In a prospective cohort study of patient outcomes following stroke, William Whiteley and colleagues find that markers of inflammatory response are associated with poor outcomes. However, addition of these markers to existing prognostic models does not improve outcome prediction.