Despite the rising incidence of heart failure (HF), there is no validated prediction model for incident HF. Each of the LA and LV parameters taken separately has a prognostic value to predict incident HF. Based on the close physiological interaction between LA and LV, the simultaneous assessment of the left atrioventricular coupling could be even more accurate to predict incident HF. However, no data has been described on the existence of such an atrioventricular index.

This study aimed to investigate the prognostic value of a novel left atrioventricular functional coupling index(LAFCI) and to assess its predictive value for incident HF occurrence.

A total of 4,124 participants in the Multi-Ethnic Study of Atherosclerosis underwent a cardiac MRI study. The LAFCI by cardiac MRI was defined by the ratio of the indexed LA end-diastolic volume divided by the LV global function index(LVGFI) already described. Cox proportional hazard models were constructed to predict the occurrence of HF. In univariable and multivariable Cox analysis, the association between LAFCI or all other LA and LV variables and time-to-event was analyzed, adjusting for traditional HF risk factors and biomarkers.

A total of 231(5.6%) patients experienced HF during a median(IQR) follow-up period of 15.6(12.7–16.3) years. In univariable analysis, LAFCI was associated with incident HF (HR 2.23, 95%CI [2.02–2.34] P<0.0001). In Cox multivariate analysis, LAFCI had a significant independent predictive value for incident HF (HR 1.58, 95%CI [1.47–1.70] P<0.001). Regarding the inter-models comparison to predict the occurrence of HF, C-statistic for LAFCI(0.73) was greater than LAVI_min(0.70), LVGFI(0.66), LA strain(0.61), LV MVR(0.61) or LVEF(0.61). LAFCI showed better predictive performance than the multivariate model (C-statistic: 0.84 versus 0.81) (Fig. 1).

The LAFCI is an independent predictor of incident HF in this large multi-ethnic cohort.