Temporal trends in incidence, clinical characteristics and outcomes among young adults with heart failure: a territory-wide study from 2014 to 2023 on 19,537 patients

Summary

Background

Heart failure (HF), traditionally considered a disease of the elderly, is increasingly common in younger people, but temporal data remain scarce. This cohort study aimed to examine temporal trends in incidence, comorbidities, risk factor profiles, and clinical outcomes among young HF patients.

Methods

Using a territory-wide administrative database in Hong Kong, we identified 19,537 young adults aged <65 years with incident HF between 2014 and 2023. Data on baseline characteristics, echocardiographic parameters, comorbidities and prescribed medications were retrieved. Annual standardised incidence rates (IRs) of HF were calculated by direct age- and sex-standardisation. Comparisons were made between two 5-year periods: 2014–2018 and 2019–2023. Multivariable regression models were applied to assess temporal shifts in risk factor profiles. The primary outcome was one-year all-cause mortality, with incidence rates reported per 100 person-years. Kaplan–Meier survival curves were plotted to illustrate survival trends.

Findings

Among the cohort (median age 57.1 years, 69% men), IRs of young HF increased by 20% (IRR 1.20, 95% CI 1.13–1.27) from 2014 to 2023. Concurrent with fewer comorbidities, young HF patients in 2019–2023 were more likely to present with obesity, cardiomyopathy, lower socioeconomic status, and be aged 45–65 years, contrasting with the conventional risk factors (including history of sudden cardiac arrest) predominant in the 2014–2018 cohort (all p < 0.001). Additionally, the 2019–2023 cohort demonstrated elevated subtype of HF with reduced ejection fraction (HFrEF) and increased use of guideline-directed medical therapy (GDMT). A modest reduction in one-year mortality was observed between the two periods (15.6 [14.8, 16.5] vs. 14.6 [13.8, 15.5] events per 100 person-years).

Interpretation

The incidence of HF among young adults increased substantially between 2014 and 2023. During this period, the risk factor profile shifted considerably, with a pronounced rise in HFrEF subtype. Despite improved therapeutic management and better use of GDMT, reductions in one-year mortality were modest. Proactive public health strategies are urgently needed to address these emerging challenges in this population.

Funding

This work was funded by grants from the National Natural Science Foundation of China (No. 82270400) and the Natural Science Foundation of Guangdong Province (No. 2023A1515010731).

Research in context.

Evidence before this study

We search PubMed for studies in English between January 2000 and October 2024 using search terms: “temporal trends”, “heart failure”, “incidence” and “the young” in the title/abstract initially. This search yielded only three articles. The search was expanded to include “risk factors” and “mortality” in all fields. We also consulted with experts on relevant studies.

Data on incident heart failure among young adults (<65 years) are scarce, owing to low number of individuals reported, especially among those aged <45 years. In contrast to older individuals, emerging evidence from the Western countries has reported an increasing trend in the incidence of HF in younger adults in recent years. So far, there is no contemporary population-based study from Asia, which comprehensively examines trends in incidence, changing risk factor profile, medication usages and outcomes of heart failure in young individuals.

Added value of this study

This study (N = 19,537 patients with young heart failure [aged 18–65 years]) provides important new information of the increasing incidence of heart failure in young individuals, despite lower comorbidity burden, with more pronounced increase in men than women. The results also highlight the changing ‘landscape’ over the period (2014–2023), with declining prevalence of traditional risk factors of heart failure, e.g. atrial fibrillation, type II diabetes mellitus, valvular heart disease, chronic kidney disease but increase in obesity, dilated cardiomyopathy and individuals with lower socioeconomic status. This study is further strengthened by examining the shifts in HF subtypes (HFpEF vs. HFrEF) and the use of medications (including guideline-directed medical therapy [GDMT] for heart failure). Reductions in one-year mortality are marginal despite an increase in usage of GDMT, which highlights the challenges in managing heart failure.

Implications of all the available evidence

These findings call for targeted preventive strategies of heart failure, early screening, and detection to reduce risk factors and address disparities in young populations.

