Neurologic improvement after thoracic, thoracolumbar, and lumbar spinal cord (conus medullaris) injuries.

STUDY DESIGN Retrospective. OBJECTIVE With approximately 10,000 new spinal cord injury (SCI) patients in the United States each year, predicting public health outcomes is an important public health concern. Combining all regions of the spine in SCI trials may be misleading if the lumbar and sacral regions (conus) have a neurologic improvement at different rates than the thoracic or thoracolumbar spinal cord. SUMMARY OF BACKGROUND DATA Over a 10-year period between January 1995 to 2005, 1746 consecutive spinal injured patients were seen, evaluated, and treated through a level 1 trauma referral center. A retrospective analysis was performed on 150 patients meeting the criteria of T4 to S5 injury, excluding gunshot wounds. One-year follow-up data were available on 95 of these patients. METHODS Contingency table analyses (chi-squared statistics) and multivariate logistic regression. Variables of interest included level of injury, initial American Spinal Injury Association (ASIA), age, race, and etiology. RESULTS A total of 92.9% of lumbar (conus) patients neurologically improved one ASIA level or more compared with 22.4% of thoracic or thoracolumbar spinal cord-injured patients. Only 7.7% of ASIA A patients showed neurologic improvement, compared with 95.2% of ASIA D patients; ASIA B patients demonstrated a 66.7% improvement rate, whereas ASIA C had a 84.6% improvement rate. When the two effects were considered jointly in a multivariate analysis, ASIA A and thoracic/thoracolumbar patients had only a 4.1% rate of improvement, compared with 96% for lumbar (conus) and incomplete patients (ASIA B-D) and 66.7% to 72.2% for the rest of the patients. All of these relationships were significant to P < 0.001 (chi-square test). There was no link to age or gender, and race and etiology were secondary to region and severity of injury. CONCLUSION Thoracic (T4-T9) SCIs have the least potential for neurologic improvement. Thoracolumbar (T10-T12) and lumbar (conus) spinal cord have a greater neurologic improvement rate, which might be related to a greater proportion of lower motor neurons. Thus, defining the exact region of injury and potential for neurologic improvement should be considered in future clinical trial design. Combining all anatomic regions of the spine in SCI trials may be misleading if different regions have neurologic improvement at different rates. Over a ten-year period, 95 complete thoracic/thoracolumbar SCI patients had only a 4.1% rate of neurologic improvement, compared with 96.0% for incomplete lumbar (conus) patients and 66.7% to 72.2% for all others.