Introduction Mortality from heart failure (HF) remains unacceptably high.¹ Left ventricular (LV) ejection fraction (EF) is a flawed, surrogate marker of function.² Pressure- Volume (PV) loops permit comprehensive study of myocardial energetics, stiffness and loading conditions. PV-loops can be assessed invasively but this is laborious and carries risk.

Materials and Methods A discovery cohort of PREFER-CMR study patients underwent same-day invasive impedance catheter and cardiac magnetic resonance (CMR) studies. CMR-derived PCWP was calculated as previously validated.³ Custom software constructed invasive beat-by-beat PV-loops (figure 1A&B), which were used to develop non-invasive approximations from CMR volumetrics and peripheral blood pressure. These were calculated in the outcome cohort of mixed HF patients. Univariable and multivariable Cox regression analysis was applied to assess all-cause and 1-year mortality. ROC-curves found the optimal cut-off values for each metric. These were tested in a validation cohort using Kaplan-Meier analysis.

Results In this multicentre study, a total of 1321 patients were recruited (15 in the discovery and 1306 for the outcomes cohort). There was no significant difference between the stroke work (SW) or static end-diastolic compliance (sEDC) between the CMR and invasive studies. For the outcomes study, patients were randomised to the derivation (n=690) or validation cohort (n=616). The mean follow-up duration was 3.54 years and age was 64.7 years. 48% were female. In the derivation cohort, univariable analysis demonstrated parameters of energetics to show the strongest association with mortality (figure 1C). Multivariable analysis selected sEDC, CMR SW and PE for predicting all-cause mortality, whilst sEDC and CMR SW were selected for 1-year mortality. Kaplan- Meier analysis in the validation cohort found the best performing individual parameter was the sEDC (figure 1D). However the multivariable risk scores showed the best performance in both all cause (figure 1E) and 1-year (²=49.87, P<0.001) mortality. 57.3% of the high-risk patients had preserved EF.

Discussion In this internally validated, multicentre study examining hard mortality endpoints, we show that HF assessment should be multifaceted and include meaningful pathophysiological measures. These are amenable to wide adoption at almost no cost. The era of LVEF guiding HF prognosis and treatment is ending.

• Download figure
• Open in new tab
• Download powerpoint

Abstract 17 Figure 1

Conclusion CMR-derived measures of cardiac energetics and myocardial stiffness predict death in HF more accurately than LVEF.

Acknowledgements We are grateful to all our collaborators at the University of Leeds and University of Sheffield for collating the outcomes data. GM is funded by the National Institute of Health Research. PS is funded by the British Heart Foundation. PG is funded by the Wellcome Trust.

References

1. Savarese G, et al. Global burden of heart failure: a comprehensive and updated review of epidemiology. Cardiovasc Res. 2023 Jan 18;118(17):3272–3287.

2. Marwick TH. Ejection fraction pros and cons: JACC state-of-the-art review. J Am Coll Cardiol. 2018 Nov 6;72(19):2360–2379.

3. Garg P, et al. Cardiac magnetic resonance identifies raised left ventricular filling pressure: prognostic implications. Eur Heart J. 2022 Jul 7;43(26):2511–2522.