The two-edged sword of large-dose steroids for spinal cord trauma.

OBJECTIVE: In 1990, large-dose steroid administration was advocated in spine-injured patients to lessen neurologic deficits. The authors undertook both prospective and retrospective studies to evaluate the response of such profound pharmacologic intervention. SUMMARY BACKGROUND DATA: Of all sources of nonfatal injury, spinal cord trauma remains the most devastating in both cost and impact on the quality of the patient's life. One study found that routine large-dose steroid administration after injury lessened the extent of neurologic injury. After uncommonly prompt and broad lay press publicity, this practice was widely accepted. Biased by knowledge of the known immunosuppressive effects of steroids, the authors suspected that pneumonia was both more frequent and severe in steroid-treated patients.
METHODS: Thirty-two patients with cervical or upper thoracic spinal injuries (C3-6, 20 patients; C6-7, 6 patients; and T1-6, 6 patients) were studied at an urban level I trauma center from January 1987 to February 1993. Complete spinal cord injury was present in 22 of 32 patients; 14 patients received steroids postinjury. There was no difference in mean age, cord level, age-adjusted injury severity score, or the percent of injury severity score caused by the spinal injury.
RESULTS: The length of hospital stay was longer in steroid-treated patients (S) than in nonsteroid (NS) patients, that is, 44.4 days versus 27.7 days, respectively (p = 0.065). Seventy-nine per cent of S patients had pneumonia compared with 50% of NS patients (p = 0.614). There was no statistical difference in the episodes of pneumonia per patient between the two groups (p > 0.05). Prospectively, the authors evaluated sequentially several parameters known to be important in human immune responses to bacterial challenges in nine S and five NS patients. In S patients, both the per cent and density of monocyte class II antigen expression and T-helper/suppressor cell ratios were lower than in NS patients. However, S patients did have an initially higher, earlier boost in some host defense parameters that rapidly declined, and their subsequent response was both blunted and delayed. These differences became even clearer when stratified according to cord level and incomplete versus complete cord status. Not surprisingly, infected patients, whether S or NS, had lower levels of monocyte antigen expression, CR3, and helper/suppressor ratios.
CONCLUSIONS: These data do not permit a judgment to be made whether neurologic status was improved by S administration. It is known that vital immune responses were adversely affected, that pneumonia was somewhat more prevalent, and that hospitalization was prolonged and costs therefore increased by an average of $51,504 per admission. Further clinical studies will be needed to determine to what extent these observations offset the putative benefits of large-dose steroids in the treatment of spinal trauma.