Author + information
- Published online May 27, 2019.
- James E. Udelson, MD∗ ()
- ↵∗Address for correspondence:
Dr. James E. Udelson, Tufts Medical Center, Cardiology, 800 Washington Street, Box 70, Boston, Massachusetts 02111-1526.
In a patient who presents with new-onset heart failure (HF), extant guidelines suggest that while initial therapeutic maneuvers are undertaken to relieve symptoms an assessment of the etiology is undertaken because HF is a syndrome not a disease (1). Such initial clinical and imaging assessment may lead to direct treatments that could completely ameliorate the HF syndrome, such as a finding of multivessel coronary disease with substantial myocardial hibernation, in which revascularization may lead to normalization of left ventricular function.
In large part, a major etiologic distinction is made between those patients with ischemic cardiomyopathy and those with nonischemic cardiomyopathy. The latter category is heterogeneous and may include entities such as familial, alcoholic, hypertensive, post-viral, peripartum, and numerous other causes. In the past, several studies suggested that this distinction, although sometimes having important therapeutic implications as noted above, also had prognostic implications because studies seemed to show that those patients with ischemic cardiomyopathy had a more unfavorable outcome trajectory compared with those patients with a nonischemic etiology. As an example, in an analysis of trials in patients with heart failure with reduced ejection fraction (HFrEF), Frazier et al. (2) reported less favorable survival and hospitalization rates with ischemic etiology. In an analysis of the MAGGIC (Meta-Analysis Global Group In Chronic Heart Failure) individual patient meta-analysis database, ischemic etiology of HF had an independent association with all-cause mortality as well as cardiovascular death (3). However, analysis of more recent data on more than 156,000 hospitalized patients with HF showed no difference in hospital mortality between patients with ischemic versus nonischemic etiology (4).
Thus, published studies do not add up to a clear answer on whether there is an important prognostic implication regarding etiology. The inconsistent results may reflect that some analyses incorporate only randomized trials, others a mix of trials and cohort studies or large registries, may involve chronic or hospitalized patients, and may focus on just patients with HFrEF or patients with HF regardless of EF. Moreover, the published reports on this point spans multiple therapeutic eras.
In the current paper in this issue of JACC: Heart Failure, Balmforth et al. (5) used the database of the PARADIGM-HF (Prospective Comparison of Angiotensin-receptor-neprilysin inhibitor (ARNI) with Angiotensin converting-enzyme inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial to assess the prognostic impact of HF etiology for patients with HFrEF in the contemporary treatment era. Among the 8,399 randomized patients, 60% were categorized as having an ischemic etiology, whereas 40% were classified as one of the nonischemic etiologies. The etiology classification used in the main analyses of this paper was assigned by the investigators at the sites. Unadjusted analysis of the trial’s primary outcome endpoint—cardiovascular death or HF hospitalization—was worse in the ischemic etiology group. However, as would be expected, there were numerous significant baseline differences between the ischemic and nonischemic groups in prognostically important parameters such as age, prevalence of diabetes, sex, and New York Heart Association functional class among others. Given the very large number of adverse outcomes, robust adjustment for these baseline differences was possible, and, in the adjusted analysis, there was no difference in outcomes between the groups. This latter finding was preserved using different definitions to parse the groups based on history of infarction or revascularization (5).
Why might these data demonstrate no influence of etiology on outcome, when the pooled analysis of 5 earlier trials (2) suggested the patients with ischemic etiology fared more poorly? One explanation may be that, with contemporary guideline-directed medical therapy, the playing field has been leveled. In Frazier et al. (2), few patients were on beta-blockers, as the trials included in that analysis largely preceded the beta-blocker era. In the PARADIGM sample, ∼93% of patients were taking beta-blockers (5). As the authors point out, the prospectively collected PARADIGM trial data also enabled more robust statistical adjustment for a greater number of factors, including the baseline level of natriuretic peptides, not available in the older studies. This aspect is a real strength of the study.
As the authors point out, a limitation of the analysis is that HF etiology was simply classified by the site investigators based on available clinical information, without specific guidance by any manual of operations, not unusual for such large trials. This issue of etiologic classification does raise a question—what exactly constitutes ischemic cardiomyopathy, if it is to be defined carefully?
