Right ventricular versus left ventricular global longitudinal strain for association with all-cause mortality after transcatheter aortic valve implantation

Abstract

Abstract Introduction Speckle-tracking echocardiography plays an increasingly important role in the assessment of aortic stenosis (AS), the most common valvular heart disease in the Western population. Right ventricular (RV) and left ventricular (LV) global longitudinal strain (GLS) were studied for their association with all-cause mortality in AS patients undergoing transcatheter aortic valve implantation (TAVI). Purpose To investigate whether simultaneous assessment of RV and LV GLS provides a better association with all-cause mortality during long-term follow-up after TAVI. Methods In our prospective registry we identified 109 patients with severe AS who underwent TAVI and had a pre-procedural echocardiography within three months of the procedure allowing complete assessment of RV and LV GLS using TomTec Image Arena. All-cause mortality was defined as endpoint. Results RV GLS was lower among non-survivors (N=36; median [IQR]: −13.94 [−16.31 to −12.62]; p<0.001) than survivors (N=73; −17.04 [−20.16 to −15.22]; p<0.001), while LV GLS did not differ (p=0.249). RV GLS >−16.6% differentiated survivors from non-survivors (sensitivity 78%; specificity 63%; ROC AUC 71%; p<0.001), while LV GLS did not (p=0.243). Kaplan Meier curves showed good differentiation of survivors and non-survivors with the RV GLS (p<0.001; Figure 1), but not the LV GLS cut-off (p=0.058). In univariable Cox regression models, RV GLS was associated with all-cause mortality (HR 1.10 [95% CI 1.02 to 1.18]; χ2=6.64; p=0.01), while LV GLS was not (HR 1.05 [0.96 to 1.16]; χ2=1.08; p=0.299). In bivariable models, association of RV GLS with all-cause mortality was independent of LV GLS or LV ejection fraction (LVEF; Table 1). ANOVA likelihood ratios revealed that inclusion of RV GLS to LV GLS or LVEF improved their model fitness, while that of LV GLS did not (Table). Conclusion RV GLS was associated with all-cause mortality after TAVI, while LV GLS and LVEF were not. The association of RV GLS was independent of LV GLS and LVEF and showed potential incremental value for assessment of outcome association, while LV GLS did not. Funding Acknowledgement Type of funding sources: None.