Table 2.

Large hemispheric infarct management: cerebral edema treatment

Trials (chronological)	Patient sample and intervention	Major findings	Reported limitations
Medical management
Qureshi et al (1998)65	27 patients with cerebral edema including patients with head trauma (n = 8), postoperative edema (n = 5), nontraumatic intracranial hemorrhage (n = 8), and cerebral infarction (n = 6) IV infusion of 3% saline/acetate to increase serum sodium concentrations to 145–155 mmol/L	A reduction in mean ICP within the first 12 h correlating with an increase in the serum sodium concentration was not in patients with nontraumatic intracranial hemorrhage or cerebral infarction In patients with head trauma, the beneficial effect of hypertonic saline on ICP was short-lasting, and after 72 h of infusion, four patients required IV pentobarbital due to poor ICP control In repeat CT scan within 72 h of treatment lateral brain displacement was not reduced in nontraumatic ICH or cerebral infarction patients	Retrospective chart review Small patient population of each group and the heterogeneity in the underlying neurological illness
Schwarz et al (1998)62	9 patients with elevated ICP after acute space-occupying hemispheric stroke (n = 8) or hypertensive putaminal hemorrhage with massive perifocal edema (n = 1) IV infusion of 100mL HS–HES or 40 g mannitol over 15 min	ICP decreased from baseline values in both groups (p < 0.01) The maximum ICP decrease was 11.4 mm Hg (after 25 min) in the HS–HES-treated group and 6.4 mm Hg (after 45 min) in the mannitol-treated group. There was no constant effect on cerebral perfusion pressure in the HS–HES-treated group, whereas cerebral perfusion pressure rose significantly in the mannitol-treated group	Small sample size
Schwarz et al (2002)63	8 patients with elevated ICP after acute space-occupying hemispheric stroke (n = 6) or supratentorial hemorrhage with massive perifocal edema (n 2) IV 75 mL hypertonic (10%) saline, if standard treatment of 200 mL of 20% mannitol was not effective	No constant effect on mean arterial blood pressure, whereas cerebral perfusion pressure was consistently increased Blood osmolarity rose by 9 mmol/L and serum sodium by 5.6 mmol/L. Potassium levels, hemoglobin, hematocrit, and pH were slightly decreased. No unexpected side effects were noted	Small sample size
Gondim et al (2005)66	95 patients treated with mannitol to determine if MI-AKI is linked to elevated osmolality	11 patients (11.6%) who developed MI-AKI did not have significant differences in patient age, sex, or race; history of cerebrovascular disease or smoking; baseline renal function; or GCS compared with those without MI-AKI Renal function spontaneously returned to baseline in all patients MI-AKI appears to be associated with history of chronic disease that affects renal function, e.g., diabetes mellitus rather than mannitol dose or osmolality	Addresses only the renal effects of increased osmolality and not the hyperosmolar effect in central nervous system function
Harutjunyan et al (2005)61	40 patients at risk of increased ICP were randomized to receive either 7.2% NaCl/HES 200/0.5 (NaCl/HES) or 15% mannitol targeting ICP < 15mm Hg. Of them, 17 patients received the NaCl/ HES regimen, 15 received the mannitol regimen, and 8 patients did not require treatment due to their ICP not going over 20mm Hg	Both treatment regimens lowered ICP below 15 mm Hg (p < 0.0001). The NaCl/HES lowered the ICP within a median of 6.0 min while the mannitol within a median of 8.7 min (p < 0.0002) The NaCl/HES caused a pronounced decrease in ICP than mannitol (57% vs 48%; p < 0.01). Cerebral perfusion pressure was increased from a median 60 to 72 mm Hg by infusion with NaCl/HES (p < 0.0001) while with the mannitol regimen increased from a median 61 to 70 mm Hg with mannitol (p < 0.0001) The mean arterial pressure was increased by 3.7% from NaCl/HES 200/0.5 and was not altered by mannitol	Small patient population of each group and the heterogeneity in the underlying neurological illness
Hauer et al (2011)64	100 patients with severe ICH signs of increased ICP received a continuous infusion of 3% hypertonic saline within ≤72 h after symptom onset over a mean period of 13 d The findings were compared previously treated control group (n = 115) with equal underlying disease	Fewer episodes of critically elevated ICP (92 vs. 167; p = 0.027) in fewer patients (50.0 vs. 60.0%; p = 0.091) and in-hospital mortality was significantly decreased (17.0 vs. 29.6%; p = 0.037)	Pilot character of study Compared findings with a previous cohort with known outcomes risk of bias Heterogeneous patient cohort with three different pathologies of cerebrovascular diseases
