The concept of decompressive surgery for treatment of elevated intracranial pressure has been developed already in the beginning of the 20th century.1 The rationale of this treatment modality consists of opening of the skull and removal of a bone flap to allow the edematous brain to swell outward, thereby preventing intracranial tissue shifts and life-threatening downward herniation. The use of decompressive hemicraniectomy (DHC) in the context of ischemic brain edema had been reported already in 1956.2 Since that time, DHC has been increasingly studied in the setting of different conditions, including traumatic brain injury, subarachnoid hemorrhage, and malignant middle cerebral artery (MCA) infarction.
Depending on the location of the affected area, different surgical decompression techniques have been developed. In the presence of diffuse brain edema without a midline shift, as commonly seen in traumatic brain injury, bilateral (eg, bifrontal) craniectomy has been advocated. Hemicraniectomy, or removal of a frontotemporoparietal bone flap, is suitable in patients with unilateral hemisphere swelling as seen after ischemic stroke.3 Accumulating experience with DHC over the years has led to increasing refinement of the surgical technique. The size of the removed bone fragment has been recognized as a factor of crucial importance for generation of a sufficient decompressive effect.4 Hemicraniectomy with a diameter of ≤10 cm, especially in combination with sharp trepanation edges, has been associated with an increased incidence of shearing injury to the herniated brain.4 Furthermore, dural opening, usually followed by insertion of a dural graft (duraplasty), has meanwhile become an integral part of the decompressive surgery technique.3
Predictors of Malignant Cerebral Edema
Early identification of patients who are most likely to develop malignant edema after MCA infarction based on clinical, radiographic, anatomic, and laboratory values can aid the clinician in offering DHC early. Previously published predictors of a National Institutes of Health Stroke Scale score of >20, thrombus at the carotid terminus location, presence of nausea and vomiting, elevations of the white blood cell count, early involvement of >50% of the MCA territory on CT, and additional involvement of the anterior cerebral artery territory and/or posterior cerebral artery territory may be clinical tools to identify high-risk patients.5,6 Involvement of the anterior choroidal artery can be subtle in the setting of a large infarct, but involvement of the uncus of the temporal lobe may lead to more rapid herniation.7 Although clinically easy to use, the positive predictive value of these variables is low.
Serum S100B is an astroglial protein that is released during neuronal injury and enters the peripheral bloodstream through an incompetent blood–brain barrier. Thresholds of S100B levels can be monitored at time points in the acute period to determine patients most likely to develop malignant edema. Single measurements obtained in the 12- to 24-hour time period may be a useful tool to identify high-risk patients. At 24 hours, a value of 1.03 μg/L has 94% sensitivity and 83% specificity for detection of malignant cerebral edema.8
The availability of MRI in the acute period may allow for more precise volumetric analysis of the infarct. A MRI diffusion-weighted imaging volume of >82 cm3 when performed <6 hours has a high specificity (98%) but low sensitivity (52%).9 A MRI diffusion-weighted imaging volume of >145 cm3 obtained before 14 hours was associated with 100% sensitivity and 94% specificity in a small cohort of patients.10 The differences in the sensitivities are likely due to the timing of obtaining the MRI. Such volumetric analysis can be complicated by the presence of an arterial occlusion that is yet to be reperfused through intravenous thrombolysis or intra-arterial treatment. Moreover, when such treatments are used and successful reperfusion occurs, there may be concerns of reperfusion injury that may potentially lead to exacerbation of the edema.11 Nonetheless, MRI volumetric analysis appears to have a high specificity to detect patients at highest risk but must be considered in the context of treatment strategies being used.
Such analyses have allowed for better determination of patients who would most benefit from DHC. Randomized controlled studies have used such predictors as inclusion criteria in trial design.
Hemicraniectomy for Malignant MCA Infarction: Randomized Controlled Trials
Based on the promising results of experimental research and nonrandomized studies on hemicraniectomy in malignant MCA infarction, 5 randomized controlled trials (RCTs) have been initiated in the past decade (Table), and meanwhile, the results of the 3 European RCTs (DECIMAL [DEcompressive Craniectomy In MALignant middle cerebral artery infarcts]12; DESTINY [DEcompressive Surgery for the Treatment of malignant Infarction of the middle cerebral artery]13; HAMLET [Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema Trial]14) and 2 pooled meta-analyses14,15 have been published. The North American HeADDFIRST study (Hemicraniectomy And Durotomy on Deterioration From Infarction Related Swelling Trial) has been completed but data have not been published yet and the Philippine HeMMI trial (Hemicraniectomy for Malignant Middle cerebral artery Infarcts) is still recruiting patients.
