Multiple biomarkers for the prediction of ischemic stroke: the PRIME study.

OBJECTIVE To simultaneously evaluate 14 biomarkers from distinct biological pathways for risk prediction of ischemic stroke, including biomarkers of hemostasis, inflammation, and endothelial activation as well as chemokines and adipocytokines. METHODS AND RESULTS The Prospective Epidemiological Study on Myocardial Infarction (PRIME) is a cohort of 9771 healthy men 50 to 59 years of age who were followed up over 10 years. In a nested case-control study, 95 ischemic stroke cases were matched with 190 controls. After multivariable adjustment for traditional risk factors, fibrinogen (odds ratio [OR], 1.53; 95% confidence interval [CI], 1.03-2.28), E-selectin (OR, 1.76; 95% CI, 1.06-2.93), interferon-γ-inducible-protein-10 (OR, 1.72; 95% CI, 1.06-2.78), resistin (OR, 2.86; 95% CI, 1.30-6.27), and total adiponectin (OR, 1.82; 95% CI, 1.04-3.19) were significantly associated with ischemic stroke. Adding E-selectin and resistin to a traditional risk factor model significantly increased the area under the receiver-operating characteristic curve from 0.679 (95% CI, 0.612-0.745) to 0.785 and 0.788, respectively, and yielded a categorical net reclassification improvement of 29.9% (P=0.001) and 28.4% (P=0.002), respectively. Their simultaneous inclusion in the traditional risk factor model increased the area under the receiver-operating characteristic curve to 0.824 (95% CI, 0.770-0.877) and resulted in an net reclassification improvement of 41.4% (P<0.001). Results were confirmed when using continuous net reclassification improvement. CONCLUSIONS Among multiple biomarkers from distinct biological pathways, E-selectin and resistin provided incremental and additive value to traditional risk factors in predicting ischemic stroke.