We measured serum antibodies to botulinum toxin (ABT) in 96 patients with focal dystonia who had been treated with type A botulinum toxin. The frequency of detectable ABT was 3% (three patients). Patients with ABT had received more than 50 ng of botulinum toxin, and the shortest time between two injections was significantly less than in patients without ABT. The clinical evolution of the three patients was heterogeneous: one had decreased effectiveness with repeated injections, another had persistent improvement, and the third never responded to toxin injections.