A Contemporary Perspective on Acute Decompensated Heart Failure Classification: A State-of-the-art Review from an International Expert Group

Acute decompensated heart failure (ADHF) remains a major cause of hospitalisation and mortality.^1,2 Despite several European Society of Cardiology (ESC) guidelines and consensus documents, the place of ADHF within the overall spectrum of heart failure (HF) has long been debated.^3–13 The definition and classification of ADHF remain controversial and insufficiently adapted to modern clinical practice.^14,15 The heterogeneity of the pathophysiology and clinical presentations of this syndrome, as well as the variable relationship between the chronic condition and the episodes of acute decompensation, remain major factors hindering a simple and thorough classification of the disease.¹⁴ The nomenclature of ‘acute’ may create confusion because in most cases, decompensation occurs gradually and does not represent a new acute disease, but rather a deterioration of the primary cardiovascular condition.¹⁴

ADHF remains a clinical diagnosis with no single biological or haemodynamic parameter being a sine qua non for its detection and classification.⁸ The current classification schemes often overlap and do not provide binary therapeutic ramifications as do other cardiovascular emergencies, such as acute coronary syndromes (ACS) or acute aortic syndromes, where the presence of ST-elevation or involvement of the ascending aorta dictates further management algorithms. A robust classification scheme should include timely, actionable items to generate immediate therapeutic interventions and appropriate disposition decisions (Figure 1). Also, the classification of ADHF patients should facilitate shared decision-making with the patient and should inform future clinical research and new epidemiological studies.

The classification of ADHF has evolved significantly in the ESC HF guidelines from 2008 to 2021, moving from frameworks purely based on haemodynamics and blood pressure (BP) toward more holistic, phenotype-oriented approaches (Table 1).^3–7 However, existing systems still fall short of fully integrating time course, aetiology, triggers, severity and in-hospital and long-term trajectories into a single, actionable model. ADHF is a dynamic syndrome characterised by episodes of acute decompensation separated by periods of relative clinical stability. These fluctuations occur not only across the lifespan of the disease but also within the timeframe of a single hospitalisation. In-hospital trajectories reflect the patient’s clinical response to initial and subsequent rescue therapies, typically IV diuretics, vasodilators, inotropes or mechanical circulatory support.¹⁶

In the long term, patients follow distinct long-term trajectories which are influenced by the interplay of cardiac function, comorbidities, precipitant events and treatment intensity (Figure 2). Some HF patients demonstrate rapid downward spirals and follow a progressive decline with possibly numerous episodes of acute decompensation.^17,18 Each hospitalisation marks a turning point: the risk of death and rehospitalisation increases incrementally with every subsequent unplanned admission.^17,18 More often, ADHF patients follow a relatively stable course with infrequent admissions.^19–23 Introducing long-term trajectories into the classification of ADHF is clinically useful because it may help identify the likelihood of future events, as well as the patients who require advanced therapies such as left ventricular assist device (LVAD) support or heart transplantation.

The concept of trajectories in HF, both in-hospital and long-term, provides a structured framework for describing patterns of clinical change, anticipating risks and guiding therapeutic decisions.

Classifications of ADHF

Classification of ADHF patients may serve several key purposes: early decision-making regarding triage and treatment intensity, targeted therapy for high-risk subgroups, setting appropriate goals aligned with disease stage, shared decision-making with the patient and family, and improvement of both clinical and patient-centred outcomes (Figure 1).

Previous Classifications

All previous guidelines considered de novo versus worsening chronic HF (WHF) classification (Table 1).^3–7 ADHF may present as a first occurrence (de novo) or more frequently, as decompensation of chronic HF, which may be caused by primary cardiac dysfunction or precipitated by extrinsic factors. Decompensation of chronic HF can occur without known precipitant factors, but more often may be accompanied by one or more factors, such as infection (particularly respiratory), uncontrolled hypertension, rhythm disturbances, or the patient is failing to adhere to drugs or diet modification.^24–27 Acute myocardial dysfunction (ischaemic, inflammatory or toxic), acute valve insufficiency, hypertensive emergency, or pericardial tamponade are among frequent acute primary cardiovascular causes of ADHF. Although in-hospital mortality was similar for the two categories, 1-year mortality was significantly higher in patients with WHF compared to de novo AHF, confirming that hospitalisation for HF represents an important predictor of long-term mortality.^28–30 However, despite its simplicity, this scheme is not helpful to guide therapy or disposition decisions.

