Measurement of cerebral blood flow (CBF) and computed tomography (CT) cisternography were performed in 37 patients with a tentative diagnosis of normal pressure hydrocephalus (NPH) to predict their surgical outcome. The mean CBF of the whole brain was measured quantitatively by single photon emission computed tomography with technetium-99m-hexamethylpropylene amine oxime before surgery. The results of CT cisternography were classified into four patterns: type I, no ventricular stasis at 24 hours; type II, no ventricular stasis with delayed clearance of cerebral blush; type III, persistent ventricular stasis with prominent cerebral blush; type IV, persistent ventricular stasis with diminished cerebral blush and/or asymmetrical filling of the sylvian fissures. The mean CBF was significantly lower than that of age-matched controls (p < 0.005). Patients with a favorable outcome had a significantly higher mean CBF than patients with an unfavorable outcome (p < 0.005). Patients with the type I pattern did not respond to shunting. Some patients with type II and III patterns responded to shunting but improvement was unsatisfactory. Patients with type IV pattern responded well to shunting, and those with a mean CBF of 35 ml/100 g/min or over achieved a favorable outcome. The combination of CBF measurement and CT cisternography can improve the prediction of surgical outcome in patients with suspected NPH.