Steroids Counteract the Secondary Mechanisms of Injury after Acute Spinal Cord Injury and Experimental Evidence for the Effectiveness of Steroids

There have been many studies of the vascular, biochemical, electrolyte, and inflammatory changes after experimental SCI that involved the therapeutic administration of steroids. Most of the early experimental animal studies were in large animals, especially cats,⁹ but after 1990, most of the studies were in rats. There was intense interest in the 1980s in oxygen free radicals and lipid peroxidation as major mechanisms of secondary injury and evidence arose that these processes could be beneficially altered by steroids.^10,11 Many of the papers showed beneficial effects in terms of improved histology, such as reduced tissue cavitation and functional improvement.^12,13 Of note is the fact that many of the beneficial results in cats were not seen in other species, including a study in rats in my laboratory.¹⁴

Clinical Evidence for the Effectiveness of Steroids in Acute Spinal Cord Injury

The main trials of steroids for acute SCI were conducted in the United States and Canada in the 1980s and 1990s. The first trials were named the National Acute Spinal Cord Injury Study (NASCIS); and to underline the basis for the selection of MP, the NASCIS papers cited the long list of positive studies of steroids in experimental SCI.^37,38 The NASCIS investigators included myself and were led by a number of prominent neurosurgeons including William Collins and neuroscientists including Wise Young. The NASCIS group made many contributions to SCI, including the refinement of the Frankel system for scoring neurologic function, which became known as the NASCIS system but was ultimately replaced by the American Spinal Injury Association (ASIA) system.¹ The NASCIS expert in trials design and analysis was Michael Bracken, and the NASCIS studies were funded by the National Institutes of Health (NIH) and by the Upjohn Corporation, the pharmaceutical company that made and supplied MP, which was delivered intravenously in very large doses comparable to the doses found to be effective in experimental SCI. In total, there have been five RPCT MP trials, including three NASCIS trials and two trials in Japan. The first NASCIS MP trial, reported in 1984, encompassed 306 patients in nine centers, showed no difference between low-dose and high-dose MP, and had no placebo group.³⁹ There was an increase in wound infections in the high-dose MP group. NASCIS 2, reported in 1990, included a placebo group; a second pharmacotherapeutic agent, Naloxone, an opioid antagonist; and 487 patients in 10 centers.³⁸ This trial generated much controversy for a number of reasons, beginning with the announcement of the trial as an NIH alert to practitioners in the United States. Partly because of this apparent endorsement of MP use in SCI by the NIH, there was a tendency from then onward to consider MP a standard of therapy for acute SCI. However, the SCI guidelines group emanating from the two major North American neurosurgical organizations⁴⁰ rejected MP as a standard of therapy for SCI. Nevertheless, in some countries, fear of legal liability for nonusage of MP in acute SCI became a factor in clinicians’ decision to use the drug.⁴¹ In NASCIS 2, the rates for wound infections and gastrointestinal bleeding were approximately twice those in the placebo group. The third and last NASCIS trial, reported in 1997, involved 499 patients in 16 centers; showed that the new 21-aminosteroid, tirilazad, was not as effective as MP; and added very little new information.⁴² The first Japanese MP study essentially confirmed NASCIS 2 but involved only 177 patients in 42 centers.⁴³ The French nimodipine trial also tested MP in a much smaller number of patients, as discussed later, and MP did not improve neurologic recovery over the placebo or nimodipine groups.⁴⁴

The NASCIS 2 and NASCIS 3 trials engendered considerable controversy concerning the use of MP in acute SCI and produced many very thoughtful and thorough analyses of the results of these two studies.^41,45–50 The criticisms of these trials included problems with the selection and interpretation of the statistical tests, lack of impact of the minimal neurologic improvements on functional deficits of importance to SCI patients such as loss of bladder control, and lack of reproducible results by subsequent studies (outlined later). There was also criticism of the use of neurologic scores from only one side of the body. The critics also stated that insufficient consideration was given to the increased complications in the steroid groups, including the increased incidence of pneumonia, wound infections, and sepsis. Many critics concluded that these increased risks do not warrant the use of MP, which, at best, only marginally improves neurologic recovery. Also, it is now recognized that large doses of MP such as those used in SCI can cause acute myopathy in SCI patients.⁵¹ These issues have resulted in a major decline in some countries in the use of steroids in acute SCI. For example, it was reported in 2008 that three quarters of neurosurgeons and orthopedic surgeons in Canada involved in the care of acute SCI patients do not use steroids, representing a complete reversal of the results from 5 years previously.⁵² Many authors have defended the results of the NASCIS studies and provided additional evidence of the effectiveness of steroids.^5,7,53 Bracken’s Cochrane review of steroids in SCI has not dispelled the controversy.⁷

Therapeutic Time Window for Steroids after Acute Spinal Cord Injury

Unfortunately, the optimal time window after SCI for steroid therapy has not been completely elucidated in either laboratory animals or patients. Some secondary injury mechanisms such as ischemia and glutamate toxicity develop within minutes of injury, but others such as edema and apoptosis develop more slowly and may continue for months. Thus, the onset and duration of therapy with neuroprotective agents such as steroids are complex issues and may be factors responsible for the disparity in some of the therapeutic results obtained with steroids. In general, it has been recommended that the agent should be administered as soon as possible after SCI in humans, certainly within 8 hours.^38,42 Treatment initiated after 8 hours was discouraged because it was ineffective. Continuing treatment beyond 24 hours has also been discouraged because of the toxicity of prolonged treatment.