The ESC HF 2012 guidelines proposed a classification based on systolic BP (SBP) at admission with three categories: <85 mmHg, 85–110 mmHg and >110 mmHg.⁴ Although easy to apply in clinical practice and showing little geographical variability across registries, these SBP thresholds have never been validated in randomised controlled trials (RCTs) or large observational studies.³¹ Particularly for ADHF, specific SBP cut-off thresholds may not be consistently clinically relevant and may not reliably direct immediate management strategies in the absence of signs/symptoms of a low output state, laboratory and/or other diagnostic variables. SBP can vary significantly during the initial hours of presentation and represents an individualised ‘process’ rather than a simple number at one point in time. Also, it is very difficult to distinguish between a high SBP that may be a precipitant for decompensation or a reactive stress response.³² However, despite these limitations, SBP >85 mmHg remains important for initiation or discontinuation of IV therapies.

The 2016 ESC HF guidelines included classification based on the congestion and perfusion state.⁵ Although mutually exclusive, this classification has several limitations. A criticism of this last classification is that 6–9% of patients in recent registries have neither congestion nor signs of hypoperfusion.^33,34 To explain this, it is important to note that some ADHF patients may be treated with IV vasoactive drugs in the pre-hospital phase (in the ambulance or in the emergency department [ED]) with resolution of signs/symptoms of HF at hospital admission, changing their trajectory. Mild signs of congestion, potentially undetected at initial evaluation, may still have caused sufficient symptoms for patients to seek acute care. Also, significant alterations in vital signs or a low-output state may occur without signs of hypoperfusion at a particular point in time, yet the overall clinical history warrants hospital admission for management of HF. Further, this classification does not differentiate between left-sided and right-sided congestion, conditions that may have a major impact on the clinical course in hospital and post-discharge. The proponents of this classification have acknowledged these limitations.³⁵ Despite these limitations, this classification offers a robust risk stratification in the ED, at hospital admission and discharge and provides immediate therapeutic implications.^6,7,33,34

As stated in the 2016 ESC HF guidelines, treatable causes of ADHF with specific management pathways and triage dispositions must be addressed as the first step of ADHF management.⁵ Some of these are considered life-threatening causes/aetiologies that must be identified and treated before initiating any other standard HF management algorithm process; they include acute coronary syndrome, hypertension, arrhythmia, acute mechanical causes and pulmonary embolism (CHAMP).⁵ Other causes may also be considered, such as tamponade and infection, leading to the CHAMPIT acronym proposed in the 2021 ESC HF guidelines.⁶

The 2021 ESC HF guidelines proposed a classification based on clinical profiles: decompensated heart failure, cardiogenic shock (CS), acute pulmonary oedema (APO) and acute right heart failure.⁶ The 2008 ESC HF guidelines also considered this classification.³ When considering a classification based on clinical profile, substantial geographical variability has been found regarding the prevalence of profiles.³¹ These differences may reflect variations in interpreting the definitions and different moments in time when the definitions are applied. Different clinical profiles may have similar clinical presentations at admission, challenging the initial clinical classification and potentially leading to misclassification and overlap among clinical phenotypes. Although clinical characteristics overlap among the different clinical profiles, clinical profile classification of ADHF patients may still facilitate early decision-making regarding appropriate triage and provide individual therapeutic approaches for certain clinical profiles.

The New York Heart Association (NYHA) functional classification remains the most frequently used tool to assess the symptom burden in patients with ADHF in all guidelines.^3–7 Previous registries have consistently shown that NYHA class at admission is strongly associated with prognosis and higher NYHA class predicts longer hospitalisation, increased risk of readmission and higher mortality.^24,28,30,31 More recent analyses from the ESC-EORP-HFA Registry³³ also confirmed its role in risk stratification, particularly when combined with congestion and perfusion profiles.

