One biceps muscle of 8 patients with Duchenne muscular dystrophy was injected at 55 sites with a total of 55 million viable, purified, and contamination-free normal myoblasts (myoblast transfer). The other biceps of each patient was injected with a placebo to serve as a control. The procedure was blinded to the patients, parents, and investigators. Myoblasts derived from a biopsy specimen of the fathers were cultured and purified under strict conditions and carefully screened for microbial contamination. All patients received cyclophosphamide for immunosuppression for 6 or 12 months. No serious complications were observed after myoblast transfer, indicating that the procedure is safe. The overall therapeutic efficiency of myoblast transfer was poor as judged by the results in maximal voluntary force generation, dystrophin content of the muscle, magnetic resonance imaging of the muscle, and the lack of donor-derived DNA and dystrophin messenger RNA in the injected muscle. An improved efficiency of the take of myoblasts might be achieved by using younger cells and injecting the myoblasts with a myonecrotic agent (to increase the prevalence of regeneration) and a basal laminal fenestrating agent.