Current Experimental Studies of Steroids in Acute Spinal Cord Injury

Current experimental studies have focused on combined therapy, with a variety of agents being paired with steroids in an attempt to produce additive effects. For example, the combination of MP and an anti-Nogo receptor antibody produced better results in rats with acute SCI than either agent alone.⁵⁴ This chapter has concentrated on MP, but there are other steroids such as progesterone derivatives that may have therapeutic potential in SCI.

Non-NASCIS and Current Clinical Trials with MP in Acute SCI

Gerndt et al.⁵⁵ analyzed consecutive cohorts of patients treated with and without steroids and found a 2.6-fold increase in pneumonia and an increase in ventilated and intensive-care days but no increase in mortality with NASCIS 2 doses of MP. However, there was a decrease in the duration of rehabilitation days, and these authors advised “the continued but cautious use” of MP. Tsutsumi et al.⁵⁶ performed a retrospective study of a small case series of acute cervical SCI patients and found that high-dose MP improved ASIA motor scores only in patients with incomplete SCI; the authors stated that only this group should receive MP. They also cautioned that patients should be screened for potential side effects such as serious infections and diabetes and recommended that these groups should not receive the drug. Matsumoto et al.⁵⁷ performed the fifth RPCT study of MP using a NASCIS 2 dosing regimen in a small number of cervical cord injuries and added prophylactic antibiotics to the treatment protocol. There were significantly more respiratory and gastrointestinal complications in the MP-treated group, especially in patients over 60 years of age. The neurologic results were not reported.

Some of the more recent steroid trials in patients with acute SCI have combined MP with other therapies. Ito et al. recently completed a trial in a small number of patients using a novel consecutive cohort design in which all patients in both cohorts also received early surgical decompression.⁵⁸ MP was administered according to the NASCIS 2 protocol, with high doses administered within 8 hours of trauma to only one cohort. There was no difference in recovery, but there were significantly more cases of pneumonia in the MP cohort. Pointillart et al.⁴⁴ devised an excellent RPCT study containing four groups of acute SCI patients treated with either MP alone, nimodipine alone, MP plus nimodipine, or placebo. Unfortunately, there were only 104 patients distributed across these four groups. There was an attempt to operate early, but there was no randomization for surgical groups. There were no differences in neurologic recovery based on the ASIA system, but there were more infectious complications and more hyperglycemia in the MP-treated group.

Inclusion of MP in Future Trials of Neuroprotection in SCI

Because of the unwarranted assumption that MP is a standard of therapy, as discussed previously, and the presumed risk of liability for failure to administer this drug in acute SCI, some investigators have felt obliged to allow inclusion of MP as either an option or a requisite in therapeutic trials of other drugs or strategies. This is indeed unfortunate and should be resisted, as it makes evaluation of other agents more difficult. For example, if one believes that there is a small possible therapeutic value of steroids, such as in incomplete cervical cases, it would make it more difficult to establish that there is an additional value to the new drug or strategy. The large number of patients required for such a study and the costs involved may prevent the trial from being undertaken. The lack of trials of neuroprotective agents in the past 20 years may be partly due to this factor.

Conclusions

It is surprising that we are still in the midst of the MP controversy. It should have been completely eclipsed by the discovery of more effective treatments for acute SCI. Sadly, this has not occurred, and one reason is the dearth of properly designed and analyzed trials of other promising agents. One can attribute this to a pipeline that has dried up because of lack of interest by pharmaceutical companies that have been bankrupted by negative or minimally positive trials in acute SCI (Upjohn and Fidia are the major examples) and by disinterest from NIH and other public funders. The criticism of NIH’s premature early announcement of the “positive” results of NASCIS 2 and the criticisms of the nontransparent analysis of its results seem to have caused NIH to shun all subsequent offers of trials in acute clinical SCI for the past 15 years or so since the last intake of patients in NASCIS 3. Fortunately, neuroprotection trials in acute SCI have resumed recently with the Riluzole trial in North America and the Minocycline trial in Canada, largely funded by noncorporate and nongovernment agencies: the Christopher and Dana Reeve Foundation and the Rick Hansen Foundation, respectively. Without effective successor drugs for acute SCI, the old controversial MP is still around.

In my opinion, the risk of complications from the judicious use of MP for 24 hours after acute SCI is not prohibitively large in patients without specific contraindications such as diabetes mellitus. Therefore, in the absence of a proven alternative therapy, one could not criticize the continued use of high-dose MP for 24 hours after acute SCI in selected patients, perhaps only those with incomplete injuries but treated in an intensive-care setting with extreme vigilance for the known and well-documented complications of steroids, especially pneumonia and septicemia.

Viewed with the passage of time, the steroid controversy highlights the role played by the NIH in causing this controversy. Not only was the “alert” announcement unwarranted and a distraction, the NIH then contributed to the lack of transparency of the results and data analysis. It was not until years later that the raw data were made available to other investigators for analysis. Furthermore, some manipulations of the raw data by the original authors of the NASCIS reports, although justifiable, such as the concept of percent possible recovery, were extremely difficult for clinicians to appreciate and should have engendered greater efforts at clarity on the part of NIH.