Abstract *Kuriko Kagitani Shimono, *Katsumi Imai, *Rie Sakakibara, *Takeshi Okinaga, and *Keiichi Ozono *[Department of Developmental Medicine (Pediatrics), Osaka University Graduate School of Medicine]. Purpose: A new antiepileptic drug (AED), clobazam (CLB), was introduced in Japan in 2000. In previous studies, CLB was effective in 30–80% of epilepsy patients. A retrospective study of the clinical efficacy and adverse effects of CLB was performed in patients with refractory epilepsy. The development of tolerance to CLB and related factors also was studied. Methods: Patients with refractory epilepsy who had one or more seizures per month despite treatment with three or more AEDs were included in the study. CLB was prescribed as an add-on therapy, once or twice per day. In patients who were taking other benzodiazepines (BZDs), the other BZDs were gradually tapered as the dosage of CLB was increased. The patients were followed up for >6 months. Results: Seventy-two patients were studied. Their ages ranged from 1 to 37 years (16.4 ± 10.2 years, mean ± SD). Twenty-one patients had symptomatic generalized epilepsy (early myoclonic encephalopathy, one; West syndrome, four; Lennox–Gastaut syndrome, 12), 48 had localization-related epilepsy (cryptogenic, one), and three had severe myoclonic epilepsy in infancy. The mean minimum dosage of CLB was 0.13 mg/kg/day in patients with body weight <20 kg, or 5.3 mg/day in those with body weight of ≥20 kg. The mean maximum dosage was 0.4 mg/kg/day in patients with body weight <20 kg, or 16.7 mg/day in those with body weight of ≥20 kg. CLB reduced the frequency of seizures by ≥50% in 48% of the subjects (18% with no seizure, 6% with 75% reduction, and 24% with 50–75% reduction). CLB was effective for both generalized epilepsy (42.5%) and focal epilepsy (50%), and especially for temporal lobe epilepsy (TLE) (60%). CLB was effective for partial seizures (55.5%), primary (57%) and secondarily generalized (40%) tonic–clonic seizures, and tonic seizures (46%). It was not effective for spasms (0), atonic seizures (0), and the majority of absence seizures (25%). CLB was effective in 10 nonresponders to clonazepam (CZP). Among the adverse effects of CLB, the incidences of somnolence, salivation, and aggressiveness were 29%, 1.3%, and 13.8%, respectively, which was lower than the incidence previously reported for patients taking CZP; and the incidence of depression was 13.8%, which was higher than that reported for patients taking CZP. Similar to several other BZDs, in some patients, tolerance to CLB developed. In 65% of patients, tolerance to CLB developed within 1 year (55% within 6 months). Tolerance was not observed in other patients who had taken CLB for >1 year. In 55% of the cases in whom tolerance to CLB developed, the frequency of seizures returned to the original frequency before starting CLB treatment, whereas in the remaining 45% of cases, the frequency of seizures increased slightly and was lower than the frequency before starting CLB treatment. The mean CLB dosage in patients in whom tolerance did not develop (16.7 mg/day) was lower than the dosage in those in whom tolerance developed (19.0 mg/day). The mean serum N-desmethyl CLB concentration in long-term users who did not develop tolerance was very high (3,071 ng/ml). Conclusions: CLB was effective for both refractory generalized and localization-related epilepsies and was especially effective for TLE. CLB was effective at a lower dosage than previously reported. The incidence of some adverse effects such as somnolence, salivation, and aggressiveness was lower for CLB than for CZP. CLB was also effective in some patients who had tolerance to other BZDs. In those who developed tolerance to CLB, tolerance appeared within 6 months in most cases. Patients who did not develop tolerance to CLB had a higher serum N-desmethyl CLB concentration, which may be an important factor.