Lin et al (2015)67	432 patients of whom 62.3% had ischemic stroke, received mannitol treatment	The incidence of MI-AKI was 6.5% (95% CI, 4.5–9.3%) in acute stroke patients, 6.3% in patients with ischemic stroke, and 6.7% in patients with ICH Multivariate analysis revealed that diabetes, lower eGFR at baseline, higher baseline NIHSS, and use of diuretics increased the risk of MI-AKI	Decision to initiate osmotic therapy and the dose depended only on the treating physician’s expertise Retrospective nature of the study that cannot exclude missing data in the analysis The causative relationship between concurrent medications and the development of MI-AKI were not easily clarified
Sheth et al (2016)69	86 patients with anterior circulation LHI within 10 h from onset assigned to placebo or IV glyburide in a double-blind, randomized, placebo-controlled trial	At 90 d, 17 (41%) patients in the IV glyburide group and 14 (39%) in the placebo group had an mRS 0–4 without surgery (adjusted OR, 0.87; 95% CI, 0.32–2.32; p = 0.77)	Enrollment was stopped because of funding reasons
Surgical management
Jüttler et al (2007)71	32 patients were randomized to either surgical plus conservative treatment or to conservative treatment alone in a prospective, multicenter, randomized, controlled, clinical trial and patients	At 30 d, 88% patients in surgery group survived compared with 47% patients in medical therapy group (p = 0.02)	81% of patients originated from two centers Possibility of bias introduced due to nonblinding of evaluation of clinical outcome
Vahedi et al (2007)72	38 patients in a multicenter, randomized trial in France involving patients with large MCA infarction comparing functional outcomes with or without surgery	The proportion of patients with a mRS ≤3 at the 6-mo and 1-y follow-up was 25 and 50%, respectively, in the surgery group compared with 5.6 and 22.2%, respectively, in the no-surgery group (p = 0.18 and 0.10, respectively) There was a 52.8% absolute reduction of death after craniectomy compared with medical therapy only (p < 0.0001)	The trial stopped because of slow recruitment
Hofmeijer et al (2009)70	64 patients were included; 32 were randomly assigned to surgical decompression and 32 to best medical treatment Patients with LHI with cerebral edema were randomly assigned within 4 d of stroke onset to surgical decompression or best medical treatment	Surgical decompression had no effect on the primary outcome measure (absolute risk reduction 0%; 95% CI, −21 to 21) but did reduce case fatality (absolute risk reduction 38%, 15–60)	Unknown number of patients who were screened Possibility of selection bias due to the small number of patients with aphasia in the referral of patients for inclusion in this trial
Slezins et al (2012)76	28 patients with a malignant MCA stroke Random assignment to surgery plus best medical therapy group or best medical therapy alone	6 patients survived: 5 in the DCE group (none of them was older than 60 y) and 1 in the BMT group (p = 0.03/0.06) All survivors in the DCE group had favorable outcomes. There was no significant difference in the NIHSS and GCS scores between the groups and survivors/nonsurvivors (p > 0.05)	Small sample size
Zhao et al (2012)77	47 patients (n = 24 surgery plus medical treatment vs. n = 23 medical treatment alone) with MCA infarction with cerebral edema in a randomized controlled trial	At 6 and 12 mo, reduced mortality was observed in surgery plus medical therapy compared with medical therapy alone (12.5 vs. 60.9%; p = 0.001 and 16.7 vs. 69.6%; p < 0.001, respectively) At 6 and 12 mo, fewer patients had mRS >4 postsurgery (33.3 vs. 82.6%; p = 0.001 and 25.0 vs. 87.0%; p < 0.001, respectively)	63.8% of patients were from one single center Unknown number of patients that have been screened Outcome measurement was based on questionnaires replied by patients’ caregivers No data quality of life, neuropsychiatric status, or aphasia
Hofmeijer et al (2013)73	20 patients underwent surgery within 48 h of MCA infarction due to life-threatening edema, underwent detailed neuropsychological examination at a median of 14.5 mo after stroke onset	Poorer cognitive performance was associated with worse functional outcomes on the mRS scale (b −0.4; 95% CI, −0.6 to −0.1) No differences were observed between operated and nonoperated patients	Small sample size Investigator who performed the examinations was not blinded to treatment assignment and functional outcome

Abbreviations: CI, confidence interval; GCS, Glasgow Coma Scale; HS–HES, hypertonic saline–hydroxyethyl starch; ICH, intracerebral hemorrhage; ICP, intracranial pressure; LHI, large hemispheric infraction; MCA, middle cerebral artery; MI-AKI, mannitol-induced acute kidney injury; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale.