Defect size is an essential determinant of shunt severity in children with ventricular septal defect (VSD). Three-dimensional echocardiography (3DE) can better estimate VSD dimensions than two-dimensional echocardiography (2DE).

To assess the accuracy of VSD area obtained by 3DE to predict shunt severity, and to compare with VSD diameters obtained from both 2DE and 3DE in children with isolated VSD.

Children with isolated VSD were prospectively included. Patients with multiple VSDs or with other extracardiac anomalies were excluded. Shunt severity was evaluated according to the presence of volume overload and the level of pulmonary arterial pressure as mild, moderate, and severe shunt. Two orthogonal diameters of VSD were measured at end diastolic frame using 2DE and then 3DE in each patient. Systolic (sVSDA) and diastolic (dVSDA) VSD areas, and systolic aortic valve (AVA) areas were also measured using a multi-planar reformatting (MPR) mode from 3DE.

Sixty patients were included. Mean age was 20.1±27.7 month old. VSDs were muscular in 20 patients (30%) and membranous in 40 patients (70%). There were 24 VSDs with mild, 21 with moderate, and 15 with severe shunt. VSD dimension and shunt severity were not influenced by the anatomical type of the defect. VSD areas were better predictors of shunt severity than VSD diameters. The best predictor of shunt severity was found to be the sVSDA/AVA ratio with a cutoff >0.33 for the prediction of severe shunt with a 93.3% sensitivity and a 95.2% specificity (Figure 1).

3DE shows great interest for the evaluation of VSD shunt severity. The use of the VSD-systolic-area-to-aortic-valve-area ratio seems to be the best predictor of shunt severity in children with isolated VSD.