Surgical decompression for cerebral oedema in acute ischaemic stroke

Abstract

Background

Large cerebral infarction has a high case fatality. Despite the use of conventional medical treatments such as hyperventilation, mannitol, diuretics, corticosteroids and barbiturates, the outcome of this condition remains poor. Decompressive surgery to relieve intracranial pressure is performed in some cases, although evidence of any clinical benefits has not been available until recently. This is an update of a Cochrane review first published in 2002.

Objectives

To examine the effects of decompressive surgery in patients with massive acute ischaemic stroke complicated with cerebral oedema, and to judge whether decompressive surgery is effective in improving survival or survival free of severe disability.

Search methods

We searched the Cochrane Stroke Group's Trials Register (last searched October 2010), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2010, Issue 7), MEDLINE (1966 to October 2010), EMBASE (1980 to October 2010) and Science Citation Index (October 2010). We also searched the reference lists of all relevant articles.

Selection criteria

Randomised controlled studies of decompressive surgery plus medical treatment versus medical treatment alone in patients with clinically and radiologically confirmed cerebral infarcts complicated with cerebral oedema.

Data collection and analysis

One author assessed the titles and retrieved the relevant studies. The same author extracted data, with discussion among all authors for clarification. Outcomes were death at the end of follow‐up, death or disability defined as the modified Rankin Scale (mRS) > 3 at the end of follow‐up, death or severe disability defined as mRS > 4 at 12 months and disability defined as mRS 4 or 5 at 12 months. The results are given using the Peto odds ratio (Peto OR) with 95% confidence intervals (CIs).

Main results

We included three trials in this review, involving 134 patients who were 60 years of age or younger. The time window for the intervention was 30 hours from stroke onset in two studies and 96 hours in one study. All trials were stopped early. Surgical decompression reduced the risk of death at the end of follow‐up (OR 0.19, 95% CI 0.09 to 0.37) and the risk of death or disability defined as mRS > 4 at 12 months (OR 0.26, 95% CI 0.13 to 0.51). Death or disability defined as mRS > 3 at the end of follow‐up was no different between the treatment arms (OR 0.56, 95% CI 0.27 to 1.15).

Authors' conclusions

Surgical decompression lowers the risk of death and death or severe disability defined as mRS > 4 in selected patients 60 years of age or younger with a massive hemispheric infarction and oedema. Optimum criteria for patient selection and for timing of decompressive surgery are yet to be defined. Since survival may be at the expense of substantial disability, surgery should be the treatment of choice only when it can be assumed, based on their preferences, that it is in the best interest of patients. Since all the trials were stopped early, an overestimation of the effect size cannot be excluded.

Plain language summary

Surgical decompression for cerebral oedema in acute ischaemic stroke

About four‐fifths of strokes are due to blockage of an artery in the brain. When the artery is blocked, part of the brain is damaged, this is called a cerebral infarct. If a large artery is blocked the area of brain damage can be large. About 24 to 48 hours after a large infarct the brain can swell, causing a dangerous rise of pressure inside the head. Surgery to remove some of the skull bone over the swollen area of brain reduces the pressure. Results from recent clinical trials showed that surgery reduced the risk of death. However, survivors were left with moderate to severe disability requiring help in their daily life activities. These results only apply to people 60 years of age or younger.

Background

Patients with a large cerebral infarction generally have a poor prognosis. Approximately 40% of patients with total anterior cerebral infarction (TACI) syndrome deteriorate during the first week, and half of them die during the first month (Bamford 1991; Tei 2000). Poor outcome is mostly explained by the volume of cerebral tissue that is damaged. Early deterioration and death is often the result of oedema in the infarcted tissue (Hacke 1996; King 1951; Ng 1970; Ropper 1984; Saito 1987). Oedema causes mass‐effect with distortion, tissue shift and increased intracranial pressure (Frank 1995; Schwab 1996). Such changes lead to cerebral herniation, further brain damage and death.

Conventional medical treatment aims at reducing oedema and intracranial pressure in stroke patients using hyperventilation, mannitol, diuretics, corticosteroids or barbiturates (Manno 1999; Schwab 1997; Wijdicks 2000). However, once brain swelling produces clinical signs and imaging features of mass effect with tissue shift, case‐fatality becomes higher, despite intensive medical treatment (Hacke 1996; Saito 1987).

