Clinical characteristics and prognostic differences among acute heart failure phenotypes in the emergency department: a real-world study

Abstract

Background

The 2021 European Society of Cardiology (ESC) guidelines classify acute heart failure (AHF) into four phenotypes: acute decompensated heart failure (ADHF), acute pulmonary oedema (APO), isolated right ventricular failure (IRVF), and cardiogenic shock (CS). However, differences in clinical characteristics and long-term outcomes among these phenotypes in the emergency department (ED) remain unclear. This study aimed to evaluate their characteristics and prognostic implications in the ED setting.

Methods

In this prospective–retrospective observational real-world study, adults presenting with AHF in the ED were consecutively enrolled and categorized according to ESC-defined phenotypes. The study endpoints were 1-year all-cause and cardiovascular mortality. Associations between phenotypes and mortality were assessed using Cox regression models with progressive adjustment.

Results

A total of 2,960 patients were included. Clinical characteristics and treatment patterns differed substantially across phenotypes. The 1-year all-cause mortality rates were 10.3% in ADHF, 11.0% in APO, 24.5% in IRVF, and 58.7% in CS (P < 0.001). In multivariable Cox analyses adjusting for clinical variables and treatments, IRVF remained independently associated with increased risks of all-cause mortality (HR 1.92, 95% CI 1.02–3.63; P = 0.044) and cardiovascular mortality (HR 2.22, 95% CI 1.10–4.48; P = 0.027). CS was also independently associated with higher risks of all-cause mortality (HR 2.22, 95% CI 1.34–3.67; P = 0.002) and cardiovascular mortality (HR 2.71, 95% CI 1.57–4.66; P < 0.001).

Conclusions

In this ED-based AHF cohort, IRVF and CS were independently associated with worse 1-year outcomes, supporting phenotype-based risk stratification in the ED.

Introduction

Acute heart failure (AHF) is a clinical syndrome characterized by the sudden deterioration of cardiac function, leading to inadequate perfusion to meet metabolic demands. The 2021 European Society of Cardiology (ESC) guidelines classify AHF patients into four phenotypes based on clinical manifestations: acute decompensated heart failure (ADHF), acute pulmonary oedema (APO), isolated right ventricular failure (IRVF), and cardiogenic shock (CS) [1]. Previous studies have demonstrated that different clinical phenotypes of acute heart failure (AHF) are associated with distinct prognoses. Based on the 2008 ESC guideline classification, Ovidiu et al. conducted a one-year follow-up of 6,629 AHF patients and reported 1-year mortality rates of 28.1% for APO, 54.0% for CS, 27.2% for ADHF, 12.8% for right heart failure, 34.0% for hypertension-related heart failure, and 20.6% for acute coronary syndrome–related heart failure, showing significant prognostic heterogeneity among phenotypes [2]. According to the 2016 ESC guideline classification of AHF, the 1-year all-cause mortality rates for the “warm-dry,” “cold-dry,” “warm-wet,” and “cold-wet” profiles were 12.1%, 22.6%, 28.0%, and 26.4%, respectively. This hemodynamic classification based on congestion and perfusion status provided an effective means of risk stratification for patients with AHF [3]. Compared with previous classification systems, the updated 2021 ESC clinical phenotyping more accurately captures the diverse pathophysiological states of patients with AHF. By characterizing the clinical features and prognostic profiles of each phenotype, this framework can effectively guide clinical decision-making and play a positive role in improving patient outcomes.

The emergency department (ED) plays a crucial role in the management of patients with acute heart failure (AHF), as more than 80% of heart failure patients receive initial evaluation and treatment in the ED before hospitalization [4]. Timely and effective therapy in this setting can rapidly alleviate acute symptoms and significantly improve clinical outcomes [5]. However, most randomized controlled trials (RCTs) on heart failure have adopted strict inclusion and exclusion criteria, focusing primarily on chronic or hospitalized populations. In contrast, patients presenting with AHF in the ED often have more comorbidities, more severe illness, and more rapid clinical deterioration, yet they are rarely included in RCTs. Consequently, real-world evidence regarding this population remains limited, underscoring the urgent need for studies to fill this critical knowledge gap in the field of AHF [6].

Therefore, this study aimed to evaluate differences in clinical characteristics and treatment patterns across acute heart failure phenotypes in the emergency department, and to assess their association with 1-year all-cause and cardiovascular mortality.

Methods

Study design

This was a single-center, prospective plus retrospective, observational, registry study conducted at Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College.

