Another Score to Predict Ischemic Stroke Mortality?

Abstract

An 84-year-old man with a history of atrial fibrillation, congestive heart failure, and chronic renal insufficiency has a moderate to severe right middle cerebral artery infarct. He was not a candidate for intravenous or intra-arterial thrombolytics because of symptom onset on awakening and a completed infarct with no appreciable penumbra on computed tomography angiography or computed tomography perfusion. Following acute stabilization of patients like this man, the primary goals of care are to prevent complications and to begin rehabilitation. In addition, discussions with family regarding code status, end-of-life care preferences, and prognosis occur soon after admission. Clinicians generally rely on their anecdotal experience when these discussions occur, and these predictions have been shown to be relatively good. In one study of ischemic stroke patients, 90% of patients predicted to do poorly were dead or dependent at 6 months.1 However, clinicians tend to overestimate the likelihood of good outcome in stroke patients. In the same study, only 65% of ischemic and hemorrhagic stroke patients predicted by clinicians to survive 1 year and be independent were actually independent (defined as modified Rankin Score ≤2).1 Having a quantitative estimate of mortality or outcome in a patient such as this gentleman can help guide family discussions, since reliance solely on clinical prediction is insufficient. Article see p 739 Multiple prognostic models for mortality and functional status in stroke patients have been developed and validated. Simple models that include only age and National Institutes of Health Stroke Scale within the first 6 hours of symptom onset were shown to predict mortality up to 150 days poststroke.2 In this stroke population, mortality was predicted better by the models than by the treating physicians' …