AIMS:Atrial fibrillation and flutter remain an important cause of morbidity in adults with atrial septal defect (ASD). This study aimed at investigating predictors for late (≥ 1 month after repair) atrial arrhythmia. METHODS:Patients who underwent ASD closure after the age of 18 years, were selected through the databases of three medical centres in Belgium. Preprocedural, periprocedural and follow-up data were extracted. Univariate and multivariate Cox-regression analysis was performed. Kaplan-Meier analysis was performed for any independent predictor of late atrial arrhythmia. RESULTS:A total of 155 patients (38 men and 117 women) was included. Twenty-four patients (median age 48.3 years, range 19.9-79.8) underwent surgical and 131 (median age 57.6 years, range 18.2-86.9) underwent transcatheter closure. Thirty-nine patients (25.2%) presented with late atrial arrhythmia. Male gender (P = 0.008), creatinine (P = 0.002), atrial arrhythmia before (P < 0.0001) and within 1 month after repair (P = 0.001) and a mean pulmonary artery pressure (mPAP) ≥ 25 mmHg (P < 0.0001) correlated with late atrial arrhythmia in univariate Cox-regression analysis. Multivariate analysis showed that mPAP ≥ 25 mm Hg (HR 3.72; 95%CI 1.82-7.59; P < 0.0001) and the presence of atrial arrhythmia before (HR 3.22; 95%CI 1.56-6.66; P = 0.002) and within 1 month after repair (HR 2.08; 95%CI 2.08-15.92; P = 0.001) were predictive of late atrial arrhythmia. Kaplan-Meier analysis showed that patients with a mPAP ≥ 25 mm Hg had a higher risk at developing late atrial arrhythmia (P < 0.0001). CONCLUSION:In patients with ASD type secundum, a mPAP ≥ 25 mmHg is an independent predictor of late atrial arrhythmia. The presence of pulmonary hypertension before repair should raise awareness for atrial arrhythmias and may be used to guide therapy.