Original Article: Kolias AG, Adams H, Timofeev IS, et al. “Evaluation of Outcomes Among Patients With Traumatic Intracranial Hypertension Treated With Decompressive Craniectomy vs Standard Medical Care at 24 Months: A Secondary Analysis of the RESCUEicp Randomized Clinical Trial.” JAMA Neurol. 2022;79(7):664–671. doi:10.1001/jamaneurol.2022.1070

Background

Guidelines for decompressive craniectomy (DC) for intracranial pressure (ICP) control after traumatic brain injury (TBI) center around the DECRA trial, published in 2011. This study looked at bifrontal decompressive hemicraniectomy compared to standard of care in patients who had sustained ICP >20mmHg for >15 minutes despite optimized non-surgical interventions^.1 DECRA found that patients who underwent DC had similar mortality rates compared to standard care and higher rates of poor functional outcomes at 6 months. The trial was criticized for aggressive treatment of patients with elevated ICP not consistent with standard practice as well as a restrictive criteria for inclusion, limiting its applicability to most patients with severe TBI.² To that end, Hutchinson et al, and the RESCUEicp investigators published their results of a second randomized controlled trial comparing DC with medical management. The RESCUEicp authors utilized methodologies that were more consistent with current practice, utilizing DC as a last-tier intervention for sustained, refractory elevations in ICP and was designed to be more generalizable. Their results demonstrated a significant mortality benefit at both 6 and 12 months in the surgical group at a cost of more patients left in a vegetative state or with severe disability at similar time points.³ This study evaluates outcomes for these same participants at 24 months to determine if there is any differential effect of DC on late-outcomes after severe TBI.⁴

Methods

Patients were eligible for inclusion if they were aged 10-65 with abnormal CT and ICP monitoring with ICP>25mmHg refractory to initial medical treatments. Patients were excluded if they had clinical findings consistent with devastating brain injury (i.e., fixed and dilated pupils) or coagulopathy.  All patients received initial ‘stage 1’ treatment for elevated ICP consisting of targeting cerebral perfusion pressures >60mmHg, hypocarbia, normothermia, and normal blood glucose.  Persistent ICP>25mmHg led to ‘Stage 2’ treatments, consisting of inotropes, ventriculostomy, hyperosmolar therapy, hypothermia and loop diuretics.  If patients continued to have ICP>25mmHg for 1-12 hours despite both stage 1 and 2 therapies, they were randomized to surgery or medical therapy. Patients randomized to surgery received either a unilateral or bilateral fronto-temporo-parietal craniectomy at the discretion of the surgeon. Those randomized to medical management continued ‘stage 2’ treatments and were placed in a barbiturate coma for refractory ICP elevation.  Barbiturates were not permitted in the stage 2 treatment phase. The primary outcome was the extended Glasgow outcome score (GOS-E) measured at 24 months assessed by mail or telephone interview.

Results

Of 408 patients initially randomized, 356 (87%) were included in the 24-month analysis. The between-group difference, initially seen at 6 and 12 months, was maintained at 24 months (x2=24.20, p=0.001). Additionally, the significant mortality benefit noted at 6 and 12 months in the surgical group was maintained at 24-months with an absolute percentage difference of -20.5% (95% CI -30.8% to -10.2).

Furthermore, in this analysis, the initial 6-month mortality benefit associated with DC was offset by an equivalent number of patients left in either a vegetative or severely disabled state. However, at 24 months the allotment was no longer similar as a disproportionate number of patients in the surgical group improved by at least 1 point on the GOS-E scale (30.4% vs 14.5%) compared to the medical group, respectively. Rates of good recovery at 24 months were similar between the surgical and medical groups, with lower good recovery in 7.7% vs 5.2% and upper 3.3% vs 5.7%, respectively.

Commentary

This study found that surgical treatment of sustained refractory intracranial hypertension demonstrated continued mortality benefits at 24 months without an increase in good recovery but with increased rates of vegetative state, severe disability and moderate disability. Patients in the surgical group were more likely to improve by at least 1 GOS-E grade.

Compared to the prior DECRA trial by Cooper et al, RESCUEicp’s methodology was more applicable to current practice of severe TBI. However, much of the criticism aimed at the initial RESCUEicp trial, evaluating 6 and 12-month outcomes, is also applicable to this secondary analysis reporting 24-month outcomes.^2-4 Bifrontal craniectomies outnumbered hemicraniectomies by approximately 2-to-1, but these rates appear inconsistent with common practice. The number of patients crossing over between groups was high with almost 40% of patients randomized to medical treatment receiving DC and 7% of patients randomized to surgery treated medically. Large-scale crossover such as this can lead to decreased statistical power favoring the null hypothesis. Fortunately, in this trial that does not appear to be the case, but nonetheless the treatment effect of DC may have been blunted due to this imbalance. Differences between those lost to follow-up between 6, 12 and 24 months may also have an unclear effect on the analysis that may bias results. Finally, the issue of timing of cranial reconstruction after DC is poorly referenced in both the trial text and supplementary appendices. Smaller studies and pooled meta-analyses have reported benefits with earlier cranioplasty (<90 days) compared to those beyond (>90 days). As an improvement in GOS-E of at least one point was noted in 30% of those randomized to surgery at 24 months’ time, future trials regarding DC should require collection of data pertaining to this variable.⁵

Walter Valesky, MD
Clinical Instructor, University of California, Irvine
Department of Neurology