Major Clinical Trials of Steroids in Trauma
Based on the most authoritative evidence, steroids should NOT be routinely used in traumatic brain injury and are NOT recommended for acute spinal cord injury, despite decades of research centered on the NASCIS trials.
Spinal Cord Injury: The NASCIS Trials
NASCIS I (1980s)
- Design: Compared low-dose vs high-dose methylprednisolone sodium succinate (MPSS)
- Results: No significant neurological difference between groups
- Complications: High-dose group had 3-times higher wound infection rate 1
NASCIS II (1990s)
- Design: Compared high-dose MPSS vs naloxone vs placebo
- Dosing regimen: 30 mg/kg bolus over 15 minutes, followed by 5.4 mg/kg/hour for 23 hours
- Critical flaw: Primary (a priori) outcome measures were all negative; any positive results came from post-hoc analysis rather than preplanned endpoints 2
- Evidence downgrade: Originally considered Class I evidence, but downgraded to Class III by the 2013 CNS/AANS guidelines due to methodological flaws 2
NASCIS III (late 1990s)
- Design: Compared 24-hour vs 48-hour MPSS infusion, with third arm testing tirilizad mesylate
- Results: No long-lasting neurological benefit of MPSS 1
- Post-hoc claim: Suggested additional benefit if treatment delayed 3-8 hours post-injury, but this was not a primary endpoint 3
Other SCI Trials
- Pointillart et al: Showed no benefit of MPSS, though limited by small sample size and methodological flaws 1
- GM-1 Ganglioside (Sygen): Two trials conducted; initial single-center study (37 patients) showed promise, but larger multicenter trial failed to demonstrate significant neurological improvement vs MPSS 1
- Nimodipine: 100-patient trial with four arms (placebo, MPSS, nimodipine, MPSS+nimodipine) showed no treatment group improved neurological function vs placebo 1
Traumatic Brain Injury: The CRASH Trial
CRASH Trial (2004)
- Design: Largest randomized controlled trial with approximately 10,000 participants (80% of all randomized participants across all TBI steroid trials)
- Results: Significant INCREASE in mortality with steroids (RR 1.18,95% CI 1.09-1.27) 4
- Impact: This single trial should carry the most weight given its size and quality 4
Meta-analyses
- 2005 Cochrane Review: 20 trials with 12,303 participants; due to significant heterogeneity, no pooled estimate calculated; concluded steroids should no longer be routinely used 4
- 1997 BMJ Systematic Review: 13 trials with 2,073 participants; pooled odds ratio 0.91 (0.74-1.12) for death, showing considerable uncertainty with neither moderate benefits nor harms excluded 5
Current Guideline Recommendations
Spinal Cord Injury
The 2019 CNS Guidelines state unequivocally: MPSS use remains controversial due to:
- Paucity of functional neurological benefits
- High rate of complications in MPSS groups
- Reliance on post-hoc analysis for any claimed benefits 1
The 2013 AANS/CNS Guidelines downgraded methylprednisolone from Level I to Level III recommendation because primary outcome measures were negative, with any positive results derived from post-hoc data mining rather than preplanned analysis 2
Traumatic Brain Injury
The evidence clearly demonstrates steroids increase mortality and should not be used 4
Critical Pitfalls to Avoid
Do not rely on NASCIS II/III post-hoc analyses: These represent retrospective data mining of prospective trials, which is Class III evidence regardless of how the data were originally collected 2
Do not confuse statistical significance with clinical significance: Even if putative gains were statistically valid, clinical benefits are questionable 6
Recognize the complications: MPSS therapy carries significant risks including gastrointestinal bleeding, hyperglycemia, wound infections (3x higher rate), and metabolic complications 1, 7
Understand the timing controversy: Claims that 48-hour infusions provide additional benefit when treatment is delayed 3-8 hours are based on post-hoc analysis, not primary endpoints 3
The Bottom Line
Despite 30+ years of intense research, no pharmacologic agent—including steroids—has been shown to improve neurological outcomes in acute spinal cord injury when evaluated by rigorous a priori outcome measures 1. For traumatic brain injury, the evidence actively demonstrates harm 4. The NASCIS trials, while historically influential, are fundamentally flawed in their statistical methodology and should not guide current practice 2, 6.