Introduction

Heart failure (HF), a major public health issue intricately associated with age-related risks, is predominantly reported in the older population, with most patients aged >65 years and some even surpassing 70 years.^1,2 Yet, emerging evidence has indicated an escalating burden of HF among the younger population since 2014.³ Studies from the UK, France, Sweden, Denmark, New Zealand, Western Australia, and the US suggest a rising trend in HF occurrence at a young age. Depending on the definitions of ‘young’ used, varying proportions of young cases are observed, ranging from 4.0% to 14.9%.4, 5, 6, 7, 8, 9, 10 Reasons for this trend of earlier diagnosis of HF can be partially attributed to the growing burden of cardiometabolic risk factors and unhealthy lifestyles that begin in early adulthood.^1,11,12 Given the evolving epidemiology of contributing comorbidities and social determinants observed in recent decades,¹³ understanding the potential implications of these risk factors in the younger population is urgently needed.

Despite the advancements in HF management over the past two decades, young patients with HF have shown significantly worse survival outcomes compared with matched controls. This mainly affects the youngest individuals, who had a mortality rate seven times higher than the overall cohort under 55 years.¹⁴ Notably, a temporal trend analysis in Denmark from 1995 to 2012 indicated a significant decrease in one-year mortality rates for patients aged 45–65 years, although no corresponding improvement was observed for those under 45.⁷ In Sweden, a persistent decline in one-year mortality rates across different young age groups were observed between 1987 and 2006.⁶ However, these studies either focused on the prevalence of comorbidities without addressing the shift patterns or were limited to earlier cohorts predating 2015, thereby lacking the incorporation of contemporary guideline-directed medical therapy (GDMT), such as angiotensin receptor-neprilysin inhibitor (ARNI) and sodium-glucose cotransporter 2 (SGLT2) inhibitor.

We leveraged the Clinical Data Analysis Reporting System (CDARS), a comprehensive administrative electronic medical records database in Hong Kong, to bridge these knowledge gaps. Our study aimed to examine contemporary trends in standardised incidence of HF among young individuals under 65 years, along with the shifts in risk factor patterns, echocardiographic characteristics, medications and adverse outcomes over the past decade.

Methods

Data source

In this territory-wide, population-based cohort study, we utilised data from the CDARS, managed by the Hong Kong Hospital Authority since 1993.¹⁵ As the sole statutory provider of public healthcare in Hong Kong, the Hospital Authority provides >80% of inpatient and >30% of outpatient care to the territory's 7.5 million residents.¹⁶ CDARS comprises comprehensive patient information, including sociodemographic data, comorbidities, medications, procedures, laboratory results, and hospital visit episodes, linked to 43 public hospitals, 49 specialist outpatient clinics, and 74 general outpatient clinics. The database uses anonymised patient identifiers to preserve confidentiality and has been widely employed in rigorous, large-scale cohort studies of cardiovascular disease.17, 18, 19 The reliability of the database has been elaborated in several earlier epidemiological investigations.^20,21 Nevertheless, we conducted further validation on a subset of our cohort to ensure the validity. This study received approval from the University of Hong Kong Ethics Committee and the West Cluster of Hong Kong Hospital Authority (UW 20–817), with informed consent waived due to the anonymised nature of the data.

Study population

Patients aged 18–65 years with newly diagnosed HF, identified using International Classification of Diseases, ninth and tenth revisions (ICD-9 and ICD-10 codes, Table S1), were included from January 1, 2014 to December 31, 2023. The index date was determined as the date of the first HF diagnosis. Patients were categorised by age, <45 years and 45–65 years, and were further divided into two 5-year calendar periods, 2014–2018 and 2019–2023. To calculate the incidence rates (IRs), we excluded individuals with a prior HF diagnosis before 2014 (n = 31,620) (Fig. S1). The study complied with the Declaration of Helsinki and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Patient characteristics

