Tolerance of long-term buflomedil was assessed by compiling safety data (adverse effects, vital signs and clinical laboratory results) from three multicentre clinical trials in patients with intermittent claudication or Alzheimer's disease-type senile dementia. The three studies were similar in design: open placebo lead-in; double-blind, placebo-controlled treatment; and open long-term treatment. Patients were randomly assigned to receive 600 mg/day buflomedil given orally for 3 or 6 months (n = 297) or placebo (n = 298). Buflomedil was continued for a further 6-12 months in 193 patients and for 12 months or more in 99 patients. Side-effects occurred in 20.5% and 18.1% of buflomedil- and placebo-treated patients, respectively, with discontinuation in 14.5% and 13.1%, respectively. In the open phase, 10.9% experienced side-effects, with 1.5% of patients discontinuing treatment. Mean changes in vital signs and laboratory tests were occasionally statistically, but not clinically, significant. Overall long-term tolerance was excellent.