Universal definition and classification of heart failure: a report of the Heart Failure Society of America, Heart Failure Association of the European Society of Cardiology, Japanese Heart Failure Society and Writing Committee of the Universal Definition of Heart Failure

1.1 Writing committee composition

The Heart Failure Society of America (HFSA), the Heart Failure Association (HFA) of the European Society of Cardiology (ESC) and the Japanese Heart Failure Society (JHFS) selected the members of the writing committee. The writing committee consisted of 38 individuals with domain expertise in HF, cardiomyopathy and cardiovascular disease.

1.2 Consensus development

On 20 August 2020, in response to the necessity for consensus for definition of HF, the HFSA, HFA/ESC and JHFS convened a virtual consensus conference to develop a universal definition of HF with participation from 14 different countries and 6 continents. The work of the writing committee was accomplished via a series of teleconference and web conference meetings, along with extensive email correspondence. The review work was distributed among subgroups of the writing committee based on interest and expertise. The proceedings of the workgroups were then assembled, resulting in the proposed universal definition. All members reviewed and approved the final vocabulary.

1.3 Peer review and approval

The 2020 Universal Definition of HF was reviewed by official reviewers nominated by the HFSA, HFA/ESC and JHFS. The writing committee anticipates that the proposed definition and classification will require review and updating in the same manner as other published universal definitions.²² The writing committee, therefore, plans to review the universal definition on a periodic basis, starting with the anniversary of publication of the definition, to ascertain whether modifications should be considered.

2.1 Definitions of heart failure used in current clinical trials and registries

The definitions and inclusion criteria used in clinical trials and registries in HF differ from those in clinical practice, guidelines and textbooks. Most trials in HF with reduced EF (HFrEF) (Table 2),^25-29 and in HF with preserved EF (HFpEF) (Table 3)^30-33 reflect inclusion criteria that usually include a left ventricular EF (LVEF) threshold, an established HF diagnosis with specific New York Heart Association (NYHA) functional class categories, certain levels of natriuretic peptides and may sometimes include a requirement of past HF hospitalizations, depending on the severity of HF targeted for the trial. HFpEF studies also may include corroborative evidence by imaging reflecting structural and/or functional changes. Nonetheless, a number of gaps remain in standardizing the criteria for clinical trials. These include sensitivity and specificity of diagnostic criteria for HF, establishing standardized natriuretic peptide criteria; the complexity of additional requirements to ascertain the diagnosis of HF; challenges with HFpEF including multiple comorbidities that are often excluded in clinical trials, how to handle patients with EF recovery or changes in clinical trajectory, competing diagnoses that may mimic findings of HF and the generalizability of the trial criteria to the ultimately intended treatment population. It is also important to distinguish between clinical trial inclusion criteria that aim to select target populations, from clinical trial endpoint definitions that facilitate measurement of outcomes secondary to the disease process. For example, natriuretic peptides, which are commonly used in entry criteria in HF trials, are not commonly required for clinical endpoint definitions.²¹

2.2 Gaps in current definitions of heart failure

3.1 Current subclassification of heart failure according to ejection fraction and its limitations

Because clinical trial inclusion criteria, and hence evidence of benefit, have often been restricted to patients with a reduced EF, HF has traditionally been subcategorized according to EF when defining recommended treatments in clinical practice guidelines.^3-5All guidelines use the terminology of HFrEF and HFpEF (Table 5) but differ in the terminology used in patients with EFs between 40% and 49%. The 2013 ACC/AHA guidelines have used the terminology of HFpEF-borderline for patients with EF between 41–49%, and HFpEF-improved for those whose EF improved from a lower level to an EF of >40% under the subgrouping of patients with HFpEF.³ The HFA/ESC and JHFS guidelines have defined a third category of HF with mid-range EF (HFmrEF) or mildly reduced EF for those with an EF of 41% to 49%.^{4, 5} The concept of HFmrEF is not necessarily accepted by all guidelines.⁴⁹

In an effort, through a public-private partnership with the US Food and Drug Administration (FDA) and with an intent to standardize terminology and LVEF cut-points used in US clinical trials, the Heart Failure Collaboratory and Academic Research Consortium proposed the following definitions and EF ranges as their most recent recommendations: (i) HFrEF: HF with LVEF ≤40%; (ii) HFpEF: HF with LVEF ≥50%; (iii) HFmrEF: HF with LVEF >40% and LVEF <50%.⁵⁰

