CONSORT indicates Consolidated Standards of Reporting Trials; RESCUEicp, Randomized Evaluation of Surgery With Craniectomy for Uncontrollable Elevation of Intracranial Pressure.
Patients with severe traumatic brain injury and sustained and refractory intracranial hypertension were randomly assigned 1:1 to decompressive craniectomy with standard care (surgical group) or to ongoing medical treatment with the option to add barbiturate infusion (medical group). Outcomes were assessed using the 8-point Extended Glasgow Outcome Scale (GOS-E), with 1 indicating death and 8 denoting upper good recovery.
eMethods 1. RESCUEicp Eligibility Criteria
eMethods 2. Therapeutic Management Protocol
eTable 1. Trial Recruiting Centers
eTable 2. Cross-Tabulation of Functional Outcomes
eTable 3. Sensitivity Analysis
eTable 4. Pairwise Comparison of Functional Outcome Groups
eFigure. Flow-Stacked Bar Charts for Extended Glasgow Outcome Scale Results of the Surgical and Medical Groups at 6, 12, and 24 Months
Nonauthor Collaborators. The RESCUEicp Trial Collaborators
Data Sharing Statement
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
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
What are the 24-month outcomes for patients with traumatic intracranial hypertension who receive a decompressive craniectomy vs standard medical treatment?
In this prespecified secondary analysis of a randomized clinical trial, 408 adults received either decompressive craniectomy or standard care. The Extended Glasgow Outcome Scale was used to assess 24-month outcomes in this secondary analysis; surgical patients had sustained reduced mortality but higher rates of vegetative state, severe disability, and moderate disability at 24 months.
At 24 months, surgical patients with traumatic intracranial hypertension were more likely to improve over time compared with patients in the standard medical treatment group.
Trials often assess primary outcomes of traumatic brain injury at 6 months. Longer-term data are needed to assess outcomes for patients receiving surgical vs medical treatment for traumatic intracranial hypertension.
To evaluate 24-month outcomes for patients with traumatic intracranial hypertension treated with decompressive craniectomy or standard medical care.
Design, Setting, and Participants
Prespecified secondary analysis of the Randomized Evaluation of Surgery With Craniectomy for Uncontrollable Elevation of Intracranial Pressure (RESCUEicp) randomized clinical trial data was performed for patients with traumatic intracranial hypertension (>25 mm Hg) from 52 centers in 20 countries. Enrollment occurred between January 2004 and March 2014. Data were analyzed between 2018 and 2021. Eligibility criteria were age 10 to 65 years, traumatic brain injury (confirmed via computed tomography), intracranial pressure monitoring, and sustained and refractory elevated intracranial pressure for 1 to 12 hours despite pressure-controlling measures. Exclusion criteria were bilateral fixed and dilated pupils, bleeding diathesis, or unsurvivable injury.
Patients were randomly assigned 1:1 to receive a decompressive craniectomy with standard care (surgical group) or to ongoing medical treatment with the option to add barbiturate infusion (medical group).
Main Outcomes and Measures
The primary outcome was measured with the 8-point Extended Glasgow Outcome Scale (1 indicates death and 8 denotes upper good recovery), and the 6- to 24-month outcome trajectory was examined.
This study enrolled 408 patients: 206 in the surgical group and 202 in the medical group. The mean (SD) age was 32.3 (13.2) and 34.8 (13.7) years, respectively, and the study population was predominantly male (165 [81.7%] and 156 [80.0%], respectively). At 24 months, patients in the surgical group had reduced mortality (61 [33.5%] vs 94 [54.0%]; absolute difference, −20.5 [95% CI, −30.8 to −10.2]) and higher rates of vegetative state (absolute difference, 4.3 [95% CI, 0.0 to 8.6]), lower or upper moderate disability (4.7 [−0.9 to 10.3] vs 2.8 [−4.2 to 9.8]), and lower or upper severe disability (2.2 [−5.4 to 9.8] vs 6.5 [1.8 to 11.2]; χ27 = 24.20, P = .001). For every 100 individuals treated surgically, 21 additional patients survived at 24 months; 4 were in a vegetative state, 2 had lower and 7 had upper severe disability, and 5 had lower and 3 had upper moderate disability, respectively. Rates of lower and upper good recovery were similar for the surgical and medical groups (20 [11.0%] vs 19 [10.9%]), and significant differences in net improvement (≥1 grade) were observed between 6 and 24 months (55 [30.0%] vs 25 [14.0%]; χ22 = 13.27, P = .001).
