Hepatic arterial blood flow (HABF) and portal blood flow (PBF) were measured in 18 dogs while awake and during isoflurane and halothane anesthesia. Surgical preparation 1 week before the measurements consisted of a left thoracotomy, placement of a left atrial catheter, and insertion of another catheter into the distal aorta via the left femoral artery. Cardiac output and liver blood flow were determined using microspheres at three stages: stage 1-awake state; stage 2-after 45 min of 1 MAC of isoflurane (eight dogs) or halothane (10 dogs) anesthesia; and stage 3-after 45 min of 2 MAC of inhalation anesthesia. Half-life and fractional clearance for indocyanine green (ICG) were determined 1 day before the experiment (awake state), and at the end of stages 2 and 3. Mean arterial pressure (MAP) and cardiac index (CI), as well as PBF, decreased during isoflurane and halothane anesthesia. HABF increased significantly during isoflurane anesthesia, remained unchanged during 1 MAC of halothane anesthesia, and significantly decreased during 2 MAC of halothane anesthesia. Apparently, hepatic oxygen supply was maintained much better during isoflurane than during halothane anesthesia. PBF correlated with CI during halothane (r = 0.97) and, to a certain extent, with MAP during isoflurane (r = 0.66). HABF correlated with CI and MAP during halothane (r = 0.74 and 0.71, respectively) but did not correlate with systemic hemodynamic variables during isoflurane. ICG half-life significantly increased during 1 and 2 MAC of halothane anesthesia. The degree of increase did not correlate with the level of anesthesia or the decrease in total hepatic blood flow.(ABSTRACT TRUNCATED AT 250 WORDS)