We collected patient characteristics including basic demographics (age, sex, smoking, alcohol, obesity, and educational attainment), socioeconomic status, comorbidities, medications, and laboratory measurements. Obesity was defined based on either ICD-coded diagnoses or a BMI ≥30 kg/m² documented within one year of HF diagnosis, where available. Low socioeconomic status was identified by receipt of the Comprehensive Social Security Assistance (CSSA) scheme, a means-tested welfare scheme in Hong Kong for residents unable to sustain themselves financially, for payment.²⁰ For CSSA beneficiaries, public healthcare expenses are waived.²² Additionally, nineteen comorbidities were ascertained within five years before HF, including hypertension, type II diabetes mellitus (T2DM), ischaemic heart disease (IHD), dyslipidemia, atrial fibrillation (AF), stroke, congenital heart disease, cardiomyopathy, rheumatic heart disease, infective endocarditis (IE), valvular heart disease, previous history of sudden cardiac arrest (SCA), peripheral vascular disease (PVD), liver cirrhosis, chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), rheumatism, depression, and cancer (Table S1). Medications, including angiotensin converting enzyme inhibitor (ACEI), angiotensin II receptor blocker (ARB), ARNI, beta blocker, calcium channel blocker (CCB), diuretics, mineralocorticoid receptor antagonist (MRA), digoxin, SGLT2 inhibitor, insulin, metformin, aspirin, other antiplatelets, and statins were identified as dispensed prescriptions with ≥90 consecutive days of use within the first year after HF. Blood glucose (fasting glucose, glycated hemoglobin A1c [HbA1c]), lipid profiles (triglyceride [TG], total cholesterol [TC], high-density lipoprotein [HDL], low-density lipoprotein [LDL]), and serum creatinine levels were also collected within one year before and 30 days after. The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.

Echocardiographic data were exclusively obtained from our center, Queen Mary Hospital, one of the Hong Kong's two university teaching hospitals, via the clinical management system (CMS) within one year before and 90 days after HF diagnosis. Left ventricular ejection fraction (LVEF) was estimated using the biplane Simpson's method. HF with reduced ejection fraction (HFrEF) was defined as HF with an LVEF <50% and HF with preserved ejection fraction (HFpEF) as HF with an LVEF ≥50%.²³ Relative wall thickness (RWT) was calculated as (2 × diastolic posterior wall thickness)/LV end-diastolic diameter. LV hypertrophy (LVH) was determined by values of LV mass indexed to body surface area >115 g/m² for men and >95 g/m² for women. LV geometry was classified according to guideline recommendations as follows: 1) normal: RWT ≤0.42 and no LVH, 2) concentric remodeling: RWT >0.42 and no LVH, 3) eccentric hypertrophy: RWT ≤0.42 and LVH, and 4) concentric hypertrophy: RWT >0.42 and LVH.²⁴

Study outcomes

The primary outcome was all-cause mortality within one year. Secondary outcomes included cause-specific mortality, HF hospitalisation, cardiovascular mortality or HF hospitalisation, major adverse cardiovascular events (MACE, defined as nonfatal myocardial, stroke, or cardiovascular mortality), and sudden death within the 1-year follow-up period.

Data validation

To evaluate the validity of recorded obesity diagnoses, we randomly selected 120 patients from our cohort and manually reviewed their original medical records from the computer-based CMS at Queen Mary Hospital, which offers highly specialised services and serves as a tertiary referral center for all complex cases in Hong Kong.²⁵ Obesity cases were identified by a comprehensive review of clinical anthropometric data, metabolic assessments, bariatric surgery records, and relevant diagnostic documentation. In this study, 3240 individuals were diagnosed with obesity, with an overall positive predictive value (PPV) of 88.3% (106 of 120) for confirmed cases. In another study cohort (QW Ren et al., 2021),¹⁷ the diagnosis of HF was validated by evaluating clinical notes and assessments that included symptoms (e.g., dyspnoea, fatigue), echocardiographic reports, and cardiac magnetic resonance findings. Among the 87,102 cases of HF included, 95 out of 100 cases (95%) were verified with a diagnosis of HF, demonstrating a high level of diagnostic accuracy.

