This study was designed to evaluate the mechanism by which propofol modifies leukocyte production of nitric oxide (NO) in humans. In vitro experiments used whole blood from healthy volunteers (n = 10 samples/experiment). Ex vivo experiments studied the effects of an intravenous dose of 2.5 mg propofol per kilogram body weight followed by intravenous infusion of 4 mg kg(-1) h(-1) in surgical patients in ASA class I or II (n = 20). In whole blood, neutrophils and plasma, we measured NO production and the activities of the enzymes nitric oxide synthase [inducible (iNOS) and constitutive (cNOS)] and cyclooxygenase [constitutive (COX-1) and inducible (COX-2)]. Concentrations of interleukins (IL-1beta, IL-6, and IL-10) and tumor necrosis factor-alpha (TNFalpha) were measured in plasma. In blood from healthy donors, propofol increased NO production and cNOS activity. The concentration of propofol that increased NO production by 50% (EC(50)) was 23.5 microM, and the EC(50) of propofol for cNOS was 18.6 microM. In blood from surgical patients, propofol increased NO production by 52% and cNOS activity by 57%. Propofol inhibited iNOS activity in vitro; the concentration that reduced activity by 50% (IC(50)) was 19.9 microM. In surgical patients propofol inhibited iNOS activity by 53%. COX-1 and COX-2 activities were inhibited in vitro (IC(50) 32.6 and 187 microM, respectively) and in surgical patients (53 and 81% inhibition, respectively). Plasma concentrations of IL-1beta, IL-6, and TNFalpha were significantly reduced in surgical patients (32, 23, and 21% inhibition, respectively). None of these parameters were modified in a group of patients (n = 10) anesthetized with sevoflurane. We conclude that propofol stimulated constitutive NO production and inhibited inducible NO production, possibly by curtailing the stimulation of iNOS by inflammatory mediators in surgical patients.