• adult congenital heart disease
• cardiac magnetic resonance
• risk score
• structural equation modeling
• tetralogy of Fallot

BACKGROUND. Tetralogy of Fallot (TOF) is the commonest cyanotic congenital heart disease (CHD), and managing adults who underwent corrective surgery remains challenging. Although several studies have identified various risk factors with good sensitivity, strong predictors of adverse events are still lacking.
AIM. We sought to identify predictors of adverse events in a cohort of repaired TOF (rTOF) patients, and we elaborated a risk score using Structural Equation Modelling (SEM).
METHODS. From a registry promoted by the Cardiac Magnetic Resonance (CMR)/Computed Tomography Working Group of the Italian Society of Pediatric Cardiology and CHD, we included in this study consecutive patients with rTOF who underwent CMR for clinical indications in a single CHD center. Clinical and surgical history, physical examination, cardiopulmonary exercise test (CPET) and echocardiography data were extracted from medical records. Any adverse event (AE) was recorded during follow-up. We defined the primary outcome as the composite of sudden cardiac death, sustained ventricular arrhythmias and appropriate defibrillator shocks, whereas atrial arrhythmias, non-sustained ventricular tachycardias, and heart failure-related hospitalization were included in the secondary outcome. The risk score (rTOF SEM score) was elaborated with SEM. In SEM methodology, hypothesized structural relationships between variables are defined as causal “paths”, then the path analysis estimates the magnitude and statistical significance of each association, and creates a latent variable which, in this case, reflected the global disease severity. Finally, the latent variable was translated into a multiparametric risk score, ranging from 0 to 10.
RESULTS. 412 patients with rTOF (age: 28.4 ± 13.8 years, 61.4% male) composed the study cohort. During a median follow-up of 19 months (interquartile range 3.0-64.3), the primary outcome occurred in 17 patients (4.2%), whereas 75 patients (18.7%) experienced at least a secondary outcome. The multivariate Cox regression model showed that the main predictors of any AE were age at CMR and right atrial end-systolic volume index (RAESVi) (HR 1.12, 95% CI 1.02-1.28, p=0.009 and HR 1.07, 95% CI 1.03-1.12, p<0.001, respectively). The goodness of fit of the “rTOF SEM score” was acceptable, according to pre-specified cut-offs (Standardized Root Mean Square Residual=0.089 and Root Mean Square Error of Approximation=0.119). At SEM analysis, the highest weighting factors for the latent variable were: age at repair, age at CMR, RAESVi, NT-proBNP plasmatic concentrations and peak VO2 during CPET. ROC curve identified a “rTOF SEM score” cut-off of 2.5 (AUC 0.76, 95% CI 0.70-0.82, p<0.001) with a sensitivity of 83% and a specificity of 60% for predicting any AE, of 2.77 (AUC 0.69, 95% CI 0.58-0.80, p=0.009) with a sensitivity of 82% and a specificity of 58% for predicting the primary outcome, and of 2.47 (AUC 0.74, 95% CI 0.69-0.80, p<0.001) with a sensitivity of 85% and a specificity of 57% for predicting the secondary outcome. The freedoms from primary and secondary outcomes at Kaplan-Meier analysis according to the “rTOF SEM score” cut-off were statistically significant (log rank 13.8; p<0.001 and log rank 58.7; p<0.001, respectively).
CONCLUSIONS. The “rTOF SEM score” elaborated in this study was a strong predictor of adverse events in our cohort of patients. A validation of this prognostic score in a larger population is warranted to confirm our results and define its role for risk stratification of patients with rTOF.