What email address or phone number would you like to use to sign in to Docs.com?
If you already have an account that you use with Office or other Microsoft services, enter it here.
Or sign in with:
Signing in allows you to download and like content, and it provides the authors analytical data about your interactions with their content.
Embed code for: Hypothermia for Intracranial Hypertension after Traumatic 2015
Select a size
The new england journal of medicine n engl j med 373;25 nejm.org December 17, 2015 2403 established in 1812 December 17, 2015 vol. 373 no. 25 From the Centre for Clinical Brain Scienc- es (P.J.D.A.), Department of Anaesthesia, Critical Care, and Pain Medicine (H.L.S., B.A.H., C.G.B., J.K.J.R.), and Centre for Population Health Sciences (A.R., G.D.M.), University of Edinburgh, and Critical Care, Western General Hospital, NHS Lothian (B.A.H., J.K.J.R.) — all in Edinburgh. Ad- dress reprint requests to Dr. Andrews at Ward 20, Intensive Care Unit, Western General Hospital, Crewe Rd., Edinburgh EH4 2XU, United Kingdom, or at p . andrews@ ed . ac . uk. * A complete list of investigators in the European Study of Therapeutic Hypo- thermia (32–35°C) for Intracranial Pres- sure Reduction after Traumatic Brain Injury (the Eurotherm3235 Trial) is pro- vided in the Supplementary Appendix, available at NEJM.org. This article was published on October 7, 2015, at NEJM.org. N Engl J Med 2015;373:2403-12. DOI: 10.1056/NEJMoa1507581 Copyright © 2015 Massachusetts Medical Society. BACKGROUND In patients with traumatic brain injury, hypothermia can reduce intracranial hypertension. The benefit of hypothermia on functional outcome is unclear. METHODS We randomly assigned adults with an intracranial pressure of more than 20 mm Hg despite stage 1 treatments (including mechanical ventilation and sedation management) to standard care (control group) or hypothermia (32 to 35°C) plus standard care. In the control group, stage 2 treatments (e.g., osmotherapy) were added as needed to control intracranial pressure. In the hypothermia group, stage 2 treatments were added only if hypothermia failed to control intracranial pressure. In both groups, stage 3 treatments (barbiturates and decompressive craniectomy) were used if all stage 2 treatments failed to control intracranial pressure. The primary outcome was the score on the Extended Glasgow Outcome Scale (GOS-E; range, 1 to 8, with lower scores indicating a worse functional outcome) at 6 months. The treatment effect was estimated with ordinal lo- gistic regression adjusted for prespecified prognostic factors and expressed as a com- mon odds ratio (with an odds ratio <1.0 favoring hypothermia). RESULTS We enrolled 387 patients at 47 centers in 18 countries from November 2009 through October 2014, at which time recruitment was suspended owing to safety concerns. Stage 3 treatments were required to control intracranial pressure in 54% of the patients in the control group and in 44% of the patients in the hypothermia group. The adjusted com- mon odds ratio for the GOS-E score was 1.53 (95% confidence interval, 1.02 to 2.30; P = 0.04), indicating a worse outcome in the hypothermia group than in the control group. A favorable outcome (GOS-E score of 5 to 8, indicating moderate disability or good recovery) occurred in 26% of the patients in the hypothermia group and in 37% of the patients in the control group (P = 0.03). CONCLUSIONS In patients with an intracranial pressure of more than 20 mm Hg after traumatic brain injury, therapeutic hypothermia plus standard care to reduce intracranial pressure did not result in outcomes better than those with standard care alone. (Funded by the Na- tional Institute for Health Research Health Technology Assessment program; Current Controlled Trials number, ISRCTN34555414.) abstract Hypothermia for Intracranial Hypertension after Traumatic Brain Injury Peter J.D. Andrews, M.D., M.B., Ch.B., H. Louise Sinclair, R.G.N., M.Sc., Aryelly Rodriguez, M.Sc., Bridget A. Harris, R.G.N., Ph.D., Claire G. Battison, R.G.N., B.A., Jonathan K.J. Rhodes, Ph.D., M.B., Ch.B., and Gordon D. Murray, Ph.D., for the Eurotherm3235 Trial Collaborators* The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 20152404 The new england journal of medicine I n Europe, traumatic brain injury is the most common cause of permanent dis- ability in people younger than 40 years of age, with the annual cost exceeding €33 billion (approximately $37.5 billion in U.S. dollars).1,2 Recent statistics show a 21% increase in the in- cidence of traumatic brain injury during the past 5 years — three times greater than the increase in population. Despite this, management of trau- matic brain injury has been underrepresented in medical research as compared with other health problems.3 Consequently, there are few data to support the commonly used stage 2 interven- tions (Fig. 1) for the management of traumatic brain injury,4-6 with even the use of intracranial- pressure monitoring being debated.7 Hypothermia is one treatment option for this patient group.8-12 Some previous trials of early induction of prophylactic hypothermia have shown benefit, but the trials of hypothermia for Figure 1. Stages of Therapeutic Management and Trial Follow-up. CSF denotes cerebrospinal fluid, EEG electroencephalographic, GOS-E Extended Glasgow Outcome Scale, ICU in- tensive care unit, and MOHS modified Oxford Handicap Scale. Intracranial