To note, several consensus documents emphasise that while the NYHA class is a cornerstone of functional assessment, it has important limitations.^9–12,16 A key issue in ADHF is variability over time: NYHA class often differs substantially between admission, when symptoms are most severe, and discharge, after symptoms have stabilised. This dynamic shift reflects both treatment response and the challenges of using a static scale in a rapidly evolving condition, underlining that NYHA is useful for broad categorisation when compared at similar clinical time points but insufficient when used in isolation. Further, in the era of more advanced prognostic assessment, including severity of clinical and haemodynamic congestion, natriuretic peptides and markers of neurohormonal activation, the discriminatory power of the NYHA class diminishes. Once these indices are considered, NYHA adds little incremental prognostic information. Another limitation is subjectivity. NYHA is typically rated from the clinician’s perspective rather than the patient’s. Goode et al. demonstrated that physician- and patient-reported NYHA class agreed in only about half of cases.¹⁵ This discrepancy underscores that NYHA may not fully capture the patient’s lived experience, especially in ADHF, where symptoms can fluctuate rapidly.

Left ventricular ejection fraction (LVEF) offers little additional information in terms of the initial decongestion strategies, since the severity of congestion, clinical course and the response to decongestive therapies are similar between HF with reduced ejection fraction (HFrEF) and HF with preserved ejection fraction (HFpEF).^36–39 Also, in ADHF, the clinical benefit of uptitrating GDMT has been observed irrespective of LVEF, distinct to chronic settings where initiation and optimisation of GDMT are strongly influenced by LVEF classification, with the most robust outcome benefits demonstrated in HFrEF.³⁹ However, initiation and up-titration of GDMT should be considered a distinct therapeutic phase following clinical stabilisation. In the recent trials that recruited ADHF patients during hospitalisation, the intervention benefits were evident regardless of LVEF. Initiation of empagliflozin in the EMPULSE trial was beneficial across the whole spectrum of LVEF.⁴⁰ Similarly, in the pooled data from the PIONEER-HF and PARAGLIDE-HF trials, the benefit of sacubitril/valsartan after the AHF episode was present in all patients and particularly in those with LVEF ≤60%.⁴¹ A recent analysis of the FINEARTS-HF trial has also demonstrated that HF with mildly reduced EF (HFmrEF) and HFpEF patients who had been recently hospitalised for HF before initiation of finerenone benefited from mineralocorticoid receptor antagonist therapy.⁴² The positive effects of the comprehensive, intensive neurohormonal blockade in patients stabilised for an AHF episode in the STRONG-HF trial were present across the whole spectrum of LVEF.^43–46

Nevertheless, LVEF retains decisional value in acute haemodynamic assessment, particularly in patients presenting with low-output states, where severely reduced LVEF may support the use of IV inotropes and guide decisions regarding temporary reduction or withholding of GDMTs during hypoperfusion, particularly β-blockers.^10,11

Proposed Classification Scheme

A more granular classification of ADHF is needed because the current schemes provide only a static snapshot of a dynamic syndrome and fail to describe multifactorial components of HF decompensation. Any approach for a new classification of ADHF should consider heterogeneity of the syndrome and there are several axes where every patient with ADHF can be placed: cardiovascular substrate (normal, structural heart disease with or without previous history of HF), potential triggers, the burden of associated non-cardiac comorbidities, clinical profiles, care settings (hospital, ED, outpatient clinic), in-hospital and long-term trajectories, and previous cardiovascular treatment.

The TACCITT classification (Figure 3) provides a multidimensional framework for ADHF, integrating Triggers, Aetiologies, Comorbidities, Clinical phenotypes, Individual settings of care, Trajectories and Therapies. A refined classification would offer several advantages: integration of acute and chronic perspectives, contextualising the acute episode within the patient’s disease history; provide actionable items, as categories would directly point toward tailored management strategies; (dynamic assessment, allowing reclassification based on the patient’s response to therapy; integration of multidisciplinary support, highlighting when input from nephrology, surgical and interventional teams is necessary; inform about futility by integrating trajectories, comorbidities, therapy tolerance and avoiding escalation of therapies; and research facilitation, providing standardised, reproducible phenotypes that improve trial design, comparability and quality improvement initiatives.

Unlike traditional classifications that overlook where the patient stands in the broader long-term journey of HF and provide only a static snapshot, TACCITT captures the acute presentation and the patient’s long-term journey. This approach allows clinicians to link the acute event with underlying causes, comorbidity burden and therapeutic tolerance. This type of classification is extremely useful for repeated events. A new decompensation in a patient with a remitting-relapsing trajectory points to identification and control of triggers, while a repeated event in a patient with a downsloping trajectory signals disease progression and prioritises referral to advanced HF programmes. Some triggers may be recurrent, suggesting personalised prevention strategies, such as better control of hypertension in patients with high SBP or immunisation in patients with repeated infections. In addition to triggers, the progression of underlying cardiac pathology, particularly coronary artery disease, plays a significant role in the recurrence of ADHF.