Prospective RCTs on Hemicraniectomy in Malignant MCA Infarction
All 3 European trials showed a significant reduction in mortality in surgically treated patients as compared with the conservatively treated groups. In DECIMAL, the absolute risk reduction for mortality at 6 months was 53% with hemicraniectomy.12 DESTINY reached significance for 30-day mortality (12% with surgery versus 53% with conservative management) after enrollment of 32 patients.13 In HAMLET, surgical treatment also led to improved survival with an absolute risk reduction of 38%.14 However, none of the 3 RCTs could demonstrate a significant benefit of hemicraniectomy considering functional outcome as defined in the primary outcome measures (Table). The most recent meta-analysis of DECIMAL, DESTINY, and HAMLET, published after the completion of the HAMLET trial in 2009, included all patients from the 3 trials who were randomized within 48 hours after symptom onset (n=109) and focused on mortality and functional outcome after 1 year.14 Of 109 patients, 58 had been assigned to hemicraniectomy and 51 to conservative treatment. With surgical treatment, the absolute risk reduction for mortality in this analysis comprised 49.9% (95% CI, 33.9 to 65.9), corresponding to a number needed to treat of 2 for prevention of death. There was also a significant absolute risk reduction of 41.9% (95% CI, 25.2 to 58.6) with hemicraniectomy for a modified Rankin scale (mRS) >4 with a number needed to treat=2. Surgery, however, did not lead to a significant benefit in functional outcome when dichotomization between a mRS of 0 to 3 and 4 to 6 after 12 months was chosen (absolute risk reduction, 16.3%; 95% CI, −0.1 to 33.1). Reduction in mortality with surgical treatment was accompanied by an increase in moderate severe disability (mRS of 4) in survivors.
Complications Associated With Hemicraniectomy
Surgical and medical complications associated with DHC may impact the clinical outcomes of patients. Immediate surgical complications include: insufficient decompression,4 surgical site infections,16 hemorrhagic complications, and contralateral subdural effusions.17 Delayed complications include the sinking flap syndrome,18 extra-axial fluid collections, hydrocephalus,19 and development of subdural hematomas.
The rates of infections at the time of DHC or replacement of the bone flap range from 5% to 10%. Infection rates at the time of cranioplasty may be related to the type of bone flap used.16 Synthetic materials used as flaps may lead to a foreign body reaction, whereas autologous bone flaps may be associated with higher rates of infection.16,20 Unfortunately there is no consensus as to which approach is associated with a higher likelihood of infections.
The sinking flap syndrome is felt to occur as a result of the pressure gradient between the atmospheric pressure and intracranial vault. This leads to severe headaches, changes in mentation, and seizures.18 Focal neurological deficits may result due to reductions in cerebral blood flow in the infarcted region that may be viable21 or due to a mass effect on the contralateral hemisphere. Rarely, this can lead to death from “paradoxical herniation.” In a recent prospective cohort, 11% of patients developed symptomatic sinking flap syndrome that was associated with delays in replacement of the bone flap, older age, and larger initial infarct volume.22 Placing patients supine often relieves the clinical symptoms with replacement of the bone flap being curative.
The development of communicating hydrocephalus as a result of DHC may be a common occurrence.19 The distinction between radiographic ventriculomegaly and clinically symptomatic hydrocephalus may account for the varied experiences among institutions.23 This phenomenon may occur due to the altered hydrodynamics of the intracranial vault that has been disrupted after DHC. Extracranial fluid collections are frequently noted and are a result of overaccumulation of cerebrospinal fluid either due to poor reabsorption or space availability for fluid accumulation. These collections often signify the presence of hydrocephalus and are associated with neurological decline in mentation. The treatment for this includes cerebrospinal fluid diversion with an external ventricular catheter or repeated lumbar punctures. Delays in replacing the bone flap appear to be the most significant predictor of the development of hydrodynamic complications but are complicated by the fact that delays in replacement may be due to the degree of swelling initially noted.19
In-hospital medical complications due to patient immobility and survival from the DHC can impact the clinical outcomes of the patient. A National Inpatient Sample database over a 6-year period in the United States found the rates of pneumonia to be 11.1%, gastrointestinal bleeding 2.4%, and sepsis 4.76% in 252 patients studied. Each of these complications was associated with an increased rate of mortality and were significantly higher compared with patients with a similar comorbidity index.24
Unanswered Questions and Future Directions
Functional Outcome and Quality of Life
Early hemicraniectomy significantly reduces mortality after malignant MCA infarction; however, it also increases the probability of survival with moderately severe disability (mRS of 4). With approximately 40% of survivors becoming disabled after decompressive surgery, the question arises if a mRS of 4 (unable to walk without assistance and unable to attend to own bodily needs without assistance) can be considered a favorable outcome. Looking at motor function, the benefit of surviving malignant MCA infarction after hemicraniectomy seems to be largely outweighed by the high incidence of moderately severe or severe disability in survivors.25 However, the more important question is if the mRS is an adequate outcome measure in those patients. From the patients' perspective, neuropsychological deficits, aphasia, or depression may have an equally strong impact on quality of life as compared with motor function. Other factors such as psychosocial environment, caregiver burden, familial support, and financial support should be additionally considered in this context. The prospective trials and pooled analyses published to date12,–,15 do not provide conclusive results on quality of life and depression in patients who survived malignant MCA infarction after surgery, and those aspects certainly deserve further investigation.
