We studied 242 patients with cervical dystonia who had adequate follow-up after botulinum toxin injections to determine which clinical variables had a predictive value in the treatment outcome. Twenty-one patients (16%) categorized as nonresponders were compared with 113 patients (47%) considered to be definite responders. On average, the nonresponders had symptoms for 14 years longer than responders. Seventy-eight of 100 patients with complications were female compared with 54% of 190 patients without complications. In addition, patients with complications weighed less than those without complications. Both findings suggest that the occurrence of complications is related to smaller mean neck muscle mass. Botulinum toxin antibodies were detected in 35.7% of the nonresponders tested and in none of the responders. This comprehensive analysis of outcome variables leads us to conclude that patients with a long duration of dystonia before their first botulinum toxin injection respond less well than those with a short duration of symptoms, that some patients lose their responsiveness because of the development of blocking antibodies, and that women are more likely to develop complications, such as dysphagia and neck weakness, than are men.