Predictive Scoring System for Spontaneous Closure of Infant Ventricular Septal Defect: The P-VSD Score

Abstract

Ventricular septal defect (VSD) is a common congenital heart disease. However, consensus on the utility of echocardiography in predicting spontaneous closure (SC) of VSD remains lacking. This study aimed to identify and validate significant predictors of SC through a predictive scoring system. This retrospective study included medical records of 712 echocardiography instances performed on 304 patients diagnosed with VSD from 2016 to 2020 in their first year of life. A novel scoring system for predicting the SC of VSD was developed and validated using another dataset from different hospitals. Of the 304 patients, 215 (70.7%) had perimembranous (PM) VSDs and 89 had muscular (29.3%) VSDs. The median follow-up periods were 36.2 (interquartile range [IQR], 13-59) months and 13.7 9 (IQR, 5-37.4) days for PM and muscular VSDs, respectively. The overall SC rate during follow-up was 29.3%. Pulmonary hypertension (HTN), concomitant left ventricle (LV)-right atrium (RA) shunt, VSD size to aortic valve (AV) annulus size ratio, and left ventricular end-diastolic dimension (LVEDD) z-score were significant risk factors affecting SC of VSD. The "P-VSD" score, a new scoring system, demonstrated an area under the curve for predictability of 0.769. Pulmonary HTN, concomitant LV-RA shunt, LVEDD z-score, and VSD size-to-AV annulus size ratio at diagnosis were significantly associated with non-SC VSD after infancy. The P-VSD score can predict the SC of VSD in clinical settings and simplify the identification and appropriate management of high-risk patients.