pressure >20 mm Hg within 10 days after injury Stage 3 Options (if required) Day 28, Hospital Discharge, or Death MOHS grade, length of stay in ICU and hospital 6-Mo Follow-up GOS-E score Stage 1 Stage 2 Stage 3 Trial Follow-up Traumatic Brain Injury Stage 1 treatment: Admission to ICU Mechanical ventilation Sedation Analgesia with or without paralysis Head of bed elevated to 30 degrees Intravenous fluids with or without inotropes to maintain mean arterial pressure ≥80 mm Hg Stage 1 options: Ventriculostomy with or without CSF drainage Surgical removal of space-occupying lesions Control Group Continue stage 1 treatments and add stage 2 treatments without therapeutic hypothermia Stage 2 treatment: Mannitol (maintain serum osmolarity <315 mOsm per kilogram of water) Hypertonic saline (avoid in hyponatremia, caution with cardiac or pulmonary problems) Inotropes to maintain cerebral perfusion pressure ≥60 mm Hg Barbiturates not permitted Hypothermia Group Continue stage 1 treatments and initiate hypothermia Add stage 2 treatments only if needed Barbiturates not permitted Continued medical care Barbiturate therapy with processed EEG monitoring Decompressive craniectomy Further surgical intervention if required The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 2015 2405 Hypothermia for Intracranial Hypertension neuroprotection that were judged to be higher in quality and to have a lower risk of bias (on the basis of assessment of randomization proce- dures, blinding, outcome assessment, and com- pleteness of the data)11 have shown trends to- ward unfavorable outcomes13,14 or were stopped for futility.15,16 Although hypothermia is routinely used to treat elevated intracranial pressure in patients with traumatic brain injury in some intensive care units (ICUs), its effect on outcome in this context has limited evaluation.17 We con- ducted a trial of therapeutic hypothermia for elevated intracranial pressure in which we tested hypothermia in the way that many clinicians currently use it.18-21 Methods Trial Design and Oversight The European Study of Therapeutic Hypothermia (32–35°C) for Intracranial Pressure Reduction after Traumatic Brain Injury (the Eurotherm3235 Trial) aimed to recruit 600 patients who had a trau- matic brain injury. The first patient was enrolled in November 2009, and the trial was stopped early in October 2014 for participant safety. The trial protocol was developed by the first, second, fourth, and last authors in consultation with an international advisory board. The trial was conducted and reported with fidelity to the study protocol. Full details of the trial protocol have been published previously,22 and the proto- col is available with the full text of this article at NEJM.org. After the pilot phase of the trial, the inclusion criteria and power calculation were refined as described below. The authors vouch for the accuracy and completeness of the data and analyses. Data were gathered by investiga- tors at the trial sites (see the Supplementary Appendix, available at NEJM.org). Ethical approval was obtained from the Scot- land A Research Ethics Committee, the Bradford Research Ethics Committee, and ethics commit- tees in another 14 countries. Owing to the inca- pacitated state of the potential participants, it was not possible to obtain consent directly from them. Written informed consent was therefore sought from each eligible patient’s nearest rela- tive or person designated to give consent. Early consent was obtained when possible to prevent a delay between a rise in intracranial pressure and potential randomization. An independent steering committee and inde- pendent data and safety monitoring committee reviewed the trial regularly, assessing conduct, progress, and safety (see the Supplementary Ap- pendix). Trial recruitment was stopped on the advice of the data and safety monitoring com- mittee after its ninth meeting (Table S11 in the Supplementary Appendix). Participants All patients admitted to the ICU after a traumatic brain injury who had intracranial-pressure mon- itoring in place were screened. Eligible patients were believed to be of legal age for consent. Other inclusion criteria were a primary, closed traumatic brain injury; an intracranial pressure of more than 20 mm Hg for at least 5 minutes after stage 1 treatments (Fig. 1), with no obvious reversible cause; an initial head injury that had occurred no more than 10 days earlier; the avail- ability of a cooling device or technique for more than 48 hours; a core temperature of at least 36°C (at the time of randomization); and an ab- normal computed tomographic scan of the brain. Patients who were already receiving therapeutic hypothermia or who were unlikely to survive for the next 24 hours were excluded. Other exclusion criteria were the administration of barbiturate infusion before randomization, a temperature of 34°C or less at hospital admission, and preg- nancy. The inclusion criteria were changed in Janu- ary 2012, on the basis of the pilot-phase find- ings,23 to remove an upper age limit (previously 65 years) and to increase the time from injury from 72 hours to 10 days. These changes allowed the enrollment of older patients and those with evolving brain swelling. ICUs in hospitals that provide specialist neu- rologic treatment for traumatic brain injury were recruited (25 centers in the United Kingdom and 39 elsewhere). Evidence of