Conclusions and Relevance
At 24 months, patients with surgically treated posttraumatic refractory intracranial hypertension had a sustained reduction in mortality and higher rates of vegetative state, severe disability, and moderate disability. Patients in the surgical group were more likely to improve over time vs patients in the medical group.
ISRCTN Identifier: 66202560
Decompressive craniectomy is a life-saving procedure to reduce critically elevated intracranial pressure (ICP) among patients with traumatic brain injury (TBI).1 The optimal timing, indications, and functional outcome benefits associated with decompressive craniectomy have been widely debated.2 Two multicenter randomized clinical trials, Decompressive Craniectomy (DECRA) and Randomized Evaluation of Surgery With Craniectomy for Uncontrollable Elevation of Intracranial Pressure (RESCUEicp), have reported contradictory findings.
The DECRA trial, published in 2011, assessed early decompressive craniectomy for patients with diffuse TBI who experienced increased ICP (>20 mm Hg for >15 minutes within 1 hour) despite optimized early interventions such as sedation, normalized arterial carbon dioxide pressure, and use of hyperosmolar therapy, neuromuscular blockade, and external ventricular drainage.3 DECRA patients were randomized to (1) early bifrontal decompressive craniectomy and standard care or (2) standard care. In 2013, the DECRA investigators reported similar mortality rates and a higher rate of unfavorable outcomes for patients who received a decompressive craniectomy at 6 months vs those who received standard care.4 The difference in outcomes, in favor of medical treatment, persisted at 12 months but was no longer statistically significant.
Contrary to DECRA, in which decompressive craniectomy was investigated for patients with less severe intracranial hypertension that was not controlled by early interventions, the RESCUEicp investigators evaluated the effectiveness of craniectomy as a last-tier intervention for patients with TBI and refractory and sustained intracranial hypertension.5 The RESCUEicp results were published in 2016, demonstrating significant differences in functional outcomes at 6 and 12 months. Patients who received a decompressive craniectomy had lower mortality rates, higher rates of vegetative state, and higher Extended Glasgow Outcome Scale (GOS-E) scores for vegetative state, lower severe disability (dependent), and upper severe disability (independent at home for at least 8 hours) compared with those who received medical treatment. Rates of moderate disability and good recovery were similar for both groups. The initial RESCUEicp trial results suggested that the reduced mortality rate as a result of decompressive craniectomy translates to both dependent and independent living at 6 and 12 months. Here, we report group outcomes for RESCUEicp participants at 24 months and examine the trajectory of functional outcomes from 6 to 24 months.
This prespecified secondary analysis of data from the RESCUEicp randomized clinical trial was approved by the Cambridgeshire 4 Research Ethics Committee; the trial protocol is provided in Supplement 1. Ethics committee approval was also obtained from all other participating institutions outside the UK. Because this trial only enrolled patients with severe TBI, their nearest relative or designated person provided written informed consent. This study adhered to the Consolidated Standards of Reporting Trials (CONSORT) guideline.
