Prediction and prevention of heart failure in high-risk elderly patients

This editorial refers to ‘Prevalence and prognosis of ischaemic and non-ischaemic myocardial fibrosis in older adults’†, by S.M. Shanbhag et al., on page 529. Late gadolinium enhancement cardiovascular magnetic resonance (LGE-CMR) offers the unique ability to characterize the myocardium and reproducibly detect myocardial infarction (MI) and non-ischaemic replacement fibrosis in a single non-invasive test without exposing the patient to ionizing radiation. The presence of ischaemic and non-ischaemic myocardial fibrosis is a well-established predictor of adverse outcomes including death, sudden cardiac death, and heart failure events in patients with acute and chronic heart failure across a spectrum of left ventricular ejection fraction (LVEF).1–3 Although the prognosis for patients with heart failure has improved over the last 30 years, overall outcomes remain worse than those associated with many cancers.4 Around a half of patients suffer from heart failure with preserved ejection fraction (HF-PEF), for which there are no proven prognostic treatments.5 Improved therapy, especially for patients with HF-PEF, and the detection of early phenotypes of disease, potentially enabling the introduction of effective treatment in the pre-symptomatic phase, remain major unmet needs. By characterizing myocardial substrate using LGE or parametric mapping, CMR has the potential to detect early phenotypes of HF in high-risk populations and guide therapy more effectively. In this issue of the European Heart Journal, Shanbhag and colleagues provide additional, important data on the association between LGE and adverse outcomes in a population of asymptomatic elderly participants recruited from the AGES-Reykjavik Study.6 Medium- and long-term outcomes of participants with unrecognized MI have been reported previously.7 , 8 In the current study, the authors focused on the association between outcome and major and minor patterns of non-ischaemic LGE, comparing patients with such patterns with those who have had a previous MI and those without LGE. The primary endpoint was a composite of all-cause mortality and hospitalization for heart failure, verified by review of hospital records. The authors used proportional hazard modelling and adjusted the models based on the inverse probability of having LGE patterns using a propensity score. Major non-ischaemic LGE included subepicardial/mid-wall patterns of LGE, patchy LGE in the presence of left ventricular hypertrophy, nodular LGE in keeping with sarcoidosis, or diffuse LGE with rapid clearance in keeping with amyloidosis. Enhancement localized to the insertion points, aortic root, and mitral annulus was classified as minor. Overall, 6% of patients had major non-ischaemic LGE, 26% had minor non-ischaemic LGE, and 23% had infarct patterns of LGE. In total, 115 patients died during follow-up and 77 were admitted to hospital with heart failure, indicating a high-risk, albeit asymptomatic population at baseline. The number of events within each subgroup was small, and the results of the proportional hazard modelling should be interpreted within this context. Nevertheless, the analyses confirmed that major non-ischaemic enhancement was associated with an increased rate of the primary composite endpoint after adjustment using inverse probability weighting [hazard ratio (HR) 3.2; 95% confidence interval (CI) 1.9–5.2; P < 0.001]. Strikingly, almost a quarter of such patients, all of whom were asymptomatic at baseline, had a heart failure admission during follow-up compared with 15% of patients with previous infarcts. While ischaemic LGE was associated with the composite endpoint on univariable analysis, the HR reduced and became non-significant after adjustment based on the propensity score, which included a number of traditional cardiovascular risk factors (HR 1.4; 95% CI 0.9–2.1; P = 0.10). This suggests that the adverse prognosis in patients with ischaemic heart disease was at least partially accounted for by traditional risk factors. This was not the case, however, amongst patients with major non-ischaemic LGE and is likely to be explained by the heterogeneous nature of the pathophysiological processes contributing to the development of non-ischaemic myocardial fibrosis, that were incompletely accounted for by the variables in the propensity score. The baseline characteristics of the entire population are enlightening and indicate an elderly population (median age 77 years) at high risk of heart failure. In these patients, 83% had a history of hypertension, 35% a history of type 2 diabetes mellitus, and many had a mildly raised creatinine. Those with major non-ischaemic LGE were even more likely to have hypertension, and had greater left ventricular mass and higher creatinine than other subgroups. The LVEF was typically preserved in those with non-ischaemic LGE at baseline [62.6%; interquartile range (IQR) 58.3–66.9]. Although follow-up LVEF at the time of heart failure admission is not available, the baseline LVEF and the familiar baseline characteristics suggest that the predominant subtype amongst those with non-ischaemic LGE admitted with heart failure was HF-PEF. Overall, the results suggest that the presence of non-ischaemic fibrosis at baseline may have represented the precursor and substrate for this syndrome. Possible approach to the diagnosis and treatment of patients deemed to be at high risk of developing heart failure. AF, atrial fibrillation; CMR, cardiovascular magnetic resonance; HF-REF, heart failure with reduced ejection fraction; HTN, hypertension; RFs, risk factors; TTR, transthyretin. Another important aspect to consider is how the baseline CMR results changed the subsequent management of patients during follow-up. Those with infarcts are likely to have been given evidence-based treatments to improve prognosis. Conversely, there are no specific, prognostic, treatments for patients with non-ischaemic fibrosis and preserved LVEF. Treatment of these patients is, therefore, unlikely to have changed as a result of their CMR. This may explain the difference in outcome between these groups and emphasizes the importance of research investigating novel treatments for such patients. Targeting asymptomatic high-risk patients, based on age and the presence of risk factors such as diabetes mellitus, hypertension, obesity, or atrial fibrillation, may prove fruitful in the early diagnosis, treatment, and possibly even prevention of HF-PEF.9 The current study suggests that the identification of myocardial substrate in the form of replacement fibrosis using LGE-CMR has the potential to improve risk stratification further. Whilst population screening using CMR is not feasible, targeting subgroups at high risk of HF-PEF and using the presence of myocardial fibrosis on CMR to guide the use of medications with anti-fibrotic properties may be beneficial. Although not available in this study, natriuretic peptide levels and a history of atrial fibrillation may have provided incremental value in identifying those patients who were at greatest risk of developing heart failure. The identification and quantification of interstitial fibrosis using T1 mapping may also be helpful, and offers the potential advantage of avoiding administration of contrast.10 Given the relatively small number of patients with major patterns of non-ischaemic LGE in the current study, the authors were unable to investigate differences in outcome based on the pattern and location of LGE.11 Accepting the heterogeneity of non-ischaemic insults, LGE occurring in different patterns and locations is likely to be the result of different pathophysiological processes. In order to personalize treatment, further in-depth characterization of the disease process using advanced imaging techniques and genetic and proteomic data may eventually result in targeting of the specific disease processes. Although no confirmed cases of amyloidosis were identified in the current cohort at baseline, considering other recent data,12 it appears possible that in a cohort of this size and advanced age, at least a small number of patients may have developed transthyretin amyloidosis during follow-up, contributing to the development of heart failure. Given the emergence of targeted treatment and the availability of non-invasive methods of diagnosis,12 , 13 this emphasizes the importance of a more comprehensive approach to the diagnosis of heart failure phenotypes in the elderly. In conclusion, the current study suggests that the presence of major non-ischaemic fibrosis in a high-risk elderly population is associated with a substantial risk of heart failure events and death. The major challenge faced by the heart failure and CMR community is now to demonstrate that by detecting the presence of risk markers such as LGE, the outcome of these patients can be improved through the better use of targeted therapies. Conflicts of interest: D.J.P. reports grants from Siemens, La Jolla, and Bayer, outside the submitted work. B.P.H. has no conflicts to declare.