Development and prognostic validation of partition values to grade right ventricular dysfunction severity using 3D echocardiography

Aims Transthoracic 3D echocardiography (3DE) has been shown to be feasible and accurate to measure right ventricular (RV) ejection fraction (EF) when compared with cardiac magnetic resonance (CMR). However, RV EF, either measured with CMR or 3DE, has always been reported as normal (RV EF > 45%) or abnormal (RV EF <_ 45%). We therefore sought to identify the partition values of RV EF to stratify RV dysfunction in mildly, moderately, or severely reduced as we are used to do with the left ventricle. Methods and results We used 3DE to measure RV EF in 412 consecutive patients (55 +/- 18 years, 65% men) with various cardiac conditions who were followed for 3.7 +/- 1.4 years to obtain the partition values which defined mild, moderate, and severe reduction of RV EF (derivation cohort). Then, the prognostic value of these partition values was tested in an independent population of 446 patients (67 +/- 14 years, 58% men) (validation cohort). During follow-up, we recorded 59 cardiac deaths (14%) in the derivation cohort. Using K-Adaptive partitioning for survival data algorithm we identified four groups of patients with significantly different mortality according to RV EF: very low > 46%, 40.9% < low <= 46%, 32.1% < moderate <= 40.9%, and high <= 32.1%. To make the partition values easier to remember, we approximated them to 45%, 40%, and 30%. During 4.1 +/- 1.2 year follow-up, 38 cardiac deaths and 88 major adverse cardiac events (MACE) (cardiac death, non-fatal myocardial infarction, ventricular fibrillation, or admission for heart failure) occurred in the validation cohort. The partition values of RV EF identified in the derivation cohort were able to stratify both the risk of cardiac death (log-rank = 100.1; P < 0.0001) and MACEs (log-rank = 117.6; P < 0.0001) in the validation cohort too. Conclusion Our study confirms the independent prognostic value of RV EF in patients with heart diseases, and identifies the partition values of RV EF to stratify the risk of cardiac death and MACE.