Statistical analysis

Baseline patient characteristics were summarised as median (interquartile range [IQR]) or frequency (%), as appropriate. Missing data were reported as the proportion of records with incomplete entries, and analyses were conducted using available data, excluding missing observations. Between-group differences were assessed using Pearson χ² test or Student t-test, as appropriate. The annual IRs of young HF were calculated by direct age- and sex-standardisation based on Hong Kong mid-year census population estimates from 2014 to 2023,²⁶ with two age bands up to 65 years. Overall and category-specific IR ratios (IRRs) with 95% confidence intervals (CIs) were further estimated using Poisson regression, with standardised IRs as the dependent variable and calendar year as the independent variable, as previously described.¹³ Temporal shifts in risk factors were analysed using multivariable logistic regression. One-year mortality and secondary outcomes were evaluated by crude incidence rates (events per 100 person-years) and Kaplan–Meier survival curves.

Subgroup analyses were performed by estimating annual IRs, examining temporal trends in risk factors, and assessing all-cause mortality within period-specific cohorts stratified by age, sex, and prespecified patient characteristics. Unadjusted and adjusted Cox regression models were further performed, incorporating variables that indicated significant shifts over time, as sensitivity analysis. Patients with a history of SARS-CoV-2 infection or who acquired the virus during follow-up were excluded owing to its independent association with increased risks of incident HF and post-infection mortality.²⁷ SARS-CoV-2 infection was confirmed by a positive PCR or rapid antigen test, as verified by the Centre of Health Protection, with COVID-19 recorded as the primary diagnosis (ICD-9 code 519.88, 519.89; ICD-10 code U07.1). The representativeness of the sub-cohort with available echocardiographic data was also evaluated. Outcomes were compared between younger and older cohorts, including individuals aged 65–90 years, as emerging evidence suggested that younger individuals experienced escalating risks of adverse outcomes, whereas risks in older populations remain stable or decline.^6,9 Statistical significance was set at p < 0·05, and all statistical analyses were executed using R software (Version 4.2.2).

Role of the funding sources

The funders had no role in the study design, data collection, analysis, interpretation, manuscript drafting, or the decision to submit the manuscript for publication.

Results

Baseline characteristics

Between 2014 and 2023, 19,537 patients with new-onset HF were identified (median age 57.1 years; 85% aged 45–65 years; 69% men). Cohort characteristics by period were summarised in Table 1. Compared to young HF patients in 2014–2018, those in 2019–2023 demonstrated a slight increase in median age (56.9 vs. 57.2 years, p = 0.007), alongside declines in smoking and alcohol consumption, but rises in obesity prevalence, CSSA beneficiaries, and higher educational attainment. The later cohort showed a lower prevalence of hypertension, T2DM, IHD, AF, stroke, congenital heart disease, rheumatic heart disease, IE, valvular heart disease, previous history of SCA, PVD, liver cirrhosis, COPD, CKD, rheumatism, and depression. Conversely, cardiomyopathy, notably dilated cardiomyopathy (DCM) was more prevalent. The usages of ARNI, SGLT2 inhibitors, MRA, and statins increased over time, accompanied by improvements in lipid profiles. Among 1134 patients with echocardiographic data (Table 2), young HF patients in 2019–2023 showed stable trends in cardiac remodeling and LVEF (45% vs. 40%, p = 0.071) compared to those in 2014–2018. Notably, the prevalence of HFrEF subtype increased significantly over time (54% vs. 61%, p = 0.022). Although valvulopathies remained comparable, the proportion of advanced left atrial dilation declined during the study period.

Table 1.

Baseline characteristics of young individuals with heart failure, 2014 through 2023.