Study population

Consecutive patients with AHF who presented to the ED of Fuwai Hospital between June 2024 and February 2025 were enrolled. Data from June to August 2024 were collected retrospectively, and data from September 2024 to February 2025 were collected prospectively. Patients in the retrospective phase were identified from electronic medical records using predefined criteria, and data were collected using the same standardized protocol as in the prospective phase. Inclusion criteria were as follows: (1) Diagnosis of AHF according to the 2021 ESC Guidelines for the Diagnosis and Treatment of Acute and Chronic Heart Failure; (2) Provision of written informed consent. Exclusion criteria were as follows: (1) Patients who visited the emergency department multiple times and had already been enrolled previously; (2) Refusal to provide informed consent; (3) After heart transplantation or left ventricular assist device implantation.

Patients were categorized at initial emergency evaluation by experienced cardiologists and emergency physicians into four clinical phenotypes (ADHF, APO, IRVF, and CS) according to standardized criteria based on ESC guidelines. ADHF was characterized by subacute onset with predominant signs of systemic congestion. APO was defined by abrupt onset of severe dyspnea and pulmonary edema. IRVF was defined by right-sided heart failure with preserved left ventricular systolic function [(Left ventricular ejection fraction LVEF) > 50%] and echocardiographic evidence of right ventricular dysfunction [e.g., tricuspid annular plane systolic excursion (TAPSE) < 15 mm], typically in the presence of underlying conditions such as pulmonary hypertension or pulmonary embolism. CS was defined as a state of systemic hypoperfusion due to severe reduction in cardiac output, supported by clinical features such as hypotension, oliguria, and elevated lactate levels.

The study protocol was approved by the Ethics Committee of Fuwai Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (approval No. 2024–2394) and was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki.

Data collection

Data collected included demographic characteristics, clinical presentation, medical history, laboratory and imaging findings, and treatments administered in the emergency department.

Baseline data included age, sex, visit and discharge dates, vital signs, and body measurements. Clinical presentation comprised symptoms of heart failure (e.g., dyspnea, orthopnea) and signs of congestion (e.g., jugular venous distention, pulmonary rales, peripheral edema, and New York Heart Association class).

Medical history included major cardiovascular comorbidities (e.g., coronary artery disease, cardiomyopathy, valvular disease, arrhythmias) and relevant non-cardiovascular conditions (e.g., pulmonary disease, renal disease, and metabolic disorders), as well as smoking and alcohol use.

Laboratory and imaging assessments included routine blood tests, cardiac biomarkers, arterial blood gas, chest X-ray, electrocardiography, and echocardiography at initial ED presentation.

Emergency treatments included pharmacological therapies (e.g., diuretics, vasodilators, inotropes, vasopressors, and guideline-directed medications) and device-based interventions [e.g., Intra-aortic balloon pump (IABP), Extra-corporeal membrane oxygenation (ECMO), Continuous renal replacement therapy (CRRT), temporary pacing, and mechanical ventilation]. Emergency treatments were defined as therapies initiated at or shortly after presentation to the ED, reflecting early management decisions, and did not include treatments administered during subsequent hospitalization.

Follow-up and study endpoints

A total of 2,960 consecutive patients with AHF were enrolled between June 2024 and February 2025 and included in the baseline analysis. This is an on-going real-world registry. The present analysis was based on a predefined cutoff date (December 2025) and represents an interim analysis of an ongoing registry. A total of 1,827 patients completed follow-up, among whom 1,246 patients had complete follow-up information and were included in the survival analyses. Time zero was defined as the time of presentation to the ED. Follow-up was conducted by trained research staff through structured telephone interviews. The study endpoints were 1-year all-cause mortality and cardiovascular mortality. All-cause mortality was defined as death from any cause. Cardiovascular mortality included death due to sudden cardiac death, heart failure, myocardial infarction, cerebrovascular events, malignant arrhythmias, and death from unknown causes or without an obvious non-cardiac cause.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation (SD) for normally distributed data or as median (interquartile range) for non-normally distributed data, while categorical variables were presented as counts (percentages). For group comparisons, analysis of variance (ANOVA) was used for normally distributed continuous variables, and the Kruskal–Wallis test was applied for non-normally distributed variables. Categorical variables were compared using the chi-square test or Fisher’s exact test as appropriate. Survival analyses for all-cause mortality and cardiovascular mortality were performed using the Kaplan–Meier method, and differences between groups were assessed using the log-rank test. Univariable and multivariable COX models were conducted to analysis the association between different phenotypes and prognosis. Underlying cardiac etiologies were collected but were not included in the multivariable models to maintain model simplicity and avoid overfitting, while focusing on acute clinical parameters at ED presentation. A two-sided P value < 0.05 was considered statistically significant.

