Can Prediction of Functional Outcome after Stroke Be Improved by Adding Fibrinogen to Prognostic Model?

Abstract

Several studies demonstrated that elevated plasma fibrinogen level is associated with poor functional outcome after ischemic stroke. It remains, however, unclear if prediction of functional outcome after stroke can be improved by adding fibrinogen to prognostic model. We aimed to determine the prognostic value of plasma fibrinogen for the prediction of functional outcome after ischemic stroke. We retrospectively analyzed the data of 727 ischemic stroke patients (median age: 70; 48% men). The functional outcome was assessed 1 month after stroke onset using modified Rankin Scale. Unfavorable outcome was defined as modified Rankin Scale score higher than 2. Using C-statistic and reclassification measures (net reclassification improvement-NRI and integrated discrimination improvement-IDI), we compared the predictive abilities of 2 models. The first model contained stroke severity and age, and the second one included fibrinogen in addition. After adjusting for age and stroke severity, plasma fibrinogen level higher than 2.95 g/L was associated with unfavorable outcome (OR: 1.80, 95% CI: 1.20-2.72, P < .01). The addition of fibrinogen did not lead to an improvement in predictive ability of the model. C-statistic did not differ between models (.90 versus .90, P = .34). The categorical NRI was .01 (P = .66) and the IDI was .006 (P = .08). The addition of fibrinogen to predictive model including age and stroke severity does not improve discrimination between favorable and unfavorable outcomes after ischemic stroke.