Surgical decompression seeks to create space to accommodate the increased volume created by the swollen brain (Cristofori 1998; Ivamoto 1974; Jourdan 1993; Kakita 1976; Kalia 1993; Koh 2000; Kondziolka 1988; Kristensen 1998; Manai 2000; Mracek 1978; Oro 2000; Sakai 1998; Sollid 1999; Ueno 1984; Van Leusen 2001; Young 1982). This can be accomplished by opening the cranial vault and dura (Carter 1997; Delashaw 1990; Ivamoto 1974; Lindegaard 1999; Rengachary 1981; Rieke 1995; Schwab 1998; Wijdicks 2000), or by removing non‐viable or non‐essential brain tissue (Fujita 1982; Martins 1993; Mori 1998; Tsuruno 1993). These strategies are described as 'external' or 'internal' decompression respectively. External decompression involves hemicraniectomy with or without duroplasty. Internal decompression involves the removal of brain tissue, either the infarcted region or non‐eloquent regions of the brain. These techniques are sometimes combined. Despite the possible benefit, surgical therapy involves risks which may include secondary cerebral haemorrhage and brain herniation through the craniectomy defect (Wagner 2001).

Objectives

The aim of this review was to examine the effects of decompressive surgery in patients with massive acute ischaemic stroke complicated with cerebral oedema, and to judge whether decompressive surgery is effective in improving survival or survival free of severe disability.

Methods

Criteria for considering studies for this review

Types of studies

We sought randomised controlled studies comparing decompressive surgery and medical treatment with medical treatment alone as control. To avoid confounding factors, comparison groups had to be comparable for at least the major prognostic factors: age, stroke location, time to admission, time to randomisation, time from stroke onset to clinical appearance of herniation and sex.

Types of participants

We sought trials involving patients with massive acute ischaemic stroke and complicating brain oedema, evident on cerebral computed tomography or magnetic resonance imaging.

Types of interventions

We sought trials of any kind of surgical decompression (external or internal). Surgical decompression refers to hemicraniectomy with duroplasty or resection of neural tissue, or both. We sought studies with an unconfounded comparison of surgery versus no surgery. Such studies should involve 'best' medical treatment (mannitol, other diuretics, corticosteroids, hyperventilation, barbiturates, etc) and nursing care in both the surgical and non‐surgical arms. If the protocol for medical therapy and nursing (including monitoring) was different between arms, then the study was biased and poorly controlled. We examined studies for such discrepancies.

Types of outcome measures

Primary outcomes

• Death at the end of follow‐up.

Secondary outcomes

• Death or disability defined as a modified Rankin Scale (mRS) > 3.

• Death or severe disability defined as mRS > 4.

• Disability defined as mRS of 4 or 5 at the end of follow‐up.

Disability was defined as a Glasgow Outcome Scale (GOS) score of 2 or 3 (Jennett 1975), a mRS score of 3, 4 or 5 (Rankin 1957) or a Barthel Index (BI) < 60 (Mahoney 1965). Where no valid scale was given, we interpreted the text description of dependence seeking further detail from authors when necessary. The GOS is often misquoted, with the scale inverted, and we took care to interpret data given in this form.

We evaluated the studies for the presence of observer bias when patients were assessed following treatment. Ideally, outcome or endpoint assessment should be blinded to the type of intervention participants were assigned to. The use of special hats to hide the patient's scalp, thereby 'blinding' the observer during assessment of all patients, was proposed by the HEADDFIRST group as a way of avoiding observer bias.

Search methods for identification of studies

See the 'Specialized register' section in the Cochrane Stroke Group module.

Electronic searches

We searched the Cochrane Stroke Group's Trials Register, which was last searched by the Managing Editor in October 2010. In addition, we searched the following electronic bibliographic databases: the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2010, Issue 7), MEDLINE (1966 to October 2010) (Appendix 1) and EMBASE (1980 to October 2010) (Appendix 2). We performed cited reference searches in Science Citation Index (October 2010) using three geographically separated, relevant, high‐profile articles as a prospective search. The articles used were Carter 1997, Mori 1998, and Schwab 1998.

