BACKGROUND The mechanism of isoflurane-induced cerebral hyperemia is poorly understood. Data from studies in vitro suggest that volatile anesthetics release a vasodilator prostanoid. We hypothesized that prostanoids and nitric oxide (NO) are mediators of this response in vivo. If true, inhibition of cyclooxygenase by indomethacin (5 mg/kg intravenously) or of nitric oxide synthase by N omega-nitro-L-arginine methyl ester (L-NAME; 40 mg/kg intravenously) should attenuate isoflurane-induced hyperemia. Any response to L-NAME occurring via nitric oxide should be competitively reversed by L-arginine. METHODS The cerebral blood flow (microsphere) response to 1 MAC isoflurane was tested at three time points (0, 90, and 180 min) in pentobarbital-anesthetized pigs. Isoflurane challenges were separated by 60-min periods of continuous intravenous pentobarbital alone. Control animals (n = 7) received no additional pharmacologic intervention. Experimental animals were randomized to receive L-NAME before the second and indomethacin before the third isoflurane challenge (n = 7); L-NAME before the second and L-arginine (400 mg/kg intravenously) before the third isoflurane challenge (n = 9); or indomethacin before the second and L-NAME before the third isoflurane challenge (n = 8). RESULTS In control animals, isoflurane reproducibly increased cerebral blood flow (whole brain; 113 +/- 18%, 120 +/- 18%, and 103 +/- 19% increase above baseline at each time point, respectively). Both indomethacin and L-NAME attenuated (10 +/- 10% and 52 +/- 11% increase, respectively) the hyperemic response to isoflurane. The effect of L-NAME was reversed by L-arginine. CONCLUSIONS We conclude that both prostanoids and nitric oxide contribute to isoflurane-induced hyperemia. We are unable to determine from our data what, if any, interaction exists between these two mechanisms.