Clinical profiles in acute heart failure: one size fits all or not at all?

This article refers to ‘Clinical phenotypes and outcome of patients hospitalized for acute heart failure: the ESC Heart Failure Long-Term Registry’ by O. Chioncel et al., published in this issue on pages 1242–1254.

Hospitalization for acute heart failure (AHF) is a global public health problem of pandemic proportions.1, 2 There are more than one million admissions annually for heart failure (HF) in Europe3 and the risk for readmission or death following index hospitalization may be in excess of 40% at 1 year post-discharge.4, 5 However, there have been no major breakthroughs in the management of AHF and the cornerstones of therapy remain i.v. diuretics and/or vasodilators and, less commonly, inotropes.3 One of the hypothesized reasons for the failure of a multitude of clinical development programmes is that HF is a heterogeneous syndrome and clinical trials have not tried to match the right drug with the right patient.6, 7 This issue of the journal features a secondary analysis of the European Society of Cardiology Heart Failure Long-Term (ESC-HF-LT) Registry by Chioncel et al.,8 which prospectively applied three common classification systems in order to better understand the clinical characteristics, management and outcomes of patients admitted for a primary diagnosis of AHF.

The ESC-HF-LT Registry is a prospective, observational study enrolling patients at 211 geographically representative ESC member sites in 21 European and/or Mediterranean countries.4, 9 The registry employs periodic consecutive enrolment and includes all outpatients with chronic HF and all patients admitted to hospital for a primary diagnosis of AHF. A diagnosis of AHF is made at the point of care by clinician-investigators and includes the presence of signs and symptoms of HF and the need for i.v. therapies. The present analysis included all patients admitted for a primary diagnosis of AHF and systematically studied patient characteristics, inpatient management, and in-hospital and long-term outcomes by clinical profile.8 Patients were divided according to three common classification systems based on the 2008 ESC Guidelines10 [i.e. acute decompensated HF (DHF), cardiogenic shock (CS), pulmonary oedema (PO), right HF (RHF), hypertensive HF (HT-HF), HF in the setting of an acute coronary syndrome (ACS-HF)], systolic blood pressure (SBP) (i.e. SBP <85 mmHg, SBP 85–110 mmHg, SBP 110–140 mmHg, and SBP >140 mmHg), and congestion and perfusion status (i.e. no congestion/no hypoperfusion, congestion/no hypoperfusion, no congestion/hypoperfusion, and congestion/hypoperfusion). The primary outcomes were death, HF readmission, and death or HF readmission at 1 year.

In total, 16 012 patients were enrolled in the ESC-HF-LT Registry over the study timeframe including 6629 patients admitted for a primary diagnosis of AHF (41.4%).8 DHF was by far the most common clinical profile, evidenced in 61.1% of patients, followed by PO and ACS-HF, which, respectively, applied to 13.2% and 14.4% of patients. Only a minority of patients were phenotyped as CS (2.9%), RHF (3.5%) or HT-HF (4.8%). Similarly, most patients were normotensive (i.e. SBP 110–140 mmHg) (42.9%) or hypertensive (i.e. SBP >140 mmHg) (30.3%) at initial presentation, and substantially fewer patients were found to have borderline SBP (i.e. SBP 85–110 mmHg) or to be frankly hypotensive (i.e. SBP <85 mmHg). Finally, when patients were classified according to congestion and perfusion status, 14.8% were found to have no congestion/no hypoperfusion, 69.7% to have congestion/no hypoperfusion, 0.9% to have no congestion/hypoperfusion and 13.6% to have congestion/hypoperfusion. Tremendous variation was observed in baseline characteristics, inpatient management, and in-hospital and long-term outcomes by clinical profile. In general, patients with high-acuity clinical profiles (i.e. CS and PO, SBP <85 mmHg and congestion/perfusion) received more i.v. therapies (including vasoactive medications such as vasodilators, inotropes and vasopressors) and procedures (revascularization, intra-aortic balloon pumps, implantable devices, etc.) and had very poor short- and long-term prognoses. For example, patients presenting with CS, SBP of <85 mmHg, or combined congestion and perfusion abnormalities had in-hospital and 1-year mortality rates, respectively, of 36.1% and 54.0%, 26.6% and 34.8%, and 16.5% and 29.8%. Interestingly, the investigators found that among patients surviving past 6 months post-discharge, differences in long-term outcomes by clinical profile were largely attenuated.4

This is by far the largest and most contemporary cohort in which the clinical profile as outlined in the 2008 ESC Guidelines10 has been prospectively applied and it offers several unique observations. Although it is well established that cardiac injury, as evidenced by elevated or rising levels of cardiac troponin, may be present in upwards of 90% of patients admitted for a primary diagnosis of AHF without clinical suspicion for ACS, the present analysis found that nearly 15% of patients were diagnosed by clinician-investigators with a concomitant ACS event.11, 12 This has important implications for management in the acute setting as there is a strong evidence base for the initiation of an early invasive approach in patients with high-risk features including signs and symptoms of HF, as well as aggressive pharmacotherapy (i.e. anticoagulation, antiplatelet agents, lipid-lowering therapy, etc.).13 In addition, recent evidence suggests that an ischaemia evaluation (i.e. stress testing or invasive angiography) is underutilized in patients with de novo or newly decompensated HF,14 despite the fact that surgical revascularization (i.e. coronary artery bypass grafting) improves long-term survival compared with medical therapy alone in patients with an ischaemic cardiomyopathy and severely reduced systolic function.15, 16 It is also notable that applying the ESC classification system at the point of care has prognostic implications as clinician-investigators with minimal additional training were readily able to identify high-acuity clinical profiles (i.e. CS and PO) with poor short- and long-term prognoses. Future research should prospectively evaluate the added utility of the ESC clinical profiles to the initial triage of patients to the appropriate level of care (i.e. telemetry vs. stepdown vs. intensive care unit).17-19

