A model simulating acute-compressive spinal cord trauma at the second lumbar spinal cord segment (100 g, 300 seconds) was used to evaluate the efficacy of a vehicle control, methylprednisolone sodium succinate (MPSS), and a 21-aminosteroid compound (U74389G). Dogs were allocated into one of five treatment groups (A to E) using ultrasonographic determination of spinal cord diameters to ensure even distribution of spinal cord diameters among the treatment groups. Initial dosages of the vehicle control (A), methylprednisolone (30 mg/kg of body weight) (B), or U74389G (30 mg/kg, 3 mg/kg, or 10 mg/kg of body weight) (C, D, or E, respectively) were administered intravenously 30 minutes after trauma. Dosages were reduced by one-half for 2 and 6 hour treatments. Then every 4 hours for 42 hours, dosages were reduced one-third and one-sixth from the original dose of methylprednisolone and U74389G, respectively. Neurological examinations were performed daily for 21 days. Histopathological examination of the traumatized spinal cord showed malacic and degenerative lesions. Although significant differences in some portions of the neurological and histopathologic examinations were observed, clinical efficacy for MPSS and U74389G could not be established in this model.