Cardiovascular Substrate and Aetiologies

ADHF episodes may occur in patients with or without previously identified cardiac dysfunction. Acute loss of myocardial tissue or function, such as acute MI, myocarditis or spontaneous rupture of mitral valve chordae, can lead to ADHF, especially in patients with previously normal cardiac structure. Usually, these patients have a clear aetiological factor that, if resolved, contributes to HF remission.

In ADHF, aetiologies are very diverse and may include ACS, valvular disease, arrhythmias, cardiomyopathies, including transient forms such as peripartum cardiomyopathy, takotsubo syndrome, as well as cardiac involvement related to malignancy itself and treatment-related cardiotoxicity from anti-cancer therapies.^47–55

Each mechanism contributes to different haemodynamic and prognostic profiles, requiring specific diagnostic and therapeutic strategies. Primary ADHF refers to ADHF as the leading cause of hospitalisation, while secondary ADHF develops in patients admitted for other medical conditions, such as infection, ACS or non-cardiac surgery. Primary ADHF, present at admission, is predominantly driven by intrinsic cardiac disease progression and is associated with higher rates of HF–related rehospitalisation. Secondary ADHF develops during hospitalisation and typically arises from acute precipitating events, such as infections or acute MI.⁵⁶ Primary ADHF necessitates comprehensive optimisation of GDMT and structured follow-up programmes to minimise the risk of recurrent hospitalisations. Conversely, secondary ADHF, often triggered by reversible causes, requires a focused approach addressing the precipitating factors.

Perioperative ADHF represents a distinct secondary form, often driven by haemodynamic stress, inflammation or fluid imbalance, with significant management and prognostic challenges.⁵³

Incorporating aetiology into ADHF classification is essential, as it contextualises the clinical presentation, refines risk stratification and guides causal therapy beyond symptomatic improvement. For instance, when ACS is the underlying cause, urgent revascularisation is immediately required, while for severe aortic stenosis, transcatheter aortic valve implant or cardiac surgery must be considered.

Triggers

ADHF has diverse precipitants, including infections, ischaemia, arrhythmias and uncontrolled hypertension. Early identification is crucial, as some acute triggers require urgent correction. The 2021 ESC guidelines recognise several life-threatening clinical conditions (referring to the CHAMPIT scheme), which require urgent specific treatments to avoid further deterioration.⁶ Distinct from these acute phase precipitants, decompensation results more commonly from a continuum of chronic overlapping mechanisms – episodes of silent ischaemia, worsening of secondary valvular regurgitation, sub-clinical AF and sodium retention – that are triggered days to weeks before presentation rather than acutely.^24–27

How intense should the precipitants be to produce HF decompensation is not clear. With notable exceptions, such as acute MI and rupture of the cardiac structure, most precipitants are mild or moderate and it is more plausible that progressive deterioration of the underlying cardiac disease makes the heart more vulnerable to diverse associated conditions. Second, some entities, considered to be triggers for decompensation, may, in reality, be direct manifestations of deterioration of cardiovascular substrate (AF as a consequence of increased filling pressure) or enhanced neurohormonal activation (high SBP as a reactive stress response).

Comorbidities

Non-cardiac comorbidities are highly prevalent in ADHF, due to demographic ageing and the growing burden of chronic diseases.^57–59 In ADHF, distinct comorbidity patterns exist across LVEF phenotypes, with higher multimorbidity observed in HFpEF. In addition, the association between each individual comorbidity and post-discharge outcomes varies substantially in patients with HFrEF, HFmrEF and HFpEF, suggesting that an LVEF-specific multidisciplinary approach with distinct comorbidity management programmes should be applied in the post-discharge phase.⁵⁹ Patients with multiple comorbidities often respond poorly to decongestion and will have residual congestion at discharge, and are less likely to receive GDMT; instead, they require more non-cardiovascular medications.^59,60 These factors contribute to longer hospital stays, therapeutic complexity and poor post-discharge outcomes, underscoring the importance of integrating comorbidity profiles into ADHF classification.^57–59 Recognising these patterns can optimise risk stratification, guide individualised therapy and address polypharmacy challenges, which are increasingly relevant in contemporary ADHF care.^61,62

