Regional hepatic blood flow has been determined by 4 methods with the aid of the 133Xe washout technique: scintisplenoportography (direct application of 133Xe into the spleen by means of a thin needle); arterial method (133Xe is injected into the A. hepatica by means of a catheter); retrograde-venous method (133Xe administered by an occluding hepatic vein catheter); percutaneous intrahepatic method (133Xe administered directly into the parenchyma by means of a Chiba needle). Ad 1.: Scintisplenoportography (SSP) was executed with 97 patients: 8 patients with a healthy liver presented a hepatic blood flow of 103.37 +/- 11.5 ml/100 g/min. 4 patients with a chronic hepatitis showed a hepatic blood flow of 105.67 +/- 10.2 ml/100 g/min. In 38 patients with compensated cirrhosis, hepatic blood flow was determined with 58.15 +/- 11.5 ml/100 g/min and 19 patients with decompensated cirrhosis showed a blood flow of 34.54 +/- 7.2 ml/100 g/min. Of the 19 patients, who did not present any liver image, 2 patients suffered from a prehepatic block, 1 patient (female) from a posthepatic block, the rest were decompensated cirrhoses. In 5 patients suffering from steatosis only collateral circulation was determined and in 4 patients the spleen could not be punctured. In the patients with compensated and decompensated cirrhosis of the liver, hepatic blood flow differentiated significantly (p less than 0.001) from patients with healthy livers and chronic hepatitis. In the patients with bioptically assured steatosis only the washout constant was determined. Reproducibility of this method was tested in 4 patients and no statistical difference of hepatic blood flow values could be found and the correlation coefficient amounted to 0.9856. The advantage of SSP lies in the possibility of recording the portal vein circulation: cranial collaterals were found in 33 patients, 2 patients had caudal collaterals exclusively and 29 patients cranial and caudal collaterals. 33 cirrhosis patients presented evidence of hepatic shunts. In nearly all patients hepatic blood flow was higher in the right lobe than in the left. Ad 2.: Arterial method was executed in 26 patients: 2 patients with healthy livers had a hepatic blood flow of 89.85 +/- 2.9 ml/100 g/min, 19 compensated cirrhoses with 49.28 +/- 11 ml/100 g/min and 3 decompensated cirrhoses with 36.43 +/- 3.4 ml/100 g/min. Patients suffering from cirrhosis demonstrated significantly lower hepatic blood flow than patients with healthy livers (p less than 0.001). In arterial application also, with the exception of a single patient, the values for hepatic blood flow were higher for the right than the left lobe of the liver.(ABSTRACT TRUNCATED AT 400 WORDS)