Objective-To review the outcomes of transcatheter closure of atrial septal defects using the Cardio-Seal implant. Design-A prospective interventional study. Setting-Tertiary referral centre. Patients-The first 50 patients (median age 9.7 years) who underwent attempted percutaneous occlusion. Interventions-Procedures were done under general anaesthesia and transoesophageal guidance between December 1996 and July 1998. Main outcome measures-Success of deployment, complications, and assessment of right ventricular end diastolic diameter, septal wall motion, and occlusion status by echocardiography. Results-The median balloon stretched diameter was 14 mm. Multiple atrial septal defects were present in 11 patients (22%) and a deficient atrial rim (< 4 mm) in 19 (38%). In four patients (8%), a second device was implanted after removal of an initially malpositioned first implant. There were no significant immediate complications. All patients except one were discharged within 24 hours. At the latest follow up (mean 9.9 months) a small shunt was present in 23 patients (46%), although right ventricular end diastolic dimensions (mean (SD)) corrected for age decreased from 137 (29)% to 105 (17)% of normal, and septal motion abnormalities normalised in all but one patient. No predictors for a residual shunt were identified. Supporting arm fractures were detected in seven patients (14%) and protrusion of one arm through the defect in 16 (32%), the latter being more common in those with smaller anterosuperior rims. No untoward eVects resulted from arm fractures or protrusion. There were no complications during follow up, although five patients (10%) experienced transient headaches. Conclusions-The implantation of the Cardio-Seal device corrects the haemodynamic disturbances secondary to the right ventricular volume overload, with good early outcome. (Heart 2000;84:320-326)