Right ventricle strain has emerged as an accurate tool for RV function assessment and is a strong predictor of survival in HFrEF. The impact of RV strain assessed by cMR in HFpEF is unknown.

We sought to analyze RV strain by cMR in controls and HFpEF and to analyze its prognostic value.

Between January 2015 and June 2017, we prospectively enrolled 100 consecutive patients with HFpEF (79±8 years, 59% women) and 64 controls (56±20 years, 52% women). All patients underwent 2D echo and 3Tesla cMR in sinus rhythm. A four-chamber view was analyzed by 2D feature tracking to evaluate the global longitudinal strain of RV (RV-GLS). HFpEF patients were followed up for a combined outcome of all-cause mortality and first HF hospitalization.

In HFpEF, mean RV-GLS was significantly lower than in controls (−14.6±4.6% vs.−17.0±3.9%; P=0.007) (Fig. 1). Due to poor tracking quality, 28 controls (44%) and 19 HFpEF patients (19%) were excluded from the analysis. In univariate linear regression analysis, NTproBNP (R=0.36, P=0.001), E wave velocity (R=0.29, P=0.001), E/e’ ratio (R=0.28, P=0.002), RV fractional area change (R=−0.29, P=0.001), TAPSE (R=−0.24, P=0.006), indexed LA volume (R=0.27, P=0.003), LVEF (R=−0.32, P=0.001), indexed LV mass (R=0.36, P=0.001), RVEF by cMR (R=−0.45, P<0.001) and extracellular volume (R=0.25, P=0.005) were associated with RV-GLS. Multiple regression analysis showed that RVEF by cMR and NTproBNP (beta=−0.20 [−0.31,−0.09] P<0.001 and 2.40 [0.46, 4.33], P=0.016 respectively) were significant predictors of RV strain. During a mean follow-up of 17±10months, 37 HFpEF patients (44%) reached the combined outcome (18 all cause deaths and 30 first HF hospitalization). In univariate analysis, RV-GLS was not significantly associated with combined outcome (HR=1.00 [0.93–1.07], P=0.90).

RV-GLS assessed by cMR feature tracking is significantly different between controls and HFpEF patients and is associated with RVEF and NTproBNP. However, RV GLS was not associated with poor outcome in HFpEF.