Massive Cerebral Infarction

ORIGINAL ARTICLE

Suresh Subramaniam, MD, MSc,* and Michael D. Hill, MD, MSc, FRCPC*†

severely disabled with poor quality of life. The incidence is Background: Massive cerebral infarcts cause brain edema with 1 midline shifts and impingement on vital structures producing coma estimated at 2 to 10% of all ischemic stroke. and death. The mortality rate is estimated at 80% with standard Massive cerebral infarction may be managed both medi- medical treatment. Surgical decompression with hemicraniectomy cally and surgically, with little available evidence from random- has proved to be life saving, but the impact on functional outcomes ized controlled studies to guide the choice. Numerous develop- is largely unknown. The focus of this review is to discuss the ments have occurred in the medical and surgical management of treatment options for massive cerebral infarcts. patients with massive cerebral infarction in the recent past. This Review Summary: Neurologic deterioration following massive article presents a narrative review of the literature concerning the cerebral infarct needs to be recognized early enough for medical and surgical interventions. Medical management includes monitoring in clinical course and medical and neurosurgical management of a neurologic intensive care unit, hyperosmolar agents, and hyper- massive cerebral infarction, highlighting speciﬁc areas in which ventilation. Surgical management includes decompressive hemi- developments are occurring rapidly. craniectomy and duraplasty with resection of infarcted tissue in some instances. Conclusion: Hemicraniectomy is emerging as a promising treatment of patients with massive cerebral infarcts, but only select patients beneﬁt from this procedure. Further information from ran- Malignant middle cerebral artery (MCA) domized controlled trials is required to elucidate the best treatment options for this kind of stroke. infarction is a form of massive cerebral Key Words: massive cerebral infarcts, hemicraniectomy, infarction that is used to describe complete hypothermia, malignant middle cerebral artery infarct MCA territory infarction resulting in significant (The Neurologist 2005;11: 150–160) space-occupying effect.

assive cerebral infarct is one of the most catastrophic Mforms of ischemic stroke with no proven treatment. METHODS Malignant middle cerebral artery (MCA) infarction is a form The literature was searched from 1970 using MEDLINE of massive cerebral infarction that is used to describe com- and EMBASE with keywords massive cerebral infarcts, hemi- plete MCA territory infarction resulting in signiﬁcant space- craniectomy, malignant MCA infarct, and hypothermia. Articles occupying effect. The resulting depression of consciousness unavailable for translation were not retrieved. Theoretically, terminates in coma and brain death within 2 to 5 days in massive cerebral infarcts are large hemispheric infarcts associ- almost 80% of patients treated with conservative medical ated with edema and tissue displacements. Only the salient therapy alone. The survivors of this form of stroke are features of management are reviewed here.

From the *Calgary Stroke Program, Department of Clinical Neurosciences, and the †Departments of Medicine and Community Health Sciences, University of Calgary, Alberta, Canada. Complete middle cerebral artery (MCA) territory Reprints: Michael D. Hill, MD, MSc, FRCPC, Calgary Stroke Program, Department of Clinical Neurosciences, University of Calgary, Foothills infarction is invariably caused by occlusion of Hospitals, Room 1242A, 1403 29th Street NW, Calgary, Alberta, T2N the distal internal carotid artery and the 2T9 Canada. E-mail: [email protected]. Copyright © 2005 by Lippincott Williams & Wilkins proximal MCA trunk. ISSN: 1074-7931/05/1103-0150 DOI: 10.1097/01.nrl.0000159987.70461.d7

150 The Neurologist • Volume 11, Number 3, May 2005 The Neurologist • Volume 11, Number 3, May 2005 Massive Cerebral Infarction

