BACKGROUND Mivacurium chloride is a bis-benzylisoquinolinium nondepolarizing neuromuscular blocking agent, hydrolyzed by butyrylcholinesterase (PCHE). The dose-response relationships for PCHE after mivacurium have not been studied. Therefore, this study was designed to establish dose-response relationships for PCHE as an antagonist of mivacurium-induced neuromuscular blockade. METHODS Forty-eight physical status 1 adults were given 0.15 mg/kg mivacurium during fentanyl-thiopental-nitrous oxide-isoflurane anesthesia. Train-of-four (TOF) stimulation was applied to the ulnar nerve every 12 s, and the force of contraction of the adductor pollicis muscle was recorded. When spontaneous recovery of first twitch height (T1) reached 10% of its initial control value, exogenous PCHE equivalent to activity present in 2.5, 5, 7.5, 15, or 25 ml/kg of human plasma was administered by random allocation to 40 patients. Neuromuscular function in another eight subjects was allowed to recover spontaneously. Two blood samples were taken for determination of plasma cholinesterase activity. The first sample was taken before induction of anesthesia, and the second sample was taken when the TOF ratio had recovered to 0.75. Dibucaine and fluoride numbers were determined from the first assay. RESULTS Administration of PCHE produced significant increases in PCHE activity in all patients. The larger the dose, the greater was the resultant plasma activity. Human PCHE produced a dose-dependent antagonism of mivacurium-induced neuromuscular blockade and the recovery times correlated inversely with PCHE activity (P < 0.01). The recovery of T1 was greater (P < 0.01) and time to attain a TOF ratio of 0.75 was shorter (P < 0.01) with any dose of PCHE than that observed in the spontaneous recovery group. After the administration of exogenous PCHE equivalent to activity present in 25 ml/kg of human plasma, recovery of TOF ratio to 0.75 or more was observed in all patients in less than 10 min and time to attain a TOF ratio of 0.75 was 55% shorter than the spontaneous recovery group (8.4 [7.1-9.7] vs. 18.7 [15.4-22] min; mean and 95% confidence intervals). CONCLUSIONS Administration of exogenous PCHE equivalent to activity present in 25 ml/kg of human plasma (in a 65-kg patient, this dose is equivalent to PCHE activity of 1,625 ml of adult human plasma) resulted in reliable antagonism of mivacurium-induced neuromuscular blockade. Nevertheless, because of the prohibitive cost of this compound, this reversal modality is unlikely to have a routine practical application at this time.