expertise with intra- cranial-pressure monitoring and therapeutic cooling were necessary. Data Collection An online case-report form (Lincoln, Paris) was used for collection of data (Fig. S8 in the Supple- mentary Appendix), including baseline demo- graphic information and data on completion of stage 1 interventions; intracranial pressure and temperature at randomization; intracranial pres- The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 20152406 The new england journal of medicine sure, mean arterial pressure, cerebral perfusion pressure, and temperature measured hourly on days 1 through 7; failure of stage 2 therapy to control intracranial pressure; new pneumonia; and functional outcome. This trial was pragmat- ic, with a focus on patient-oriented outcomes; therefore, we did not collect data on which stage 2 therapies were delivered to patients.5 Randomization and Study Treatment Participants were randomly assigned to standard care (control group) or therapeutic hypothermia plus standard care (intervention group). Random- ization was performed with the use of a minimi- zation procedure to balance assignments accord- ing to center, age, Glasgow Coma Scale (GCS) motor score, time from injury, and pupillary re- sponse. The online case-report form ensured minimization (with a random element) and concealment of allocation to study groups. The trial had an open-label design, with patients, families, and treating clinicians aware of the study-group assignments. Scoring of the primary outcome measure (described below) was blinded. According to the study protocol, hypothermia was induced by a bolus of intravenous, refriger- ated 0.9% sodium chloride (20 to 30 ml per kilo- gram of body weight) and thereafter maintained with the usual cooling technique of each site. Guidelines were provided for induction and maintenance of hypothermia, rewarming, and detection and treatment of shivering in the inter- vention group (Fig. S1, S2, and S3 in the Supple- mentary Appendix). Core temperature in the hypothermia group was reduced by the minimum required to main- tain an intracranial pressure of 20 mm Hg or less (in keeping with guidelines of the Brain Trauma Foundation24), within the limits of 32 to 35°C. Stage 2 treatments were added if hypothermia failed to control intracranial pressure. Stage 3 treatments were used for patients whose intra- cranial pressure was not controlled by hypother- mia and all other stage 2 treatments. Hypothermia was maintained for at least 48 hours in the intervention group and continued for as long as necessary to control intracranial pressure. Rewarming was considered after 48 hours at a rate of 0.25°C per hour, provided that intracranial pressure was 20 mm Hg or less. The control group also received stage 2 and 3 treat- ments but without hypothermia (Fig. 1). Outcomes The primary outcome measure was the score on the Extended Glasgow Outcome Scale (GOS-E) at 6 months after injury.25,26 The eight-point scale assesses the effects of traumatic brain injury on function in major areas of life. A GOS-E score of 1 indicates death, 2 indicates a vegetative state, 3 or 4 indicates severe disability, 5 or 6 indicates moderate disability, and 7 or 8 indicates good recovery (Table S2 in the Supplementary Appen- dix). The GOS-E questionnaire (Fig. S4 in the Supplementary Appendix) was sent by mail to surviving participants from the trial office in Edinburgh. When this was not possible, a local staff member contacted the patient by telephone to complete the questionnaire. An investigator who was unaware of the study-group assign- ments scored all outcomes according to the stan- dardized approach (Fig. S4 in the Supplementary Appendix). The manually calculated scores were automatically checked in the trial database with the use of a specially developed algorithm. An independent expert was consulted in the few cases in which adjudication was needed. Secondary outcomes were 6-month mortality, lack of intracranial-pressure control (failure of all stage 2 therapies to control intracranial pressure to ≤20 mm Hg), incidence of pneumonia during days 1 through 7 after randomization, length of ICU stay, and grade on the modified Oxford Handicap Scale (MOHS; a score of 0 indicates no symptoms, 1 minor symptoms, 2 some re- striction, 3 dependent, 4 fully dependent, and 5 death) (Table S3 in the Supplementary Appen- dix)27 at 28 days or discharge from an acute-care hospital (whichever came first). Data were collected on serious adverse events, including bleeding, cardiovascular instability, thermal burns, and a cerebral perfusion pressure of less than 50 mm Hg. Data on other adverse events were not collected, because many un- toward events are expected in patients with trau- matic brain injury who are admitted to the ICU. Statistical Analysis As a result of the internal pilot phase, the sam- ple size for the full trial was reduced from 1800 to 600 patients.23 Two factors contributed to this decision: our original sample size may have under- estimated the possible benefit of hypothermia because, unlike participants in most previous trials, participants in the Eurotherm3235 Trial The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 2015 2407 Hypothermia for Intracranial Hypertension had evidence of brain swelling (raised intracra- nial pressure); and we showed that an enhanced cooling intervention could be delivered, as de- scribed by