This analysis assessed functional outcomes using GOS-E scores at 24 months. Detailed study population characteristics, recruitment, and primary results were published previously.5 In brief, participants were recruited from 52 centers in 20 countries. Enrollment occurred between January 2004 and March 2014. Data analysis was performed between 2018 and 2021. Eligibility criteria were as follows: age 10 to 65 years, TBI confirmed with computed tomography, ongoing ICP monitoring, and increased ICP (>25 mm Hg for 1-12 hours) despite tier 1 and 2 interventions (eMethods 1 in Supplement 2). Patients with traumatic intracranial hematomas requiring immediate evacuation were included if the bone flap had been replaced (ie, they did not have a primary decompressive craniectomy). Exclusion criteria were as follows: bilateral fixed and dilated pupils, bleeding diathesis, or unsurvivable injury. All participating hospitals provide 24-hour neurosurgical services and have acute neurosciences services for patients with severe TBI. Patients were treated according to a tiered treatment protocol as described in eMethods 2 in Supplement 2. Included patients arrived needing tier 3 treatment and had elevated ICP (>25 mm Hg) for 1 to 12 hours despite tier 1 and 2 interventions. They were randomly assigned 1:1 to receive (1) a decompressive craniectomy with medical treatment (surgical group) or (2) ongoing medical treatment with the option to add barbiturate infusion (medical group).
The 24-month outcome data were part of the end points described in the trial protocol (Supplement 1) but were not included in the primary publication5 because their collection was ongoing at the time. The 24-month outcome data were collected and analyzed with the same methodology as the 6- and 12-month outcome data.
In brief, follow-up for UK participants was performed by the trial office in Cambridge initially by mail. If no response was received, a telephone interview was conducted with patients and/or their family members to complete the GOS-E questionnaire. For non-UK sites, local members of the research team conducted the interview. Investigators blinded to treatment centrally adjudicated outcomes, as detailed by Hutchinson et al.5
Similar to the primary outcome measure at 6 months, outcomes in this study were assessed at 24 months with the GOS-E, a 19-item questionnaire that measures an ordinal scale of the global outcome and is commonly used as the primary outcome in trials involving neurotrauma patients. Scores range from 1 (death) to 8 (upper good recovery). The GOS-E defines 8 categories of possible outcomes based on functional independence, work, social and leisure activities, and relationships: 1 indicates death; 2, vegetative state (unable to obey commands); 3, lower severe disability (requires frequent help in daily living); 4, upper severe disability (independent at home for at least 8 hours); 5, lower moderate disability (independent at home and outside the home but with some physical or mental disability); 6, upper moderate disability (independent at home and outside the home but with some physical or mental disability, with less disruption than lower moderate disability); 7, lower good recovery (able to resume normal activities with some injury-related problems); and 8, upper good recovery (no problems).
All outcomes were reported in the intention-to-treat population, excluding patients who withdrew consent or were declared lost to follow-up. The 24-month outcomes analysis was performed on the intention-to-treat population with no imputation of missing data, in accordance with the RESCUEicp statistical analysis plan (Supplement 1) and the primary end point data published previously.5
The previously published 6- and 12-month outcomes were analyzed with an unordered χ2 test because the proportional-odds model was rejected, suggesting a GOS-E distribution difference between the 2 randomized groups.5 Therefore, we followed the same analysis using an unordered χ2 test to assess GOS-E differences at 24 months.
To determine GOS-E changes over time, Dunn-Bonferroni post hoc tests with Bonferroni adjustments were performed after the Friedman test to assess paired comparisons between differences between individual time points (6 vs 12, 6 vs 24, and 12 vs 24 months).
We also analyzed net change in GOS-E scores between the 6- and 24-month time points. We used the unordered χ2 test to compare the proportions of patients with an unchanged GOS-E score and improvement or worsening by at least 1 grade between randomized groups. As a sensitivity analysis, the proportion of patients achieving upper severe disability or better on the GOS-E scale was compared between randomized groups using a χ2 test. P < .05 (2-tailed) was considered significant. Statistical analysis was performed with Stata MP version 16 (StataCorp LLC).
A total of 408 patients were enrolled in the RESCUEicp trial, with 206 in the surgical group and 202 in the medical group; after exclusions, there were 202 patients in the surgical group and 196 patients in the medical group (Figure 1 and eTable 1 in Supplement 2). The mean (SD) age was 32.3 (13.2) vs 34.8 (13.7) years. In the surgical group, there were 165 men (81.7%) and 37 women (18.3%); in the medical group, there were 156 men (79.6%) and 39 women (20.0%; sex was not available for 1 patient). For the surgical vs medical groups, intention-to-treat analyses of outcomes at 6, 12, and 24 months included 389 (201 vs 188), 373 (194 vs 179), and 356 (182 vs 174) patients, respectively.