The dichotomization of LVEF of above or below, for example, 40% has been helpful to apply therapies that have been shown to work in patients with reduced EF. Further classification into HFmrEF has potential utility as well as challenges due to its ambiguity, uncertainty and dynamic trajectory.^{15, 51} Post-hoc analyses of certain HF trials have suggested that standard therapy for HFrEF may be effective and extended to patients with HFmrEF,^52-55 but meta-analyses report diverse findings with neurohormonal antagonism in patients with HFmrEF, specifying benefit in certain subgroups and underlining heterogeneity of this category^{15, 55-57} The characteristics of HFmrEF overlap with HFrEF and HFpEF, straddling either category, sometimes one more than the other depending on the clinical circumstance or patients studied.¹⁵ In population-based studies, usually without exclusions of specific aetiologies, HFmrEF comprises 10–20% of the HF population,^{54, 58} resembles the HFrEF group, but with similar⁵⁷ or better survival than HFrEF patients.^{15, 58} Although some patient characteristics of HFmrEF are just between those of HFrEF and HFpEF, the prognosis of patients is not necessarily related to EF,⁵⁹ and the relation between mortality and BNP is not affected by EF.^{59, 60} In many patients, HFmrEF reflects a transitional trajectory for a dynamic temporal change; either to improvement or recovery from HFrEF,^{57, 61} or to deterioration to HFrEF.^{15, 57, 61, 62} Although HFrEF and HFpEF have different clinical spectrums and proposed pathophysiological mechanisms, there is no clear defining syndrome recognized or postulated for HFmrEF. It is likely that patients in this range may have aetiologies that are similar to those in lower or higher LVEF groups, and may be in transition from higher to lower LVEF or vice versa. Persistent HFmrEF can be seen in some patients, including heterogeneous aetiologies such as those with ischaemic, infiltrative, restrictive or hypertrophic cardiomyopathies.^{57, 61, 62} Therefore, lower than a normal EF does not necessarily represent one phenotype and does not always entail the maladaptive deleterious mechanisms seen in patients with HFrEF. Furthermore, patients with restrictive, infiltrative and hypertrophic cardiomyopathies, who may have HFmrEF, have traditionally been excluded from some clinical trials, emphasizing the necessity to focus on aetiology rather than LVEF. The prevalence of HFmrEF, without overlap of other categories, has posed a major challenge for recruitment in trials, resulting in termination due to enrolment futility⁶³ and in some clinical trials and epidemiological studies, patients with LVEF 40–49% have been categorized as HFpEF.

Another criticism is the accuracy of the measurement of EF in clinical practice. Echocardiography is widely used to assess EF in patients with cardiovascular diseases, but the inter-observer and intra-observer variability are not small enough to allow precise quantification of differences in one integer place values such as 39% vs. 41%. Although other cardiovascular imaging modalities can be used to assess EF, there is substantial variation between modalities as well.⁶⁴ Furthermore, EF is not a reliable measure of contractile performance, is load-dependent, and can vary according to haemodynamic status and loading conditions. Other imaging modalities such as global longitudinal strain are evolving to better characterize the ventricle, structural abnormalities, contractile performance, reverse remodelling, response to therapy, and will likely expand the structural phenotyping beyond EF.

Finally, the trajectory of EF over time in addition to a single absolute value of EF, and severity of left ventricular dysfunction even among HFrEF may need to be taken into account to further classify patients with HF. Despite all these reservations, classification by EF has proven to be clinically and epidemiologically useful.

3.2 Current classification according to stages of heart failure and its limitations

The ACC/AHA stages are categorized as Stage A, patients at high risk for HF but without structural heart disease or symptoms of HF; Stage B, structural heart disease but without signs or symptoms of HF; Stage C, structural heart disease with prior or current symptoms of HF; Stage D, refractory HF requiring specialized interventions..^{3, 4, 46} The original ACC/AHA definition of stages of HF⁴⁶ has been ubiquitously adapted throughout other HF guidelines globally.^3-5 Although these stages of HF are well recognized amongst health care professionals, they are not standard nomenclature for general practitioners, patients, payers, or among the literature or education platforms provided by patient advocacy groups. Patients living with HF are less likely to identify with stages of HF in comparison to the familiarity with EF and subjective symptom burden. Contemporary clinical trials have not enrolled or randomized based on stages of HF and most treatment strategies are not guided by the stages in HF.

