The benzodiazepine midazolam is rapidly eliminated by oxidative metabolism. In young healthy volunteers elimination half-life (t1/2) is about 2.4 hours. A recent study showed a prolonged t1/2 from 8 to 22 hours in 6.5% of surgical patients, and a genetic polymorphism of midazolam's metabolism has been suggested. Therefore, we measured in 168 surgical patients the elimination of midazolam and its major hydroxylated metabolite (alpha-OH-midazolam) in blood and urine. Co-medication, disease status, smoking habits and alcohol intake were recorded; normal liver and kidney functions were assessed by routine laboratory tests. Midazolam was administered intravenously (0.1 to 0.2 mg/kg) for the induction of anaesthesia. Blood was drawn 1.5, 3, 4.5 and 6 hours after application and urine was collected for 6 hours. Plasma protein binding of midazolam was determined by equilibrium dialysis. Midazolam and alpha-OH-midazolam were measured in plasma by specific gas-liquid chromatography and in urine by high performance liquid chromatography. Data for the dose-corrected area under plasma-level curve of midazolam (AUC-midazolam/dose: 1.23 +/- 961 x 10(5) h/ml; mean +/- SD) and for the metabolic plasma ratio (AUC of alpha-OH-midazolam/AUC-midazolam: 0.52 +/- 0.28) demonstrated a log-normal distribution. Likewise, the percentage of the unbound fraction of midazolam in plasma (5.0 +/- 2.4%), urinary excretion of alpha-OH-midazolam (55.9 +/- 22.7% of dose) and the values for t1/2 (2.9 +/- 1.1 hours) did indicate a unimodal distribution. Age, comedication and smoking habits did not affect the disposition of midazolam. However, patients with regular intake of alcohol had a higher (p less than 0.05) metabolic ratio. Only in 3 patients could a prolonged t1/2 of midazolam from 7.5 to 10.2 hours be detected, but plasma levels and urinary excretion of alpha-OH-midazolam in those individuals were found to be normal. Therefore it is very unlikely that the oxidative metabolism of midazolam exhibits a genetic polymorphism.