Cardiovascular Magnetic Resonance-Estimated Pulmonary Capillary Wedge Pressure, Congestion Markers, and Effect of Empagliflozin in Patients with Heart Failure with Reduced Ejection Fraction and Dysglycaemia (SUGAR-DM-HF)

Aims A cardiovascular magnetic resonance (CMR) approach to non-invasively estimate left ventricular (LV) filling pressure was recently developed and shown to correlate with invasively measured pulmonary capillary wedge pressure (PCWP). We examined the association between CMR-estimated PCWP (CMR-PCWP) and other imaging and biomarker measures of congestion, and the effect of empagliflozin on these, in the SUGAR-DM-HF trial (NCT03485092). Methods and results SUGAR-DM-HF enrolled 105 patients with heart failure with reduced ejection fraction (HFrEF) and pre-diabetes or type 2 diabetes who were randomly assigned to empagliflozin 10 mg or placebo once daily for 36 weeks. Transthoracic echocardiography (TTE) was performed at screening and physical examination, lung ultrasound (LUS), CMR, and blood sample collection at baseline and 36 weeks. The mean age of participants was 69 years, mean LV ejection fraction was 33%, and mean CMR-PCWP was 15.1 mmHg (standard deviation 2.5 mmHg). CMR-PCWP correlated with log N-terminal pro-B-type natriuretic peptide (Pearson's r = 0.31, p = 0.0012), total number of B-lines on LUS (Spearman's rho = 0.34, p < 0.001), TTE average E/e' (r = 0.22, p = 0.03) and CMR ventricular volumes (LV end-systolic volume: r = 0.51, p < 0.001; right ventricular end-systolic volume: r = 0.55, p < 0.001). Mean CMR-PCWP decreased from baseline to 36 weeks by −0.37 (95% confidence interval −0.78 to 0.05) mmHg and −0.17 (−0.69 to 0.36) mmHg in the empagliflozin and placebo groups, respectively; placebo-corrected difference −0.43 mmHg (−1.09 to 0.23, p = 0.20). Conclusion Cardiovascular magnetic resonance-estimated PCWP correlates with clinical and ultrasound markers of congestion in HFrEF. CMR not only enables assessment of cardiac structure and function but also provides a non-invasive assessment of LV filling pressure.