The spleen plays a pivotal role in the pathogenesis and maintenance of portal hypertension. Few data exist about splenic hemodynamics evaluated by duplex sonography in this condition. Twenty-six normal subjects, 207 patients with portal hypertension of various causes, and in different splenoportal hemodynamic conditions, and 31 patients with liver transplantation were evaluated. In each patient the splenic resistive index (RI = peak systolic--end diastolic velocity/peak systolic velocity) and pulsatility index (PI = peak systolic--end diastolic velocity/mean velocity) were measured. In 17 cirrhotic patients, splenic indices were compared with portal hemodynamics as invasively evaluated by hepatic vein catheterization. In the various groups, RI and PI were respectively: normal subjects, 0.51 +/- 0.05 and 0.72 +/- 0.11; cirrhotic patients with hepatopetal portal blood flow (n = 167), 0.64 +/- 0.08 and 1.03 +/- 0.24; cirrhotic patients with hepatofugal portal flow (n = 3), 0.74 +/- 0.08 and 1.27 +/- 0.08; cirrhotic patients with portal vein thrombosis (n = 9), 0.74 +/- 0.08 and 1.36 +/- 0.34; patients with noncirrhotic obstruction of the portal system (n = 7), 0.69 +/- 0.11 and 1.16 +/- 0.28; cirrhotic patients with surgical decompression of splenic vein system (n = 21), 0.54 +/- 0.07 and 0.76 +/- 0.15; patients with liver transplantation (n = 31), 0.50 +/- 0.08 and 0.70 +/- 0.15. Both RI and PI were significantly higher in cirrhotic patients with hepatopetal portal flow compared with controls (P < .0001), and even higher in cirrhotic patients with portal vein thrombosis (P < .004 and P < .001 in comparison with RI and PI values of cirrhotic patients). In patients with noncirrhotic portal vein thrombosis, splenic impedance indices were higher than those in controls (RI and PI P < .0001). Cirrhotic patients who underwent surgery for the therapy of portal hypertension showed splenic impedance indices significantly decreased compared with other cirrhotic patients (RI and PI P < .0001). In patients who underwent liver transplantation, splenic impedance indices were the same as those in controls. In 23 of the 52 patients surgically treated (surgical shunt or liver transplantation), impedance indices were evaluated both before and after surgical treatment. All these patients showed a decrease in splenic impedance indices (RI and PI, P < .0001) after surgical treatment. RI and PI values were higher in patients with large esophageal varices as compared with patients without or with small varices (P < .02 and P < .01). RI and PI values were not related to age, mean arterial pressure, sex, Child-Turcotte-Pugh score, presence of ascites, or cause. A significant correlation was found between splenic impedance indices and portal resistance as evaluated by hepatic vein catheterization (r = .80, P < .001 for RI values; r = .87, P < .001 for PI values). In conclusion, this study shows that splenic impedance indices are increased in cirrhotic patients, and seems to demonstrate that in patients with cirrhosis these indices reflect portal vein blood flow resistance.