Clinical Profiles

Several ADHF registries applied a clinical profile-based classification, grouping patients into four main syndromes: decompensated heart failure, CS, APO and isolated acute right heart failure.^31,63–65 The 2021 ESC HF guidelines used the same classification.⁶ This approach reflects real-world bedside assessment and allows meaningful comparisons of outcomes across diverse populations. This classification individualises distinct lines of management, with a stepped approach (initial and then escalation therapies) based on the severity of the congestion/hypoperfusion.

CS encompasses heterogeneous aetiologies with distinct clinical implications. ACS-related CS results from acute ischaemic myocardial injury leading to pump failure, has a rapid onset and is characterised by profound hypotension and low cardiac output.^66,67 HF-related CS arises from chronic or progressive ventricular dysfunction and is now the most frequent CS phenotype in many contemporary registries.⁶⁷ HF-related CS is characterised by more gradual onset, higher prevalence of comorbidities, more pronounced congestion and more common right ventricular (RV) dysfunction. HF-related CS can be further subclassified based on a history of HF into acute de novo HF-CS and acute-on-chronic HF-CS.⁶⁶ Acute de novo HF-CS occurs in patients without prior HF, often triggered by arrhythmias, infection or ischaemia, with severe perfusion deficits but fewer chronic comorbidities. In contrast, acute-on-chronic HF-CS arises in patients with established HF, frequently associated with baseline congestion and comorbidities and these patients experience higher rates of rehospitalisation and long-term mortality.⁶⁶

Although acute RV failure is a distinct profile, RV dysfunction may contribute to the other profiles, except APO. RV dysfunction in ADHF reflects complex pathophysiology involving pressure and volume overload, impaired contractility and ventricular interdependence.⁶⁸ It carries major prognostic implications in ADHF, particularly in valvular heart disease, CS and advanced HF, where it predicts mortality and poor response to therapy.^50,69–72 Early recognition is crucial because acute RV dysfunction may be responsible for life-threatening presentations and requires more intense decongestion, complicating mechanical support decisions and affecting candidacy for advanced therapies.^6,73

Settings of Care

Venue of care is not a biological threshold and ADHF patients may have a substantial risk of death irrespective of the decision to hospitalise versus administer IV diuretics in the ED or outpatient settings.^74–76

WHF refers to the clinical deterioration of a patient with established HF, manifesting as new or progressive symptoms and signs that require escalation of therapy.⁷⁷ Importantly, WHF is distinct from de novo acute HF, which occurs in patients without a prior HF diagnosis and from advanced HF, which represents a chronic, end-stage phase characterised by persistent severe symptoms, marked functional limitation and frequent hospitalisations despite optimal therapy.⁷⁸

Traditionally, WHF has been defined as being the need for hospitalisation for worsening signs and symptoms, but contemporary definitions have expanded to encompass ED and outpatient settings, including events managed by intensifying oral diuretics.^6,79 To note, performance bar measures aimed to decrease 30-day hospitalisations may result in an increase in ED and ambulatory visits and were not associated with a decrease in long-term mortality.^80,81

The 2023 ESC/Heart Failure Association consensus document further defines WHF as an episode in which patients with chronic HF exhibit congestion and/or hypoperfusion requiring treatment intensification beyond baseline therapy, regardless of the care setting.¹² WHF leading to hospitalisation has been linked to a two- to threefold higher short-term mortality compared with stable patients.^12,79 While WHF episodes requiring hospitalisation have a worse prognosis than WHF in the ED or outpatient settings, on average, a substantial proportion of outpatients with WHF had similar or higher event rates compared to inpatients.⁷⁶ This underscores the continuum of WHF across care settings and the need for documentation of the history of HF for any episode of decompensation (regardless of treatment location). In the context of ADHF classification, WHF is best conceptualised as a trajectory modifier reflecting disease progression and clinical instability.

