Left ventricular end-diastolic diameter to body surface area ratio as a continuous risk marker in left ventricular non-compaction with a dilated phenotype: a multicenter cross-sectional study

Dilated cardiomyopathy (DCM) is a risk factor for poor outcomes in patients with left ventricular non-compaction (LVNC). This study aims to identify key factors associated with the DCM phenotype in LVNC and develop a risk stratification model. This multicenter cross-sectional study included 448 patients with imaging-confirmed LVNC. Multivariable logistic regression was used to identify factors independently associated with the DCM phenotype in LVNC. Risk stratification was performed using tertile grouping, and restricted cubic spline (RCS) models were applied to analyze the dose-response relationship with DCM risk. Extensive subgroup analyses and interaction tests were conducted to evaluate the robustness and consistency of the associations. In both children and adult LVNC cohorts, patients with DCM showed significant cardiac structural remodeling and functional deterioration. The most common pattern of non-compacted myocardial involvement was both the apex and the left ventricular lateral wall (LVLW). Multivariable logistic regression identified left ventricular end-diastolic diameter (LVEDD) as a factor independently associated with the DCM phenotype in LVNC (overall cohort OR = 1.065, P = 0.001), with an area under the curve (AUC) of 0.694 in receiver-operating characteristic (ROC) analysis for LVEDD. After body surface area (BSA) adjustment, tertile stratification of LVEDD/BSA (T1: <30.89 mm/m²; T2: 30.89–37.27 mm/m²; T3: >37.27 mm/m²) showed a significant stepwise increase in DCM risk (T3 vs. T1: OR = 4.242, P = 0.001). ROC analysis for LVEDD/BSA showed an AUC of 0.817, indicating superior predictive performance compared to LVEDD alone. RCS analysis after adjustment indicated a continuous positive correlation between LVEDD/BSA and the odds of DCM without a clear threshold effect. Subgroup analyses confirmed consistent associations across populations with different clinical characteristics (all P-interaction > 0.05). LVEDD/BSA is a strong continuous risk marker that quantifies the severity of the dilated phenotype in LVNC. The tertile stratification system based on LVEDD/BSA can help identify high-risk patients early, facilitating a shift from binary diagnosis to continuous risk monitoring in LVNC management.