Patients receiving chronic anticonvulsant therapy have been reported to show resistance to certain nondepolarizing neuromuscular blockers. In this study, the effects of chronic phenytoin therapy on the onset, duration, and recovery of rocuronium action was assessed. Thirty-six patients scheduled for various neurosurgical procedures were studied: 18 receiving chronic phenytoin therapy (Group I) and 18 controls (Group II). Rocuronium 0.6 mg/kg (2 x DE95) was administered after induction of general anesthesia with 4-6 mg/kg thiopental sodium and 3-5 microg/kg intravenous (IV) fentanyl. Maintenance anesthesia consisted of N2O in O2, 0.5% end-tidal isoflurane, and a fentanyl infusion. Neuromuscular block was monitored with acceleromyography of the adductor pollicis-brevis muscle by using a TOF-GUARD Biometer monitor (Biometer International A/S, Odense, Denmark). According to the amplitude of the first response of train-of-four, neither the lag time nor the onset time differed between the two groups. However, the recovery index was significantly shorter in patients chronically treated with phenytoin (mean recovery index: control group, 8.3 +/- 1.7 minutes; phenytoin group, 6.7 +/- 2.3 minutes; P < .05). In addition, the times of recovery to 10%, 25%, 75%, and 90% of the baseline response were also significantly shorter in the phenytoin group than in the control group. We conclude that the duration of action of rocuronium and the recovery index were affected by chronic phenytoin therapy.