CLINICAL COURSE AND PREDICTORS OF than a global increase in intracranial pressure (ICP). Regional CLINICAL DETERIORATION differences in ICP and cerebral perfusion pressure (CPP) play a The clinical course of massive cerebral infarction is major role in the exacerbation of cerebral ischemia. A focal determined by the location of vascular occlusion, intrinsic increase in ICP in areas of impaired autoregulation around the tissue susceptibility, and duration of ischemia. Complete infarction can cause paradoxic reduction in cerebral perfusion MCA territory infarction is invariably caused by occlusion of pressure, resulting in ischemia. Focal ICP can lead to focal the distal internal carotid artery and the proximal MCA trunk ischemia when ICP is greater than 20 mmHg, and global (commonly referred to as a T-occlusion or L-occlusion).1 The ischemia occurs when ICP is greater than 50 mmHg. Patients at etiology of malignant MCA infarction is almost always em- risk for ischemic brain edema are those with large hemispheric bolic, either from a proximal cardiac source or carotid disease strokes involving more than 50% of the MCA territory (Fig. 1). (dissection, atheroembolic). Anatomic factors such as an The clinical predictors of early deterioration have been largely incomplete circle of Willis and impaired collateral circulation derived from numerous case reports. Headache, nausea, vomit- play a critical role in determining the extent of infarction. The ing, early somnolence. and respiratory disturbances occur in hemodynamic effects of collateral circulation are important in approximately 50% of patients with malignant MCA infarction maintaining perfusion to penumbral regions. The extent of as early as 3 hours after stroke onset and may herald mass effect infarction is also determined by brain tissue susceptibility to from significant brain edema and early deterioration. These ischemia. For example, susceptibility is augmented by diabe- symptoms ideally warrant neurologic intensive care unit (ICU) tes mellitus. Diabetes mellitus enhances apoptosis induced by admission and monitoring. Mass effect with edema usually cerebral ischemia in rodent models.2 Reperfusion injury may occurs 1 to 5 days following the ictus, leading to herniation of be another important component cause of massive MCA brain tissue through dural spaces. This can produce subfalcine, infarction. Reperfusion of irreversibly damaged brain may transtentorial, uncal, and tonsillar herniation, often resulting in enhance edema formation and result in rapid development of rapid death.1,6–8 swelling (within 24 hours), with hemorrhagic transformation leading to mass effect. These factors commonly result in a fatal outcome. There is a mortality rate of 80% with standard 3 medical treatment alone. Mass effect with edema usually occurs 1 to 5 Clinically, carotid T-occlusion produces ischemia of the frontoparietal regions, leading to hemianopia and flaccid days following the ictus, leading to herniation hemiplegia. Often, these patients have head turning and gaze of brain tissue through dural spaces. deviation to the affected hemisphere secondary to involvement of frontal eye fields. Global aphasia due to involvement of the dominant hemisphere or hemispatial neglect from nondominant hemisphere involvement can all be highly dis- abling.1,4,5 Early identification of patients susceptible for Neuroimaging is an invaluable tool for identifying pa- malignant MCA infarction is crucial for appropriate selection tients at risk for developing ischemic brain edema. Neurologic of medical and surgical interventions. deterioration correlates with horizontal displacement of the an- terior septum and pineal gland on computed tomographic (CT) scan rather than with ICP elevation. The presence of low-density infarction and edema on CT scan that occupies more than 50% of MCA territory is a reliable predictor of subsequent edema The clinical symptoms of worsening neurologic formation following a large hemispheric infarct. von Kummer et status following a large hemispheric infarct are al9 have shown that early (Ͻ6 hours) CT scan changes of greater than 50% MCA hypodensity or local brain swelling (effacement mainly due to focal ischemic brain edema and of sulci, compression of lateral ventricle) are associated with focal displacement of brain rather than a fatal outcomes in 85% of patients (94% specificity and 61% sensitivity). In addition, hemorrhagic transformation of large global increase in intracranial pressure (ICP). hemispheric infarcts may worsen the brain edema and tissue shifts. Barber et al10 proposed the use of early radiologic signs on follow-up CT scans performed within 48 hours of stroke onset in predicting mortality at 30 days. The CT parameters The clinical symptoms of worsening neurologic status anteroseptal shift (Ͼ5 mm), pineal shift (Ͼ2 mm), hydroceph- following a large hemispheric infarct are mainly due to focal alus, temporal lobe infarction, and the presence of other vascular ischemic brain edema and focal displacement of brain rather territory infarction were predictive of fatal outcomes.10 Case

TREATMENT CONSIDERATIONS Management options for massive cerebral infarction can be classiﬁed into medical and surgical. Currently, there are no strict guidelines available from randomized controlled trials. Patients susceptible to the development of massive cerebral infarction should be admitted to a neuro-ICU setting with facilities available for early recognition of brain edema and intervention with an integrated approach between the neurointensivist and the neurosurgeon.

Medical management should focus on prevention of brain edema and tissue displacements.

MEDICAL MANAGEMENT FIGURE 1. (A, B) A case of massive cerebral infarct involving Medical management should focus on prevention of the left middle cerebral artery (MCA) territory caused by brain edema and tissue displacements. Detailed hemody- cardiac emboli in a 50-year-old man. The hypodense areas namic, neuroimaging, and ICP monitoring are the main tools represent infarcted brain tissue with edema and some midline shift at the level of septum pellucidum (A–D). that should guide management in the neuro-ICU. Current medical therapies have a very limited role in the management of massive cerebral infarction and largely fail to prevent the reports have also shown that a pineal shift of 2.5 to 4 mm from formation of brain edema and tissue displacement. It should midline is associated with drowsiness; 6 to 9 mm, with stupor; be remembered that early neurologic deterioration is not due and greater than 9 mm, with coma.11–13 In the European Coop- to an increase in global ICP and, as a result, therapies that erative Acute Stroke Study, fatal brain edema occurred in pa- focus on lowering global ICP are not effective. Ideal therapy tients with baseline (Ͻ6 hours) CT scans showing greater than should be aimed at preventing brain edema, tissue shifts, and 33% of MCA involvement.14 Krieger et al7 have shown that secondary increases in ICP. patients with a baseline National Institutes of Health Stroke General Measures Scale (NIHSS) score greater than or equal to 20 points with left Initial management is aimed at reducing systemic com- hemispheric infarctions, or greater than or equal to 15 points plications. Patients with reduced level of consciousness may with right hemispheric infarctions within 6 hours of symptom be unable to protect their airway, and endotracheal intubation onset accompanied by nausea/vomiting, or more than 50% is indicated. There is no evidence to support prophylactic MCA territory hypodensity on CT scan are at high risk for intubation in these patients.15,16 Elevation of the head to a 30° developing fatal brain edema. angle has been shown to reduce ICP, but this may be negated by the reduction in cerebral perfusion.17 Blood pressure management is a controversial issue in the setting of acute ischemic stroke. There is no evidence to suggest that acute Early (Ͻ6 hours) CT scan changes of greater blood pressure lowering in large hemispheric infarcts reduces edema and tissue displacements. Theoretically, lowering than 50% middle cerebral artery hypodensity or blood pressure may reduce disruption of the blood-brain local brain swelling (effacement of sulci, barrier, vasogenic edema, hemorrhagic transformation, and further progression. Conversely, perfusion to the ischemic compression of lateral ventricle) are associated penumbra (a region where autoregulation of blood ﬂow is no with fatal outcomes in 85% of patients. longer active) may be reduced, further exacerbating infarction. A reasonable approach to blood pressure management would be to keep systolic blood pressure less than 220 mmHg

