Quantitative vectorcardiographic criteria for the differentiation between atrial septal defects of primum and secundum types.

From a population of 304 patients who were operated on for atrial septal defects, 25 patients with primum defects and 78 patients with secundum defects were studied with regard to the ability of various quantitative axial lead vectorcardiographic (VCG) data to predict the anatomical type of defect. The patients included all in whom there had been preoperative doubt concerning the classification. By applying the electrocardiographic (ECG) mean frontal QRS axis, 8 patients were erroneously classified, and 14 could not be classified because no definite QRS resultant was present in the frontal plane (diagnostic performance index 0.85). Several VCG criteria showed a better performance, the best being sums of 10 msec Y lead amplitudes from 20 to 50 msec after QRS onset. These criteria, which reflect the superior shift of the middle part of the QRS complex in primum defects, reduced the number of misclassifications to a minimum of 4 (performance 0.96). A combination of this criterion with one based on the duration of the initial inferior QRS vectors gave some further improvement. This simple combination was not surpassed by classification performed by a stepwise discriminant analysis computer program (BMD 07M).