Characteristics and long-term outcomes of patients with reduced ejection fraction referred for adenosine stress perfusion cardiac magnetic resonance imaging

Funding Acknowledgements Type of funding sources: None. Background Adenosine stress perfusion has been shown to be of minimal incremental benefit in distinguishing between ischaemic and non-ischaemic aetiology of severe left ventricular systolic dysfunction (LVSD) over and above that obtained from Cardiac Magnetic Resonance (CMR) with Late Gadolinium Enhancement (LGE). Stress CMR has, however, been shown to be effective in risk-stratifying LVSD patients, with ischaemia being an independent predictor of cardiovascular death or myocardial infarction (MI) and associated with higher rates of further intervention. Purpose Evaluate real world data from a single tertiary UK cardiac MRI centre to determine the characteristics and long-term clinical outcomes of patients with LVSD referred for stress CMR. Methods As part of an ongoing registry, all consenting patients with Ejection Fraction (EF) ≤40% and a completed adenosine stress perfusion CMR between January 2015 and December 2019 were included with prospective baseline data collection. All-cause mortality and cardiac hospitalisation, coronary angiography/revascularisation was determined from electronic hospital records. Outcomes were compared between the inducible ischaemia vs. no ischaemia groups, and LGE present vs. no LGE groups using chi square. Results The sample included 86 patients. The mean EF was 32 ± 6%. Median follow up was 3.8 years (range 41-2222 days). The indications for CMR were: 30 (35%) assess ischaemia, 35 (41%) assess LVSD aetiology and 21 (24%) LVSD assess viability. Inducible ischemia was present in 30 (35%) patients and absent in 56 (65%). Patient characteristics and outcomes are shown in Table 1. Baseline characteristics were similar between the groups but there was a higher rate of hypertension and ischaemic heart disease in the ischaemia group. There was a non-significant difference in combined mortality and cardiac hospitalisation rates between the groups (40% vs. 27% p = 0.20). LGE was present in 69 (80%) patients (28 with ischaemia; 41 without) and absent in 17 (20%, 2 with ischaemia, 15 without). The event rate was 23 (33%) vs. 4 (24%) between LGE vs. No LGE groups (p = 0.44). Of the 15 patients (17%) with no LGE or ischaemia; 2 died and 1 was hospitalised, there were no MI"s and no Percutaneous Coronary Intervention (PCI). The lack of statistical difference in event rates between ischaemia and no ischaemia groups may be due to our relatively small sample size or could reflect the effectiveness of contemporary disease modifying treatment for Heart Failure with reduced EF. Conclusion This real-world data supports published findings that in patients with LVSD and no LGE on CMR, ischaemia is very uncommon and stress CMR is unlikely to increase diagnostic yield. Conversely, if stress CMR is performed and ischaemia is absent, incidence of subsequent angiography and revascularisation is very low, which is reassuring in clinical practice. In those patients without ischaemia and LGE, likelihood of MI is low.