Peterson et al.28 These data therefore informed the revised power calculation. Using an ordinal analysis of the GOS-E scores together with covariate adjustment (primary effi- cacy analysis), we were able to increase the statis- tical efficiency of the analysis,29,30 so that a trial involving 600 patients would have power equiva- lent to that of a trial involving 1000 patients that assessed a binary outcome. We calculated that with such an analysis, the study would have the equivalent of 80% power to detect a rate of un- favorable outcome (GOS-E score of 1 to 4) that was 9 percentage points lower with hypothermia than with standard care (51% vs. 60%), at the 5% significance level (two-sided). All analyses were performed with SAS soft- ware, version 9.3 (SAS Institute). Analyses were performed on an intention-to-treat basis, incor- porating all patients who underwent randomiza- tion and for whom outcome data were available, with patients evaluated according to their as- signed intervention. For the primary analysis, the distribution of the 6-month GOS-E scores between the two groups (hypothermia vs. control) was compared with the use of ordinal logistic regression30 and with adjustment for the following baseline co- variates: age (included as a continuous variable, with the use of a linear term in the regression model), postresuscitation GCS motor score (1 or 2 [no or extensor response] vs. 3 to 6 [flexion or better response]) (Table S1 in the Supplementary Appendix), time from injury (<12 hours vs. ≥12 hours), and pupillary response (both reacting vs. one reacting vs. neither reacting; included as an unordered categorical variable in the regression model). For this analysis, we collapsed the eight-point GOS-E to six categories by pooling death with a vegetative state and lower severe disability. This ensured that the analysis would not favor an intervention that reduced mortality at the expense of increasing the proportion of severely disabled survivors. Prespecified subgroups for the primary analy- sis were defined on the basis of the baseline co- variates described above, the location of the center (United Kingdom vs. elsewhere), and the volume of the center (≥10 vs. <10 patients). We performed these analyses by including an interaction term between intervention and the relevant covariate in the ordinal logistic-regres- sion model; a stricter level of statistical signifi- cance (P<0.01) was used owing to their explor- atory nature. MOHS grades were analyzed in the same way as GOS-E scores, but we collapsed the six grades to four categories by grouping dependent, fully dependent, and death (Table S9B in the Supple- mentary Appendix). In the analysis of the between- group difference in mortality, Cox proportional- hazards regression was used to estimate the intervention effect. Other continuous outcomes were tested with an analysis of covariance; for binary outcomes, logistic regression was used. Intracranial pres- sure, core temperature, mean arterial pressure, and cerebral perfusion pressure on days 1 through 7 were analyzed post hoc with the use of a linear model, with study days as repeated measure- ments with a compound-symmetry covariance matrix. All these analyses used the same covari- ates as were prespecified for GOS-E scores to- gether with the baseline value of the relevant variable. Results Recruitment A total of 2498 patients at 55 centers in 18 coun- tries were assessed for trial eligibility, and 387 patients at 47 centers in 18 countries underwent randomization, of whom 205 (53.0%) were re- cruited in the United Kingdom (Table S4 in the Supplementary Appendix). Patients underwent randomization between November 2009 (pilot phase to September 15, 2011) and October 2014, at which time recruitment was stopped (Fig. S6 in the Supplementary Appendix). The most com- mon reasons for exclusion from the trial were an intracranial pressure of 20 mm Hg or less (41% of 2111 exclusions), the unlikelihood of survival (8%), and current receipt of therapeutic hypo- thermia (6%). Recruitment was stopped after the steering committee concluded that there were signs of harm with the treatment being evaluated and that a result of futility, at best, would be ex- pected if the trial were to continue. These find- ings became apparent when the committee ex- amined the designated primary outcome measure analyzed according to the prespecified statistical The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 20152408 The new england journal of medicine analysis plan (Table S11 in the Supplementary Appendix). On an intention-to-treat basis, 195 participants were randomly assigned to the hypothermia group and 192 to the control group. Of the 387 randomly assigned patients, 386 received the intended treatment (1 patient in the hypother- mia group was withdrawn before receiving the intervention), and 376 (188 in each group) were evaluated for the primary outcome (Fig. S6 in the Supplementary Appendix). Baseline charac- teristics of the two groups were similar in all respects (Table 1, and Table S5 in the Supple- mentary Appendix). There were no significant differences between groups with respect to pre- randomization neurosurgery for single or multi- ple procedures (Table S6 in the Supplementary Appendix). Intracranial Pressure and Core Temperature Mean daily intracranial pressure was similar in the two groups (Fig. 2 and Table 2). Core tem- perature was