Prior to proximal aortic cross-clamping, baseline measurements of spinal cord blood flow and function were done. Blood flow was evaluated with radioactive microspheres and function determined by assessment of somatosensory evoked potential (SEP). Group 1 (N = 6) animals had aortic cross-clamping for 5 minutes after ischemic spinal cord dysfunction (SEP loss) was documented. Group 2 (N = 9) underwent aortic cross-clamping for 10 minutes after loss of SEP. Group 3 (N = 6) also underwent 10 minutes of cross-clamping after initial SEP loss, but were treated intravenously with methylprednisolone (30 mg per kilogram of body weight) 10 minutes prior to cross-clamping and again 4 hours postoperatively. After release of the cross-clamp, the animals were allowed to recover and serial evaluations of spinal cord blood flow and neurological status were carried out for seven days. Group 1 animals recovered uneventfully without evidence of neurological injury. Group 2 animals sustained a 67% incidence of permanent spastic paraplegia (p = 0.02 versus Group 1). In contrast, methylprednisolone-treated animals sustained no clinically detectable neurological injury (p = 0.02 versus Group 2). Measurements of spinal cord blood flow at the time of SEP loss revealed similar degrees of spinal cord ischemia in all groups. No significant differences were observed in the duration of aortic cross-clamping prior to SEP loss among the three groups. The data indicate that short periods of cross-clamping (5 minutes) following SEP loss are well tolerated, whereas longer periods (10 minutes) are associated with a high incidence of paraplegia.(ABSTRACT TRUNCATED AT 250 WORDS)