the treatment of large hemispheric infarcts is not convincing TABLE 1. General Measures in the Management of Large Hemispheric Infarcts and there have been conflicting reports that suggest these agents may even worsen the ICP and exacerbate the infarc- 1. Endotracheal intubation for patients unable to protect airway; tion. Hyperosmolar agents are more effective in counteract- keep PaCO2 at 28 to 35 mmHg ing global increases in ICP compared with focal. In areas 2. Control fever (use antipyretics, wet wraps, cooling blankets) where the blood-brain barrier is disrupted, hyperosmolar 3. Elevation of head to 30° angle to reduce ICP agents have been shown to worsen tissue shifts and aggravate 4. Maintain normovolemia (central venous pressure between 4–8 brain edema.25 Rebound edema may occur following discon- mmHg) tinuation of these agents.26,27 The most common hyperosmo- 5. Maintain normoglycemia (treat with insulin/glucose/potassium infusion) lar agents used are mannitol, glycerol, hypertonic saline, 6. Keep blood pressure less than 220/120 mmHg (give labetalol, albumin, THAM buffer, indomethacin, hetastarch, high-dose enalapril, avoid nitrates) barbiturates, and hyperventilation. The optimal timing of 7. Intermittent compression stockings to prevent deep vein therapy with these agents in relation to evolving brain edema thrombosis is largely unknown. Moreover, differences may exist among 8. Close observation for aspiration pneumonia and gastrointestinal these agents. Table 2 outlines the use of hyperosmolar agents bleeding and techniques employed in ICP reduction in large hemispheric infarcts. and diastolic blood pressure less than 120 mmHg.18 Hyper- Hypothermia Inducing moderate hypothermia (31–33°C) is an glycemia should be managed aggressively with insulin be- emerging therapy that may effectively decrease brain edema cause it has been shown to aggravate brain edema and worsen and swelling following large hemispheric infarcts. A number outcomes. However, this approach is an empiric one because of animal and clinical studies have shown that delayed evidence from randomized trials is lacking. Blood glucose hypothermia has a neuroprotective role in both animal models levels should be kept below 5 mmol/L by giving intravenous and clinical research following ischemic stroke by increasing insulin.19,20 Fever is associated with poor outcomes in is- the resistance to ischemia, prolonging the survival of the chemic stroke patients at baseline and should be treated penumbra, reducing ICP, and minimizing blood-brain barrier promptly by investigating its underlying cause.21 Fluid status disruption.28–34 Current evidence suggests that hypothermia should be maintained normovolemic with 0.9% sodium chlo- reduces cerebral blood flow and metabolism, thereby reduc- ride and should be monitored using hematocrit, osmolality, ing cerebral edema following head trauma.35–37 Multicenter and serum sodium. Steroids have no proven role in the prospective studies have suggested that hypothermia is fea- management of cerebral edema following ischemic stroke.22 sible and safe to use in large hemispheric infarcts.31,32 One The general measures are outlined in Table 1. such study has shown that moderate hypothermia can help control critically elevated ICP values in large hemispheric MCA stroke and may improve clinical outcome in these patients. In this study, hypothermia was induced within 14 A reasonable approach to blood pressure hours from stroke onset and was maintained over 72 hours to minimize brain edema. External cooling was induced using management would be to keep systolic blood cooling blankets, bladder and stomach cold infusions, wash- pressure less than 220 mmHg and diastolic ings, etc. The mortality rate was 44%. Potential complications encountered following hypothermia are a rebound blood pressure less than 120 mmHg. increase in ICP on rewarming, ventricular ectopy and fibril- lation, coagulopathy, bradycardia, and pneumonia. Hernia- tion from increased ICP during rewarming can occur, neces- sitating treatment with other agents to lower ICP. Cerebral Osmotherapy perfusion pressure, ICP, and brain temperature need to be Hyperosmolar therapy reduces ICP mainly by dehy- monitored during induction of hypothermia. Hypothermia drating and shrinking the undamaged brain tissue by allowing can also be combined with other therapies, including hyper- water to diffuse out in the direction of the osmolar gradient. osmolar agents and decompressive hemicraniectomy. Table 3 The duration and degree of ICP reduction depends on the lists the current clinical trials using hypothermia in stroke. volume of the remaining normal brain tissue, the amount of Hypothermia is a promising therapy for large hemispheric undamaged blood-brain barrier, initial ICP, and the dose of strokes; however, further randomized clinical trials are the agent used.23,24 The efficacy of hyperosmolar agents in needed.

TABLE 2. Agents Used for Reducing ICP in Large Hemispheric Infarcts

Agent Dose Indications Mechanism of Action Side Effects/Notes

Mannitol 20% solution; 0.5 g/kg every Impending transtentorial Decreases viscosity, tissue water Initial hypervolemia followed by 4–6 hours; keep plasma herniation; increased content; increases CSF absorption extensive diuresis; watch for osmolality between 310– ICP pulmonary edema, electrolyte 320 mOsm/L imbalance, dehydration, etc. Glycerol 0.25–1.5 g/kg every 6 hours Increased ICP Free radical scavenger, antioxidant; Nausea, vomiting, diarrhea, causes vasodilation, inhibits hemoglobinuria, bleeding leukocyte adherence; diathesis antiinﬂammatory; improves cerebral blood ﬂow Hypertonic saline plus 6% hetastarch in 0.9% NaCl Increased ICP; should be Shift of water from extravascular to Anaphylactic shock, volume hydroxyethyl starch injection, 100–250 mL every reserved until mannitol intravascular compartment overload, metabolic acidosis, 8 hours is ineffective; more hemodilution with coagulopathy hemodilution when hydroxyethyl starch is used THAM buffer (Tris- 1 mmol/kg in 100 mL 5% Increased ICP Neutralization of acidosis-related ICP monitoring should be hydroxymethyl glucose over 45 minutes vasodilation available aminomethane) until pH is between 7.5 and 7.55 and base excess is Ͻ10