Characteristic	Overall (n = 19537)	Sex		Age		Time period
		Women (n = 6061)	Men (n = 13476)	<45 years (n = 2843)	45–65 years (n = 16694)	2014–2018 (n = 9427)	2019–2023 (n = 10110)	P-valuea
Demographics
Age, mean (SD), year	57.1 (50.1, 61.5)	57.2 (49.4, 61.8)	57.0 (50.4, 61.4)	39.1 (34.2, 42.5)	58.5 (53.7, 62.1)	56.9 (50.0, 61.4)	57.2 (50.2, 61.6)	0.007
45–65 years, n (%)	16694 (85.4)	5078 (83.8)	11616 (86.2)	–	–	8032 (85.2)	8662 (85.7)	0.357
Male, n (%)	13476 (69.0)	–	–	1860 (65.4)	11616 (69.6)	6514 (69.1)	6962 (68.9)	0.732
Smoking, n (%)	575 (2.9)	59 (1.0)	516 (3.8)	22 (0.8)	553 (3.3)	309 (3.3)	266 (2.6)	0.009
Alcohol, n (%)	281 (1.4)	30 (0.5)	251 (1.9)	43 (1.5)	238 (1.4)	204 (2.2)	77 (0.8)	<0.001
Obesity, n (%)	3240 (16.6)	1052 (17.4)	2188 (16.2)	546 (19.2)	2694 (16.1)	1254 (13.3)	1986 (19.6)	<0.001
CSSA allowance, n (%)	1620 (8.3)	514 (8.5)	1106 (8.2)	153 (5.4)	1467 (8.8)	673 (7.1)	947 (9.4)	<0.001
Education, n (%)a								<0.001
Less than Primary	649 (4.2)	371 (7.8)	278 (2.6)	33 (1.5)	616 (4.6)	368 (5.3)	281 (3.3)
Primary	4053 (26.2)	1462 (30.9)	2591 (24.1)	155 (7.1)	3898 (29.3)	2136 (30.6)	1917 (22.6)
Secondary	8985 (58.1)	2422 (51.2)	6563 (61.1)	1484 (68.1)	7501 (56.4)	3795 (54.3)	5190 (61.2)
Tertiary or above	1783 (11.5)	478 (10.1)	1305 (12.2)	507 (23.3)	1276 (9.6)	691 (9.9)	1092 (12.9)
Missing data (%)	4067 (20.8)	1328 (21.9)	2739 (20.3)	664 (23.4)	3403 (20.4)	2437 (25.8)	1630 (16.1)
Comorbidities, n (%)
Hypertension	5087 (26.0)	1451 (23.9)	3636 (27.0)	665 (23.4)	4422 (26.5)	2611 (27.7)	2476 (24.5)	<0.001
Diabetes mellitus	4072 (20.8)	1146 (18.9)	2926 (21.7)	289 (10.2)	3783 (22.7)	2328 (24.7)	1744 (17.3)	<0.001
Ischemic heart disease	5178 (26.5)	978 (16.1)	4200 (31.2)	281 (9.9)	4897 (29.3)	2562 (27.2)	2616 (25.9)	0.041
Dyslipidemia	2241 (11.5)	590 (9.7)	1651 (12.3)	154 (5.4)	2087 (12.5)	1090 (11.6)	1151 (11.4)	0.714
Atrial fibrillation	3594 (18.4)	1306 (21.5)	2288 (17.0)	274 (9.6)	3320 (19.9)	2006 (21.3)	1588 (15.7)	<0.001
Stroke	1382 (7.1)	394 (6.5)	988 (7.3)	129 (4.5)	1253 (7.5)	820 (8.7)	562 (5.6)	<0.001
Congenital heart disease	201 (1.0)	94 (1.6)	107 (0.8)	93 (3.3)	108 (0.6)	126 (1.3)	75 (0.7)	<0.001