When guidance is given in trials or registries on classification of etiology, the definition of ischemic cardiomyopathy is usually pragmatic, based on the presence of coronary artery disease (CAD), known from a prior catheterization, a history of myocardial infarction (MI), or a history of revascularization. However, a patient with left ventricular (LV) dysfunction “out of proportion” to the extent of CAD likely does not have an ischemic cardiomyopathy. Felker et al. (6) attempted to operationalize this concept, and proposed an outcomes-derived definition of ischemic cardiomyopathy based on extensive CAD, in which a patient with single-vessel disease but no history of MI or revascularization would be classified as nonischemic.
The seminal work of McCrohon et al. (7) examining patients with HFrEF using cardiac magnetic resonance with late gadolinium enhancement suggests that the identification of an ischemic cardiomyopathy might be best made by examining the myocardial substrate rather than the status of the coronary arteries. In that study, 13% of patients with LV dysfunction and normal epicardial arteries had a pattern of late gadolinium enhancement indicative of extensive subendocardial or transmural infarct, indistinguishable from patients with extensive CAD and known prior MI. Although clearly not practical for use in trials or registries, it is a reminder that etiology-based outcome differences, when identified, may to some degree be related to misclassification of the true etiology.
A second interesting finding in the paper by Balmforth et al. (5) is that there was no influence of HF etiology on the treatment effect of sacubitril/valsartan on the primary endpoint outcome or on cardiovascular death alone. This finding is entirely consistent with all other major pharmacological neurohormonal modulation therapies for HFrEF, as reported in every large outcome trial.
What lessons do we learn from the 2 major findings of this paper, that etiology in patients with HFrEF influences neither prognostic trajectory nor the treatment response to neurohormonal modulation therapy? These data as well as similar results from prior studies suggest that once LV dilatation and dysfunction have ensued, no matter what the prior etiologic journey may have been to get to that point, the progression of the HF syndrome is driven predominantly by neurohormonal forces (Figure 1) (8). Although events related to an etiologic disease may occur, for example MI in those with underlying CAD, therapies directed at prevention of such events (aspirin, statins) have not had an impact on outcomes.
Although much is written about the need and potential for more personalized therapies (9), it is remarkable that our major pharmacotherapies for HFrEF—angiotensin-converting enzyme–inhibitors, angiotensin receptor blockers, evidence-based beta-blockers, mineralocorticoid antagonists, and ARNIs—have not shown differential effects in virtually any subgroup examined in large clinical trials, including etiologic subgroups, and their use is recommended across the board (1). In the coming 2 years we will have clinical trial results of several mechanistically distinct therapies such as sodium-glucose cotransporter-2 inhibitors, a cardiac myosin activator and soluble guanylate cyclase stimulators, and we will learn whether that pattern continues or whether we have plateaued with the broad-based approach to HFrEF therapies. In the meantime, the HF community should celebrate the advances resulting from our “impersonalized” approach to HFrEF, as a recent network meta-analysis has shown that the combination of a beta-blocker, a mineralocorticoid antagonist, and an ARNI is associated with a 63% reduction in mortality compared with putative placebo (10). Not bad for something so impersonal!
↵∗ Editorials published in JACC: Heart Failure reflect the views of the author and do not necessarily represent the views of JACC: Heart Failure or the American College of Cardiology.
Dr. Udelson has received research funding from Ironwood Pharmaceuticals and served on their Science Advisory Board for a clinical trial.
- 2019 American College of Cardiology Foundation
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.
- Frazier C.G.,
- Alexander K.P.,
- Newby L.K.,
- et al.
- Martínez-Sellés M.,
- Doughty R.N.,
- Poppe K.,
- et al.,
- Meta-Analysis Global Group In Chronic Heart Failure (MAGGIC)
- Shore S.,
- Grau-Sepulveda M.V.,
- Bhatt D.L.,
- et al.
- Balmforth C.,
- Simpson J.,
- Shen L.,
- et al.
- Felker G.M.,
- Shaw L.K.,
- O'Connor C.M.
- McCrohon J.A.,
- Moon J.C.,
- Prasad S.K.,
- et al.
- Udelson J.E.,
- Stevenson L.W.
- Califf R.M.
- Burnett H.,
- Earley A.,
- Voors A.A.,
- et al.