was determined by indirect calorimetry and body composition by dual-energy X-ray absorptiometry (DXA), bioelectrical impedance analysis, and skinfold-thickness analysis. Hormonal indexes included leptin, testosterone, estradiol, and insulin-like growth factor I. RESULTS: HIV-infected subjects had a higher REE than control subjects [6794 +/- 1374 compared with 6011 +/- 607 kJ/d (1624 +/- 329 compared with 1437 +/- 145 kcal/d), P = 0.0096]. On average, REE was 119 +/- 23% of Harris-Benedict predictions in HIV-infected subjects compared with 102 +/- 9% for control subjects (P = 0.0007). In HIV-infected subjects, REE was highly correlated with fat-free mass (FFM) by DXA (R = 0.641, P < 0.001), but not with weight or disease status. The slope of the regression equation for REE and FFM was significantly greater (P = 0.027, analysis of covariance) for HIV-infected subjects [REE (kJ/d) = 203.5 (kg FFM) - 1237] than for control subjects [REE (kJ/d) = 77.4 (kg FFM) + 2923]. In a stepwise regression analysis, FFM was the most significant variable (P = 0.005), followed by free testosterone (P = 0.029), which together explained 49% of the variation in REE. The final equation was REE (kJ/d) = 230.8 (kg FFM) + 395.9 (free testosterone, pmol/L) - 3304. CONCLUSIONS: Energy expenditure was higher in HIV-infected women than in control women. FFM is the primary determinant of REE in HIV-infected women, but energy expenditure is greater per kg FFM in HIV-infected subjects than in control subjects, which may contribute to the wasting syndrome.