We made a long term prospective study of 66 patients with juvenile myoclonic epilepsy (JME). Prevalence was 10.2% among 672 patients with epilepsies. Sex distribution was equal. Sixty-three were not diagnosed on referral; JME was not initially recognized in the epilepsy clinic in 22. Clinical typical absence seizures were reported in 33.3%, myoclonic jerks in 97% and generalized tonic-clonic seizures (GTC) in 78.8% of the patients. Mean age (+/- SD) at onset was 10.5 +/- 3.4 years (range 5-16 years) for absence seizures, 15 +/- 3.5 years (range 8-26 years) for myoclonic jerks, and 16 +/- 3.5 years (9-28) years (range 1-9 years) and GTC by 4.4 +/- 2.7 years (range 1-8 years) in 14 (21.2%) patients who manifested all three types of seizure. Absence were never antedated by myoclonic jerks or GTC. Myoclonic jerks occurred on awakening in 87.5% of the patients. GTC occurred mainly on awakening, but other patients had nocturnal or diurnal GTC with no circadian distribution. Neurologic examination was normal for all patients except for tremor of the hands similar to essential tremor, noted in 35% of patients. Computed tomography (CT) brain scans were normal: 93% of patients had precipitating factors: sleep deprivation (89.5%), fatigue (73.7%), photosensitivity (36.8%; television and video games 8.8%), menstruation (24.1% of women), mental concentration (22.8%), and stress (12.3%). Incidence of JME among siblings (13 of 41 examined families) implies an autosomal recessive mode of inheritance for this Arab population. EEGs were frequently normal in treated patients. At least one abnormal EEG was recorded in 56 (84.9%) patients. Abnormalities consisted mainly of generalized discharges of spike/double spike and/or polyspike and slow wave. Frequent multiple spikes and discharge fragmentations varied from 0.5- to 20-s duration (mean 6.8 s). Twenty (30.3%) had focal abnormalities, and 18 (27.3%) had photoconvulsive discharges. Eighty-eight percent of patients remained seizure-free for > or = 3 years of follow-up. Effective treatment was achieved with valproate (VPA); control of myoclonic jerks was improved with clonazepam (CZP). CZP monotherapy did not consistently prevent GTC. Adding small doses of CZP with simultaneous reduction of VPA was the most effective and better tolerated form of medication, particularly in patients demonstrating an adverse reaction or requiring a large VPA dosage. VPA dosage was successfully reduced in 15 patients who were seizure-free for > 2 years and had infrequent seizures before treatment, but 9 of 11 patients relapsed after VPA discontinuation.(ABSTRACT TRUNCATED AT 400 WORDS)