Twenty-five patients with hand tremor of 2+ (moderate) to 4+ (severe) on the tremor severity rating scale were randomized to receive either 50 U of botulinum toxin (BTX) type A (Allergan, Irvine, CA) or placebo injections into the wrist flexors and extensors of the dominant limb. If patients failed to respond to the initial injection, they were eligible to receive another injection of 100 U 4 weeks later. Rest, postural, and kinetic tremors were evaluated at 2- to 4-week intervals over a 16-week study period using tremor severity rating scales, accelerometry, and assessments of improvement and disability. A significant improvement (p < 0.05) was observed on the tremor severity rating scale 4 weeks after injection in patients treated with BTX as compared with placebo, and this effect was maintained for the duration of the study. Four weeks after injection, 75% of BTX-treated patients vs. 27% of placebo-treated patients (p < 0.05) reported mild to moderate improvement (peak effect rating > or = 2). There were no significant improvements in functional rating scales, although trends were observed for some items. Postural accelerometry measurements showed a > or = 30% reduction in amplitude in nine of 12 BTX-treated subjects and in one of nine placebo-treated subjects (p < 0.05). Although all patients treated with BTX reported some degree of finger weakness, no severe, irreversible, or unexpected adverse events occurred. Chemodenervation with BTX may significantly ameliorate essential hand tremor in patients who fail to improve with conventional pharmacologic therapy.