There is evidence that diabetic amyotrophy is caused by a microvasculitis of the vasa nervorum. We compared the outcome of patients treated with pulsed methylprednisolone to the published natural history of diabetic amyotrophy and assessed the safety of this treatment in patients with diabetes. We retrospectively reviewed the case records of 10 episodes of diabetic amyotrophy in 9 patients treated with pulsed oral or intravenous methylprednisolone. In 6 episodes there was marked improvement in pain within days of starting treatment. Strength improved more slowly but faster than the natural history of the disease. Treatment started within 2 months of symptom onset was associated with rapid improvement in pain; and very early treatment, started within 4 weeks of symptom onset, resulted in rapid improvement of both strength and pain. Blood glucose increased on treatment days but no patient required lasting changes in diabetic treatment as the result of this therapy and no other serious adverse effects were seen. We conclude that pulsed methylprednisolone appears to be a safe and effective treatment for diabetic amyotrophy.