Pentobarbital 10 mg/kg/day over 30 minutes Increased ICP Reduce cerebral metabolic rate, Hypotension, severe infection, Neurologist The blood ﬂow, blood volume, free decreased cardiac performance; radical scavengers depth of barbiturate coma can be monitored with continuous EEG Indomethacin 30–50 mg intravenously Increased ICP Cerebral vasoconstrictor May reduce cerebral blood ﬂow; no conclusive data oue1,Nme ,My2005 May 3, Number 11, Volume • 05Lpict ilas&Wilkins & Williams Lippincott 2005 © Albumin 0.63–2.5 g/kg Cerebral edema, increased Increases plasma oncotic pressure; Pulmonary edema; no conclusive ICP hemodilution, free radical evidence so far scavenger Hyperventilation PCO2 target 30 mmHg Used only when Hypocarbia-induced cerebral Prolonged hyperventilation should achieved by raising osmotherapy is vasoconstriction be avoided due to cerebral ventilation rate 14–18 ineffective; effect can vasoconstriction; other side breaths/minute at a constant be prolonged by using effects include barotrauma, tidal volume; this lowers in conjunction with hypokalemia ICP by 25–30% THAM buffer 05Lpict ilas&Wilkins & Williams Lippincott 2005 © Neurologist The

TABLE 3. Clinical Trials of Hypothermia in Stroke

Trial Status Design Results 2005 May 3, Number 11, Volume •

CHILI Ongoing Multicenter, prospective, nonrandomized phase II trial; Not available (Controlled Hypothermia patients enrolled within 72 hours of a large ischemic in Large Infarction) stroke; hypothermia induced with external cooling device to a temperature of 35°C for 48 hours; controlled rewarming by 0.5°C every 12 hours until 37°C is reached COOL AID 1 Completed 40 patients; 5 centers; randomized, controlled study; 18 patients were randomized to hypothermia (Cooling Acute Ischemic patients randomized to receive either standard treatment, with 22 in a control group; of Brain Damage; safety normothermic stroke treatment or standard care in those in the hypothermia group, 12 were and feasibility study) conjunction with controlled moderate hypothermia; successfully cooled; there was no difference patients in the hypothermic group were cooled to a in outcome among the 2 groups at 30 days; target temperature of 33°C using a heat-exchange on MRI examination, infarct growth was catheter inserted into the inferior vena cava via the slightly lower in the hypothermia group than femoral vein; hypothermia was maintained for 24 hours in the control group (90.0% versus 108.4%, and then was slowly rewarmed to 37°C P ϭ 0.71); however, this difference was not statistically signiﬁcant COOL BRAIN–STROKE Ongoing Prospective, nonrandomized pilot trial of 10 patients; 1 Not available center; cooling helmet put into place immediately by the emergency personnel following ischemic stroke; all patients transported to hospital, where they all receive standard care according to the severity of their stroke; helmet remains on for 72 hours, vital signs and any complications monitored frequently during that interval Hemicraniectomy and Completed Retrospective study comparing moderate hypothermia Mortality was 12% for hemicraniectomy group Moderate Hypothermia with hemicraniectomy (n ϭ 36) versus 47% for hypothermia group in Patients With Severe Ischemic Stroke NOCSS Ongoing Multicenter, multinational, randomized, controlled trial; Not available (Nordic Cooling Stroke patients randomized to treatment groups are cooled to Study) 35°C over a period of 9 hours using surface cooling technique; patients receive meperidine to control

shivering and discomfort, and are passively rewarmed Infarction Cerebral Massive following the treatment period NOTHOT Completed Pilot clinical trial of 16 patients; patients with ischemic Mean temperature was 36.9°C for the (Normothermia and Stroke stroke given acetaminophen (650 mg every 4 hours) for acetaminophen group and 37.0°C for the Outcome) 72 hours or until discharge; cooling blankets ordered control group (P ϭ 0.05) for temperature Ͼ37.6°C 155 Subramaniam and Hill The Neurologist • Volume 11, Number 3, May 2005