substantially lower in the hypo- thermia group than in the control group during the first 4 days after randomization. During that time period, there were fewer first occurrences of failure of stage 2 therapy to control intracra- nial pressure in the hypothermia group than in the control group (57 vs. 84) (Table S7 in the Supplementary Appendix). This resulted in more frequent use of stage 3 treatments on days 1 through 7 in the control group than in the hypothermia group (102 of 189 patients [54.0%] vs. 84 of 192 patients [43.8%]). Barbiturate-infu- sion therapy was used more often in the control group than in the hypothermia group (41 patients vs. 20 patients) during days 1 through 4 after randomization, but decompressive craniectomy was not used more often (27 patients in each group) (Table S8 in the Supplementary Appendix). A repeated-measures analysis was performed to compare the difference between the groups with respect to the change from day 1 to day 7 after randomization in core temperature, intra- cranial pressure, mean arterial pressure, and cerebral perfusion pressure. There was a signifi- cant difference with respect to core temperature only (Table 2 and Fig. 2). Primary Outcome Six months after injury, the distribution of GOS-E scores was shifted in an unfavorable direction in the hypothermia group (adjusted common odds ratio, 1.53; 95% confidence interval [CI], 1.02 to 2.30; P = 0.04) (Table 2, and Table S9A in the Supplementary Appendix). Favorable outcomes (GOS-E score of 5 to 8, indicating moderate dis- ability or good recovery) occurred in 49 of 191 patients (25.7%) in the hypothermia group and in 69 of 189 patients (36.5%) in the control group (P = 0.03). The results without adjustment for prespecified covariates were similar for the GOS-E score (unadjusted common odds ratio, 1.55; 95% CI, 1.05 to 2.29; P = 0.03). The risk of death (hazard ratio, 1.45; 95% CI, 1.01 to 2.10; P = 0.047) favored the control group (Table 2, and Table S9A and Fig. S7 in the Supplementary Appendix). Subgroup analysis showed no significant in- teraction effect between the intervention and prespecified subgroups (Fig. S5 in the Supple- mentary Appendix). The rate of adherence, de- fined as more than 80% of core temperature measurements within range for days 1 through 4, was 64.8% in the hypothermia group (32 to 35°C) and 68.8% in the control group (≥36°C). Serious adverse events were reported more often in the hypothermia group than in the control group (33 events vs. 10 events) (Table S10 in the Sup- plementary Appendix). Discussion In this trial involving patients with traumatic brain injury and an intracranial pressure of more than 20 mm Hg for at least 5 minutes despite stage 1 therapy, hypothermia plus standard care did not result in outcomes better than those with standard care alone. The trial was stopped early owing to safety concerns, which introduces the risk of bias, but the results suggest that out- comes were worse with hypothermia than with standard care alone. The Eurotherm3235 Trial was a large random- ized, controlled trial that tested therapeutic hy- pothermia as the primary (stage 2) intervention to reduce intracranial pressure after brain trauma. Literature at the time of protocol development showed that at least one episode of intracranial pressure of more than 20 mm Hg occurred in 50% of patients with traumatic brain injury who received mechanical ventilation and intercranial- pressure monitoring.31 In contrast, data collected during the screening of patients for this trial The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 2015 2409 Hypothermia for Intracranial Hypertension indicated that fewer patients than expected had a rise in intracranial pressure. Standard care followed best practice (Brain Trauma Foundation guidelines) but was not pre- scribed in the protocol. There were guidelines for hypothermia maintenance and control of shivering, but only induction of hypothermia, rather than a specific maintenance technique, was prescribed in the protocol. Centers used whichever cooling technique they would normally Characteristic Hypothermia (N = 195) Control (N = 192) Age <45 yr — no. (%) 131 (67.2) 131 (68.2) Age — yr 37.4±15.4 36.7±14.9 GCS motor score — no. (%)† 1 or 2 56 (28.7) 51 (26.6) 3–6 139 (71.3) 141 (73.4) Pupillary response — no. (%) Both reacting 144 (73.8) 143 (74.5) One or neither reacting 51 (26.2) 49 (25.5) Time from injury — no. (%) <12 hr 19 (9.7) 15 (7.8) ≥12 hr 176 (90.3) 177 (92.2) Intracranial pressure at randomization — mm Hg 25.2±4.8 25.5±6.4 Core temperature at randomization — °C 37.0±0.72 37.1±0.72 Isolated TBI — no. (%) 123 (63.1) 133 (69.3) Marshall classification — no. (%)‡ Diffuse axonal injury I–III 72 (36.9) 78 (40.6) Diffuse axonal injury IV 21 (10.8) 15 (7.8) Any lesion surgically removed 46 (23.6) 52 (27.1) High-density or mixed-density lesion 56 (28.7) 47 (24.5) Mechanism of injury — no. (%)§ Road-traffic accident, pedestrian 22 (11.3) 31 (16.1) Road-traffic accident, motor vehicle 68 (35.1) 51 (26.6) Bicycling accident 7 (3.6) 10 (5.2) Fall 78 (40.2) 78 (40.6) Sports injury 1 (0.5) 1 (0.5) Assault 18 (9.3) 21 (10.9) * Plus–minus values are means ±SD. There were no significant differences between groups for these baseline measures. Other baseline characteristics are presented in Tables S5 and S6 in the Supplementary Appendix. TBI denotes traumatic brain injury. † The Glasgow Coma