Midazolam (Mid) is widely used as an anesthetic adjunct. To test its anesthetic effect vs. concentration relationships, it is desirable to establish stable and predictable Mid concentrations in plasma (and brain). Therefore, the pharmacokinetics of Mid in the enflurane-anesthetized dog were determined, and the ability of Mid to reduce the enflurane concentration required for anesthesia was measured and correlated with the Mid concentration in plasma [MID]. Mongrel dogs (n = 9) were anesthetized with enflurane and the enflurane EC50 (MAC--the end-tidal concentration at which one-half of the dogs respond to the noxious stimulation of clamping of the tail, and one-half do not) was determined. Group 1 (n = 5) received Mid 2.5 mg/kg iv over 60 sec. Plasma for determination of [MID] was collected and the enflurane EC50 was determined repeatedly over the 7-8-hr period following injection. Based on the pharmacokinetic parameters determined for Group 1, dogs in Group 2 (n = 4) received Mid as a continuous infusion of 21 micrograms kg-1 min-1 for 5 hr accompanied by an initial loading dose (3 mg/kg infused over 20 min) designed to produce a stable [MID] of 1000 ng/ml in plasma. Enflurane MAC and [MID] were determined regularly during the infusion and for 6 hr after discontinuation of the infusion. There were no important differences in the pharmacokinetic parameters determined for Group 1 vs. Group 2: t1/2,z = 98 +/- 5 vs. 95 +/- 10 min (mean +/- SEM); V = 3.94 +/- 0.27 vs. 2.98 +/- 25 L/kg; Cl = 28.5 +/- 3.1 vs. 22.3 +/- 1.1 ml kg-1 min-1, respectively. When administered as a continuous intravenous infusion (Group 2), [MID] remained stable at 949 +/- 53 ng/ml for more than 5 hr. The enflurane EC50 was reduced by 55% and the reduction remained stable during the 5 hours of Mid infusion. After a single iv bolus dose or after discontinuation of the continuous infusion, the degree of enflurane EC50 reduction diminished toward the control (i.e., enflurane alone) value as [MID] declined. Mid-azolam's pharmacokinetics and plasma concentration vs. effect relationships have been determined to be consistent under two different experimental conditions.