Table 1 compares population characteristics of the surgical and medical groups.5 Outcomes classification according to full GOS-E distributions at 6, 12, and 24 months after randomization (primary and secondary analyses, modified intention-to-treat population) is detailed in Table 2 and Figure 2. eTable 2 in Supplement 2 presents cross-tabulations of functional outcomes.
Between-group differences in GOS-E score distribution at 6 months (χ27 = 30.69, P < .001) and 12 months (χ27 = 29.16, P < .001) have been published previously (Table 2). This distribution was also sustained at 24 months (χ27 = 24.20, P = .001). For every 100 patients at 24 months who were treated with surgical rather than medical intent, there were 21 additional survivors (4 in a vegetative state, 2 with lower and 7 with upper severe disability, and 5 with lower and 3 with upper moderate disability). Rates of lower and upper good recovery were similar for the surgical and medical groups (20 [11.0%] vs 19 [10.9%]) (Table 2). In the sensitivity analysis, 82 of 182 patients (45.1%) in the surgical group and 54 of 174 (31.0%) in the medical group had a GOS-E outcome of upper severe disability or better (χ2 = 7.41, P = .006) (eTable 3 in Supplement 2).
Dunn-Bonferroni post hoc tests were performed to adjust for multiple testing between the various time points. Significant differences were observed between 6- and 24-month GOS-E outcomes in the surgical group (P = .004) (Table 3). No statistically significant between-group differences were found for any other time points after Bonferroni adjustments (eTable 4 in Supplement 2). Temporal changes of patient transitions across the full range of GOS-E scores for each treatment group were calculated from Table 2 and are presented in the eFigure in Supplement 2.
Significant differences in the analysis of net change in improvement or worsening by 1 GOS-E grade or more or unchanged GOS-E scores were also seen across the 2 treatment groups between 6 and 24 months (χ22 = 13.27, P = .001). A total of 55 patients (30.4%) in the surgical group achieved net improvement of 1 grade or more compared with 25 patients (14.5%) in the medical group (Table 3). The net percentage of worsening of both treatment groups was equal, at 8.2%. Of the study participants, 111 (61.3%) in the surgical group and 134 (77.5%) in the medical group had no change in GOS-E score.
This randomized clinical trial found that patients with posttraumatic intractable intracranial hypertension treated with decompressive craniectomy had a sustained reduction in mortality at 24 months compared with those treated with standard medical care. However, we observed higher proportions of patients in the surgical group who were in a vegetative state or had severe or moderate disability. Additional analyses of GOS-E changes over time showed that patients in the surgical group were more likely to improve over time than patients in the medical group, as demonstrated by the paired time-point analyses. However, it is important to note that GOS-E scores for the majority of patients in both groups remained unchanged over the period from 6 to 24 months. Improvement by at least 1 grade was seen for 30.4% of the surgical group vs 14.5% of the medical group.
Previously published 6- and 12-month RESCUEicp outcome data demonstrate that mortality was 22.1% and 21.5% lower, respectively, for the surgical group than for the medical group. The proportions of patients in the surgical group who were in a vegetative state or had lower severe or upper severe disability were higher than those of the medical group; rates of moderate disability and good recovery were similar for both groups. However, at 24 months, the rate of moderate disability was also higher in the surgical group than the medical group, with similar rates of good recovery. This observation may be because patients in the surgical group are more likely to improve over time, and as a result, shift to the better end of the GOS-E spectrum.