Scale (GCS) motor score was measured on hospital admission. A score of 1 indicates that the pa- tient makes no movements, a score of 2 indicates extension to painful stimuli, a score of 3 indicates abnormal flexion, a score of 4 indicates normal flexion, a score of 5 indicates that the patient localizes painful stimuli, and a score of 6 in- dicates that the patient obeys commands. ‡ The Marshall classification of traumatic brain injury is based on a review of computed tomographic scans, which were ob- tained at the screening visit. A diffuse injury indicates that no high-density or mixed-density lesions of more than 25 mm3 are present. Diffuse injury I indicates no visible intracranial pathologic features, diffuse injury II indicates that cisterns are present with a midline shift of 0 to 5 mm or that lesion densities are present, diffuse injury III indicates that cisterns are compressed or absent with a midline shift of 0 to 5 mm, and diffuse injury IV indicates a midline shift of more than 5 mm. § Data were missing for one patient in the hypothermia group. Table 1. Baseline Characteristics of the Study Patients.* The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 20152410 The new england journal of medicine use. Therefore, the results are not due to any one cooling method or to any treatment prescribed as part of the trial protocol. We believe this en- hances the validity and generalizability of the trial and its results, because the intervention studied is already used in clinical practice and was tested in the way that centers currently use it. In this trial, barbiturate infusion was reserved for patients who had uncontrolled intracranial pressure despite all stage 1 and stage 2 treat- ments; barbiturate infusion to reduce intracra- nial pressure32 was used more frequently and earlier in the control group than in the hypo- thermia group (Table S8 in the Supplementary Appendix). It is plausible that barbiturate infu- sion may have been beneficial, but that hypoth- esis requires further testing. There was no dif- ference in the use of decompressive craniectomy33 between the two groups. We found no significant between-group dif- ference according to the time from injury to initiation of hypothermia (<12 or ≥12 hours), a finding that is contrary to that of a previous review.34 However, there were too few patients who underwent randomization less than 12 hours after injury to be confident of having excluded a subgroup effect for the time from injury. The trials of hypothermia for neuroprotection that were judged to be of higher quality and to have a lower risk of bias11 have shown trends toward unfavorable outcomes13,14 or were stopped for futility.15,16 The trial sponsor and steering committee accepted the recommendation of the data and safety monitoring committee in full and termi- nated recruitment early. Early stopping of any trial can potentially reduce the external validity of the results; however, the burden of proof required for early stopping for possible harm is considerably lower than that for overwhelming evidence of efficacy.35 In this case, the remain- ing GOS-E scores collected after the “stopping” Figure 2. Physiological Measurements. Shown are estimated means and 95% confidence intervals (I bars), calculated with the use of a repeated-measures linear analysis. Intracranial Pressure (mm Hg) 35 25 30 20 15 5 10 0 0 2 1 3 4 5 6 7 8 Study Day A Intracranial Pressure Hypothermia Control Core Temperature (°C) 40 38 36 34 30 32 39 37 35 31 33 0 0 2 1 3 4 5 6 7 8 Study Day B Core Temperature Hypothermia Control Mean Arterial Pressure (mm Hg) 105 95 100 90 85 75 70 80 0 0 2 1 3 4 5 6 7 8 Study Day C Mean Arterial Pressure Hypothermia Control Cerebral Perfusion Pressure (mm Hg) 85 80 70 75 65 60 50 55 0 0 2 1 3 4 5 6 7 8 Study Day D Cerebral Perfusion Pressure Hypothermia Control The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 2015 2411 Hypothermia for Intracranial Hypertension decision confirmed the result and did not show regression to the mean and a resultant lack of evidence (Table S11 in the Supplementary Ap- pendix). A limitation of the study is the lack of blind- ing to the intervention, which is problematic in all trials of therapeutic hypothermia. However, because cooling to normothermia was permitted in the standard-care group, it is possible that there was masking of the intervention to pa- tients and relatives in some cases. Outcome scoring was blinded. Lack of blinding was why, in our opinion, more serious adverse events were reported in the hypothermia group. In the con- trol group, the same clinical events may have been considered expected and may not have been reported as serious adverse events. The trial was designed to be pragmatic, fo- cused on functional outcome rather than on detailed mechanistic pathways. The intensity of a stage 2 therapy is adjusted according to the effect on intracranial pressure, mean arterial pressure, and cerebral perfusion pressure. There were no clinically important differences in these variables between the two groups. Given that there were no or limited data on the benefits and harms of standard stage 2 interventions, we elected not to record which stage 2 therapies were delivered to patients. The findings suggest- ing possible harm of hypothermia could be due to a biologic effect of hypothermia or due to the harms or benefits of the other therapies used differentially