Results

Patient enrollment

The patient enrollment process was shown in Fig. 1. A total of 3,038 patients presenting to the ED with AHF were initially screened. After excluding 54 patients with duplicate visits, 9 patients with a history of heart transplantation, and 15 patients with left ventricular assist device implantation at presentation, 2,960 patients were included in the baseline analysis. Based on a predefined follow-up cutoff date (December 2025), 1,827 patients completed follow-up. Among them, 1,246 patients had complete follow-up information (including both survival time and outcome status) and were therefore included in the survival analysis.

Fig. 1.

Study flowchart of patient enrollment and follow-up

Demographic characteristics

Among the 2,960 patients, 65.9% were male, and the median age was 65 years. There were 1,716 cases of ADHF (58.0%), 918 cases of APO (31.0%), 131 cases of IRVF (4.4%), and 195 cases of CS (6.6%). Significant differences in baseline demographic characteristics and vital signs were observed among the four phenotypes (P < 0.05, Table 1). Compared with the ADHF group, the IRVF group had a higher proportion of female patients, the APO group had an older median age, while the CS group presented with lower blood pressure and higher heart rate (Fig. 2A - C).

Table 1.

Baseline characteristics across different phenotypes of AHF patients

Types	Total	ADHF	APO	IRVF	CS	P
	N = 2960	N = 1716	N = 918	N = 131	N = 195
Male sex (%)	1950 (65.9)	1188 (69.2)	572 (62.3)	64 (48.9)	126 (64.6)	< 0.001
Age (years)	65.00 [54.00, 74.00]	64.00 [54.00, 73.00]	67.00 [55.00, 76.00]	63.00 [51.00, 75.00]	66.00 [51.50, 74.00]	0.004
Systolic blood pressure (mmHg)	126.00 [107.00, 150.00]	127.00 [109.00, 151.00]	130.00 [110.00, 152.00]	120.00 [104.00, 139.50]	101.00 [80.50, 119.00]	< 0.001
Heart rate (beats/min)	93.00 [76.00, 115.00]	90.00 [73.00, 112.00]	97.00 [82.00, 116.00]	95.00 [81.00, 116.00]	102.00 [80.00, 120.00]	< 0.001
Oxygen saturation (%)	95.00 [92.00, 97.00]	96.00 [94.00, 97.00]	93.00 [88.00, 96.00]	94.00 [91.00, 96.00]	93.00 [85.00, 97.00]	< 0.001
Coronary artery disease (%)	1636 (55.3)	965 (56.2)	496 (54.0)	35 (26.7)	140 (71.8)	<0.001
Valvular heart disease (%)	483 (16.3)	294 (17.1)	166 (18.1)	10 (7.6)	13 (6.7)	<0.001
Cardiomyopathy (%)	516 (17.4)	337 (19.6)	141 (15.4)	21 (16.0)	17 (8.7)	<0.001
Myocarditis (%)	49 (1.7)	29 (1.7)	6 (0.7)	3 (2.3)	11 (5.6)	<0.001
Congenital heart disease (%)	141 (4.8)	58 (3.4)	60 (6.5)	12 (9.2)	11 (5.6)	<0.001
Hypertension (%)	1620 (54.7)	966 (56.3)	518 (56.4)	43 (32.8)	93 (47.7)	<0.001
Hyperlipidemia (%)	1937 (65.4)	1152 (67.1)	597 (65.0)	50 (38.2)	138 (70.8)	<0.001
Diabetes mellitus (%)	1294 (43.7)	745 (43.4)	435 (47.4)	33 (25.2)	81 (41.5)	<0.001
Atrial fibrillation (%)	957 (32.3)	554 (32.3)	319 (34.7)	43 (32.8)	41 (21.0)	0.003
White blood cells (×10⁹/L)	7.78 [6.00, 10.00]	7.48 [5.94, 9.33]	7.92 [6.07, 10.10]	7.32 [5.11, 9.84]	11.80 [8.85, 17.10]	< 0.001
Hemoglobin (g/L)	133.00 [113.00, 150.00]	135.00 [116.00, 151.00]	131.00 [111.00, 147.75]	134.00 [116.00, 153.50]	127.00 [110.00, 144.00]	< 0.001
Creatinine (µmol/L)	85.83 [66.85, 123.32]	84.79 [66.35, 121.74]	88.21 [68.26, 130.72]	81.47 [64.75, 120.51]	85.67 [71.00, 118.67]	0.071
NT-pro BNP (pg/mL)	3495.55 [1403.42, 7664.23]	2627.00 [1031.85, 6122.50]	4969.80 [2387.00, 9531.65]	2703.45 [1422.25, 6630.02]	6520.30 [1972.55, 16911.00]	< 0.001
Lactate (mmol/L)	1.63 [1.25, 2.28]	1.55 [1.20, 2.09]	1.70 [1.26, 2.30]	1.50 [1.20, 2.21]	3.59 [2.06, 7.74]	< 0.001
Troponin I (ng/mL)	0.02 [0.01, 0.09]	0.02 [0.01, 0.10]	0.02 [0.01, 0.09]	0.01 [0.00, 0.04]	0.02 [0.01, 0.11]	0.037
Left ventricular ejection fraction (%)	42.00 [31.00, 56.00]	42.00 [30.00, 55.00]	40.00 [30.00, 56.00]	55.00 [40.00, 60.00]	40.00 [30.00, 56.00]	< 0.001
Left atrial diameter (mm)	44.00 [39.00, 50.00]	44.00 [39.00, 49.00]	44.00 [39.00, 50.00]	44.00 [37.75, 49.25]	44.00 [39.25, 50.00]	0.364
Left ventricular end-diastolic diameter (mm)	55.00 [48.00, 63.00]	55.00 [48.00, 63.00]	56.00 [48.00, 63.00]	49.50 [40.75, 57.00]	56.00 [48.25, 63.00]	< 0.001
Right ventricular diameter (mm)	25.00 [23.00, 28.00]	25.00 [23.00, 27.00]	25.00 [23.00, 28.00]	28.00 [24.00, 36.00]	25.00 [23.00, 29.00]	< 0.001
Estimated pulmonary artery systolic pressure*(mmHg)	54.00 [46.00, 64.00]	54.00 [46.00, 62.00]	54.00 [46.00, 62.00]	69.00 [55.00, 88.00]	56.50 [46.00, 66.50]	< 0.001