during the procedure can be stored in antibiotic solution at Ϫ70°C and replaced 1 to 3 months later after the brain edema Hemicraniectomy is often used as a life-saving has subsided. Consideration to multiple factors should be procedure with no clear impact on long-term given prior to offering hemicraniectomy, including patient age, timing of surgery, dominant versus nondominant hemi- disability and functional outcomes. sphere involvement, and expected long-term disability. Hemicraniectomy is often used as a life-saving procedure with no clear impact on long-term disability and functional outcomes. Surgical Management Evidence that hemicraniectomy reduces mortality is Surgical techniques are based on an empiric approach largely derived from uncontrolled case series and retrospec- rather than evidence from randomized, controlled studies. tive studies. Early animal studies have revealed that trephi- Ventriculostomy may reduce ICP by draining cerebrospinal nation surgery, a form of hemicraniectomy in rats, reduces fluid, but may increase focal brain shifts. Resection of the the 7-day mortality from 35% in the nonsurgical group to 38,39 infarcted brain tissue is another option, but separating in- almost 0% in the surgical group. Neurologic behavior farcted brain from potentially viable brain tissue may be and infarct volume were favorable in those treated with impossible. In the past, temporal lobectomy has been tried ultra-early surgery within 4 hours of occlusion, implying the with limited success. Theoretically, temporal lobectomy may importance of early intervention. Compared with control reduce uncal herniation. Decompressive hemicraniectomy groups, hemicraniectomy in rats increases perfusion in the with duraplasty is a controversial approach to reduce the cortex and reduces the infarct volume as shown by magnetic 40 catastrophic mass effect of brain edema and tissue displace- resonance imaging (MRI). A review of human case series ments (Fig. 2). The procedure involves generous removal of reports on decompressive hemicraniectomy for large space- 41–57 a frontotemporoparietal bone flap ipsilateral to the side of occupying cerebral infarcts is shown in Table 4. infarct followed by opening of the dura to allow outward Hemicraniectomy can also be combined with other herniation of the brain tissue.1,6 The bone flap removed techniques that may help synergistically to reduce brain edema. Comparison of moderate hypothermia with hemicraniectomy in massive cerebral infarcts reveals that hemicraniectomy is in fact better, with no difference in intensive medical treatment and duration of stay in the neuro-ICU. In 1 retrospective study, the mortality rate was 47% for the hypothermia group versus 12% for the hemicraniectomy group.58 Robertson et al57 reported a case series that combines hemicraniectomy with resection of infarcted tissue. Ten of 12 patients who underwent the procedure survived, 5 ended up with independent or moderate disability, and 5 were left with severe disability. The mortality rate was 16% but the disability rate was high, at 41%, compared with other published case reports. A number of questions arise when one considers hemicraniectomy for massive cerebral infarcts. When should hemicraniectomy be offered to patients with large hemispheric infarcts? Hemicraniectomy is usually considered when there are unequivocal clinical and radiologic signs suggestive of impending brain edema and when all other medical measures have failed. Common predictors of brain edema following large hemispheric infarcts include a younger age, nausea, and vomiting within the first 24 hours of admission, National Institutes of Health Stroke Scale score of Ն15 points Ն FIGURE 2. Decompressive hemicraniectomy and duraplasty in for right-hemisphere MCA infarction and 20 points for left- a 44-year-old man following carotid dissection from motor hemisphere MCA infarction, early hypodensity on the CT scan, vehicle accident. The ischemic/infarcted brain tissue on the midline shift of more than 10 mm at the septum pellucidum right hemisphere has room for herniation outward and pre- level, and MCA infarcts involving more than 50% of the vents downward displacement of vital structures (A–D). hemisphere.7 Some authors advocate repeating the CT scan

TABLE 4. Review of Case Series Reports on Decompressive Hemicraniectomy for Large Hemispheric Cerebral Infarcts

No. No. No. Independent or Severe Patients With Author Year Patients Survivors Nonsurvivors Moderate Disability Disability Early Surgery

Ivamoto et al41 1974 1 1 0 1 NA 0 Rengachary et al42 1981 3 3 0 1 2 0 Young et al43 1982 1 1 0 1 NA 0 Saito et al44 1987 7 4 3 0 4 NA Ojemann et al45 1988 2 2 0 2 NA NA Kondziolka and Fazl46 1988 5 5 0 5 NA 2 Delashaw et al47 1990 9 8 1 4 4 3 Steiger8 1991 8 6 2 4 2 NA Tsuruno et al48 1993 14 9 5 6 3 NA Kalia and Yonas49 1993 4 4 0 4 NA 1 Rieke et al50 1995 32 21 11 16 5 8 Carter et al51 1997 14 11 3 8 3 5 Schwab et al52 1998 31 26 5 17 9 NA Koh et al53 2000 10 8 2 4 4 NA Mori et al54 2001 19 16 3 3 13 NA Walz et al55 2002 18 12 6 11 1 9 Leonhardt et al56 2002 26 20 6 11 NA 11 Robertson et al57 2004 12 10 2 5 5 7 Total 77.3% 22.6% 47.6% 25.4% 46 (21.2%)

within 6 hours for those at high risk for brain edema, so that care. Some authors strongly recommend hemicraniectomy for early hemicraniectomy can be offered before brain herniation younger patients (age Ͻ45 years) on the assumption that and irreparable brain damage occur.5,59 young patients are better able to tolerate brain edema and have better functional outcomes given fewer comorbid con- ditions.61 Some critics argue that this could be a result of selection bias. At the same time, younger patients have less Hemicraniectomy is usually considered when room for expansion due to absence of age-related atrophy and this could increase the risk of herniation and death. In one there are unequivocal clinical and radiologic analysis of pooled data from case reports and retrospective signs suggestive of impending brain edema and studies, 14% of patients younger than 50 years died despite surgery compared with 32% older than 50 years of age. Thus, when all other medical measures have failed. age is a crucial factor in predicting functional outcome after hemicraniectomy in patients with large hemisphere infarcts.62