in the two groups. This trial did not assess the benefits and risks of hypothermia used in patients with traumatic brain injury who have severe intracranial hypertension that is re- fractory to all stage 2 treatments before initia- tion of hypothermia. The benefits and harms of other interven- tions that successfully reduce intracranial pres- sure have not been assessed. More adequately powered clinical trials of hypertonic therapy, barbiturates, and hyperventilation are required.5 In patients with traumatic brain injury, thera- peutic hypothermia plus standard care success- fully reduced intracranial pressure. This inter- vention, however, did not improve functional recovery as compared with standard care alone. Variable Estimate (95% CI) P Value Physiological measurements† Adjusted mean difference in ICP on days 1–7 — mm Hg −0.48 (−2.04 to 1.08) 0.55 Adjusted mean difference in core temperature on days 1–7 — °C −2.14 (−2.34 to −1.94) <0.001 Adjusted mean difference in mean arterial pressure on days 1–7 — mm Hg 1.20 (−0.46 to 2.86) 0.16 Adjusted mean difference in cerebral perfusion pressure on days 1–7 — mm Hg 1.61 (−0.36 to 3.58) 0.11 Primary analysis: adjusted common odds ratio for GOS-E score at 6 mo‡§ 1.53 (1.02 to 2.30)¶ 0.04 Adjusted odds ratio for unfavorable outcome‡‖ 1.69 (1.06 to 2.70)¶ 0.03 Unadjusted hazard ratio for death at 6 mo 1.45 (1.01 to 2.10) 0.047 Adjusted mean difference in squared proportion of ICP measurements of ≤20 mm Hg on days 1–7‡ 440 (−160 to 1000) 0.47 Adjusted odds ratio for presence of pneumonia on days 3–7‡ 1.04 (0.69 to 1.58)¶ 0.84 Adjusted mean difference in log-transformed length of ICU stay — log hours‡ 0.05 (0.11 to 0.22) 0.54 Adjusted common odds ratio for MOHS grade at 28 days‡** 1.65 (0.91 to 3.02)¶ 0.10 * CI denotes confidence interval, ICP intracranial pressure, and ICU intensive care unit. † Values were calculated with the use of a repeated-measures model adjusted for age, postresuscitation GCS motor score, time from injury, pupillary response, study day, and (when available) baseline value. These are post hoc analyses. ‡ Results were adjusted for age, postresuscitation GCS motor score, time from injury, and pupillary response. § The eight-point Extended Glasgow Outcome Scale (GOS-E) was collapsed to six categories by pooling death (score of 1) with vegetative state (score of 2) and lower severe disability (score of 3) (Table S9A in the Supplementary Appendix). ¶ An odds ratio or common odds ratio of less than 1 corresponds to a benefit for hypothermia over control. ‖ The eight-point GOS-E was collapsed to two categories: favorable outcome (score of 5 to 8) and unfavorable outcome (score of 1 to 4) (Table S9A in the Supplementary Appendix). ** The six grades of the modified Oxford Handicap Scale (MOHS) were collapsed to four categories by pooling dependent, fully dependent, and death (Table S9B in the Supplementary Appendix). Table 2. Analysis of Primary and Secondary Outcomes for Hypothermia versus Control.* The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. n engl j med 373;25 nejm.org December 17, 20152412 Hypothermia for Intracranial Hypertension Supported by the National Institute for Health Research Health Technology Assessment program, which funded the main phase of the study. The European Society of Intensive Care Medicine funded the pilot phase. The trial sponsors, the University of Edin- burgh and NHS Lothian, provided research governance. Dr. Andrews reports receiving lecture fees from C.R. Bard and Integra LifeSciences; and Dr. Rhodes, lecture fees from C.R. Bard. No other potential conflict of interest relevant to this article was reported. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. We thank the patients who participated in this trial and their families, Lindsay Wilson, Ph.D., for support with training and adjudication of Extended Glasgow Outcome Scale scoring, Cor- rie Roberts and Dawn Gray for assistance with trial manage- ment and administration, and Jill Harris and the staff of the Wellcome Trust Clinical Research Facility for research nurse support. References 1. Olesen J, Gustavsson A, Svensson M, et al. The economic cost of brain disorders in Europe. Eur J Neurol 2012; 19: 155-62. 2. Gustavsson A, Svensson M, Jacobi F, et al. Cost of disorders of the brain in Eu- rope 2010. Eur Neuropsychopharmacol 2011; 21: 718-79. 3. The Lancet Neurology. A rally for traumatic brain injury research. Lancet Neurol 2013; 12: 1127. 4. Roberts I, Schierhout G, Alderson P. Absence of evidence for the effectiveness of five interventions routinely used in the intensive care management of severe head injury: a systematic review. J Neurol Neu- rosurg Psychiatry 1998; 65: 729-33. 5. Ker K, Perel P, Blackhall K, Roberts I. How effective are some common treat- ments for traumatic brain injury? BMJ 2008; 337: a865. 6. Le Roux P, Menon DK, Citerio G, et al. Consensus summary statement of the In- ternational Multidisciplinary Consensus Conference on Multimodality Monitoring in Neurocritical Care: a statement for healthcare professionals from the Neuro- critical Care Society and the European Society of Intensive Care Medicine. Inten- sive Care Med 2014; 40: 1189-209. 7. Chesnut RM, Temkin N, Carney N, et al. A trial of intracranial-pressure mon- itoring in traumatic brain injury. N Engl J Med 2012; 367: 2471-81. 