Searching other resources

We searched the reference lists of all relevant articles for references to trials. We contacted a selection of specialists, both medical and surgical, enquiring about relevant published, unpublished, pending or recently commenced trials not already identified. They were: Eric Juettler (Heidelberg, Germany), Jeannette Hofmeijer (Utretch, Netherlands) and Roland Jamora (Manila, Philippines).

Data collection and analysis

One review author (SCF) inspected the titles identified by the searching and examined the abstracts of references with titles of interest to determine relevance. When relevance was not clear from the abstract, or when no abstract was available, we obtained a copy of the article. We sought contact with the study author if further clarification was necessary.

We critically appraised any trial deemed relevant, using the checklist developed by Fowkes and Fulton (Fowkes 1991). One review author (SCF) reviewed and considered the limitations of studies with a view to inclusion (see Criteria for considering studies for this review).

One review author (SCF) extracted data independently from the included studies onto a pro‐forma. Primary outcomes of interest were death at the end of follow‐up, death or moderately severe disability defined as mRS > 3 at 6 and 12 months, and death or severe disability defined as mRS > 4 at 12 months. We also analysed the effect of surgical decompression on survival with severe disability defined as mRS 4 and 5. We calculated relative treatment effects using a fixed‐effect model, expressed as Peto ORs with 95% (CIs). We did not perform subgroup analyses.

Results

Description of studies

In the initial review the database searches from 1966 through to 2002 identified over 9000 references which we screened for relevance. For this update we searched MEDLINE and EMBASE from 2002 to October 2010, and after duplicates were removed we identified 5467 new references. A full search of the Cochrane Central Register of Controlled Trials (CENTRAL) identified 103 references and we received data on 10 trials from the Cochrane Stroke Group's Trials Register for assessment. Seven randomised controlled trials (RCTs) were found (DECIMAL 2007; DEMITUR Trial; DESTINY 2007; DESTINY 2; HAMLET 2009; HEADDFIRST; HeMMI 2004). Five were completed but only three were published and met criteria for inclusion in the analysis (DECIMAL 2007; DESTINY 2007; HAMLET 2009) (see Characteristics of included studies), the other two studies are awaiting assessment (DEMITUR Trial; HEADDFIRST ). Two trials are ongoing (DESTINY 2; HeMMI 2004).

Included studies

See Characteristics of included studies.

The three included studies were carried out in Europe, recruiting a combined total of 134 patients, aged 60 years or younger with a large hemispheric ischaemic stroke. Sixty‐nine patients in total were randomised to surgical decompression that was performed within 30 hours from symptom onset in two studies (DECIMAL 2007; DESTINY 2007), and within 96 hours in one study (HAMLET 2009). A study required the presence of an involvement of ≥ 50% of the middle cerebral artery (MCA) distribution and an infarct volume > 145cm³ in addition to a score ≥ 1 in the National Institutes of Health Stroke Scale (NIHSS) for item 1a (level of consciousness) (DECIMAL 2007). DESTINY 2007 required an involvement of two‐thirds of the MCA territory with NIHSS item 1a > 1 (level of consciousness); while HAMLET 2009 required involvement of two‐thirds of MCA territory associated with mass effect and midline shift. All three trials excluded patients with a pre stroke mRS ≥ 2. They shared the primary effect measure of favourable outcome defined as mRS ≤ 3 measured at 6 months in one study (DECIMAL 2007), at 6 and 12 months in the second (DESTINY 2007) and 12 months in the third study (HAMLET 2009). All three studies were stopped prematurely: two considering that data would be pooled with the other trials (DECIMAL 2007; DESTINY 2007), and one because the data safety committee established that no statistically significant difference would be found considering the sample size (HAMLET 2009).

Excluded studies

There are no excluded studies.

Completed clinical trials with data not yet available

See Characteristics of studies awaiting classification.

Risk of bias in included studies

• DECIMAL 2007: the randomisation method was unclear, it did not have concealment of allocation but blinded outcome assessment by covering patients' heads with a hat.

• DESTINY 2007: the study did not have concealment of allocation nor blinded outcome assessment.

• HAMLET 2009: the randomisation method was not clear, there was no concealment of allocation, a local investigator provided a narrative of function, and a panel of three blinded investigators adjudicated outcome based on the narrative provided.