Timing of Surgery
From the pathophysiological point of view, earlier decompression should prevent brain tissue damage by avoiding or reducing exposition to increased intracranial pressure in the course of development of ischemic brain edema. On the other hand, poststroke edema often peaks later than 48 hours after symptom onset. Therefore, there might be a wider time window within which decompressive surgery may be beneficial for such patients. This aspect has not been sufficiently addressed in the 3 European RCTs. The pooled analysis from 2007 could not demonstrate any difference in functional outcome, comparing patients treated earlier versus later than 24 hours after symptom onset15; however, all patients included in that analysis were treated within 48 hours. The HAMLET study allowed delayed surgery up to 96 hours after stroke onset, and secondary outcome analyses showed that surgery within 48 hours significantly reduced the probability of severe disability or death (mRS 5 or 6), whereas delayed hemicraniectomy did not influence outcome.14 However, considering the small number of patients who received surgery beyond 48 hours (n=11), no final conclusion can be drawn. Further data on timing of decompressive surgery are derived from observational studies, which have brought up contradictory results. Although some studies report reduced mortality and improved outcome with early treatment, as compared with treatment after clinical deterioration,26,–,28 a systematic review published in 2004, including all data reported up to that date, could not confirm this finding.29 This issue certainly deserves further investigation to identify the optimal time window for decompressive surgery after malignant MCA infarction. In the absence of other conclusive data and considering the findings reported from RCTs as well as the pathophysiological background, at present, early decompression (<48 hours after symptom onset) seems to be beneficial.
Age Limit for Surgery
None of the RCTs investigating hemicraniectomy in malignant MCA infarction included patients >60 years. Because a considerable proportion of the patients experiencing this type of stroke belong to this age cohort,30 it still remains unclear if those patients would benefit from surgical treatment. Data from observational studies indicate that hemicraniectomy may lead to improved survival, however, at the cost of poor outcome and functional dependency in patients >60 years of age.30,31 Moreover, age was identified as a major factor influencing outcome in a systematic review of 138 patients treated with hemicraniectomy.29 This finding could not be confirmed in the pooled meta-analysis of the 3 European RCTs published in 2007.15 The HAMLET trial even found a trend toward better outcome in the upper age range (51 to 60 years) as compared with younger patients treated with hemicraniectomy.14 In light of those data, the results of the ongoing DESTINY 2 trial, studying hemicraniectomy in patients >60 years, are awaited and will hopefully provide more information on this issue.
Treatment of Dominant Hemisphere Infarction
The debate whether to perform decompressive craniectomy in patients with malignant MCA infarction on the speech-dominant hemisphere is based on the assumption that in the presence of aphasia, functional outcome and quality of life may be worse as compared with patients with nondominant infarction. This assumption, however, is currently not supported by data available in the literature, because mortality, functional outcome, and quality of life do not seem to depend on whether the dominant hemisphere is involved.15,29 On the contrary, neuropsychological deficits seen in patients with infarcts on the nondominant hemisphere, as attention deficits or depression, may be as disabling as aphasia.32 However, this question has not been sufficiently elucidated yet and certainly deserves further study.
Conclusions
Predictive models of patients who may require DHC are improving through volumetric analysis based on MRI and serum markers to assess for neuronal injury. Although several RCTs have not been completed, DHC is a life-saving surgery that appears to benefit younger patients the most. Further study is required to better elucidate quality-of-life outcome measures, timing of surgery, and treatment of the dominant hemisphere.
Disclosures
R.G. is a consultant/scientific advisory board for Concentric Medical, Rapid Medical, Neurointerventions, and CoAxia Inc.
- Received October 14, 2010.
- Revision received November 19, 2010.
- Accepted November 22, 2010.
- © 2011 American Heart Association, Inc.
References
Jump to
This Issue
Article Tools
Share this Article
- Hemicraniectomy in Malignant Middle Cerebral Artery Infarction