In-hospital Trajectories

In-hospital trajectories are based on the clinical response to the initial and subsequent ‘rescue’ decongestive therapies (Figure 3).^10,11,16 Four trajectories are identifiable:

• Patients who respond to initial decongestive therapies with clinical improvement and enhanced diuresis. This category accounts for 70–80% of hospitalisations in registries and clinical trials and the long-term prognosis is favourable.
• Patients who initially respond and then deteriorate. This was previously described as in-hospital worsening heart failure (InH-WHF) and occurs in 10–15% of patients hospitalised for HF.^16,82–85 InH-WHF may be associated with increased mortality up to 180 days regardless of the time of occurrence or intensity of therapy required.⁸³ Importantly, RCTs (ASCEND-HF, PROTECT, VERITAS, RELAX-AHF) using InH-WHF as an endpoint have had neutral results, raising concerns about the validity of InH-WHF to predict long-term outcomes.^80,82,84,85 However, in a systematic review including RCTs and observational studies, the risk for all-cause death at 6 months post-discharge was approximately two- to fourfold greater in patients who had experienced InH-WHF and it continued to be significantly elevated throughout the first year.⁸⁶ InH-WHF remains a marker of disease severity and a therapeutic challenge, underscoring the need for more effective interventions and refined trial endpoints.^82,83,85
• Patients presenting with refractory symptoms requiring further intensification of therapies from the beginning of the treatment. These patients often have advanced HF physiology and require adjunctive therapies, such as continuous IV diuretic infusion, combinations of diuretics, ultrafiltration, vasopressors or inotropes.^87–89
• Patients presenting with a continuous downward course and continuous worsening, many of those dying in the first hours of hospitalisation.

While the first trajectory may benefit from early initiation and up-titration of GDMT, the last two categories are typical presentations of advanced HF, in which many patients are intolerant of GDMT. Not uncommonly, some patients with CS may present in the third and fourth trajectory and may require escalation beyond pharmacological therapy. In this setting, temporary mechanical circulatory support (MCS) may be considered to stabilise haemodynamics, maintain end-organ perfusion and allow time for recovery or decision-making regarding definitive therapies.⁶⁶ Escalation to MCS represents a trajectory-modifying intervention within the TACCITT framework and aligns acute management with long-term pathways as patients requiring temporary support may transition either toward myocardial recovery, durable left ventricular assist device implantation or heart transplantation.

Describing these trajectories is clinically relevant because they may guide the type and intensity of therapy escalation, as well as the timing of initiation and up-titration of GDMT (Figure 3).^87–89 These trajectories should consider background GDMT and, notably, be synchronised with long-term trajectories (Figure 3). In-hospital trajectories should not be viewed in isolation because the in-hospital course is a critical node in the broader long-term trajectory of HF. Poor inpatient response to therapies may suggest transition to a higher long-term risk state. Patients who are early/sustained responders will very likely follow a stable or improving long-term trajectory if GDMT is optimised. For patients with InH-WHF or refractory patients, long-term transition into worsening or advanced HF trajectories is frequent. Finally, patients with a downward course often have a terminal trajectory unless they are rescued by advanced therapies.

Patient trajectories during inpatient management should be informed by treatment response, as this response could affect the patient’s phenotype. Sodium loss is the primary driver of water loss in response to diuretics and spot urinary sodium is a better surrogate of natriuresis and decongestion than urine output and weight loss.^90,91 The blunted tubular responsiveness to diuretics is the major mechanism of poor diuretic response and sodium retention in ADHF, which could be overcome by the use of higher loop diuretic doses or the addition of another class of diuretics.^92–96 However, with the notable exception of empagliflozin, it is important to mention that even if ‘add-on’ diuretic strategies studied in these trials improved decongestion metrics, they were associated with more adverse events, including worsening renal function, hypokalaemia and mortality.^95–98

Diuretic resistance, an inadequate diuretic response or the need for escalation of diuretic therapy represents an important marker of an unfavourable in-hospital trajectory in ADHF, reflecting persistent congestion and higher clinical risk.¹⁶ The requirement for intensified or ‘add-on’ diuretic strategies should therefore be interpreted primarily as a sign of disease severity and incomplete decongestion rather than a therapeutic endpoint.

