A multiparameter model predicting in-hospital mortality in malignant cerebral infarction.

The early identification of patients with large hemisphere infarctions (LHIs) at risk of fatal brain edema may result in better outcomes. A quantitative model using parameters obtained at admission may be a predictor of in-hospital mortality from LHI.This prospective study enrolled all patients with LHI involving >50% of the middle cerebral artery (MCA) admitted to our neurological intensive care unit within 48 hours of symptom onset. Early clinical and radiographic parameters and the baseline CHADS2 score (congestive heart failure, hypertension, age ≥ 75 years, diabetes mellitus, stroke [double weight]) were analyzed regarding their ability to predict patient outcomes.Seventy-seven patients with LHIs were identified, 33 (42.9%) with complete MCA infarction (CMCA), and 44 (57.1%) with incomplete MCA infarction (IMCA). The predictors of CMCA score included: >1/3 early hypodensity in computed tomography findings, hyperdense MCA sign, brain edema, initial National Institutes of Health Stroke Scale (NIHSS) score ≥17, and stroke in progression during the 1st 5 days of admission. The cutoff CMCA score was 2, with a sensitivity of 81.8% and specificity of 70.5%. Mortality score 1, used for predicting in-hospital mortality from LHI, included CMCA and CHADS2 scores ≥4 (sensitivity 100.0%, specificity 57.4%), and mortality score 2 included CMCA and CHADS2 scores ≥4, and NIHSS score ≥26, during the 1st 5 days (sensitivity 100.0%, specificity 91.7%).Patients qualifying for a mortality score of 2 were at high-risk of in-hospital mortality from LHI. These findings may aid in identifying patients who may benefit from invasive therapeutic strategies, and in better describing the characteristics of those at risk of mortality.