What is the impact of age on outcomes in patients undergoing hemicraniectomy? Case reports have clearly shown that older patients undergoing hemicraniectomy have Age is a crucial factor in predicting functional 51,53,55 poor functional outcomes and increased mortality. This outcome after hemicraniectomy in patients with may be because older patients have other comorbid condi- tions that increase the risk of poor outcomes and mortality. In large hemisphere infarcts. 1 recent study, age older than 60 years, severe neurologic deficit on admission, longer duration of intensive care treatment, and mechanical ventilation were significantly associated with more severe disability (Barthel Index Ͻ50).60 Is timing of hemicraniectomy important on outcome? Health-related quality of life was considerably impaired in Animal and clinical studies have provided evidence for ben- the subscales of mobility, household management, and body efit from early surgery. In animal MCA occlusion models,

early hemicraniectomy at 1 hour versus 24 hours reduces require randomized, controlled trials to determine the exact infarct volume significantly.39 Schwab et al52 analyzed the timing for hemicraniectomy. Current ongoing clinical trials influence of early hemicraniectomy (Ͻ24 hours after symp- randomize patients between 12 to 96 hours from symptom tom onset) based on clinical status on admission and initial onset (Table 5). CT findings versus late surgery (Ͼ24 hours after first revers- Can hemicraniectomy be offered to large, dominant- ible signs of herniation) on mortality, and functional out- hemisphere infarcts? This is again a very controversial topic. come. From 31 patients who underwent early hemicraniec- The benefit of hemicraniectomy in dominant hemisphere tomy, mortality was 16% and 84% had a Barthel index more infarcts appears less convincing due to the fact that the than 60 at the 3-month follow-up. Early hemicraniectomy led procedure may leave patients with severe aphasia and poor to a significant reduction in ICU admission (7.4 days versus quality of life. Following stroke, there is no cortical reorga- 13.3 days, P Ͻ 0.05). On the other hand, late intervention nization of adjacent areas of normal brain tissue surrounding (Ͼ24 hours) has neither been shown to affect outcome nor language areas.63 However, this is all based on case series recovery. Early decompressive surgery should be performed and we clearly need a randomized, controlled study to answer to prevent irreversible damage to adjacent brain tissues. this question. A list of current ongoing and completed ran- Several other case reports and retrospective studies have domized clinical trials of hemicraniectomy in stroke is shown revealed the same results.50,51,55 Analysis of pooled pub- in Table 5. lished reports suggest that signs of herniation have no impact Hemicraniectomy is a life-saving procedure in patients on the timing of surgery and outcomes.60 We certainly with massive cerebral infarcts. At this time it is not very clear

TABLE 5. Clinical Trials of Hemicraniectomy in Stroke

Trial Status Design Results

DESTINY Ongoing 68 patients, 9 centers; prospective, randomized, Not available (Decompressive Surgery for the open, controlled, multicenter study; to compare Treatment of Malignant the efﬁcacy of decompressive surgery combined Infarction of the Middle with conservative treatment to conservative Cerebral Artery) treatment alone after malignant hemispheric ischemic cerebral infarction; patients randomized between 12 and 36 hours of deterioration symptoms; treatment/surgery within 6 hours of randomization HAMLET Ongoing 112 patients, 7 centers; multicenter, open, Not available (Hemicraniectomy After MCA randomized clinical trial; patients randomized to infarction with Life-threatening either decompressive surgery, consisting of a Edema Trial) large hemicraniectomy and a duraplasty, followed by intensive care treatment, or conservative treatment, consisting of either intensive care treatment or “standard” therapy on a stroke unit; randomization stratiﬁed according to the intended mode of conservative treatment HeaDDFIRST Completed Multicenter pilot clinical trial with a planned Not available (Hemicraniectomy and Durotomy enrollment of 75 patients; all patients receive for Deterioration From standardized medical therapy (SMT); patients Infarction Relating Swelling who develop severe brain swelling within 96 Trial) hours of stroke onset are randomized to receive SMT alone or SMT ϩ standardized hemicraniectomy and durotomy Hemicraniectomy and Moderate Completed Retrospective study comparing moderate Mortality was 12% for Hypothermia in Patients With hypothermia with hemicraniectomy (n ϭ 36) hemicraniectomy Severe Ischemic Stroke group versus 47% for hypothermia group HeMMI Ongoing 56 patients; open, randomized clinical trial; Not available (Hemicraniectomy For Malignant patients randomized to receive either standard Middle Cerebral Artery medical treatment or hemicraniectomy with Infarcts) duraplasty within 72 hours from symptom onset

which age group of patients may have a better functional 12. Ropper AH. A preliminary MRI study of the geometry of brain dis- outcome. Younger patients with massive cerebral infarcts placement and level of consciousness with acute intracranial masses. Neurology. 1989;39:622–627. should be monitored carefully for brain edema and should be 13. Pullicino PM, Alexandrov AV, Shelton JA, et al. Mass effect and death strongly considered for hemicraniectomy. We require a better from severe acute stroke. Neurology. 1997;49:1090–1095. understanding of the pathophysiology of massive cerebral 14. von Kummer R, Allen KL, Holle R, et al. Acute stroke: usefulness of early CT findings before thrombolytic therapy. Radiology. 1997;205: infarcts in this group of patients. 327–333. 15. Bushnell CD, Phillips-Bute BG, Laskowitz DT, et al. Survival and outcome after endotracheal intubation for acute stroke. Neurology. CONCLUSION 1999;52:1374–1381. There is no proven therapy for massive cerebral in- 16. Gujjar AR, Deibert E, Manno EM, et al. Mechanical ventilation for farcts. Case series and retrospective studies are difficult to put ischemic stroke and intracerebral hemorrhage: indications, timing, and together to form coherent inference because selection bias outcome. Neurology. 1998;51:447–451. 17. Feldman Z, Kanter MJ, Robertson CS, et al. Effect of head elevation on prevents valid conclusions. The key pathophysiologic con- intracranial pressure, cerebral perfusion pressure, and cerebral blood cept is that focal increases in ICP are different from global flow in head-injured patients. J Neurosurg. 1992;76:207–211. increases in ICP. Although surgical decompressive craniec- 18. Adams HP Jr, Adams RJ, Brott T, et al. Guidelines for the early management of patients with ischemic stroke: a scientific statement from tomy can save a life and appears more promising than the Stroke Council of the American Stroke Association. Stroke. 2003; moderate hypothermia, it remains unclear whether this is 34:1056–1083. appropriate given the expected severe disability. Quality of 19. Berger L, Hakim AM. The association of hyperglycemia with cerebral edema in stroke. Stroke. 1986;17:865–871. life and social disability measures may be the most important 20. Yip PK, He YY, Hsu CY, et al. Effect of plasma glucose on infarct size outcomes for future randomized clinical trials. Finally, ran- in focal cerebral ischemia-reperfusion. Neurology. 1991;41:899–905. domized trials are difficult to perform in both because exper- 21. Rordorf G, Koroshetz W, Efird JT, et al. Predictors of mortality in stroke tise in the treatment of this disease, either with hypothermia patients admitted to an intensive care unit. Crit Care Med. 2000;28: 1301–1315. or surgical management is just developing and because the 22. Norris JW, Hachinski VC. High dose steroid treatment in cerebral condition is relatively rare. infarction. BMJ. 1986;292:21–23. 23. Manno EM, Adams RE, Derdeyn CP, et al. The effects of mannitol on cerebral edema after large hemispheric cerebral infarct. Neurology. ACKNOWLEDGMENT 1999;52:583–587. The authors thank Ms. Rae A. Kokotailo for her help in 24. Paczynski RP, He YY, Diringer MN, et al. Multiple-dose mannitol reduces brain water content in a rat model of cortical infarction. Stroke. preparing this manuscript. 1997;28:1437–1443. Discussion. Stroke. 1997;28:1444. 25. Kaufmann AM, Cardoso ER. Aggravation of vasogenic cerebral edema REFERENCES by multiple-dose mannitol. J Neurosurg. 1992;77:584–589. 26. Node Y, Yajima K, Nakazawa S. ͓Rebound phenomenon of mannitol 1. Hacke W, Schwab S, Horn M, et al. “Malignant” middle cerebral artery and glycerol: clinical studies͔ . No To Shinkei.1983;35:1241–126. territory infarction: clinical course and prognostic signs. Arch Neurol. 27. Node Y, Yajima K, Nakazawa S. ͓A study of mannitol and glycerol on 1996;53:309–315. the reduction of raised intracranial pressure and on their rebound 2. Li ZG, Britton M, Sima AA, et al. Diabetes enhances apoptosis induced phenomenon͔. No Shinkei Geka. 1983;11:259–267. by cerebral ischemia. Life Sci. 2004;76:249–262. 28. Busto R, Dietrich WD, Globus MY, et al. Postischemic moderate 3. Berrouschot J, Sterker M, Bettin S, et al. Mortality of space-occupying hypothermia inhibits CA1 hippocampal ischemic neuronal injury. Neu- (“malignant”) middle cerebral artery infarction under conservative in- rosci Lett. 1989;101:299–304. tensive care. Intensive Care Med. 1998;24:620–623. 29. Karibe H, Chen J, Zarow GJ, et al. Delayed induction of mild hypother- 4. Demchuk A, Krieger D. Large middle cerebral artery infarction. Neu- mia to reduce infarct volume after temporary middle cerebral artery rology. 1998;51:1514–1515. 5. Demchuk AM, Krieger DW. Mass effect with cerebral infarction. Curr occlusion in rats. J Neurosurg. 1994;80:112–119. Treat Options Neurol. 1999;1:189–199. 30. Xue D, Huang ZG, Smith KE, et al. Immediate or delayed mild hypother- 6. Kasner SE, Demchuk AM, Berrouschot J, et al. Predictors of fatal brain mia prevents focal cerebral infarction. Brain Res. 1992;587:66–72. ͓ edema in massive hemispheric ischemic stroke. Stroke. 2001;32:2117– 31. Schwab S, Schwarz S, Bertram M, et al. Moderate hypothermia for the ͔ 2123. treatment of malignant middle cerebral artery infarct . Nervenarzt. 7. Krieger DW, Demchuk AM, Kasner SE, et al. Early clinical and 1999;70:539–546. radiological predictors of fatal brain swelling in ischemic stroke. Stroke. 32. Schwab S, Georgiadis D, Berrouschot J, et al. Feasibility and safety of 1999;30:287–292. moderate hypothermia after massive hemispheric infarction. Stroke. 8. Steiger HJ. Outcome of acute supratentorial cerebral infarction in pa- 2001;32:2033–2035. tients under 60. Development of a prognostic grading system. Acta 33. Steiner T, Friede T, Aschoff A, et al. Effect and feasibility of controlled Neurochir (Wien). 1991;111:73–79. rewarming after moderate hypothermia in stroke patients with malignant 9. von Kummer R, Meyding-Lamade U, Forsting M, et al. Sensitivity and infarction of the middle cerebral artery. Stroke. 2001;32:2833–2835. prognostic value of early CT in occlusion of the middle cerebral artery 34. Su J, Qiu YM, Chen ZH , et al. Clinical analysis of massive hemispheric trunk. AJNR Am J Neuroradiol. 1994;15:9–15. Discussion. AJNR infarction treated with moderate hypothermia. Chin J Traumatol. 2003; Am J Neuroradiol. 1994;15:16–18. 6:318–3120. 10. Barber PA, Demchuk AM, Zhang J, et al. Computed tomographic 35. Marion DW, Penrod LE, Kelsey SF, et al. Treatment of traumatic brain parameters predicting fatal outcome in large middle cerebral artery injury with moderate hypothermia. N Engl J Med. 1997;336:540–546. infarction. Cerebrovasc Dis. 2003;16:230–235. 36. Peng RY, Bongard FS. Hypothermia in trauma patients. J Am Coll Surg. 11. Ropper AH. Lateral displacement of the brain and level of consciousness 1999;188:685–696. in patients with an acute hemispheral mass. N Engl J Med. 1986;314: 37. Kirkpatrick AW, Chun R, Brown R, et al. Hypothermia and the trauma 953–958. patient. Can J Surg. 1999;42:333–343.

38. Doerfler A, Forsting M, Reith W, et al. Decompressive craniectomy in a 52. Schwab S, Steiner T, Aschoff A, et al. Early hemicraniectomy in patients rat model of “malignant” cerebral hemispheric stroke: experimental support with complete middle cerebral artery infarction. Stroke. 1998;29:1888– for an aggressive therapeutic approach. J Neurosurg. 1996;85:853–859. 1893. 39. Forsting M, Reith W, Schabitz WR, et al. Decompressive craniectomy 53. Koh MS, Goh KY, Tung MY, et al. Is decompressive craniectomy for for cerebral infarction. An experimental study in rats. Stroke. 1995;26: acute cerebral infarction of any benefit? Surg Neurol. 2000;53:225–230. 259–264. 54. Mori K, Aoki A, Yamamoto T, et al. Aggressive decompressive surgery 40. Engelhorn T, von Kummer R, Reith W, et al. What is effective in in patients with massive hemispheric embolic cerebral infarction asso- malignant middle cerebral artery infarction: reperfusion, craniectomy, or ciated with severe brain swelling. Acta Neurochir (Wien). 2001;143: both? An experimental study in rats. Stroke. 2002;33:617–622. 483–491. Discussion. Acta Neurochir (Wien). 2001;143:491–492. 41. Ivamoto HS, Numoto M, Donaghy RM. Surgical decompression for 55. Walz B, Zimmermann C, Bottger S, et al. Prognosis of patients after cerebral and cerebellar infarcts. Stroke. 1974;5:365–370. hemicraniectomy in malignant middle cerebral artery infarction. J Neu- 42. Rengachary SS, Batnitzky S, Morantz RA, et al. Hemicraniectomy for rol. 2002;249:1183–1190. acute massive cerebral infarction. Neurosurgery. 1981;8:321–328. 56. Leonhardt G, Wilhelm H, Doerfler A, et al. Clinical outcome and 43. Young PH, Smith KR, Dunn RC. Surgical decompression after cerebral neuropsychological deficits after right decompressive hemicraniectomy hemispheric stroke: indications and patient selection. South Med J. in MCA infarction. J Neurol. 2002;249:1433–1440. 1982;75:473–475. 57. Robertson SC, Lennarson P, Hasan DM, et al. Clinical course and 44. Saito I, Segawa H, Shiokawa Y, et al. Middle cerebral artery occlusion: surgical management of massive cerebral infarction. Neurosurgery. correlation of computed tomography and angiography with clinical 2004;55:55–61. Discussion. Neurosurgery. 2004;55:61–62. outcome. Stroke. 1987;18:863–868. 58. Georgiadis D, Schwarz S, Aschoff A, et al. Hemicraniectomy and 45. Ojemann R, Heros R, Crowell R. Surgical Management of Cerebrovas- moderate hypothermia in patients with severe ischemic stroke. Stroke. cular Disease. Baltimore: Williams & Wilkins; 1988. 46. Kondziolka D, Fazl M. Functional recovery after decompressive crani- 2002;33:1584–1588. ectomy for cerebral infarction. Neurosurgery. 1988;23:143–147. 59. Demchuk AM. Hemicraniectomy is a promising treatment in ischemic 47. Delashaw JB, Broaddus WC, Kassell NF, et al. Treatment of right hemi- stroke. Can J Neurol Sci. 2000;27:274–277. spheric cerebral infarction by hemicraniectomy. Stroke. 1990;21:874–881. 60. Foerch C, Lang JM, Krause J, et al. Functional impairment, disability, 48. Tsuruno T, Takeda M, Imaizumi T, et al. ͓Internal decompression with and quality of life outcome after decompressive hemicraniectomy in hippocampectomy for massive cerebral infarction͔. No Shinkei Geka. malignant middle cerebral artery infarction. J Neurosurg. 2004;101: 1993;21:823–827. 248–254. 49. Kalia KK, Yonas H. An aggressive approach to massive middle cerebral 61. Wijdicks EF, Diringer MN. Middle cerebral artery territory infarction artery infarction. Arch Neurol. 1993;50:1293–1297. and early brain swelling: progression and effect of age on outcome. 50. Rieke K, Schwab S, Krieger D, et al. Decompressive surgery in space- Mayo Clin Proc. 1998;73:829–836. occupying hemispheric infarction: results of an open, prospective trial. 62. Gupta R, Connolly ES, Mayer S, et al. Hemicraniectomy for massive Crit Care Med. 1995;23:1576–1587. middle cerebral artery territory infarction: a systematic review. Stroke. 51. Carter BS, Ogilvy CS, Candia GJ, et al. One-year outcome after 2004;35:539–543. decompressive surgery for massive nondominant hemispheric infarction. 63. Cao Y, Vikingstad EM, George KP, et al. Cortical language activation in Neurosurgery. 1997;40:1168–1175. Discussion. Neurosurgery. 1997; stroke patients recovering from aphasia with functional MRI. Stroke. 40:1175–1176. 1999;30:2331–2340.