During 1991-1995, 223 patients were investigated in the Department of Neurosurgery, Kuopio University Hospital because of a clinical and CT diagnosis of NPH. All patients underwent intracranial pressure measurements and were formed into 3 biopsy groups. Group A included incidentally biopsied patients (104 patients, 34 biopsies) seen during 1991-1992; Group B was a prospective study group from 1993-1995 (all 51 patients biopsied); and Group C patients excluded from Group B (68 patients, 34 biopsies) by age and concomitant diseases. A cortical biopsy was taken before intracranial pressure recording altogether in 118 of the 223 patients. The biopsy revealed normal brain tissue in 66 patients. Prevalence of Alzheimer's disease (AD) in biopsied patients was 42% in Group A, 31.3% in Group B and 50% in Group C. A shunt was placed according to pressure measurement in 110 patients; of these, 8 had both AD and raised ICP. Two patients with both AD and raised ICP improved after shunt placement during the first follow-up year, 4 patients deteriorated and the condition of 2 was similar to that before shunting. The frequency of haematomas after biopsy was 2.9% in groups A and C; in Group B patients had no postoperative haematomas. There was no difference in the incidence of complications in patients who had or did not have a biopsy. The relatively high prevalence of AD in patients with NPH may explain the unsuccessful recovery of many patients after shunt placement. Cortical biopsy is an effective and safe method for finding the co-existence of AD and thus improving the diagnosis of NPH and may prevent unnecessary shunt surgery.