This study4 also confirms prior research investigating the epidemiology and predictive value of SBP at initial presentation.20, 21 Despite previous misconceptions based on clinical trial experiences, it is now well established that the vast majority of patients admitted for a primary diagnosis of AHF are normotensive or hypertensive.20, 21 This is confirmed by the present analysis,4 which found that more than 70% of patients had SBP of >110 mmHg at initial presentation. In addition, although hypertension is a strong independent risk factor for HF, following the genesis of HF there is a paradoxical relationship between SBP and outcomes.22, 23 For example, in this study, in-hospital mortality reached an astonishing 26.6% in patients presenting with SBP of <85 mmHg, whereas in-hospital mortality was a mere 2.7% in patients with SBP of >140 mmHg.4 The pathophysiological basis for the paradoxical relationship seen between SBP and outcomes is likely to reflect the presence of cardiac reserve in patients presenting with resting hypertension. However, it is important to recognize that patients with HF and a reduced ejection fraction, particularly those patients presenting with elevated SBP, may be undertreated or underdosed with guideline-directed medical therapies and hospitalization represents an opportune time to initiate or up-titrate evidence-based medications.9 Clinical trials24 and registries25 have demonstrated that in-hospital initiation of medications is associated with improved adherence and persistence, and possibly with better clinical outcomes.

Finally, it is noteworthy that the vast majority of patients presented with signs and symptoms of congestion and that relatively few patients had evidence of inadequate end-organ perfusion by history or physical examination. This is not unexpected given that signs and symptoms of congestion, particularly dyspnoea, are the most common precipitants for hospitalization, and it suggests that only a minority of patients truly present with end-stage HF refractory to medical therapy that requires evaluation for advanced therapies (i.e. mechanical circulatory support and/or cardiac transplantation).26, 27 This classification system was first popularized by Nohria and colleagues and has more commonly referred to patients as being wet vs. dry and warm vs. cold.28 The findings from the ESC-HF-LT Registry are largely consistent with this initial publication, which found approximately 70% of patients were wet and warm.28 However, a notable difference is that close to 15% of patients enrolled in the ESC-HF-LT Registry were found to have no signs and symptoms of either congestion or hypoperfusion (i.e. normal). There are a couple of potential explanations for this somewhat discrepant finding. It is possible that patients had already experienced rapid and dramatic improvement in response to standard therapy before they were assessed by the enrolling clinician-investigator.29, 30 Alternatively, the signs and symptoms of HF are neither sensitive nor specific and it is possible that patients were initially misclassified. Prevalences of cardiac and non-cardiac co-morbidities are high in HF and, following an index hospitalization for HF, the causes of more than half of readmissions may be adjudicated as non-HF-related.31

In conclusion, Chioncel and colleagues8 should be commended for the ambitious undertaking of the prospective phenotyping of patients admitted for a primary diagnosis of AHF in the ESC-HF-LT Registry. HF is a heterogeneous syndrome and improving collective understanding of its pathophysiology and management requires a nuanced and tailored approach. Notably, patients with chronic HF can be classified according to demographic factors (i.e. age and gender), medical co-morbidities (i.e. coronary artery disease, atrial fibrillation, diabetes mellitus, etc.) and systolic function (i.e. reduced vs. preserved), which have important diagnostic, prognostic and therapeutic implications. Although the proposed approaches to patient phenotyping in the acute setting (i.e. clinical profiles, SBP and congestion/perfusion status) are informative and complementary, none of these classification systems provides a comprehensive assessment in isolation and additional research is required before their use to guide treatment decisions in routine practice can be recommended. In an era of ‘omics’ and personalized medicine, future research should focus on the role of a multimarker approach and the application of cluster analysis to identify novel and clinically meaningful patient profiles in an effort to develop new therapies and improve the quality of care.

Conflict of interest: A.P.A. declares no conflicts of interest. M.G. has been a consultant for Abbott Laboratories, Astellas, AstraZeneca, Bayer HealthCare AG, CorThera, Cytokinetics, DebioPharm SA, Errekappa Terapeutici, GlaxoSmithKline, Ikaria, Johnson & Johnson, Medtronic, Merck, Novartis Pharma AG, Otsuka Pharmaceuticals, Palatin Technologies, Pericor Therapeutics, Protein Design Laboratories, Sanofi-Aventis, Sigma Tau, Solvay Pharmaceuticals, Takeda Pharmaceutical and Trevena Therapeutics.