P2.012: Maximum ASD size Indexed to Total Septal Length Outperforms Expert Opinion in Predicting Future Candidacy for Percutaneous ASD Closure

Background: Secundum atrial septal defects (ASDs) can be closed surgically or percutaneously. Traditionally, surgery occurs between 2-5 years of age (1). Early surgery < 2 years allows for a “mini sternotomy” or thoracotomy approach (2) but requires determining interventional approach before the patient is large enough for device closure. Thus, predictive tools in assessing future device candidacy are needed.

Methods: Secundum ASD closures after 3 years of age from 2012 to 2023 at a single center were reviewed. “Early” (12-24 months) and “pre-intervention” (most recent within eleven months of closure) echocardiogram pairs were independently reviewed by echocardiographers who performed standardized measurements. The catheterization team then made binary recommendations on device closure candidacy. Reviewers were blinded to the paired echocardiogram and eventual interventions. Given evolving device options, the primary outcome was perceived device candidacy based on the pre-intervention echocardiogram rather than actual therapy received. The predictive utility of early catheterization team’s device candidacy recommendation was assessed using Cohen’s kappa, sensitivity, and specificity. The association of early echocardiogram measurements with eventual device candidacy was assessed. Significant predictors were included in a multivariable model. Absolute and indexed ASD sizes were considered as candidate predictors. Receiver operating characteristic curves and optimal thresholds using Youden’s method were determined.

Results: Fifty-five subjects with paired echocardiograms were included. Based on early echocardiogram, 34/55 (62%) were considered eventual device candidates. Based on pre-intervention echocardiogram occurring at a median of 4.0 years (IQR 3.3-4.8 years), 37/55 (67%) subjects were deemed device candidates (Figure 1). There was moderate agreement in perceived device candidacy between echocardiograms (kappa = 0.44), corresponding to a sensitivity/specificity of 78% / 72%. Smaller maximum ASD size indexed to total septal length (ASD/TSL, p<.001), smaller atrioventricular valve rim (p=.001), and larger posterior rim (p =.011) on early echocardiogram were associated with eventual perceived device candidacy. In the multivariable model, only ASD/TSL remained significant (p < 0.001). Early ASD/TSL was an excellent predictor of future perceived device candidacy (Area Under the Curve = 0.92), corresponding to a sensitivity/specificity of 89% / 89% at the optimal threshold of 0.46 (Figure 2).

Conclusion: ASD/TSL between 1 and 2 years of age was strongly predictive of eventual perceived candidacy for device closure, outperforming expert consensus. Patients with ASD/TSL < 0.46 on early echocardiogram are likely to be device candidates later in life. Further research is needed to validate this finding in a larger multi-center cohort.

References:
1. Driscoll et al (1994). Guidelines for evaluation and management of common congenital cardiac problems in infants, children, and adolescents. American Heart Association. Circulation, 90(4), 2180-2188.

2. Bichell, D et al (2000). Minimal access approach for the repair of atrial septal defect: the initial 135 patients. The Annals of thoracic surgery, 70(1), 115-118.