Cardiomyopathy	3115 (15.9)	848 (14.0)	2267 (16.8)	734 (25.8)	2381 (14.3)	1176 (12.5)	1939 (19.2)	<0.001
Hypertrophic cardiomyopathy	405 (2.1)	119 (2.0)	286 (2.1)	75 (2.6)	330 (2.0)	184 (2.0)	221 (2.2)	0.272
Dilated cardiomyopathy	2512 (12.9)	659 (10.9)	1853 (13.8)	602 (21.2)	1910 (11.4)	893 (9.5)	1619 (16.0)	<0.001
Rheumatic heart disease	553 (2.8)	361 (6.0)	192 (1.4)	58 (2.0)	495 (3.0)	330 (3.5)	223 (2.2)	<0.001
Infective endocarditis	326 (1.7)	89 (1.5)	237 (1.8)	81 (2.8)	245 (1.5)	225 (2.4)	101 (1.0)	<0.001
Valvular heart disease	1829 (9.4)	642 (10.6)	1187 (8.8)	255 (9.0)	1574 (9.4)	1149 (12.2)	680 (6.7)	<0.001
Previous history of SCA	615 (3.1)	169 (2.8)	446 (3.3)	104 (3.7)	511 (3.1)	329 (3.5)	286 (2.8)	0.009
Peripheral vascular disease	677 (3.5)	182 (3.0)	495 (3.7)	61 (2.1)	616 (3.7)	468 (5.0)	209 (2.1)	<0.001
Liver cirrhosis	282 (1.4)	87 (1.4)	195 (1.4)	45 (1.6)	237 (1.4)	157 (1.7)	125 (1.2)	0.014
COPD	488 (2.5)	48 (0.8)	440 (3.3)	16 (0.6)	472 (2.8)	258 (2.7)	230 (2.3)	0.043
Chronic kidney disease	3124 (16.0)	893 (14.7)	2231 (16.6)	468 (16.5)	2656 (15.9)	1785 (18.9)	1339 (13.2)	<0.001
Rheumatism	385 (2.0)	221 (3.6)	164 (1.2)	53 (1.9)	332 (2.0)	215 (2.3)	170 (1.7)	0.003
Depression	346 (1.8)	181 (3.0)	165 (1.2)	36 (1.3)	310 (1.9)	188 (2.0)	158 (1.6)	0.026
Cancer	2148 (11.0)	1166 (19.2)	982 (7.3)	174 (6.1)	1974 (11.8)	1001 (10.6)	1147 (11.3)	0.110
Medications, n (%)
ACE inhibitor/ARB	11491 (58.8)	3057 (50.4)	8434 (62.6)	1493 (52.5)	9998 (59.9)	5526 (58.6)	5965 (59.0)	0.598
ARNI	2012 (10.3)	367 (6.1)	1645 (12.2)	354 (12.5)	1658 (9.9)	108 (1.1)	1904 (18.8)	<0.001
SGLT2 inhibitor	1883 (9.6)	390 (6.4)	1493 (11.1)	238 (8.4)	1645 (9.9)	113 (1.2)	1770 (17.5)	<0.001
Beta blocker	10555 (54.0)	2922 (48.2)	7633 (56.6)	1468 (51.6)	9087 (54.4)	5032 (53.4)	5523 (54.6)	0.082
CCB	4715 (24.1)	1456 (24.0)	3259 (24.2)	651 (22.9)	4064 (24.3)	2455 (26.0)	2260 (22.4)	<0.001
Diuretics	8106 (41.5)	2441 (40.3)	5665 (42.0)	902 (31.7)	7204 (43.2)	4502 (47.8)	3604 (35.6)	<0.001
MRA	3629 (18.6)	808 (13.3)	2821 (20.9)	567 (19.9)	3062 (18.3)	1357 (14.4)	2272 (22.5)	<0.001
Digoxin	1583 (8.1)	617 (10.2)	966 (7.2)	137 (4.8)	1446 (8.7)	949 (10.1)	634 (6.3)	<0.001
Insulin	1810 (9.3)	583 (9.6)	1227 (9.1)	136 (4.8)	1674 (10.0)	907 (9.6)	903 (8.9)	0.102
