Abstract 1254: Development of a Risk Score for Incident Atrial Fibrillation in the Community; the Framingham Heart Study

Atrial fibrillation (AF) contributes to increased morbidity and mortality. An easily applicable AF risk score for incidence of AF in the community is unavailable. We examined 8044 person-exams among subjects who were 46 –95 years of age and attended routine Framingham Heart Study (FHS) exams (n=4764 individuals; mean age 61 years, 55% women). Multivariable Cox regression related clinical data to 10-year AF incidence (n=457). Secondary analyses incorporated routine echo data (n=7156, 445 events) for left atrial size, left ventricular wall thickness, and fractional shortening. Age, sex, body mass index, significant murmur, systolic blood pressure, hypertension treatment, PR interval, and heart failure were associated with incident AF (C-statistic 0.78; 95% CI 0.76 – 0.80; p<0.05). Predicted 10-year risks of AF were similar to observed risks (calibration Chi 2 statistic=4.16; p=0.90). In a later FHS sample, additional incorporation of echo variables slightly increased the C-statistic from 0.78 (0.75– 0.80) to 0.79 (95% CI 0.77– 0.82), p=0.005. Reclassification by echo variables was only modest (integrated discrimination improvement of 0.021; p=0.003). We provide a point score (and excel spreadsheet) for incident AF. The figure displays AF risk by selected risk factor levels. The FHS AF risk profile stratifies individuals in the community, and may provide a framework to evaluate new biological or genetic markers for AF risk prediction. Further studies are needed to validate the risk score in other populations. Predicted 10-Year Risk of AF by Risk Factors & Sex