The ACC/AHA stages are based on symptoms and the presence or absence of structural heart disease and are applicable to both HFrEF and HFpEF. Certainly, there are prognostic nuances that are missed in such a broad staging classification, and opinions also vary as to whether those individuals solely identified with risk factors should be labelled as having a disease state, especially given that they have risk factors for many different diseases (not just HF risk factors). In comparison, classification schemas such as the Society for Cardiovascular Angiography and Interventions (SCAI) cardiogenic shock stages⁶⁵ classified their stages based on detailed parameters of laboratory values, and haemodynamics as well as physical examination findings and exemplifies a more detailed approach to staging. Furthermore, the definitional progression along the ACC/AHA stages A through D is a unidirectional path with little appreciation of a possibility to revert to a lower stage with appropriate GDMT.

If the HF process were to be defined as a continuum from Stage A through D, the highest number of patients would be in Stage A or Stage B.^66-69 This is due to the fact that the prevalence of hypertension, diabetes, coronary artery disease, obesity/metabolic syndrome – the risk factors with significant relative risk and population attributable risk for development of HF – are present in approximately one-third of the US population.¹⁰ By population-based registries, more than 40 to 50% of the adult population has been categorized to be in Stages A or B.^66-68 The high prevalence of HF risk in the general population raises the question of whether Stage A patients should really be defined to have HF. From the public and health care perspective, being called HF, regardless of such an early status as Stage A, raises important concerns, since HF is usually perceived as an advanced chronic disease with symptoms and very adverse outcomes, and may have implications for health and life insurance. Of course, it is critical to focus on prevention, with recognition, prevention and treatment of these risk factors, but it is also important to differentiate those who have HF from those at risk for HF. Similarly, clinicians in general or HF practice have not adopted the terminology of Stage A HF beyond academic circles, partly due to lack of actionable specific treatment recommendations according to stages, and most of their assessment and management focuses on management of left ventricular dysfunction (Stage B) or symptomatic HF (Stages C/D). When clinicians address risk factors such as hypertension, diabetes, obesity or coronary artery disease, they do not refer to those as Stage A HF or pre-HF but rather independent diagnoses. Furthermore, despite recognized increased adverse outcome risk and possibility of progress to symptomatic HF in some patients,^{66, 69, 70} the data on the likelihood of progression from Stages A/B to C/D are limited.^{67, 69, 70} Thus most clinicians do not commonly use the HF terminology for Stage A patients, and do not commonly educate patients regarding risk of progression from Stages A/B to C.

Another important development that needs to be taken into consideration of stages in HF is the advances in prevention of future risk of HF by specific therapies. Although in the past, prevention and holistic treatment of risk factors by standard treatment strategies were felt to prevent HF,³ there is growing evidence that certain treatment strategies are better for the prevention of HF, and not all treatment strategies of hypertension and diabetes prevent HF equally or at all. For example, in the treatment of hypertension, diuretic-based antihypertensive therapies, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers have been shown to prevent HF in a wide range of target populations, whereas calcium channel blockers have not.⁷¹ There is growing evidence that treatment with sodium–glucose co-transporter 2 (SGLT2) inhibitors prevent HF hospitalizations among patients with type 2 diabetes^72-74 or in patients with HFrEF regardless of diabetes^{27, 28} whereas other glucose treatment strategies do not. It is also interesting to note that patients with a higher future HF risk identified by risk scores that include biomarkers such as albuminuria, seem to derive greater benefit from SGLT2 inhibitor therapy among patients with type 2 diabetes.⁷⁵ The biomarker profile may identify patients with cardiometabolic, cardiovascular and cardiac structural changes in patients predestined to develop HF or in other words pre-HF. Supporting this concept was the STOP-HF trial which provided evidence that screening with natriuretic peptides among individuals with cardiovascular disease or with cardiovascular risk factors such as diabetes and hypertension, can be helpful to prevent development of HF or left ventricular systolic or diastolic dysfunction.⁷⁶ Accordingly, the 2017 ACC/AHA/HFSA focused update of the guidelines for the management of HF incorporated recommendations for natriuretic peptide-based screening in the prevention of HF as a Class IIa recommendation.³⁸ Similarly, high-sensitivity cardiac troponin levels are associated with future development of incident HF in the general population^{77, 78} and in those with evidence of cardiotoxicity or cardiac injury in high-risk populations⁷⁹ allowing for treatment strategies to prevent development of HF. Thus, biomarker elaboration can further identify risk and presence of ultrastructural abnormalities in HF among asymptomatic patients and could be a marker for Stage B HF without development of macroscopic structural changes detectable by imaging or electrocardiography.