Metformin	3020 (15.5)	868 (14.3)	2152 (16.0)	215 (7.6)	2805 (16.8)	1426 (15.1)	1594 (15.8)	0.224
Aspirin	7094 (36.3)	1460 (24.1)	5634 (41.8)	592 (20.8)	6502 (38.9)	3914 (41.5)	3180 (31.5)	<0.001
Other antiplatelets	3005 (15.4)	473 (7.8)	2532 (18.8)	202 (7.1)	2803 (16.8)	1499 (15.9)	1506 (14.9)	0.054
Statin	8610 (44.1)	2076 (34.3)	6534 (48.5)	690 (24.3)	7920 (47.4)	4085 (43.3)	4525 (44.8)	0.047
Laboratory, median (IQR)
TG, mmol/La	1.2 (0.9, 1.7)	1.2 (0.9, 1.7)	1.2 (0.9, 1.6)	1.2 (0.9, 1.7)	1.2 (0.9, 1.7)	1.2 (0.9, 1.7)	1.2 (0.9, 1.6)	0.050
Missing data (%)	2101 (10.8)	883 (14.6)	1218 (9.0)	554 (19.5)	1547 (9.3)	1195 (12.7)	906 (9.0)
TC, mmol/La	4.2 (3.5, 5.1)	4.3 (3.6, 5.2)	4.2 (3.4, 5.0)	4.3 (3.6, 5.2)	4.2 (3.5, 5.0)	4.3 (3.6, 5.1)	4.1 (3.4, 5.0)	<0.001
Missing data (%)	4529 (23.2)	1723 (28.4)	2806 (20.8)	928 (32.6)	3601 (21.6)	1958 (20.8)	2571 (25.4)
HDL, mmol/La	1.1 (1.0, 1.4)	1.2 (1.0, 1.5)	1.1 (1.0, 1.3)	1.1 (1.0, 1.3)	1.1 (1.0, 1.4)	1.1 (1.0, 1.4)	1.1 (1.0, 1.4)	0.678
Missing data (%)	6299 (32.2)	2067 (34.1)	4232 (31.4)	1374 (48.3)	4925 (29.5)	3440 (36.5)	2859 (28.3)
LDL, mmol/La	2.5 (1.8, 3.2)	2.5 (1.8, 3.2)	2.5 (1.9, 3.2)	2.7 (2.1, 3.3)	2.4 (1.8, 3.2)	2.6 (2.0, 3.3)	2.4 (1.8, 3.1)	<0.001
Missing data	4681 (24.0)	1766 (29.1)	2915 (21.6)	954 (33.6)	3727 (22.3)	2065 (21.9)	2616 (25.9)
Fasting glucose, mmol/La	5.6 (5.0, 6.3)	5.5 (4.9, 6.2)	5.6 (5.0, 6.3)	5.3 (4.8, 6.0)	5.6 (5.0, 6.3)	5.6 (5.0, 6.3)	5.6 (5.0, 6.3)	0.528
Missing data (%)	6280 (32.1)	2137 (35.3)	4143 (30.7)	1137 (40.0)	5143 (30.8)	2702 (28.7)	3578 (35.4)
HbA1c, %a	5.8 (5.5, 6.2)	5.8 (5.5, 6.1)	5.9 (5.5, 6.2)	5.7 (5.4, 6.0)	5.8 (5.6, 6.2)	5.8 (5.6, 6.2)	5.8 (5.5, 6.2)	0.790
Missing data	8111 (41.5)	2382 (39.3)	5729 (42.5)	1231 (43.3)	6880 (41.2)	4315 (45.8)	3796 (37.5)
Creatine, μmol/La	90 (72, 113)	71 (59, 89)	98 (82, 118)	88 (69, 113)	91 (73, 113)	90 (73, 113)	89 (72, 113)	0.046
Missing data (%)	51 (0.3)	18 (0.3)	33 (0.2)	17 (0.6)	34 (0.2)	9 (0.1)	42 (0.4)
eGFR, mL/min/1.73m2a	84 (66, 103)	90 (68, 107)	82 (65, 102)	97 (72, 118)	82 (65, 102)	84 (66, 103)	84 (66, 104)	0.063
Missing data (%)	109 (0.6)	28 (0.5)	81 (0.6)	25 (0.9)	84 (0.5)	33 (0.4)	76 (0.8)