* Estimated pulmonary artery systolic pressure (PASP) was estimated using the formula: PASP = 4 × (tricuspid regurgitation velocity [TRV])² + right atrial pressure (RAP)

Fig. 2.

Violin plots of key clinical parameters across different clinical phenotypes

Etiology and comorbidities

Regarding the etiology of heart failure, the most common cause was coronary artery disease (55.3%), followed by cardiomyopathy (17.4%) and valvular heart disease (16.3%). Significant differences in etiology were observed among the four phenotypes (Table 1). Compared with other groups, the CS group had a higher proportion of coronary artery disease and myocarditis, the APO group had a higher proportion of valvular disease, the ADHF group had a higher proportion of cardiomyopathy, and the IRVF group showed a higher proportion of congenital heart disease (P < 0.05). With respect to comorbidities, 54.7% of patients had hypertension, 65.4% had hyperlipidemia, 43.7% had diabetes mellitus, 32.3% had atrial fibrillation. The APO group showed higher proportions of hypertension, diabetes, and atrial fibrillation, whereas the CS group had a higher prevalence of hyperlipidemia (P < 0.05).

Laboratory findings

The main laboratory parameters of the study population are summarized in Table 1. Among hematologic indices, the CS group showed higher white blood cell and lower red hemoglobin levels (P < 0.05). As for biomarkers reflecting cardiac function and tissue perfusion, N-terminal pro-brain natriuretic peptide (NT-pro BNP) and lactate levels were significantly higher in CS patients (P < 0.05), whereas troponin-I did not differ significantly from those in the ADHF group (Fig. 2D - F).

Echocardiographic findings

As shown in Table 1, there were no significant differences among the four clinical phenotypes in left atrial diameter. Compared with other groups, the IRVF group demonstrated a smaller left ventricular end-diastolic diameter, higher LVEF, and larger right ventricular diameter (P < 0.05) (Fig. 2G - I).

Therapeutic interventions

The emergency treatment strategies administered to the study population are summarized in Table 2. Among medications for rapid symptom relief, 72.5% of patients received diuretics, 28.7% received vasodilators, 18.4% received inotropes, and 3.2% received vasoconstrictors. Across the four phenotypes, the APO group had a higher proportion of diuretic and vasodilator use, whereas the CS group had a higher proportion of inotrope and vasoconstrictor use (P < 0.05).