Compared with the DECRA trial that investigated decompressive craniectomy when early interventions failed to control intracranial hypertension, we observed substantial differences in patient outcomes in the RESCUEicp trial that examined last-tier use of decompressive craniectomy for refractory and sustained intracranial hypertension.3 The DECRA investigators reported similar mortality rates in both groups at 6 months (19.0% and 18.0%, respectively) and 12 months (21.0% and 19.0%, respectively), whereas our findings showed a decreased mortality rate at 6 and 12 months in the surgical group with last-tier decompressive craniectomy.4,6 Our current analysis supports these findings; we report that mortality rates at 24 months were 54.0% and 33.5% and were lower in the surgical group.
The RESCUEicp trial has other notable differences from the DECRA trial, including a larger sample size (408 vs 155), which indicates that this trial may have had greater power to detect between-group differences.3,5 In addition, RESCUEicp had a longer follow-up time (24 vs 12 months).6 Because of the shorter follow-up period, it is not possible to analyze the GOS-E trajectories in the DECRA trial, which is important given the findings of our study.
Much of the controversy regarding the use of decompressive craniectomy for patients with severe TBI and intracranial hypertension is possibly attributable to the paucity of good-quality, long-term outcome data from controlled studies.7-9 Here we present outcomes at 24 months for 1 of 2 main randomized clinical trials comparing decompressive craniectomy vs medical treatment for patients with TBI. In addition, we provide data indicating that the scope of functional improvement up to 24 months after decompressive craniectomy is significant. This is an important addition to the debate and potentially raises questions regarding the assessment of the primary outcome at 6 months in similar trials.8,10 Ongoing functional recovery past 12 months after decompressive craniectomy, as well as improved health-related quality of life up to 10 years after neurorehabilitation, has been recognized in TBI.11,12
In this analysis, the improvement by at least 1 point over the GOS-E scale in the surgical group from 6 to 24 months was more than double that in the medical group (30.4% vs 14.5%). This improvement could have resulted from the physiologic effects of the cranioplasty, which is usually undertaken a few months after the primary surgery, or possibly more frequent health care contact, including the provision of rehabilitation. It is possible that patients in the surgical group who remained on a surgical or hospital path as a result of cranioplasty requirements received more rehabilitation. However, the trial did not systematically collect data regarding rehabilitation, so this hypothesis cannot be corroborated. The improvement in the decompressive craniectomy group from 6 to 24 months may also indicate that cranioplasty, although it has its own attendant risks, had a positive effect on the long-term outcomes of these patients.
The current longer-term analysis reinforces our previous recommendation that clinicians should carefully counsel families of patients regarding decompressive craniectomy, taking into account the possible detrimental effects of the various treatment strategies with regard to survival and long-term outcomes.5,13 In addition to the clinical indications, the decision to perform a decompressive craniectomy should be made by the multidisciplinary team in conjunction with the patient’s closest relatives to consider the patient’s wishes.
Our study has a few limitations. More than one-third of the medical group received a decompressive craniectomy as a result of medical treatment failure, which could mitigate the observed treatment effect. Furthermore, we analyzed GOS-E scores in this trial to explore global functional outcomes after TBI; the granularity of the 8 outcome categories may not have allowed us to adequately capture functional outcome differences between the 2 treatment groups in the context of severe TBI. Because of the trial’s pragmatic nature, long-term data on cranioplasty were not collected systematically. Cranioplasty could influence the trajectory of global functional outcomes because the procedure restores ICP to a normal physiologic range and has been shown to improve cerebrospinal fluid flow and cerebral metabolism.13,14 This trial also did not evaluate the effectiveness of primary decompressive craniectomy, which is currently being investigated as part of the RESCUE-ASDH trial.
This secondary analysis of the RESCUEicp randomized clinical trial found that patients with intractable intracranial hypertension after TBI who received a decompressive craniectomy had a sustained reduction in mortality at 24 months compared with those who received standard medical care. For every 100 patients treated with surgical rather than medical intent, 21 additional patients survived at 24 months (4 were in a vegetative state, 2 had lower and 7 had upper severe disability, and 5 had lower and 3 had upper moderate disability). Rates of good recovery were similar in both groups. Patients in the surgical group were more likely to improve over time compared with patients in the medical group. These data support the use and potential benefit of longer-term follow-up for TBI clinical trials. There is a wide spectrum of outcomes among patients undergoing decompressive craniectomy for intractable intracranial hypertension. These findings support the notion that careful patient selection, following the principles of multidisciplinary consensus and shared decision-making with the closest relatives, is required.