Long-term Trajectories

The long-term evolution of patient condition following an ADHF episode is highly heterogeneous, reflecting distinct underlying pathophysiology (Figure 2 ). Some patients follow a descending/downsloping pathway culminating in advanced HF, characterised by progressive deterioration, recurrent events and poor outcomes.^16–18 Others follow a remitting-relapsing course, where episodes of decompensation occur but long-term stability can be maintained, supporting the view that ADHF is not always an inevitable step toward end-stage disease.^19–23 The stability of the cardiovascular substrate varies across LVEF phenotypes. In HFpEF, long-term remodelling appears limited, and in a long-term follow-up study, including 126 HFpEF patients, only 12% transitioned to reduced LVEF after 11 years.²⁰ By contrast, about 50% of HFrEF patients demonstrated EF improvement at follow-up, highlighting the dynamic nature of systolic dysfunction and the potential for recovery under optimal therapy.

Despite this heterogeneity, the magnitude of congestion at admission is similar across LVEF categories, both haemodynamically and clinically.³⁶ Although 1-year all-cause mortality and HF hospitalisation are significantly higher in ADHF patients with HFrEF compared to HFpEF, about a half of patients with HFrEF may experience an improvement in LVEF and may follow another trajectory.⁹⁸ In terms of clinical status at discharge, registry data suggest that advanced functional limitation at discharge is rare, with <5% of patients leaving the hospital with NYHA class IV symptoms, suggesting that most patients stabilise after decongestion.^31,63 In addition, over the long term, transition to advanced HF is not the rule. In a retrospective cohort study, including patients with incident HF with a mean age of 74 years, the cumulative incidence of advanced HF was 6.0%, 9.1% and 11.5% at 2, 4 and 6 years after incident HF diagnosis.²³ Whether these patients are identified in a timely manner and offered appropriate therapy remains an open question.

Hospitalisation burden also reflects these trajectories. In a study with a 3-year follow-up involving 10,363 patients, only 9% of patients, classified as ‘frequent admitters’, followed a downsloping trajectory, averaging 2.35 hospitalisations per year.²¹ In comparison, 90% of patients had a lesser burden, with just 0.5 hospitalisations per year.²¹ All these data suggest that two broad categories of ADHF patients can be distinguished with distinct trajectories (Table 2).

Several predictors of trajectory have been identified. Subramaniam et al., in a population-based cohort, showed that the cumulative incidence of advanced HF was strongly associated with older age, male sex, lower baseline EF, ischaemic aetiology, renal dysfunction and higher comorbidity burden.²³ These factors define patients at greater risk for a downsloping course. Importantly, the presence of multiple uncorrected aetiologies, such as ischaemic heart disease and valvular pathology, higher myocardial fibrosis burden, and suboptimal use or poor GDMT tolerance further accelerate progression toward advanced HF.

Lupón et al. showed that predictors of stability included higher baseline LVEF, female sex, lower natriuretic peptide levels and absence of extensive comorbidities.²⁰ This group illustrates the remitting-relapsing phenotype, where decompensations occur but do not necessarily imply irreversible decline. Clinically, classifying patients by trajectory can guide intensity of follow-up, consideration of early referral for devices or advanced therapies in downslopers and a focus on trigger control and decongestion quality in remitting-relapsing patients.

Background Therapies

In the classification of ADHF, background GDMT is a critical determinant of prognosis and therapeutic strategy. Patients hospitalised with AHF should ideally be on quadruple therapy, angiotensin receptor-neprilysin Inhibitor/angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, β-blocker, mineralocorticoid receptor antagonists and sodium-glucose co-transporter-2 inhibitors, at maximally tolerated doses.¹⁸ Accordingly, patients may be categorised as optimally treated, sub-optimally treated or treatment-intolerant. This distinction reflects both baseline risk and residual therapeutic opportunity. The STRONG-HF trial demonstrated that rapid initiation and up-titration of GDMT in the ADHF setting is feasible and improves outcomes, provided safety markers are closely monitored.⁹⁹ Thus, integrating background GDMT status into ADHF classifications aligns disease severity with therapeutic gaps and intensification of therapy and highlights modifiable risk through optimised therapy.

Clinical Practice Perspective

In clinical practice, the TACCITT classification is feasible as it integrates routinely available clinical, haemodynamic and patient history data. Components, triggers, comorbidities, phenotypes and therapy tolerance are systematically assessed during ADHF care, allowing structured application. Its layered design promotes real-time, dynamic patient profiling without requiring complex investigations.

