Validation of Clinical Prediction Rules for Stroke: Results from the National Registry of Atrial Fibrillation (NRAF) Project

Abstract

38 Background: Patients who have atrial fibrillation (AF) have an increased risk of stroke, but the absolute rate of stroke depends on age and comorbid conditions. Several classification schemes have been proposed to estimate the risk of stroke in the AF population. We combined 2 existing classification schemes into a new stroke-risk index and then validated all 3 classifications schemes. Methods: We formed the new stroke-risk index, CHADS 2 by assigning 1 point for the presence of each of the following conditions-Congestive heart failure, Hypertension, Age greater than 75 years, Diabetes mellitus-and 2 points for a history of Stroke or transient ischemic attack. To test the 3 classification schemes we collaborated with Peer Review Organizations representing 7 states to assemble a National Registry of AF (NRAF) consisting of Medicare beneficiaries who have non-rheumatic AF. The NRAF dataset contained health outcomes based on Medicare part A claims data. Results: Of 1733 NRAF cohort members who had not been prescribed warfarin at the time of hospital discharge, 94 were readmitted for an ischemic stroke (stroke rate, 4.4 per 100 patient-years). As indicated by a c-statistic significantly greater that 0.5, the 2 existing classification schemes predicted stroke better than chance: 0.78 for the scheme developed by the Atrial Fibrillation Investigators and 0.82 for the SPAF 3 scheme. However, the new CHADS 2 index performed better, c-statistic = 0.87. The stroke rate per 100 patient-years without antithrombotic therapy increased by a factor of 1.5 (95% CI, 1.3–1.7) for each 1-point rise in the CHADS 2 score: 1.9 for CHADS 2 =0; 2.8 for CHADS 2 =1; 4.0 for CHADS 2 =2; 5.9 for CHADS 2 =3; 8.5 for CHADS 2 =4; 12.5 for CHADS 2 =5; and 18.2 for CHADS 2 =6. Conclusions: The 2 existing classification schemes and especially a new stroke-risk index, CHADS 2 , can quantify the risk of stroke for patients who have AF. Use of the CHADS 2 index should aid in the selection of antithrombotic therapy for patients who have non-rheumatic AF.