Methodological discordance between apical four-chamber and biplane Simpson’s method for left ventricular ejection fraction: a retrospective study of a credentialed echocardiographic dataset

BACKGROUND: Left ventricular ejection fraction (LVEF) remains central to heart failure phenotyping and device-based decision-making, yet the degree to which apical four-chamber (A4C) and biplane Simpson measurements diverge at clinically actionable thresholds is not well defined. METHODS: We analysed 1,022 unique algorithmically derived echocardiographic studies from 784 patients in the credentialed MIMIC-IV-ECHO-Ext-LVVOLUMES-A4C-ROI resource. Each study contained paired A4C and biplane volumetric labels derived from the same annotated DICOM sequence. Discordance was defined primarily at the HFrEF threshold (LVEF < 40%). Agreement was assessed with Bland-Altman analysis, and independent predictors were evaluated using multivariable logistic regression with cluster-robust standard errors. RESULTS: LVEF discordance at the HFrEF threshold occurred in 48 of 1,022 studies (4.7%, 95% CI 3.5-6.2%). At the ICD threshold (LVEF < 35%), discordance was present in 32 studies (3.1%). In the prespecified borderline zone (A4C LVEF 35-45%; n = 81), discordance rose to 30.9% (95% CI 21.9-41.6%). Mean bias was 0.11%, but the 95% limits of agreement were wide (- 13.5% to + 13.7%). LV end-diastolic volume was the only independent predictor of discordance (OR 1.61 per SD, 95% CI 1.27-2.05; p = 0.0001), and this association persisted after adjustment for acquisition variables. CONCLUSIONS: Discordance between A4C and biplane Simpson LVEF is uncommon across an unselected cohort but becomes frequent near therapeutic cut-offs. LV dilatation is the dominant driver. These findings support continued preference for biplane quantification when the ventricle is enlarged or the measured LVEF falls near a treatment threshold.