Table 2.

Treatments across different phenotypes of AHF patients in the EDs

Types	Total	ADHF	APO	IRVF	CS	P
	N = 2960	N = 1716	N = 918	N = 131	N = 195
Diuretics (%)	2147 (72.5)	1115 (65.0)	805 (87.7)	95 (72.5)	132 (67.7)	< 0.001
Vasoconstrictors (%)	95 (3.2)	11 (0.6)	18 (2.0)	1 (0.8)	65 (33.3)	< 0.001
Inotropes (%)	545 (18.4)	202 (11.8)	178 (19.4)	33 (25.2)	132 (67.7)	< 0.001
Vasodilators (%)	849 (28.7)	471 (27.4)	329 (35.8)	15 (11.5)	34 (17.4)	< 0.001
RAAS inhibitors (%)	1016 (34.3)	618 (36.0)	353 (38.5)	28 (21.4)	17 (8.7)	< 0.001
β-blockers (%)	509 (17.2)	311 (18.1)	162 (17.6)	17 (13.0)	19 (9.7)	0.015
Mineralocorticoid receptor antagonists (%)	1382 (46.7)	765 (44.6)	493 (53.7)	66 (50.4)	58 (29.7)	< 0.001
SGLT2 inhibitors (%)	1045 (35.3)	626 (36.5)	375 (40.8)	23 (17.6)	21 (10.8)	< 0.001
Vericiguat (%)	204 (6.9)	125 (7.3)	66 (7.2)	6 (4.6)	7 (3.6)	0.176
Digoxin (%)	481 (16.2)	228 (13.3)	200 (21.8)	32 (24.4)	21 (10.8)	< 0.001
Intra-aortic balloon pump (%)	121 (4.1)	42 (2.4)	28 (3.1)	3 (2.3)	48 (24.6)	< 0.001
Extracorporeal membrane oxygenation (%)	11 (0.4)	3 (0.2)	2 (0.2)	0 (0.0)	6 (3.1)	< 0.001
Non-invasive ventilation (%)	382 (12.9)	199 (11.6)	113 (12.3)	20 (15.3)	50 (25.6)	< 0.001
Invasive mechanical ventilation (%)	52 (1.8)	7 (0.4)	9 (1.0)	0 (0.0)	36 (18.5)	< 0.001
Continuous renal replacement therapy (%)	24 (0.8)	10 (0.6)	3 (0.3)	0 (0.0)	11 (5.6)	< 0.001
Temporary pacemaker (%)	15 (0.5)	10 (0.6)	3 (0.3)	0 (0.0)	2 (1.0)	0.453

Regarding guideline-directed medical therapy (GDMT), 34.3% of patients received RAAS inhibitors (RAASi), 17.2% received β-blockers, 46.7% received mineralocorticoid receptor antagonists (MRA), and 35.3% received sodium–glucose cotransporter 2 inhibitors (SGLT2i). The APO group had higher rates of RAASi, MRA, and SGLT2i use than other phenotypes. In addition, 6.9% of patients were treated with vericiguat, and 16.2% received digoxin, with a significantly higher use of digoxin observed in the IRVF group (P < 0.05).

For device-based interventions, In terms of mechanical circulatory and respiratory support, patients with the CS phenotype had significantly higher utilization rates of IABP (24.6%), ECMO (3.1%), noninvasive ventilation (25.6%), invasive mechanical ventilation (18.5%), and CRRT (5.6%) compared with the other phenotypes (all P < 0.05). The use of temporary pacing did not differ significantly among the four phenotypes.

Prognostic differences

A total of 1,246 patients completed one-year follow-up. During follow-up, 205 deaths occurred, corresponding to an overall one-year all-cause mortality rate of 16.5%. The mortality rates differed significantly among the four phenotypes (P < 0.001), with rates of 10.3% in the ADHF group, 11.0% in the APO group, 24.5% in the IRVF group, and 58.7% in the CS group. Most deaths in the CS phenotype occurred within the first 60 days of follow-up (Fig. 3A).For cardiovascular mortality, 1,228 patients were included in the analysis after excluding 18 patients with missing cardiovascular death data. A total of 180 cardiovascular deaths occurred (14.7%). The corresponding cardiovascular mortality rates were 7.9% in ADHF, 9.6% in APO, 21.6% in IRVF, and 58.7% in CS, with significant differences among the four phenotypes (P < 0.001) (Fig. 3B).

Fig. 3.