4.1 Trajectory changes in ejection fraction

Guideline-directed management and therapy can result in improvement in LVEF and reverse remodelling in patients with HFrEF.⁸⁰ The phenomenon of improvement and recovery of LVEF has led to a growing interest in the long-term outcomes and management of these patients and how they differ from ‘non-responders’, or individuals whose LVEF does not improve with treatment. Currently, there is no consensus definition for patients with HFrEF whose LVEF improves, which has led to a variety of terms describing this phenotype, including patients with ‘improved’ LVEF, HFpEF (borderline), HFpEF, and HF with recovered EF (HFrecEF). The magnitude of change that defines ‘recovery’ of LVEF is not standardized, but it is recognized that distinguishing HFrecEF from HFmrEF requires serial measurements of the LVEF to appropriately capture change over time as this group might represent HFrecEF or deteriorated HFpEF. Moreover, because the measurement of the LVEF is subject to significant intra- or inter-reader variability, small changes in the LVEF need to be interpreted cautiously. Thus, a recent scientific panel put forth a working definition of HFrecEF that includes: a baseline LVEF of ≤40%, a ≥10% increase from baseline LVEF, and a second measurement of LVEF of >40%.⁸⁰ In this formulation, recovered EF signifies improvement of LVEF to over 40% but not necessarily totally normalized. There have been other attempts to characterize improvement in EF as an increase in LVEF by more than 10%.⁸⁰ It is also important to recognize that the trajectory might not be linear and unidirectional, and a patient may have improvement followed by a decline in EF or vice versa depending on the underlying aetiology, duration of disease, adherence to GDMT, comorbidities or re-exposure to cardiotoxins.

4.2 Trajectory changes in clinical status

Another method that captures the HF trajectory relies on an assessment of the patient's clinical status, which can inform the risk for hospitalization for HF or for mortality. A de novo diagnosis of HF, also referred to as ‘new-onset HF’, carries an increased risk for adverse clinical outcomes since the patient is not likely to be treated with optimal GDMT at the time of diagnosis.

Most patients with HF have episodes of clinical worsening of HF, which has been previously defined as worsening signs or symptoms in concert with a hospitalization.⁸¹ Data from more contemporary studies resulted in expansion of worsening HF to also include patients who require escalation of outpatient therapies, such as diuretics, even without a hospitalization.⁸² This is because the need for intensifying diuretic therapy, regardless of location (inpatient or outpatient), portends a worse prognosis than a patient who does not require intensification of therapy. Worsening HF implies a period of stability preceding a deterioration of signs and symptoms. However, the phrase ‘stable’ HF may be a misnomer, as patients with HF always carry a residual risk for hospitalization or sudden cardiac death, even when minimally symptomatic or asymptomatic receiving optimal treatment. For such patients remission may be a more suitable term.⁸³ When a patient with worsening HF does not improve with therapy escalation and continues to decline, she or he can be referred to as refractory to treatment. These patients are often assessed for advanced therapies such as mechanical circulatory support or cardiac transplant, or if they do not qualify for advanced therapies, clinicians can consider referral for palliative care.

Patients may have improvement in HF symptoms, functional capacity, quality of life and exercise performance with GDMT. Some patients with reversible or treatable causes of HF such as cardiomyopathy due to hypertensive heart disease, alcoholic cardiomyopathy, peripartum cardiomyopathy, or tachycardia-induced cardiomyopathy may even recover from HF with treatment and manifest resolution of HF symptoms, as well as normalization of the EF and cardiac structure. These patients require close follow-up and continuation of treatment to ascertain that HF symptoms or left ventricular dysfunction do not re-occur in the future.⁸⁴

6.1 Symptoms

Heart failure, like many non-categorical diseases, is widely held to be a clinical syndrome, devoid of any single pathognomonic histological or biochemical signal, and being the possible end result of many quite distinct and varied clinical disease states. Common symptoms and signs of HF are listed in Table 6.