Accepted for Publication: January 28, 2022.
Published Online: June 6, 2022. doi:10.1001/jamaneurol.2022.1070
Corresponding Author: Angelos G. Kolias, MD, PhD, Department of Clinical Neurosciences, University of Cambridge, Box 167, Biomedical Campus, Cambridge CB2 0QQ, United Kingdom (email@example.com).
Author Contributions: Drs Kolias and Hutchinson had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Kolias and Adams served as co–first authors.
Concept and design: Kolias, Adams, Timofeev, Hossain, Czosnyka, Belli, Mendelow, Mitchell, Sahuquillo, Unterberg, Servadei, Teasdale, Pickard, Menon, Murray, Kirkpatrick, Hutchinson.
Acquisition, analysis, or interpretation of data: Kolias, Adams, Timofeev, Corteen, Hossain, Timothy, Anderson, Bulters, Eynon, Wadley, Mitchell, Wilson, Critchley, Posti, Teasdale, Menon, Murray, Hutchinson.
Drafting of the manuscript: Kolias, Adams, Hossain, Belli, Posti, Pickard, Hutchinson.
Critical revision of the manuscript for important intellectual content: Kolias, Adams, Timofeev, Corteen, Hossain, Czosnyka, Timothy, Anderson, Bulters, Belli, Eynon, Wadley, Mendelow, Mitchell, Wilson, Critchley, Sahuquillo, Unterberg, Posti, Servadei, Teasdale, Menon, Murray, Kirkpatrick, Hutchinson.
Statistical analysis: Kolias, Adams, Hossain, Mitchell, Murray, Hutchinson.
Obtained funding: Timofeev, Hossain, Bulters, Belli, Pickard, Menon, Kirkpatrick, Hutchinson.
Administrative, technical, or material support: Adams, Timofeev, Corteen, Hossain, Mitchell, Critchley, Sahuquillo, Hutchinson.
Supervision: Kolias, Czosnyka, Timothy, Mendelow, Mitchell, Critchley, Sahuquillo, Unterberg, Servadei, Kirkpatrick, Hutchinson.
Conflict of Interest Disclosures: Dr Kolias reported receiving grants from the Medical Research Council during the conduct of the study. Dr Hossain reported receiving the Integra EANS grant from the European Association of Neurosurgical Societies and a research grant from the Finnish Medical Foundation during the conduct of the study. Dr Belli reported receiving grants from the UK National Institute for Health and Care Research (NIHR) during the conduct of the study. Dr Mendelow reported serving as the director of Newcastle Neurosurgery Foundation Ltd outside the submitted work. Dr Posti reported receiving grants from the Academy of Finland and the Maire Taponen Foundation during the conduct of the study. Dr Pickard reported receiving grants from the University of Cambridge during the conduct of the study. Dr Menon reported receiving grants from NIHR during the conduct of the study as well as grants from PressuraNeuro and GlaxoSmithKline and personal fees from Lantmannen AB, NeuroTrauma Sciences, Calico, and Integra Neurosciences (for educational activities) outside the submitted work. No other disclosures were reported.
Funding/Support: This work was supported by grant 09/800/16 from the Medical Research Council (MRC) and managed by the UK National Institute for Health and Care Research (NIHR) on behalf of the MRC-NIHR partnership. This work was also supported by the NIHR Cambridge Biomedical Research Centre, the Academy of Medical Sciences and Health Foundation (Senior Fellowship to Dr Hutchinson), and the Evelyn Trust.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study, collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Group Information: The RESCUEicp Trial Collaborators are listed in Supplement 3.
Data Sharing Statement: See Supplement 4.
Additional Contributions: We thank Max Pitches for administrative support throughout the trial. There was no financial compensation for this contribution.