Value of 3D-echocardiography to improve prediction of mitral valve reoperation in patients with congenital mitral valve disease

BACKGROUNDCongenital mitral valve disease (CMVD) presents major challenges in medical and surgical management. Data on predictors of mitral valve (MV) reoperation and value of 3D echocardiography (3DE) in this context are currently limited. Aim of our study was to identify predictors of MV reoperation and to investigate the value of 3DE in risk stratification. METHODSAll children <18 years old who underwent MV reconstruction for CMVD from 2002 to 2018 were included. Pre and postoperative 2D-echocardiogram (2DE) and 3DE data were collected. Competing risks and Cox regression analyses were used to identify independent predictors of MV reoperation. Receiver operating characteristics curve (ROC) analysis was used to assess the predictive values of 3DE versus 2DE. Decision tree analysis was used to compute decision algorithms. RESULTSOver the course of the study period, 206 children underwent MV reconstruction for CMVD: 105 had mitral stenosis (MS), 75 had mitral regurgitation (MR), and 26 had mixed disease (MD). Of them, 64 (31%) required a MV-reoperation. At multivariable analysis, age <1 year (hazard ratio [HR]=2.65; 95% confidence interval [CI]:1.13-6.21), tethered leaflets (HR=2.00; 95% CI:1.05- 3.82;), moderate or greater 2DE postoperative MR (HR=4.26; 95% CI:2.45-7.4), as well as changes in 3D-effective orifice area (3D-EOA) and in 3D-vena contracta regurgitant area (3D-VCRA) were all found to be independent predictors of MV reoperation. By ROC analysis, changes in 3D-EOA and 3D-VCRA were found to have significantly higher predictive value than changes in mean gradients (AUC=0.847 vs AUC=0.676, P=0.006) and 2D-VCRA (AUC=0.969 vs AUC=0.720, P=0.012), respectively. Decision-tree analysis found that a <30% increase in 3D-EOA had 80% accuracy (HR=8.50; 95% CI:2.9-25.1) and a <40% decrease in 3D-VCRA had 93% accuracy (HR=22.50; 95% CI:2.9-175) in predicting MV reoperation for stenotic and regurgitant MV, respectively. CONCLUSIONSAge <1 year, tethered leaflets, 2DE postoperative MR, changes in 3D-EOA and 3D-VCRA were all independent predictors of MV-reoperation. 3DE parameters had higher predictive value than 2DE. 3DE-based decision-tree algorithms may help risk stratification and serve as a support tool for clinical decision-making in patients with CMVD. CLINICAL PERSPECTIVESO_ST_ABSWhat is new?C_ST_ABSO_LIIn congenital mitral valve (MV) disease patients undergoing MV reconstruction, changes in 3D effective orifice area (3D-EOA) and 3D vena contracta regurgitant area (3D-VCRA) were found to be independent predictors of MV reoperation. C_LIO_LIChanges in 3D effective orifice area (3D-EOA) and 3D vena contracta regurgitant area (3D-VCRA) had significantly higher predictive value compared to 2D-echocardiography parameters, i.e. changes in mean gradient, 2D-VCRA and moderate or greater postoperative 2D mitral regurgitation. C_LIO_LIDecision tree analysis identified cut-off points in changes of 3D-EOA and 3D-VCRA to predict MV reoperation. C_LI What are the clinical implications?O_LIChanges in 3D-EOA and 3D-VCRA can be measured in the operating room to consider return to bypass, or at discharge, to help inform need for early follow-up and MV reoperation. C_LIO_LI3DE-based decision tree predictive algorithms may help in risk stratification and serve as a support tool to facilitate clinical decision-making. C_LI