P5662Development of a prognostic model for prevalent atrial fibrillation using individual patient data: Results of CATCH ME

Abstract Background/Introduction Atrial fibrillation (AF) can be challenging to diagnose due to asymptomatic and paroxysmal presentation. Identifying prognostic factors of AF would elucidate potential mechanisms causing AF and refine screening for at risk patients. Purpose To identify the main predictors of AF and to develop a prognostic model for prevalent AF. Methods Data of 120 potential predictors were harmonised in individual patient data from 4 independent European studies. A three stage Delphi expert consensus process identified predictors based on clinical knowledge. The predictors were further reduced using statistical selection (backward elimination), and a logistic regression model was fitted. We calculated odds ratios (OR) for each of the selected predictors and evaluated model performance using the C-statistic. Results Overall, 2420 patients (mean [standard deviation] age = 62.7 [14.5] years, 35.6% female, 43.1% with AF) were included in the analysis. Thirty-one potential predictors identified from the Delphi process which had sufficient data across all datasets were modelled. Of these 14 were deemed prognostic in predicting AF (age, sex, BMI, height, hypertension, diabetes, history of coronary artery disease, left atrial volume, left ventricular end systolic diameter, abnormality on echo, tricuspid valve disease of at least moderate intensity, aldosterone-antagonists, beta-blockers and P2Y12 blockers; see Figure 1). There was a clear interaction between age and sex indicating that males are at higher risk than females early in life, while females are at increased risk of AF at older age (Figure 1). The risk prediction model combining these prognostic factors performed well (C-statistic 0.79; 95% CI 0.77–0.81). Figure 1. (a) Forest plot; (b) Interaction Conclusion(s) Our preliminary analysis identified important prognostic factors and a complex relationship between age and sex, which predicts prevalent AF, highlighting the different potential causes of AF in different patients. There is a clear need to validate these factors in external datasets and for further investigation into the molecular mechanism underlying these factors. Acknowledgement/Funding European Commission H2020 framework