The current ACCF/AHA classification of HF³ includes two pre-symptomatic stages, A and B. Although we restrict the definition of the syndrome of HF to being a symptomatic clinical condition, our proposed revised stages still straddle the pre-symptomatic stages. To not lose the advantage that the A/B/C/D staging system offered, to incorporate the asymptomatic phases under the HF umbrella, and to enhance understandability of these asymptomatic phases, we propose a new categorization of Stages A and B into ‘at risk’ and ‘pre-HF’ in Section 9 below.

6.2 Objective marker

In learning from other disease states that incorporated a core and frequently measured variables in their definition, such as acute MI, eGFR in CKD, glycated haemoglobin in diabetes, bone mineral density in osteoporosis or forced expiratory volume in the first second in COPD, making the diagnosis more accessible to non-specialists and more reliable and consistent between observers, hospitals and health care systems, we propose the incorporation of an objective measurement in addition to the symptoms in the HF definition.

In HF, possible candidates for such a measurement might theoretically be haemodynamic measures such as elevated pulmonary capillary wedge pressure and right atrial pressure by right heart catheterization, biomarkers associated with congestion such as natriuretic peptides, measures of neurohormonal overactivity or measures of exercise limitation such as maximal oxygen consumption. None of these are commonly or reliably associated with the disease states of HF; for example, the LVEF can vary from low through normal to high and still be part of an HF syndrome; no single haemodynamic measure is adequate to serve as a practical, non-invasive and reliable measurement; measurement of exercise limitation with cardiopulmonary exercise testing with expired gas exchange is not practical or universally available; and to date, neurohormone levels have not universally been considered reliable measures of the disease state. The closest have been the natriuretic peptides, which are recommended in modern guidelines as both diagnostic tests of reasonable clinical usefulness with prognostic utility and as good tests to rule out HF as a cause of breathlessness in certain settings.^{4, 38} Contemporary guidelines already state that natriuretic peptides can be used as an initial diagnostic test, and that patients with normal plasma natriuretic peptide concentrations are unlikely to have HF.^{4, 38} A detailed diagnostic algorithm will require specific operational thresholds based on individual natriuretic peptides and assay systems, as well as detailing other clinical features which can affect natriuretic peptide levels, but for common clinical purposes simple thresholds can be established that have sufficient operational accuracy to be incorporated usefully into a universal definition of HF.

6.3 Proposed new heart failure definition

Such a definition is comprehensive and practical enough to form the base which allows further subclassifications and which can encompass formal disease stages, both with universal applicability, prognostic and therapeutic validity, and an acceptable sensitivity and specificity. Please note that the definition of HF requires not only symptoms or signs (Table 6) but also presence of either elevated natriuretic peptides or objective evidence of pulmonary or systemic congestion by diagnostic modalities. For example, it would be important for peripheral oedema or ascites (Table 6) to be corroborated by presence of elevated right-sided cardiac filling pressures or rales by presence of elevated left-sided cardiac filling pressures; or elevated natriuretic peptides. It is also important to note that elevated jugular venous pressure estimated by an experienced clinician could be accepted as an objective evidence.

Please also note that in certain patients, congestion and haemodynamic abnormalities may become manifest with provocation such as exercise, especially in patients with HFpEF. This can support the diagnosis of HF. It is also critical to note that in patients with low perfusion and hypovolaemic state, there may not be any evidence of congestion or elevated filling pressures, but rather decreased cardiac output accompanied with low or normal ventricular filling pressures⁹⁴ (e.g. in the setting of over-diuresis in patients with HF). Once the hypovolaemic state is corrected, patients with HF usually have elevated filling pressures.

In the definition above, we did not specify left or right HF. Though left heart HF, and in advanced stages, biventricular HF are common, right HF can also be recognized as part of the above definition when patients present with symptoms or signs (Table 6) caused by a cardiac abnormality and have elevated natriuretic peptide levels or objective evidence of cardiogenic pulmonary or systemic congestion. Right HF primarily due to cardiac abnormalities such as arrhythmogenic right ventricular cardiomyopathy (ARVC) would be part of this definition.

We recognize that asymptomatic stages with patients at risk (former Stage A HF), or patients with structural heart disease or cardiomyopathies (former Stage B HF) would not be covered under the above definition as having HF, which emphasizes symptoms and signs of HF, but we conceptualize the HF syndrome as a continuum of disease with certain stages, such as pre-HF. This is similar to the approach with other disease states such as cancer, which defines those at risk and pre-cancer. The stages preceding the symptomatic phases as those at risk and pre-HF will be discussed in the following section.