The TACCITT classification could be made quantifiable by assigning weighted scores to each domain, such as specific triggers, comorbidities, care settings, trajectory stage and GDMT tolerance. This approach could generate a composite score reflecting both acute risk and long-term prognosis.

Research Perspective

The substantial investments in research and development have not yielded proof of efficacy and safety for most of the therapies tested, and outcomes for people with ADHF remain poor.^1,2

Despite substantial investment, most ADHF trials have failed to deliver therapies with consistent efficacy or safety, in part because inclusion criteria reflect a static admission phenotype rather than the dynamic, multifactorial reality of decompensation.

The TACCITT classification provides a framework for transforming clinical trial design by capturing the temporal, mechanistic and therapeutic dimensions of ADHF. Integrating trajectory, comorbidity and trigger domains allows trials to enrol patients at defined inflexion points in their disease continuum and to match interventions to dominant pathophysiological drivers, such as arrhythmic, ischaemic, renal and inflammatory. This promotes both mechanistic precision and temporal alignment of therapy delivery. Early enrolment trials, particularly those investigating decongestive strategies or device-based interventions, may rely primarily on initial domains (triggers, clinical phenotype, care setting). Other domains, such as trajectories and background therapy, often evolve during hospitalisation and may only become fully apparent after initial stabilisation. TACCITT should therefore be interpreted as a dynamic classification, allowing progressive refinement of patient phenotyping as additional clinical information emerges.

Further, TACCITT provides a quantifiable basis for dynamic endpoints, such as improvement in trajectory domain, trigger control or therapy tolerance, thus capturing true therapeutic benefit beyond short-term congestion relief. As its domains align with routinely collected data, TACCITT can serve as a standardised ontology for registry-based or adaptive trial platforms, enhancing interoperability and external validity. Ultimately, embedding TACCITT into future RCTs would support reproducible patient phenotyping, foster targeted intervention development and create a bridge between real-world registries and mechanistic clinical research, accelerating progress in ADHF therapeutics.

Epidemiological Perspective

From an epidemiological perspective, consistent case definitions and classifications of ADHF are critical for tracking temporal trends and comparisons across regions. The precipitants, aetiologies and phenotypes are variable based on temporal and regional changes in underlying comorbidities and HF evaluations.²⁴ Recent data from over 18,000 international patients suggests that signs and symptoms and aetiologies and precipitants vary significantly at presentation based on geographical region.²⁴ These differences in signs and symptoms may be influenced by several factors, including time from symptom onset to hospital presentation, underlying precipitants, aetiologies and comorbidities, and whether the presentation is de novo or worsening chronic heart failure. A classification of ADHF relying only on dyspnoea severity may misclassify patients and make it difficult to compare over time and across regions. Finally, less than 50% of patients in eastern Europe, south-east Asia, Central and South America, eastern Mediterranean and Africa have natriuretic peptide testing routinely performed at the time of hospital and ED presentation.²⁴ Thus, an ADHF classification needs to balance the increased accuracy associated with diagnostic testing and the generalisability of signs and symptoms.

Patient Perspective

It is critical for patients to understand both their stage of disease and point in the clinical trajectory. By anticipating disease trajectories, TACCITT classification may help to empower patients and families to engage in realistic planning and align care with personal values.

Regulatory Perspective

Since ADHF is a multi-event disease, any clinical event associated with worsening signs/symptoms should be acknowledged. Including only hospitalisations for ADHF, the burden of ADHF in terms of cost is largely underestimated, since many patients present to the ED or an ambulatory office for worsening signs and symptoms of ADHF. Regulatory bodies rely on clear definitions and reproducible classifications to evaluate therapeutic interventions. Incorporating comorbidity burden and disease progression trajectories will further align regulatory evaluation with real-world needs. From a regulatory perspective, the TACCITT classification offers a structured, multidimensional framework aligning acute and chronic perspectives.

Conclusion

Current classifications are limited, relying on static admission data that fail to capture evolving in-hospital and long-term trajectories. A multidimensional classification system integrating presentation, precipitants, comorbidities, clinical profile and trajectory of response offers a path forward. Understanding the patient’s position on the disease progression trajectory and recognising distinct response patterns allows for a more dynamic and tailored management strategy, adapting to the diversity of clinical scenarios encountered during the in-hospital and post-discharge course. Such an approach could enhance acute decision-making, improve prognostication, optimise clinical trial design and align patient care with regulatory standards.