• Azmil H Abdul-Rahim, Terence J Quinn, Sarah Alder, Allan B Clark, Stanley D Musgrave, Peter Langhorne, John F Potter, and Phyo Kyaw Myint.
• From the Institute of Cardiovascular and Medical Sciences, University of Glasgow, Glasgow, United Kingdom (A.H.A.-R., T.J.Q., P.L.); Aberdeen Royal Infirmary, NHS Grampian, Aberdeen, United Kingdom (S.A.); Norwich Medical School, University of East Anglia, Norfolk, United Kingdom (A.B.C., S.D.M., J.F.P.); and Epidemiology Group, School of Medicine and Dentistry, University of Aberdeen, Aberdeen, United Kingdom (P.K.M.). Azmil.Abdul-Rahim@glasgow.ac.uk.
• Stroke. 2016 Jan 1; 47 (1): 74-9.

Background And PurposeThe stroke subtype, Oxfordshire Community Stroke Project classification, age, and prestroke modified Rankin (SOAR) score is a prognostic scale proposed for early mortality prediction after acute stroke. We aimed to evaluate whether including a measure of initial stroke severity (National Institutes of Health Stroke Scale and modified-SOAR [mSOAR] scores) would improve the prognostic accuracy.MethodsUsing Anglia Stroke and Heart Clinical Network data, 2008 to 2011, we assessed the performance of SOAR and mSOAR against in-hospital mortality using area under the receiver operating curve statistics. We externally validated the prognostic utility of SOAR and mSOAR using an independent cohort data set from Glasgow. We described calibration using Hosmer-Lemeshow goodness-of-fit test.ResultsA total of 1002 patients were included in the derivation cohort, and 105 (10.5%) died as inpatients. The area under the receiver operating curves for outcome of early mortality derived from the SOAR and mSOAR scores were 0.79 (95% confidence interval, 0.75-0.84) and 0.83 (95% confidence interval, 0.79-0.86), respectively (P=0.001). The external validation data set contained 1012 patients with stroke; of which, 121 (12.0%) patients died within 90 days. The mSOAR scores identified the risk of early mortality ranging from 3% to 42%. External validation of mSOAR score yielded an area under the receiver operating curve of 0.84 (95% confidence interval, 0.82-0.88) for outcome of early mortality. Calibration was good (P=0.70 for the Hosmer-Lemeshow test).ConclusionsAdding National Institutes of Health Stroke Scale data to create a modified-SOAR score improved prognostic utility in both derivation and validation data sets. The mSOAR may have clinical utility by using easily available data to predict mortality.© 2015 American Heart Association, Inc.

25 keywords...

hide…