We also realize certain patients with competing diagnoses such as CKD with marked volume overload, can present with symptoms and signs of HF, have elevated natriuretic peptides, and may even have evidence of congestion by imaging or elevated filling pressures. Although some of these patients may have concomitant HF, these patients have a primary abnormality that may require a specific treatment beyond that for HF. In the following section, we will address such other syndromes.

7.1 Right heart failure

The most common cause of right HF is left HF. However, right HF is characterized not only by signs and symptoms of right-sided HF but also by right atrial enlargement or right ventricular dysfunction. The presence of right HF in the setting of left HF is typically due to post-capillary, WHO group 2 pulmonary hypertension and may require modified treatment approaches and portends a poor prognosis; therefore, recognition of biventricular HF is important.⁹² Given the importance of these distinctions, the classification of types of ventricular failure in HF commonly includes three categories: left ventricular failure, right ventricular failure and combined left and right ventricular failure usually termed as biventricular failure. We believe isolated right HF due to primary pulmonary hypertension aetiologies (WHO groups 1, 3, 4), though may have symptoms or signs which may mimic HF and may have elevated natriuretic peptide levels, would not be categorized under HF, as the signs and/or symptoms are not caused primarily by a structural and/or functional cardiac abnormality. On the other hand, right HF due to primary right ventricular conditions such as ARVC would be categorized under HF.

7.2 Acute myocardial infarction/acute coronary syndrome

Acute MI may be complicated by HF. Given its acuity, specific pathophysiology and specific treatment strategies, we believe acute MI would be the overarching definition for the episode in proximity to acute MI. It is also possible that these patients may recover with timely treatment strategies and not progress to chronic HF, but also many may progress to chronic HF. In clinical trials, patients with acute MI or acute coronary syndrome within 6 weeks are usually excluded from clinical trials in HF. These patients may present with asymptomatic left ventricular dysfunction or pre-HF or symptoms and signs of HF due to a cardiac abnormality and may have elevated natriuretic peptides or evidence of congestion by imaging or haemodynamics. During the acute phase, these patients are diagnosed as having an MI complicated by HF rather than with HF alone. This does not mean acute MI should be replaced by HF alone, but it does mean the setting and specific aetiology of HF can be an important feature that determines specific therapeutic approaches. This setting has also been subject to specific clinical trial evaluation.^95-97 In addition to specific therapies for acute MI, these patients have indications for specific treatment for asymptomatic left ventricular dysfunction (pre-HF or Stage B HF) or symptomatic HF complicating acute MI during the acute phase, or as primary diagnoses in the chronic phase post-MI.

7.3 Cardiogenic shock

Another important form of HF is cardiogenic shock, which is the clinical state of organ hypoperfusion due to severe cardiac dysfunction. In cardiogenic shock, the symptoms and signs reflecting HF include hypotension unresponsive to volume repletion, altered mental status, cool extremities, and laboratory evidence of end-organ dysfunction such as elevated lactate levels due to hypoperfusion.⁶⁵ Cardiogenic shock is an extreme form of HF which requires some form of definitive therapy such as intravenous inotropes, vasopressors, or mechanical circulatory support. Cardiogenic shock is a type of HF, but due to its specific haemodynamic and clinical characterization requiring specific therapies such as vasoactive agents, circulatory support and/or revascularization depending on the aetiology, we believe keeping the descriptor ‘cardiogenic shock’ will help to identify a patient cohort with specific and urgent treatment needs. Cardiogenic shock may occur as an acute de novo presentation (e.g. large acute MI, fulminant myocarditis) or with progressive deterioration in a patient with chronic HF. Subacute cardiogenic shock may be in continuum of the wet and cold advanced HF patient with low cardiac output state. Such patients may meet the criteria for cardiogenic shock especially when they have evidence of end-organ dysfunction. A system describing stages of cardiogenic shock has been proposed by SCAI and other societies and characterizes the patients as Stage A ‘at risk’ for cardiogenic shock, stage B ‘beginning’ shock, stage C ‘classic’ cardiogenic shock, stage D ‘deteriorating’, and E ‘extremis’.⁶⁵ Such classification is important to characterize the severity and stage of shock, but it is also important to acknowledge the presence of HF as the preceding cause of shock in such patients, and identify advanced HF complicated with cardiogenic shock as the diagnosis.

