Development of the new risk score to predict occurrence of atrial fibrillation early after acute myocardial infarction

Background/Introduction New onset of atrial fibrillation (AF) is associated with adverse short- and long-term outcomes after acute myocardial infarction (AMI), and its prediction is relevant for the risk stratification in patients with AMI. Although several risk scores were developed for AF in the general population, there is no established risk score for AF occurrence after AMI. Purpose To develop a risk score to predict occurrence of AF early after AMI. Methods We enrolled consecutive 751 patients with AMI who admitted to our hospital between April 2006 and September 2012 for the present study. New occurrence of AF was defined as AF detected during hospital stay in a patient showing normal regular sinus rhythm at admission. Parameters relevant to the occurrence of AF was selected from the clinical characteristics, physical status and blood test data at admission, and peak CK/CK-MB, by stepwise logistic regression analysis. We constructed a risk score model to predict the new occurrence of AF, using selected parameters and their logistic regression coefficients. C-statistics was determined by constructing a receiver operating characteristic curve to evaluate the accuracy of the risk score for prediction of AF occurrence. Results We excluded 48 patients (6.4%) who had AF at admission, and 208 patients without sufficient data at admission, and thus, the study group consisted of 459 patients (age; 65±13 years, male gender; 79.6%). New AF occurrence was observed in 72 patients (14.5%). The following 7 parameters was selected as parameters related with AF (as p<0.1); Inferior/posterior AMI, use of β blockers, use of diuretics, single vessel disease, absence of reperfusion therapy, systolic blood pressure (sBP) at admission, and smoking. One point was given to sBP>128mmHg, 2 points to absence of reperfusion, and one point to other parameters. Sum of these points was calculated as the AF risk score (Table 1). AF occurred in 27.1% of patients with ≥5 points whereas it was observed 5.2% of those with <5 points. C-statistics of the risk score was 0.75 (95% CI 0.68–0.83). Conclusion We developed a novel risk score to estimate the risk of AF occurrence early after AMI, which can be a useful tool for the risk stratification after AMI.