Over a 5-year period, 40 patients, 11 with musician's and 29 with writer's cramp, were treated with botulinum toxin A using a precise injection technique in which the hollow-bore electromyography (EMG) needle was positioned by both standard EMG and by muscle twitch evoked by stimulating current passed through it. Moderate to complete improvement in dystonia occurred in 28 patients (70%) after the first injection and in 34 patients (85%) after the second injection with better outcome in nonmusicians than in musicians. Of note, weakness of uninjected muscles, immediately adjacent to those injected, was found in 25/40 patients (63%). The most common patterns of toxin spread were from flexor digitorum sublimis to profundus, extensor carpi radialis to extensor digitorum communis, and extensor indicis proprius to extensor pollicis brevis. Spread to, and weakness of, adjacent uninjected muscles was a major factor contributing to suboptimal outcome in 6/39 (15%) such patients.