Several authors reported a decrease in metabolism of drugs during inhalational anesthesia. In this study we investigated the influence of several kinds of anesthesia on the metabolism of midazolam. Methods. In 43 patients who underwent minor surgery, anesthesia was induced by injecting 0.2 mg fentanyl followed by 0.15 mg/kg midazolam. Anesthesia was maintained by either halothane/nitrous oxide (group 1), isoflurane/nitrous oxide (group 2), fentanyl/droperidol/nitrous oxide (group 3), or halothane/air/oxygen (group 4). Venous blood was drawn after 5, 15, 30, 60, 90, 120, 180, 240, 300 and 360 min. Using the two-compartment model we estimated distribution half-life (t1/2a), elimination half-life (t1/2el), clearance (Cl), and volumes of distribution (Vz = volume in elimination phase). Results. There were no significant differences of elimination parameters between the four groups. Elimination half-life ranged from 0.72 to 15.06 h. Cl ranged from 2.5 to 12.8 ml/min per kilogram. Four patients (= 9%) had a t1/2el of 7-15 h; in 16 cases we found secondary concentration peaks. There was no correlation between drowsiness in the postoperative period and midazolam concentration. Conclusion. Even if anesthesia should influence the metabolism of midazolam, for example by reducing liver blood flow, we did not find any difference between the four groups. In accordance with others, we found some patients with a half-life of up to 15 h, which might reflect the influence of the anesthesia itself.(ABSTRACT TRUNCATED AT 250 WORDS)