Approximately 5 to 10% of patients with cervical, segmental or multifocal dystonia receiving repetitive local injections with botulinum toxin A (BTX A) are estimated to develop secondary loss of treatment benefit (nonresponding) because of the formation of circulating serum antibodies against the neurotoxin. Because other reasons may account for loss of benefit during the course of treatment, the group of secondary nonresponders because of antibody formation need to be separated from the antibody-negative group by appropriate testing. We present an easy clinical antibody test based on a test injection of BTX A into an indicator muscle, the extensor digitorum brevis (EDB), combined with amplitude measurements of compound muscle action potentials (CMAPs) elicited by electrical nerve stimulation of the peroneal nerve before and after the injection. The results show that in a group of clinically defined secondary nonresponders, who were serologically proven to be antibody negative, a marked decrease in CMAP amplitude can consistently be detected in the injected EDB 4 weeks after BTX A injection. In contrast, antibody-positive patients are characterized by a lack of such decrease in amplitude. In all cases, results of the EDB test were in keeping with the standard mouse bioassay test. We conclude that the EDB test is a useful tool for clinicians faced with the question of whether a secondary nonresponding patient has in fact developed antibodies to BTX A.