7.4 Hypertensive emergency and hypertensive heart disease

Hypertensive emergencies encompass a spectrum of clinical presentations of uncontrolled blood pressure associated with end-organ damage that can include acute left ventricular dysfunction, pulmonary oedema, MI/ischemia and/or aortic dissection. All of these complications may result in or be complicated with an acute presentation of HF. Hypertension increases HF risk by two to threefold⁹⁸ and accounts for almost half of the HF cases in the US population as a population attributable risk.⁹⁹ Thus, both acutely hypertensive emergency and chronically, hypertensive heart disease can be complicated with HF. Treatment of hypertension is of utmost importance in the prevention and treatment of HF, underlined as a Class I recommendation with strong level of evidence in guidelines.^{4, 38}

7.5 Valvular heart disease

Aortic stenosis and mitral regurgitation can result in HF. Valvular heart disease is acknowledged as a specific disease, as it results in specific haemodynamic and ventricular alterations and requires specific treatment strategies targeting valvular abnormality. Most HF clinical trials exclude significant valvular heart disease for these reasons.

7.6 Congenital heart disease

Some types of congenital heart disease can result in HF. Incomplete or palliative correction of a congenital lesion leading to a chronic state of haemodynamic stress may result in subsequent HF, especially in complex congenital heart diseases such as tetralogy of Fallot, single ventricle defects and transposition of the great arteries. Additional myocardial, coronary or conduction system injury can occur due to complications of corrective surgery and can lead to progressive contractile dysfunction in some patients. The treatment should target the underlying anomaly and specific haemodynamic conditions.

7.7 High-output failure

High-output HF presents with similar symptoms and signs of systemic or pulmonary congestion, frequently associated with rapid heart rate and signs of peripheral vasodilatation. Cardiac dysfunction may be represented by pathologically elevated cardiac output, echocardiographic signs of right ventricular dilatation or dysfunction, and elevated natriuretic peptide concentrations. High-output HF is a response to extracardiac causes including liver disease, arteriovenous shunt, lung disease, thiamine deficiency, anaemia, thyroid disease or myeloproliferative disorders. Treatment is generally directed to the underlying causes. Given the unique nature of high-output failure, it is appropriate that it has a separate classification.

7.8 Other overlapping and competing diagnoses with heart failure

Patients can experience clinical deterioration as specific events that may not necessarily meet the universal definition of a diagnosis of HF. Such occurrences consist of events of a primary disease process that may be associated with signs and symptoms of HF as a result of the primary cause that is not HF at that encounter. These can include cardiovascular causes such as acute MI or acute coronary syndrome, hypertensive emergency as mentioned above, and also other cardiovascular primary diagnoses such as atrial fibrillation with rapid ventricular response, prolonged ventricular arrhythmias, pulmonary embolus, pericardial diseases, and acute valvular dysfunction. In these cardiovascular diagnoses, complication with HF is associated with worse prognosis and outcomes and underlines the urgency of addressing the underlying problem as well as the HF.

Other non-cardiovascular entities such as renal failure, liver failure, morbid obesity with peripheral oedema and chronic respiratory failure hypoventilation syndrome may present with symptoms and signs that mimic HF. Due to volume overload and neurohormonal compensatory mechanisms involved in some of these disease states, symptoms, signs and even haemodynamic characterization and biomarker profile can overlap with HF, and these patients may indeed also have concomitant HF. In these cases, the proximate cause of the signs and symptoms of volume overload is a distinct entity to which treatment is often primarily directed, in addition to HF. These events are often of significant interest to clinical events committees of clinical trials, where they may be considered as an event ‘with HF’ rather than a primary HF event. Another important concept that supports the principality of these competing diagnoses are that the symptoms and signs of HF may disappear once the underlying primary cause is treated, for example symptoms and signs that mimic HF may resolve with haemodialysis in a patient with end-stage CKD who may have missed a dialysis appointment. Thus, it is important not to catalogue every presentation with shortness of breath and oedema that requires treatment with fluid management strategies or diuretics as HF. It is, however, also important to not miss the complication with HF which requires timely management of HF as well as the proximate cause. Many of these factors can contribute to worsening outcomes in a complementary fashion in patients with HF. For example, patients with HF associated with CKD or diabetes are at much higher risk than those without. Rather than ‘competing’, these diagnoses can become complementary comorbidity risk factors to HF for worse outcomes.