Effects of interventions

The three studies with available data for analysis included a total of 134 patients. Surgical decompression reduced the risk of death at end of follow‐up (OR 0.19, 95% CI 0.09 to 0.37) (Analysis 1.1). Death or disability defined as mRS > 3 was not significantly different between the two treatment arms at end of follow‐up (OR 0.56, 95% CI 0.27 to 1.15) (Analysis 2.1). Surgical decompression reduced the risk of death and severe disability defined as mRS > 4 at 12 months (OR 0.26, 95% CI 0.13 to 0.51) (Analysis 2.2). Surgical decompression was associated with a non‐significant trend to survival with disability defined as mRS of 4 or 5 (OR 2.45, 95% CI 0.92 to 6.55) (Analysis 2.3).

1.1. Analysis.

Comparison 1 Primary outcome measure: Death at the end of follow‐up, Outcome 1 Death at the end of follow‐up.

2.1. Analysis.

Comparison 2 Secondary outcome measures, Outcome 1 Death or disability defined as mRS > 3 at end of follow‐up.

2.2. Analysis.

Comparison 2 Secondary outcome measures, Outcome 2 Death or severe disability defined as mRS > 4 at 12 months.

2.3. Analysis.

Comparison 2 Secondary outcome measures, Outcome 3 Severe disability among survivors defined as mRS 4 to 5 at 12 months.

Discussion

The pooled analysis of the three completed trials show that surgical decompression reduces the risk of death, and the risk of death or severe disability (mRS > 4) in patients 60 years of age or younger. However, it showed no benefit for disability or death defined as mRS > 3. These findings suggest that the intervention improves survival at the expense of an increased proportion of survivors with severe disability. These results stand in contrast to the results of the pooled analysis of individual patient data of the same clinical trials analysed here (Vahedi 2007), which showed that surgical decompression reduced the risk of death or disability defined as mRS > 3. In the analysis Vahedi 2007 included only those patients operated on within 48 hours from symptom onset; moreover, the mean age in that population was 45 years of age. These two factors may account for the difference between their results and the analyses presented here. In fact, HAMLET 2009, which had a window of intervention of 96 hours, has similar results to those of this review. These differences in findings support the idea of an earlier rather than a later intervention.

These results have two important limitations in terms of their applicability: they cannot be generalised to patients older than 60 years of age and they do not consider patients' values and views on quality of life at different levels of disability after stroke (Berge 2007). In addition, since all three trials were stopped early, an overestimation of the effect size of the intervention cannot be excluded.

Authors' conclusions

Implications for practice.

Decompressive surgery reduces the risk of death and the combined outcome of death or very severe disability. Therefore, it should be the preferred therapy in cases where this can also be assumed to be in the best interest of the patient, given that increased survival may occur at the expense of disability.

Implications for research.

Further research is needed to establish whether this intervention is beneficial in individuals older than 60 years of age. There is also a need for more information about patients' utility values for different levels of disability after stroke.

What's new

Date	Event	Description
17 September 2011	New citation required and conclusions have changed	New first author. As a result of the new trials, the review conclusions have changed significantly
17 September 2011	New search has been performed	Since the publication of the original review in 2002, three clinical trials including a total of 134 patients have been completed. These trials are included in this review

History

Protocol first published: Issue 1, 2002
 Review first published: Issue 3, 2002

Date	Event	Description
8 October 2008	New search has been performed	Converted to new review format.

Appendices

Appendix 1. MEDLINE search strategy

We used the following search strategy for MEDLINE (Ovid) and adapted it for CENTRAL

