Abstract 12985: Addition of Smoking Status Improves Risk Prediction of Ischemic Stroke in Atrial Fibrillation: The Atherosclerosis Risk in Communities (ARIC) Study
Background: Atrial fibrillation (AF) is associated with increased risk of ischemic stroke and death. Current risk scores for stroke in AF do not include modifiable lifestyle risk factors. We studied whether smoking, heavy alcohol consumption or high body mass index (BMI) are associated with increased risk of incident ischemic stroke or cardiovascular (CV) death and whether addition of these factors can improve risk prediction of ischemic stroke over and above the CHA2DS2-VASc score in participants with incident AF in the ARIC study.
Methods: We included participants who developed AF after the baseline exam, who were not on warfarin, and had no prior stroke at the time of AF diagnosis. Incident AF was ascertained from hospitalization discharge codes and study ECGs. Incident ischemic stroke and CV death were physician-adjudicated using data from community surveillance, hospitalization discharges, and death certificates. Cox model was used to study the association. C statistic, the net reclassification improvement (NRI) and relative integrated discrimination improvement (IDI) for 5-year risk prediction were calculated.
Results: Among 1222 participants (mean age: 63.4 ± 6.2 yrs; 43 % female; 18.9% black) with incident AF, during mean follow-up of 6.9 yrs, there were 332 incident ischemic stroke and CV deaths. Compared with never smokers, current smokers had higher incidence of ischemic stroke and CV death (Table). Addition of current smoking improved model discrimination marginally: C-statistic (95% CI) increased from 0.676 (0.604-0.747) to 0.691 (0.623-0.758). NRI (95% CI), 0.134 (0.020 to 0.253) and relative IDI (95% CI), 0.331 (0.181 to 0.487) were statistically significant for current smoking.
Conclusions: Current smoking is associated with an increased risk of stroke and CV death in AF and modestly improves risk prediction of stroke. Further research is needed to evaluate whether smoking status should be considered in prediction of those at high risk for stroke in AF.
Author Disclosures: Y. Kwon: None. F.L. Lopez: None. S.K. Agarwal: None. E.Z. Soliman: None. G.Y. Lip: None. A. Alonso: None. L.Y. Chen: None.
- © 2014 by American Heart Association, Inc.