Development of a Risk Score for Atrial Fibrillation in Adults With Diabetes Mellitus (from the ACCORD Study)

Abstract

We aimed to develop a novel risk score predicting 5-year atrial fibrillation (AF) risk for diabetes mellitus (DM) patients. We included subjects from the Action to Control Cardiovascular Risk in Diabetes study cohort without AF at baseline. Potential risk factor and demographic predictors were collected at baseline and incident AF was defined from ECG during follow-up. A 5-year risk score for incident AF was developed using Cox regression with internal validation. We studied 9,240 subjects with DM (62% male, mean age 62.6 years) of which 1.8% (n = 165) developed AF over a median follow-up of 4.9 years. Subjects developing AF were more likely male, of white ethnicity and with more obesity and poorer kidney function, but with lower diastolic blood pressure and low density-lipoprotein cholesterol. In the risk prediction model, age, gender, race, body mass index, heart failure, diastolic blood pressure, triglycerides, hemoglobin A1c, duration of DM, serum creatinine and hypertension medication were included as important predictors. The Harrell's C-statistic was 0.79 with excellent internal calibration (goodness-of-fit test p = 0.99 and calibration slope = 1.01). Our risk model may be useful for assess future AF risk in DM patients.