1. cerebrovascular disorders/ or basal ganglia cerebrovascular disease/ or brain ischemia/ or hypoxia‐ischemia, brain/ or carotid artery diseases/ or carotid artery, internal, dissection/ or intracranial arterial diseases/ or cerebral arterial diseases/ or "intracranial embolism and thrombosis"/ or intracranial embolism/ or intracranial thrombosis/ or stroke/ or exp brain infarction/ 
 2. (stroke$ or cerebral vascular or cerebrovasc$ or cva).tw. 
 3. ((brain or cerebr$ or hemisph$ or intracranial or mca or anterior circulat$) adj5 (isch?emi$ or infarct$ or emboli$ or thrombo$ or occlus$ or hypoxi$ or apople$)).tw. 
 4. 1 or 2 or 3 
 5. decompression, surgical/ or neurosurgical procedures/ or craniotomy/ or trephining/ 
 6. (decompress$ or craniectom$ or craniotom$ or hemi?craniect$ or trepa$ or treph$).tw. 
 7. (hippocampectom$ or lobectom$ or strokectom$).tw. 
 8. 5 or 6 or 7 
 9. 4 and 8 
 10. cerebrovascular disorders/su or basal ganglia cerebrovascular disease/su or brain ischemia/su or hypoxia‐ischemia, brain/su or carotid artery diseases/su or carotid artery, internal, dissection/su or intracranial arterial diseases/su or cerebral arterial diseases/su or "intracranial embolism and thrombosis"/su or intracranial embolism/su or intracranial thrombosis/su or stroke/su or exp brain infarction/su 
 11. exp vascular surgical procedures/ 
 12. (aneur$ or avm).tw. 
 13. 10 not (11 or 12) 
 14. brain edema/ 
 15. ((brain or cerebr$ or hemispher$ or intracranial or mca or anterior circulat$) adj5 (oedema or odema or edema or swell$ or swollen)).tw. 
 16. 14 or 15 
 17. 16 and 8 
 18. 9 or 13 or 17 
 19 limit 18 to human

Appendix 2. EMBASE search strategy

We used the following search strategy for EMBASE (Ovid)

1. cerebrovascular disease/ or cerebral artery disease/ or cerebrovascular accident/ or stroke/ or carotid artery disease/ or exp carotid artery obstruction/ or brain infarction/ or brain stem infarction/ or cerebellum infarction/ or brain ischemia/ or occlusive cerebrovascular disease/ or middle cerebral artery occlusion/ 
 2. (stroke$ or cerebral vascular or cerebrovasc$ or cva).tw. 
 3. ((brain or cerebr$ or hemisph$ or intracranial or mca or anterior circulat$) adj5 (isch?emi$ or infarct$ or emboli$ or thrombo$ or occlus$ or hypoxi$ or apople$)).tw. 
 4. 1 or 2 or 3 
 5. brain decompression/ or decompression surgery/ or decompression/ 
 6. skull surgery/ or craniectomy/ or cranioplasty/ or craniotomy/ or neurosurgery/ 
 7. (decompress$ or craniectom$ or craniotom$ or hemi?craniect$ or trepa$ or treph$).tw. 
 8. (hippocampectom$ or lobectom$ or strokectom$).tw. 
 9. 5 or 6 or 7 or 8 
 10. 4 and 9 
 11. cerebrovascular disease/su or cerebral artery disease/su or cerebrovascular accident/su or stroke/su or carotid artery disease/su or exp carotid artery obstruction/su or brain infarction/su or brain stem infarction/su or cerebellum infarction/su or brain ischemia/su or occlusive cerebrovascular disease/su or middle cerebral artery occlusion/su 
 12. exp vascular surgery/ 
 13. (aneurysm$ or avm).tw. 
 14. 11 not (12 or 13) 
 15. Brain edema/ 
 16. ((brain or cerebr$ or hemisper$ or intracranial or mca or anterior circulat$) adj5 (oedema or odema or edema or swell$ or swollen)).tw. 
 17. 15 or 16 
 18. 17 and 9 
 19. 10 or 14 or 18 
 20. limit 19 to human

Data and analyses

Comparison 1. Primary outcome measure: Death at the end of follow‐up.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Death at the end of follow‐up	3	134	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.19 [0.09, 0.37]

Comparison 2. Secondary outcome measures.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Death or disability defined as mRS > 3 at end of follow‐up	3	134	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.56 [0.27, 1.15]
2 Death or severe disability defined as mRS > 4 at 12 months	3	134	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.26 [0.13, 0.51]
3 Severe disability among survivors defined as mRS 4 to 5 at 12 months	3	78	Peto Odds Ratio (Peto, Fixed, 95% CI)	2.45 [0.92, 6.55]
4 Death or moderate to severe disability	3	134	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.53 [0.16, 1.72]

2.4. Analysis.

Comparison 2 Secondary outcome measures, Outcome 4 Death or moderate to severe disability.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

DECIMAL 2007.

