Tumor necrosis factor alpha (TNF-alpha) stimulates lipolysis in man. We examined whether plasma TNF-alpha is associated with the degree by which insulin suppresses markers of lipolysis, for example, plasma free fatty acid (FFA) and net lipid oxidation (LIPOX) rate in HIV-infected patients with lipodystrophy (LIPO) and those without (controls). LIPOX was estimated by indirect calorimetry during fasting and steady state of a hyperinsulinemic euglycemic clamp in 36 (18 LIPO and 18 controls) normoglycemic HIV-infected men on highly active antiretroviral therapy. In LIPO, TNF-alpha correlated with clamp FFA (r = 0.67, P < .01), clamp LIPOX (r = 0.47, P < .05), incremental FFA (r = 0.69, P < .01), and incremental LIPOX (r = 0.64, P < .01), all of which, but not the clamp LIPOX correlation (r = 0.29, P > .05), remained significant after correction for insulin sensitivity. None of these correlations were significant in controls. In all patients, TNF-alpha correlated with clamp FFA (r = 0.61, P < .001), clamp LIPOX (r = 0.43, P < .01), and incremental FFA (r = 0.43, P < .01), with the 2 former correlations remaining significant after correction for insulin sensitivity. LIPOX and FFA (fasting and clamp values combined) correlated strongly and positively in both LIPO (R2 = 0.43, P < .001) and controls (R2 = 0.60, P < .0001). Fasting FFA and LIPOX did not differ between study groups; however, the insulin-mediated suppression of FFA and LIPOX was attenuated in LIPO (P's < .05). Our data indicate that higher TNF-alpha, independently of insulin sensitivity, is associated with attenuated insulin-mediated suppression of FFA and LIPOX in HIV-LIPO, suggesting in turn that TNF-alpha stimulates lipolysis in this syndrome. Furthermore, FFA may be a major determinant of LIPOX in HIV-infected patients on highly active antiretroviral therapy.