Abstract
Background: Modern personalised medicine requires patient-tailored decisions. This is particularly important when considering pharmacological cardioversion for the acute treatment of haemodynamically stable atrial fibrillation and atrial flutter in a shared decision-making process. We aimed to develop and validate a predictive model to estimate the individual probability of successful pharmacological cardioversion using different intravenous antiarrhythmic agents. Methods: We analysed data from a prospective atrial fibrillation registry comprising 3053 cases of first-detected or recurrent haemodynamically stable, non-permanent, symptomatic atrial fibrillation presenting to an Austrian academic emergency department between January 2012 and December 2017. Using multivariable analysis, a prediction score was developed and externally validated. The clinical utility of the score was assessed using decision curve analysis. Results: A total of 1528 cases were included in the development cohort (median age 69 years, IQR 58–76; 43.9% female), and 1525 cases were included in the validation cohort (median age 68 years, IQR (58–75); 39.5% female). Finally, 421 cases were available for score development and 330 cases for score validation The weighted score included atrial flutter (8 points), duration of symptoms associated with AF (<24 h; 8 points), absence of previous electrical cardioversion (10 points), and the specific intravenous antiarrhythmic drug (amiodarone 10 points, vernakalant 11 points, ibutilide 13 points). The final score, the “Successful Intravenous Cardioversion for Atrial Fibrillation (SIC-AF) score,” showed good calibration (R2 = 0.955 and R2 = 0.954) and discrimination in both sets (c-indices: 0.68 and 0.66) and net clinical benefit. Conclusions: A predictive model was developed to estimate the success of intravenous pharmacological cardioversion using different antiarrhythmic agents in a cohort of patients with haemodynamically stable, non-permanent, symptomatic atrial fibrillation. External temporal validation confirmed good calibration, discrimination, and clinical usefulness. The SIC-AF score may help patients and physicians jointly decide on the appropriate treatment strategy for acute symptomatic atrial fibrillation. Registration: NCT03272620.