Kaplan–Meier survival curves for 1-year (A) all-cause mortality and (B) cardiovascular mortality across different clinical phenotypes

In the univariable Cox regression model (crude model), using ADHF as the reference group, both IRVF and CS phenotypes were significantly associated with increased risks of all-cause and cardiovascular mortality. The IRVF phenotype was associated with higher risks of all-cause mortality (HR 2.49, 95% CI 1.38–4.48, P = 0.003) and cardiovascular mortality (HR 2.79, 95% CI 1.46–5.32, P = 0.002). The CS phenotype showed a markedly increased risk of all-cause mortality (HR 10.76, 95% CI 7.83–14.78, P < 0.001) and cardiovascular mortality (HR 13.65, 95% CI 9.68–19.25, P < 0.001). No significant associations were observed between the APO phenotype and either mortality outcome.

In multivariable Cox regression Model 1, adjusting for five variables (age, systolic blood pressure, oxygen saturation, NT-proBNP, and lactate), both IRVF and CS phenotypes remained significantly associated with increased mortality risks. The IRVF phenotype was associated with higher risks of all-cause mortality (HR 2.02, 95% CI 1.09–3.77, P = 0.027) and cardiovascular mortality (HR 2.38, 95% CI 1.20–4.73, P = 0.014). Similarly, the CS phenotype remained strongly associated with all-cause mortality (HR 4.79, 95% CI 3.14–7.29, P < 0.001) and cardiovascular mortality (HR 6.24, 95% CI 3.98–9.81, P < 0.001).

In multivariable Cox regression Model 2, after further adjustment for treatment variables including diuretics, vasoconstrictors, inotropes, and vasodilators, the associations between IRVF and CS phenotypes and mortality remained statistically significant. The IRVF phenotype was independently associated with increased risks of all-cause mortality (HR 1.92, 95% CI 1.02–3.63, P = 0.044) and cardiovascular mortality (HR 2.22, 95% CI 1.10–4.48, P = 0.027). The CS phenotype was also independently associated with higher risks of all-cause mortality (HR 2.22, 95% CI 1.34–3.67, P = 0.002) and cardiovascular mortality (HR 2.71, 95% CI 1.57–4.66, P < 0.001) (Table 3).

Table 3.

Multivariable cox models for associations of clinical phenotypes and endpoints

		All-cause mortality			Cardiovascular mortality
		HR	95%CI	P	HR	95%CI	P
Crude model	ADHF
	APO	1.09	0.73–1.61	0.679	1.23	0.80–1.90	0.351
	IRVF	2.49	1.38–4.48	0.003	2.79	1.46–5.32	0.002
	CS	10.76	7.83–14.78	< 0.001	13.65	9.68–19.25	< 0.001
Model 1	ADHF
	APO	0.87	0.57–1.33	0.521	1.02	0.64–1.62	0.942
	IRVF	2.02	1.09–3.77	0.027	2.38	1.20–4.73	0.014
	CS	4.79	3.14–7.29	< 0.001	6.24	3.98–9.81	< 0.001
Model 2	ADHF
	APO	0.73	0.47–1.13	0.154	0.83	0.51–1.34	0.449
	IRVF	1.92	1.02–3.63	0.044	2.22	1.10–4.48	0.027
	CS	2.22	1.34–3.67	0.002	2.71	1.57–4.66	< 0.001

Model 1: adjusted for age, systolic blood pressure, oxygen saturation, lactate, and NT-pro BNP

Model 2: additionally adjusted diuretics, vasopressors, inotropes and vasodilator

Discussion

This study, based on a single-center real-world cohort of consecutively enrolled patients with AHF presenting to the ED, systematically compared baseline characteristics and treatment patterns across different clinical phenotypes (ADHF, APO, IRVF, and CS), and evaluated their association with 1-year mortality outcomes. Significant differences were observed among phenotypes in demographic characteristics, vital signs, laboratory findings, and echocardiographic parameters. Among these groups, ADHF exhibited the lowest one-year mortality, whereas CS had the highest one-year all-cause mortality. After adjustment for age, systolic blood pressure, oxygen saturation, NT-pro BNP, and lactate, both IRVF and CS remained significantly associated with adverse outcomes. Further adjustment for emergency treatment variables did not materially change these associations, suggesting that clinical phenotyping provides meaningful prognostic stratification in patients with AHF in the ED.

