Severity Scores for Ebstein Anomaly: Credibility and Usefulness of Echocardiographic vs Magnetic Resonance Assessments of the Celermajer Index

Abstract

Background: The severity score of Ebstein anomaly (EA) that corresponds to clinical status is still under research, with the Celermajer index (Cel-ind) being one of those. The agreement between echocardiographic and cardiac magnetic resonance (CMR) assessment of Cel-ind is not known. We determined the agreement between echocardiography- and CMR-derived Cel-ind and its relationship with heart failure markers.

Methods: A total of 37 unoperated adults with EA (mean age, 43.0 ± 14.4 years) underwent echocardiography, CMR, and cardiopulmonary tests. For the Cel-ind, end-diastolic areas in echocardiography or end-diastolic volumes in CMR were used according to the following formula: Cel-ind = (right atrium + atrialized right ventricle)/(functional right ventricle + left atrium + left ventricle). On the basis of this assumption, patients were classified as follows: grade 1 = Cel-ind < 0.5, grade 2 = 0.5 to 0.99, grade 3 = 1.0 to 1.49, grade 4 > 1.5. The agreement between echocardiographic and CMR was determined with the intraclass correlation coefficient or Cohen's kappa (<0.2 poor agreement; 0.2-0.4 fair agreement; 0.4-0.6 moderate agreement; 0.6-0.8 good agreement; 0.8-1.0 very good agreement).

Results: The median echoCel-ind was 0.9 (range, 0.4-2.3), and the median cmrCel-ind was 0.7 (range, 0.3-5.3). Grade 1 or 2 was found in 19 patients (51.3%) by echocardiography and in 27 patients (72.9%) by CMR. The agreement between imaging methods was only fair (kappa = 0.39, P = 0.002) for the 4-grade classification and moderate (intra-class correlation coefficient = 0.43; 95% confidence interval, 0.13-0.66) for Cel-ind calculation. Significant correlations between Cel-ind in CMR and cardiopulmonary parameters were found (for peak oxygen uptake: R = -0.35, P = 0.034; for the ventilation/carbon dioxide slope: R = 0.46, P = 0.005). Neither of them correlated with echocardiographic severity score.

Conclusions: The agreement between echocardiographic and CMR assessment of the Cel-ind is at most moderate; echocardiography usually overestimates, but rarely underestimates, EA severity. Cel-ind by CMR seems to be more valuable, because it is associated with heart failure markers.