Methods	Prospective, randomised, controlled, open study comparing surgical decompression plus medical treatment versus medical treatment alone The study had blinded evaluation of the primary end point
Participants	38 patients < 55 years of age All participants must have had involvement of > 50% of the MCA distribution and an infarct volume > 145 cm3 and a score > 1 for item 1a of the NIHSS (level of consciousness) 18 received medical therapy alone and 20 received medical therapy and decompressive craniectomy Patients were excluded if they had pre‐existing disability defined as mRS ≥ 2, a significant contralateral infarction, severe hemorrhagic transformation in > 50% of the MCA territory, coagulopathy, life expectancy < 3 years or any serious illness that could confound treatment assessment
Interventions	Early decompressive craniectomy plus standard medical therapy versus standard medical therapy alone in patients with malignant MCA infarction The decompressive surgery had to be a large hemicraniectomy that removed a bone flap that included temporal, frontal, parietal and occipital bones Surgery had to be performed no later that 6 hours from randomisation and < 30 hours from symptom onset The dura had to be opened Medical therapy included blood pressure, glucose control, mechanical ventilation and mannitol with clinical deterioration due to brain edema
Outcomes	Primary effect measure: favourable functional outcome defined as survival with mRS ≤ 3 at 6 months Secondary effect measures included survival, favourable functional outcome defined as mRS ≤ 3 or BI > 85 at 12 months, NIHSS and quality of life as defined by the Stroke Impact Scale 2.0 at 12 months
Notes	The study was stopped prematurely due to a slow recruitment and to the fact that data was to be combined with the two other European trials Blinded assessment was accomplished by covering the patients' heads with a surgical cap
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	Low risk	Blinded assessment by covering patients' heads with a surgical cap
Selective reporting (reporting bias)	Low risk

DESTINY 2007.

Methods	Prospective, randomised, controlled, open trial Blocked randomisation, stratified for each centre
Participants	32 patients < 60 years of age Patients included had an infarct involving > 2/3 of the MCA territory with score > 1 in item 1a of NIHSS 15 received medical therapy and 17 received medical therapy and surgical decompression Patients were excluded if they had a mRS ≥ 2, a BI < 95, a GCS < 6, dilated pupils, any other brain lesion that might affect outcome, haemorrhagic transformation, life expectancy < 3 years, coagulopathy or any other serious illness that might affect outcome
Interventions	Medical treatment alone versus medical treatment and surgical decompression Medical treatment included the use of osmotic agents on schedule to a target osmolality 315 to 320 mOsm, mechanical ventilation, blood pressure control, temperature and glucose control maintaining euvolemia Surgical decompression included a bone flap > 12 cm in diameter including frontal, temporal parietal and occipital bones and durotomy Surgery had to be performed > 12 hours but < 30 hours from symptom onset and no more than 6 hours from randomisation
Outcomes	Primary effect measure: favourable functional outcome defined as mRS ≤ 3 at 6 months and 12 months Secondary effect measures: mRS ≤ 4 and BI at 6 and 12 months Mortality at 30 days was also assessed but not as primary outcome
Notes	The study was stopped prematurely given the results of the pooled analysis of the three European trials
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	High risk

HAMLET 2009.

Methods	Prospective, randomised, controlled open trial
Participants	64 patients < 60 years of age, all had an infarct involving > 2/3 of MCA territory with evidence of mass effect and midline shift as well as changes in the level of consciousness as demonstrated by a GCS < 13 32 patients were randomised to each group Patients were excluded if they had an ischaemic stroke of the whole cerebral hemisphere, decrease in consciousness partially because of causes other than the formation of oedema, such as metabolic, disturbances or medication, pupils fixed and dilated, thrombolysis in the 12 hours before randomisation, coagulopathy, pre‐stroke score on the mRS > 1 or < 95 on the BI, life expectancy less than 3 years or other serious illness that might confound treatment assessment
Interventions	Medical treatment alone versus medical treatment plus surgical decompression performed < 96 hours from symptom onset and < 3 hours from randomisation Medical treatment included the use of osmotic agents on schedule to a target osmolality 315 to 320 mOsm, mechanical ventilation, blood pressure control, temperature and glucose control maintaining euvolemia Surgical decompression consisted of removal of a flap of bone of at least 12 cm in diameter including frontal, parietal, temporal and occipital bones The dura was opened and a dural patch was placed Infarcted brain tissue was not resected
Outcomes	Primary effect measure: favourable outcome mRS ≤ 3 at 1 year Secondary effect variable: favourable outcome defined as mRS score of 0 to 4, case fatality, functional dependence expressed as BI, symptoms of depression measured by the Montgomery and Asberg depression rating scale (MADRS), and quality of life measured with the Medical Outcomes Study 36‐item short‐form health survey (SF‐36) and a visual analogue scale (VAS)15 at 1 year
Notes	The study was stopped prematurely as the Data Monitoring Committee established that no statistically significant difference would be found between both arms with the pre‐established sample size
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk
Blinding (performance bias and detection bias) All outcomes	Low risk	An unblinded investigator provided narrative description that was submitted to three blinded investigators to ascertain mRS The final score was obtained by consensus from these blinded observers

