The National Acute Spinal Cord Injury Study II concluded in 1990 that high-dose methylprednisolone (MP) improved neurologic recovery after acute spinal cord injury (ASCI). We tested this conclusion by analysis of 54 patients with ASCI; 25 patients were treated without MP before 1990 whereas 29 patients were treated with MP after 1990. Neurologic deficit was assessed regularly, in most cases daily. Motor and sensory scores on admission, and best results at one-half week (days 2 to 4), 1 week (days 6 to 10), 2 weeks (days 11 to 21), 1 month, and 2 months were noted for both groups. Motor assessment was recorded in 22 muscle segments on a scale of 0 (complete deficit) to 5 (normal); the range, thus, was 0 to 110. The 23 patients with closed injuries demonstrated no difference in improvement with or without MP. In contrast, MP was associated with impaired improvement in the patients with penetrating wounds; the 15 patients with no MP therapy had an admission motor score of 49, which increased by 6.9 at one-half week, whereas the 16 patients treated with MP had an admission motor score of 48, which decreased by 0.3 at one-half week (p = 0.03). The neural status seen by day 4 persisted throughout the next 2 months. Changes in sensation paralleled the changes in motor function. We conclude that MP therapy for penetrating ASCI may impair recovery of neurologic function.