8. McIntyre LA, Fergusson DA, Hébert PC, Moher D, Hutchison JS. Prolonged therapeutic hypothermia after traumatic brain injury in adults: a systematic review. JAMA 2003; 289: 2992-9. 9. Sydenham E, Roberts I, Alderson P. Hypothermia for traumatic head injury. Cochrane Database Syst Rev 2009; 2: CD001048. 10. Henderson WR, Dhingra VK, Chit- tock DR, Fenwick JC, Ronco JJ. Hypother- mia in the management of traumatic brain injury: a systematic review and meta- analysis. Intensive Care Med 2003; 29: 1637-44. 11. Crossley S, Reid J, McLatchie R, et al. A systematic review of therapeutic hypo- thermia for adult patients following trau- matic brain injury. Crit Care 2014; 18: R75. 12. Harris OA, Colford JM Jr, Good MC, Matz PG. The role of hypothermia in the management of severe brain injury: a meta- analysis. Arch Neurol 2002; 59: 1077-83. 13. Hutchison JS, Ward RE, Lacroix J, et al. Hypothermia therapy after traumatic brain injury in children. N Engl J Med 2008; 358: 2447-56. 14. Clifton GL, Valadka A, Zygun D, et al. Very early hypothermia induction in pa- tients with severe brain injury (the National Acute Brain Injury Study: Hypothermia II): a randomised trial. Lancet Neurol 2011; 10: 131-9. 15. Clifton GL, Miller ER, Choi SC, et al. Lack of effect of induction of hypother- mia after acute brain injury. N Engl J Med 2001; 344: 556-63. 16. Maekawa T, Yamashita S, Nagao S, Hayashi N, Ohashi Y. Prolonged mild ther- apeutic hypothermia versus fever control with tight hemodynamic monitoring and slow rewarming in patients with severe traumatic brain injury: a randomized con- trolled trial. J Neurotrauma 2015; 32: 422-9. 17. Shiozaki T, Sugimoto H, Taneda M, et al. Effect of mild hypothermia on uncon- trollable intracranial hypertension after severe head injury. J Neurosurg 1993; 79: 363-8. 18. Marion DW, Obrist WD, Carlier PM, Penrod LE, Darby JM. The use of moderate therapeutic hypothermia for patients with severe head injuries: a preliminary report. J Neurosurg 1993; 79: 354-62. 19. Zhi D, Zhang S, Lin X. Study on thera- peutic mechanism and clinical effect of mild hypothermia in patients with severe head injury. Surg Neurol 2003; 59: 381-5. 20. Jiang JY, Xu W, Li WP, et al. Effect of long-term mild hypothermia or short- term mild hypothermia on outcome of pa- tients with severe traumatic brain injury. J Cereb Blood Flow Metab 2006; 26: 771-6. 21. Tokutomi T, Miyagi T, Takeuchi Y, Karukaya T, Katsuki H, Shigemori M. Effect of 35 degrees C hypothermia on in- tracranial pressure and clinical outcome in patients with severe traumatic brain injury. J Trauma 2009; 66: 166-73. 22. Andrews PJ, Sinclair HL, Battison CG, et al. European Society of Intensive Care Medicine study of therapeutic hypother- mia (32-35°C) for intracranial pressure reduction after traumatic brain injury (the Eurotherm3235Trial). Trials 2011; 12: 8. 23. Andrews PJ, Sinclair LH, Harris B, et al. Study of therapeutic hypothermia (32 to 35°C) for intracranial pressure reduction after traumatic brain injury (the Eurotherm 3235Trial): outcome of the pilot phase of the trial. Trials 2013; 14: 277. 24. Bratton SL, Chestnut RM, Ghajar J, et al. Guidelines for the management of severe traumatic brain injury. VIII. Intra- cranial pressure thresholds. J Neurotrauma 2007; 24: Suppl 1: S55-S58. 25. Jennett B, Snoek J, Bond MR, Brooks N. Disability after severe head injury: obser- vations on the use of the Glasgow Out- come Scale. J Neurol Neurosurg Psychia- try 1981; 44: 285-93. 26. Wilson JT, Edwards P, Fiddes H, Stew- art E, Teasdale GM. Reliability of postal questionnaires for the Glasgow Outcome Scale. J Neurotrauma 2002; 19: 999-1005. 27. Bamford JM, Sandercock PA, Warlow CP, Slattery J. Interobserver agreement for the assessment of handicap in stroke pa- tients. Stroke 1989; 20: 828. 28. Peterson K, Carson S, Carney N. Hypo- thermia treatment for traumatic brain in- jury: a systematic review and meta-analy- sis. J Neurotrauma 2008; 25: 62-71. 29. Murray GD, Barer D, Choi S, et al. Design and analysis of phase III trials with ordered outcome scales: the concept of the sliding dichotomy. J Neurotrauma 2005; 22: 511-7. 30. Roozenbeek B, Lingsma HF, Perel P, et al. The added value of ordinal analysis in clinical trials: an example in traumatic brain injury. Crit Care 2011; 15: R127. 31. Stocchetti N, Colombo A, Ortolano F, et al. Time course of intracranial hyper- tension after traumatic brain injury. J Neurotrauma 2007; 24: 1339-46. 32. Nordby HK, Nesbakken R. The effect of high dose barbiturate decompression after severe head injury: a controlled clin- ical trial. Acta Neurochir (Wien) 1984; 72: 157-66. 33. Timofeev I, Czosnyka M, Nortje J, et al. Effect of decompressive craniectomy on intracranial pressure and cerebrospinal compensation following traumatic brain injury. J Neurosurg 2008; 108: 66-73. 34. Polderman KH. Induced hypothermia and fever control for prevention and treat- ment of neurological injuries. Lancet 2008; 371: 1955-69. 35. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clini- cal trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 1977; 35: 1-39. Copyright © 2015 Massachusetts Medical Society. The New England Journal of Medicine Downloaded from nejm.org by EDER ARAIZA on December 14, 2016. For personal use only. No other uses without permission. Copyright © 2015 Massachusetts Medical Society. All rights reserved. Maekawa T, Yamashita S, Nagao S, Hayashi N, Ohashi Y. Prolonged mild ther- apeutic hypothermia versus fever control with