BI: Barthel Index 
 GCS: Glasgow Coma Score 
 MCA: middle cerebral artery 
 mRS: modified Rankin Scale 
 NIHSS: National Institutes of Health Stroke Scale

Characteristics of studies awaiting assessment [ordered by study ID]

DEMITUR Trial.

Methods	Patients aged 40 to 80 years were prospectively randomised to surgical decompression vs standard medical treatment
Participants	151 patients were included
Interventions	Surgical decompression within 48 hours from symptom onset
Outcomes	Primary outcome dichotomised mRS score 0 to 3 (good outcome) versus > 3 (poor outcome) at 6 and 12 months Secondary outcome: BI > 85 at 6 and 12 months, NIHSS and quality of life by Turkish version of Stroke Impact Scale 2.0 (SIS) at 6 and 12 months
Notes	Principal investigator: Emre Kumral Trial was carried out between January 2003 and December 2007 ‐ data not yet published It is unclear if outcome assessment was blinded A sample size of 106 was calculated on the assumption that 40% of patients in the surgical group and 70% in each of the medical treatment groups would have a poor outcome Could not contact principal investigator

HEADDFIRST.

Methods	Patients were randomised to surgical decompression versus standard medical care
Participants	75 patients were included in this study
Interventions	Surgical decompression carried out within 96 hours from symptom onset
Outcomes	Case fatality, functional outcome, quality of life, caregiver burden, patient perceptions of survivorship, acute health care utilisation measured all at 21, 90 and 180 days
Notes	Principal investigator: Jeffrey I Frank Trial carried out from February 2000 through to January 2002 ‐ data not yet published

BI: Barthel Index 
 mRS: modified Rankin Scale 
 NIHSS: National Institutes of Health Stroke Scale

Characteristics of ongoing studies [ordered by study ID]

DESTINY 2.

Trial name or title	Decompressive surgery for the treatment of malignant infarction of the middle cerebral artery 2
Methods	Prospective randomised open controlled multicentre phase III comparative trial
Participants	Participants older than 60 years of age with a large MCA infarction to be treated within 48 hours from symptom onset Sample size 160
Interventions	Surgical decompression plus medical treatment versus medical treatment alone
Outcomes	mRS score, dichotomised 0 to 4 versus 5 to 6 after 6 months
Starting date	2009
Contact information	Eric Juettler MD eric.juettler@med.uni‐heidelberg.de
Notes	ISRCTN# 21702227

HeMMI 2004.

Trial name or title	Hemicraniectomy for Malignant Middle cerebral artery Infarcts
Methods	Randomised controlled trial
Participants	Participants with MCA infarction, aged 18 to 65 years to be treated within 72 hours from onset Sample size 56
Interventions	Surgical decompression versus medical treatment
Outcomes	Primary effect measure: mRS < 4 or BI > 60
Starting date	2004
Contact information	Principal investigator: Roland D Jamora dominicjamora@gmail.com
Notes

BI: Barthel Index 
 MCA: middle cerebral artery 
 mRS: modified Rankin Scale

Contributions of authors

Dr S Cruz‐Flores: literature searching, data collection and data extraction and writing the first draft of the present review. 
 Dr E Berge: participated in writing the protocol and writing the first version of the review. In the second version of the review he participated in the assessment of data and writing the review. 
 Prof I Whittle: reviewed both versions of the review.

Sources of support

Internal sources

• University of Edinburgh, UK.

External sources

• No sources of support supplied

Declarations of interest

Dr S Cruz‐Flores was a site principal investigator for the HEADDFIRST trial supported by the National Institute of Neurological Diseases and Stroke. The trial is mentioned in this review. However, its results have not been published and therefore are not included in the review.

New search for studies and content updated (conclusions changed)