Abstract 17273: Phenotypic Clustering of Patients With Heterotaxy Syndrome Predicts Outcomes of Mortality and Transplantation

Abstract

Background: Current classifications of heterotaxy syndrome (HS) (e.g. left/right isomerism, polysplenia/asplenia) do not capture the heterogeneity of cardiac anomalies observed nor provide insight on risk stratification. We aimed to improve these classifications using cluster analysis to identify anatomic phenotypes in HS patients. Methods: Using a retrospective database of all HS patients seen at Boston Children’s Hospital from 1985-2014 (n=264), we performed principal component and cluster analyses on 47 cardiac anatomic variables and compared rates of mortality/transplantation across clusters. Results: Five clusters were identified ranging in size from 19 to 91 patients. Cluster #1 had a high prevalence of right-dominant complete atrioventricular canal defects (CAVC), double-outlet right ventricle (DORV), pulmonary stenosis/atresia (PS/PA), and total anomalous pulmonary venous connection (TAPVC); Cluster #2 of left-dominant CAVC, DORV, and TAPVC; Cluster #3 of left-sided obstructive lesions and interrupted IVC; Cluster #4 of transposition of the great arteries with PS/PA; and Cluster #5 of interrupted IVC with miscellaneous cardiac anomalies (Table). Over 80% of patients in clusters 1 and 2 underwent single ventricle palliation compared with only 16% of patients in cluster 5. Transplant-free survival was lowest in clusters 1, 2, and 3 but differed by single versus biventricular repair (Figure). Conclusions: Cluster analysis of cardiac anatomic variables identified 5 phenotypes of HS including 3 higher risk and 2 lower risk phenotypes. These findings highlight the need for improved phenotyping of HS to account for the high degree of disease heterogeneity.