Prognostic Value of LV Global Longitudinal Strain by 2D and 3D Speckle-Tracking Echocardiography in Patients With HFpEF

BACKGROUND: In patients with heart failure with preserved ejection fraction (HFpEF), the impact of type 2 diabetes (T2D) on left ventricular global longitudinal strain (LV GLS) and its prognostic implications remains unclear. We aimed to evaluate LV function using two-dimensional (2D) and three-dimensional (3D) speckle-tracking echocardiography in patients with HFpEF with and without T2D, and to investigate its prognostic significance. METHODS: A total of 335 patients with HFpEF were prospectively enrolled for echocardiographic evaluation. LV GLS was obtained using 2D- and 3D-speckle-tracking echocardiography. Cox proportional hazards regression was used to determine predictors of adverse outcomes. The C-index, Akaike information criterion, integrated discrimination improvement, and net reclassification improvement were used to assess model performance and discriminative ability. RESULTS: LV 2D-GLS and 3D-GLS were impaired in patients with HFpEF and T2D compared with those without T2D. After a median follow-up of 17.6 months, 150 patients experienced adverse outcomes. Both 2D-GLS (hazard ratio, 1.323 [95% CI, 1.225–1.429]; P <0.001) and 3D-GLS (hazard ratio, 1.412 [95% CI, 1.316–1.515]; P <0.001) were independent predictors of adverse outcomes in patients with HFpEF after adjustment for confounders. The predictive accuracy of a model incorporating 3D-GLS (Akaike information criterion=−583.9, C-index=0.775 [95% CI, 0.742–0.808]) was superior to models using 2D-GLS (Akaike information criterion=−533.3, C-index=0.719 [95% CI, 0.678–0.760], ΔC-index=0.056; P =0.034) and LV ejection fraction (Akaike information criterion=−498.9, C-index=0.659 [95% CI, 0.610–0.708], ΔC-index=0.116; P <0.001). The addition of 2D-GLS and 3D-GLS to the base model significantly enhanced its discriminatory and predictive abilities (integrated discrimination improvement=0.225 and 0.280; net reclassification improvement=0.612 and 0.734, respectively, P <0.001 for all). CONCLUSIONS: LV 2D-GLS and 3D-GLS are impaired in patients with HFpEF and T2D, and are independent predictors of adverse outcomes. Moreover, 3D-GLS provides incremental prognostic value over 2D-GLS. REGISTRATION: URL: https://www.chictr.org.cn/ ; Unique identifier: ChiCTR2100047487.