Prognostic Relevance of Speckle Tracking-Derived Biatrial Stiffness Index in Patients with Dilated Cardiomyopathy

Background: Atrial stiffness can be estimated non-invasively using speckle-tracking echocardiography (STE) and has recently emerged as an outcome predictor. We aimed to assess left atrial (LA) and right atrial (RA) phasic function; the LA stiffness index (LASI), the RA stiffness index (RASI) and their sum; and the biatrial stiffness index (BASI) in dilated cardiomyopathy (DCM), and to test whether combining the two atria adds prognostic information over either index alone. Methods: A total of 121 patients with non-ischaemic DCM in sinus rhythm were followed prospectively for a composite endpoint of all-cause death, non-fatal cardiac arrest, or hospitalisation for heart failure decompensation. LASI was defined as the mitral E/e′ ratio divided by LA reservoir strain, RASI as the tricuspid Et/e′t ratio divided by RA reservoir strain, and BASI as the sum of the two. Cox models were adjusted for NYHA class, LV ejection fraction (LVEF), maximal LA volume (LAVmax) and pulmonary artery systolic pressure (PASP). Results: After 19 ± 11 months, 55 patients reached the endpoint. LA reservoir and contraction strain, all three components of RA strain and all three stiffness indices were significantly impaired in patients with events. All stiffness indices were independent outcome predictors in multivariable Cox regression (HR 2.79 [95% CI, 1.35–5.75], p = 0.006 for LASI, HR 1.84 [95% CI, 1.02–3.29], p = 0.04 for RASI and HR 2.74 [95% CI, 1.36–5.51], p = 0.005 for BASI). BASI showed the greatest increase in risk prediction (Δ likelihood ratio χ2 test = 10.3, p = 0.001) over NYHA class, left ventricular ejection fraction, LA maximal volume and pulmonary artery systolic pressure. BASI showed the highest discrimination (AUC = 0.73); however, it was not significantly better than LASI or RASI alone. Conclusions: Left and right atrial stiffness are both associated with adverse outcome in DCM and add prognostic information to an LV-centred risk model. Their unweighted sum performs at least as well as either component and offers a single parsimonious measure, whose incremental clinical value remains to be established in larger, externally validated cohorts.