Stretched diameter of the atrial septal defect (ASD), determined by balloon sizing at cardiac catheterization, is commonly used to select the sizes of the devices used for transcatheter closure of the secundum ASD. We have previously evaluated the utility of pulmonary/systemic flow ratio and angiographic and echocardiographic (echo) sizes of the ASD in estimating stretched ASD diameter in a group of 16 patients and determined that echo diameter had the best correlation with stretched diameter (r = 0.82; p less than 0.001). The stretched diameter can be estimated: 1.05 x echo diameter in millimeters + 5.49. In this study we have prospectively evaluated this formula in estimating the stretched ASD diameter by two-dimensional echo measurements obtained in two (long and short-axis) subcostal views in another group of 21 patients aged 2.5 to 29 years (median 4.5 years). The echo size of the ASD was 9.7 +/- 3.0 mm, whereas the measured stretched diameter was 15.3 +/- 4.0 mm. The predicted stretched ASD diameter was calculated according to the above formula and was 15.7 +/- 3.1 mm, not significantly different (p greater than 0.1) from the measured stretched diameter. The correlation between predicted and measured stretched ASD sizes was excellent (r = 0.9; p less than 0.001). The mean squared error was 2.4. The differences between measured and predicted values were within 2 mm in all but three patients. It is concluded that stretched ASD diameter can be estimated accurately by two-dimensional subcostal echo measurements, which in turn could be used for selection of device size for occlusion of the ASD.