In the present study, ADHF accounted for the largest proportion of cases, with a higher prevalence of cardiomyopathy compared with other phenotypes. Pathophysiologically, ADHF is characterized by a gradual decline in cardiac output leading to progressive elevation of venous pressure and systemic fluid retention, consistent with the chronic and insidious nature of cardiomyopathy. The SMILE-HF prospective registry from Korea reported that tachycardia-induced and idiopathic dilated cardiomyopathy together accounted for more than 20% of hospitalized AHF patients, underscoring cardiomyopathy as a major etiology of AHF in Asian populations [7]. Similarly, Kajimoto et al., based on the Japanese ATTEND registry of 4,842 ADHF patients, found that dilated cardiomyopathy was one of the leading causes of ADHF. Interestingly, its long-term composite event risk was lower than that of valvular or ischemic heart failure, suggesting that although cardiomyopathy-related ADHF often presents with advanced hemodynamic deterioration at admission, its long-term prognosis may not be the poorest, reflecting etiology-specific trajectories in disease progression and outcome [8]. From a therapeutic perspective, management of ADHF primarily focuses on decongestion with diuretics and optimization of guideline-directed medical therapy, tailored to the underlying etiology.

APO patients in this study were older and presented with faster heart rates. Hemodynamically, APO is characterized by an abrupt rise in pulmonary capillary wedge pressure, rapid increase in afterload, and swift redistribution of intravascular fluid to the lungs, which triggers sympathetic activation. Consequently, diuretics and vasodilators were more frequently used in this group. The 2025 Heart Failure Association scientific statement of ESC also emphasized that vasodilators, through combined venous and arterial dilation, can effectively reduce left ventricular preload and afterload, making them particularly suitable for the APO phenotype, which typically presents with acute hypertension and rapidly developing pulmonary congestion [9]. However, recent RCTs, including RELAX-AHF-2 and TRUE-AHF, failed to demonstrate long-term survival benefits of diuretics and vasodilators, suggesting that these agents primarily provide short-term hemodynamic relief rather than improving long-term outcomes [10, 11]. In addition, APO patients in this study had a higher prevalence of hypertension and diabetes, indicating a heavier cardiometabolic comorbidity burden. They also exhibited a greater incidence of atrial fibrillation and anemia, both of which are recognized precipitating factors for acute heart failure exacerbation. These findings highlight the complex comorbidity profile and clinical heterogeneity of emergency AHF patients, which may complicate acute management and therapeutic decision-making. Previous studies have underscored that multidisciplinary and integrated management approaches play a crucial role in improving outcomes among patients with acute cardiovascular emergencies [12, 13].

Compared with patients with left heart failure, studies focusing on acute IRVF in the emergency setting are relatively limited. In our cohort, a higher proportion of IRVF patients had congenital heart disease, and this subgroup exhibited a worse long-term prognosis. Recent evidence suggests that right ventricular (RV) dysfunction, rather than impaired left ventricular systolic function, is a key determinant of prognosis in acute heart failure. In a prospective study enrolling 418 patients with predominantly ischemic heart disease, Berrill et al. demonstrated that indices of RV systolic function — including TAPSE, RV ejection fraction, fractional area change, and global longitudinal strain — were all significantly associated with two-year mortality, whereas parameters of left ventricular function such as LVEF were not prognostic [14]. Similarly, Nicholas et al. found in a prospective study of patients presenting to two large U.S. ED that early RV dysfunction detected during the initial evaluation was a powerful predictor of death and rehospitalization, outperforming LVEF in prognostic value [15]. These findings collectively suggest that the poor outcomes observed in IRVF patients may primarily result from persistent RV dysfunction, elevated pulmonary vascular resistance, and impaired right-sided circulatory adaptation. Such mechanisms are often underrecognized during acute decompensation, highlighting the need for greater clinical attention and targeted management for this high-risk phenotype. In addition, increased pulmonary vascular resistance may lead to elevated right ventricular afterload, further compromising hemodynamic stability. Systemic venous congestion associated with right heart failure may contribute to multiorgan dysfunction, particularly renal and hepatic impairment. Moreover, in the emergency setting, IRVF may be underrecognized due to its heterogeneous presentation, potentially resulting in delayed diagnosis and management. From a management perspective, treatment of IRVF should focus on careful volume optimization, reduction of right ventricular afterload, and targeted therapy for underlying causes such as pulmonary hypertension or pulmonary embolism.