8.1 NYHA classification

The NYHA funtional classification is important to characterize symptoms and functional capacity of patients with symptomatic (Stage C) HF or advanced HF (Stage D). The NYHA classification system categorizes HF on a scale of I to IV; Class I: no limitation of physical activity, Class II: slight limitation of physical activity, Class III: marked limitation of physical activity, and Class IV: symptoms occur even at rest; discomfort with any physical activity. We believe it is important to specify NYHA class at baseline after the initial diagnosis, and after treatment through the continuum of care of a patient with HF. A patient with symptomatic HF (Stage C) may become asymptomatic with treatment. Since that patient will still be categorized as HF Stage C, NYHA class I can further specify his/her absence of current symptoms. Worsening NYHA class is associated with worse prognosis and any symptomatic patient with HF (NYHA class II–IV HF) should have further optimization of GDMT.

8.2 Recognition of clinical trajectory in heart failure

It is well-recognized that the natural history of HF encompasses changes in the clinical risk of hospitalization and death over time, with risk increasing from ‘pre-HF’ to ‘new onset/de novo HF’, and further increasing with each episode of ‘worsening HF’ where there is deterioration of HF signs and symptoms despite ongoing therapy, requiring hospitalization or outpatient escalation of therapy.¹⁰¹ It is crucial to identify both the stage of the patient's natural history, as well as recognize the patient's clinical trajectory (improving vs. stalled or persistent vs. worsening)¹⁰² for optimal treatment, risk mitigation strategies, and patient-centred discussions. Gaining perspective of not only where the patient stands at the point in time, but in which direction the patient is headed, is a critical element of determining whether to continue along the current therapeutic course or to change direction. Thus, a patient with ‘worsening chronic HF’ following initial stabilization of ‘new onset/de novo HF’ would alert a physician of the immediate high risk for recurrent hospitalization or death, particularly in the period of close proximity to the worsening HF event, and trigger escalation of disease-modifying therapies rather than a focus on decongestion with diuretics alone. Of note, we caution against a terminology of ‘stable HF’, as patients are expected to improve with GDMT. These patients should have further optimization of therapies despite perceived stability or improvement, as there is evidence for significant improvement in outcomes with additional therapies in these patients. Lack of improvement is a marker of worse prognosis and should be termed as ‘persistent’ rather than ‘stable’, and prompt clinicians to further optimize therapy. For those patients who have resolution of symptoms and signs of HF along with resolution of previously present structural and functional heart disease after a phase of symptomatic HF, we recommend ‘HF in remission’ or NYHA class I HF status rather than ‘recovered HF’, which should be reserved for patients who have persistent resolution of HF symptoms and signs, normalization of cardiac structure, function and biomarker profile following resolution and treatment of a fully reversible cause, especially in view of the TRED-HF trial results which demonstrated that many patient deemed to have ‘recovered’ from dilated cardiomyopathy will relapse following treatment withdrawal suggesting remission rather than recovery⁸⁴ (Figure 2). Full and persistent recovery is rare, and even in the setting of reversible causes, patients may have recurrence of symptoms and or develop left ventricular dysfunction in the future.

8.3 Acute versus decompensated heart failure

In this document, we do not use the terms ‘acute new-onset HF’ or ‘acute decompensated HF’, which are the terminologies commonly used to describe patients requiring hospitalization or urgent care. The indications for hospitalization and/or urgent care utilization vary, and most patients who require hospitalization for HF may have chronic progressively worsening HF, rather than an acute singular event. We realize these patients may present with rapid onset or progressively escalating symptoms and/or signs of HF which are associated with adverse outcomes, requiring urgent evaluation and treatment. We have elected to characterize these patients as having ‘decompensated HF’ which may represent acutely decompensated patients due to an inciting event (e.g. atrial fibrillation with rapid ventricular response) or chronically and progressively worsening patients with marked deterioration of HF signs and symptoms despite ongoing therapy requiring urgent intervention, hospitalization, or rapid escalation of therapies including advanced therapies.

We recognize that there are a variety of acute presentations of HF (e.g. myocarditis, peripartum, cardiotoxicity, stress cardiomyopathy, etc.) and other entities associated with acute presentations of HF such as hypertensive emergency and acute MI, that will require specialized treatment strategies targeting the underlying aetiology. These have been addressed by other investigators^103-105 and are beyond the scope of this document.