CS represented the phenotype with the highest mortality in our cohort and is widely recognized as the most severe stage of AHF. Patients with CS typically present with profound hypotension and tachycardia, most commonly resulting from acute-onset conditions such as acute coronary syndrome or myocarditis. The predominant pathophysiological feature is a marked reduction in cardiac output accompanied by systemic hypoperfusion and multiorgan dysfunction. Compared with other subtypes, the management of CS primarily focuses on restoring effective circulatory perfusion, mainly through the use of inotropic and vasopressor agents. However, in the present study, the proportion of CS patients receiving mechanical circulatory support (MCS) was relatively low. In contrast, data from international cohorts such as the CZECH-SHOCK registry reported an overall MCS utilization rate (including IABP, Impella, and ECMO) of approximately 28%, with a 30-day mortality rate of around 40% [16]. An analysis of the nationwide inpatient database in Japan indicated that among CS patients requiring mechanical circulatory support, IABP alone remained the most frequently used modality, accounting for approximately 65%–81% of cases. The proportion of patients treated with ECMO alone has increased in recent years to about 30%. Although in-hospital mortality in the ECMO subgroup remains high, a downward trend has been observed, decreasing from roughly 74% to 64% [17]. It is noteworthy that the present study was conducted in the EDs setting, which differs substantially from previous studies that primarily included hospitalized or intensive care unit patients. In future clinical practice, greater emphasis should be placed on timely assessment of MCS indications and initiation timing in CS patients presenting to the ED. From a research perspective, there is also an urgent need for more evidence derived from emergency-based clinical scenarios to optimize early intervention strategies and improve patient outcomes. These findings underscore the importance of early recognition and timely initiation of vasoactive agents and mechanical circulatory support in patients with CS.

This study had several limitations. First, due to the real-world nature of the emergency department setting, invasive hemodynamic data were not routinely available, and patients’ circulatory status was primarily assessed through integrated clinical judgment rather than precise quantitative measurements. For instance, right heart function was mainly evaluated by echocardiography, while right heart catheterization was not routinely performed. Although this reflects routine clinical practice in the emergency setting, future studies incorporating invasive hemodynamic data, particularly in intensive care unit populations using pulmonary artery catheters and related monitoring techniques, may provide additional insights. Second, the present analysis focused primarily on mortality outcomes. Ongoing follow-up will include rehospitalization for heart failure, enabling more comprehensive post hoc analyses of composite endpoints. Information on prior chronic heart failure history was not consistently available, particularly in the retrospective phase of the study, and was therefore not included in the analysis. Additionally, the single-center nature of this study introduces potential bias and limits generalizability. As this is an ongoing registry, future analyses with extended follow-up will further validate and update these findings.

These ED–based findings complement inpatient registries and suggest that phenotype-guided early decisions may matter most within the first weeks after disposition. Despite these limitations, this study provides real-world evidence that clinical phenotyping at the emergency stage offers direct guidance for risk stratification and therapeutic decision-making in AHF, aligning with the ESC-recommended AHF management pathway. However, for long-term prognostication, reliance on phenotypic classification alone appears insufficient. Future risk assessment should integrate multimodal information, combining structural and imaging markers (e.g., late gadolinium enhancement on cardiac MRI) with key risk indicators such as blood pressure, renal function, lactate, right ventricular performance, and NT-pro BNP, to establish a comprehensive risk prediction model.

Conclusion

Significant differences in baseline characteristics and 1-year mortality were observed among the different clinical phenotypes of AHF. In our cohort, the CS phenotype exhibited the highest 1-year all-cause and cardiovascular mortality. Compared with other phenotypes, both IRVF and CS were significantly associated with adverse outcomes.

These findings highlight the clinical relevance of phenotype-based risk stratification at the emergency stage and its association with 1-year mortality in patients with AHF. Further studies are warranted to validate these findings in larger, multicenter cohorts and to explore phenotype-guided management strategies in the emergency setting.

Author contributions

LHY: analysis and interpretation of the data, drafting of the manuscript, statistical expertise, WYM, XYJ, LYL, ZCJ: study concept and design, acquisition of the data, YYM: critical revision of the manuscript for important intellectual content, and acquisition of funding.

Funding

This study is funded by High-Level Hospital Clinical Research Funding (ID: 2022‐GSP‐GG‐26) and Beijing Key Clinical Specialty Construction Project.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of Fuwai Hospital, Chinese Academy of Medical Sciences (approval No. 2024–2394) and was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Informed consent was obtained from all individual participants included in the study.

Consent for publication

Written informed consent for publication was obtained from all participants.

Competing interests

The authors declare no competing interests.

Running Title

Acute heart failure phenotypes and prognosis.