Data are median (IQR) or n (%) as stated.

CSSA, comprehensive social security assistance scheme; SCA, sudden cardiac arrest; COPD, chronic obstructive pulmonary disease; ACE, Angiotensin-converting enzyme; ARB, Angiotensin II receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor; SGLT2, sodium-glucose cotransporter 2; CCB, Calcium channel blocker; MRA, mineralocorticoid receptor antagonist; TG, triglyceride; TC, serum total cholesterol; HDL, high-density lipoprotein; LDL, low-density lipoprotein; HbA1c, glycated hemoglobin A1c; eGFR, estimated glomerular filtration rate.

^aFor these variables, missing values are reported as the proportion of records with incomplete entries, and analyses are conducted using available data, excluding missing observations.

Table 2.

Echocardiographic characteristics of young individuals with heart failure, 2014 through 2023.

Echocardiographic characteristic	Overall (n = 1134)	Sex		Age		Time period
		Women (n = 364)	Men (n = 770)	<45 years (n = 210)	45–65 years (n = 924)	2014–2018 (n = 587)	2019–2023 (n = 547)	P-valuea
IVSd, mean (SD), mma	12.0 (10.0, 14.0)	11.0 (10.0, 13.0)	12.0 (10.0, 14.0)	11.0 (9.0, 13.0)	12.0 (10.0, 14.0)	12.0 (10.0, 14.0)	12.0 (10.0, 13.0)	0.049
Missing data (%)	151 (13.3)	42 (11.5)	109 (14.2)	27 (12.9)	124 (13.4)	112 (19.1)	39 (7.1)
LVPWd, mean (SD), mma	11.0 (10.0, 13.0)	11.0 (9.0, 12.0)	11.0 (10.0, 13.0)	11.0 (10.0, 12.0)	11.0 (10.0, 13.0)	11.0 (10.0, 13.0)	11.0 (10.0, 12.8)	0.044
Missing data (%)	155 (13.7)	44 (12.1)	111 (14.4)	27 (12.9)	128 (13.9)	114 (19.4)	41 (7.5)
RWT, median (IQR)a	0.44 (0.36, 0.53)	0.47 (0.38, 0.57)	0.43 (0.35, 0.53)	0.43 (0.35, 0.52)	0.44 (0.36, 0.54)	0.46 (0.36, 0.54)	0.43 (0.35, 0.53)	0.102
Missing data (%)	169 (14.9)	46 (12.6)	123 (16.0)	31 (14.8)	138 (14.9)	123 (21.0)	46 (8.4)
RWT>0.42, n (%)a	598 (52.7)	211 (58.0)	387 (50.3)	102 (48.6)	496 (53.7)	301 (51.3)	297 (54.3)	0.338
Missing data (%)	169 (14.9)	46 (12.6)	123 (16.0)	31 (14.8)	138 (14.9)	123 (21.0)	46 (8.4)
LVM, median (IQR), ga	230 (178, 288)	191 (147, 239)	249 (194, 309)	216 (170, 287)	233 (180, 288)	233 (180, 295)	226 (176, 283)	0.214
Missing data (%)	169 (14.9)	46 (12.6)	123 (16.0)	31 (14.8)	138 (14.9)	123 (21.0)	46 (8.4)
LVM index to BSA, median (IQR), g/m2a	129 (103, 163)	118 (92, 147)	138 (112, 167)	120 (93, 145)	133 (105, 165)	134 (107, 168)	129 (102, 161)	0.322
Missing data (%)	498 (43.9)	147 (40.4)	351 (45.6)	103 (49.0)	395 (42.7)	388 (66.1)	110 (20.1)
LV Geometry, n (%)a								0.895
Normal	442 (42.2)	130 (38.1)	312 (44.2)	93 (47.7)	349 (41.0)	223 (42.3)	219 (42.1)
Concentric remodeling	279 (26.6)	116 (34.0)	163 (23.1)	49 (25.1)	230 (27.0)	136 (25.8)	143 (27.5)
Concentric hypertrophy	299 (28.6)	88 (25.8)	211 (29.9)	50 (25.6)	249 (29.2)	155 (29.4)	144 (27.7)
Eccentric hypertrophy	27 (2.6)	7 (2.1)	20 (2.8)	3 (1.5)	24 (2.8)	13 (2.5)	14 (2.7)
Missing data (%)	87 (7.7)	23 (6.3)	64 (8.3)	15 (7.1)	72 (7.8)	60 (10.2)	27 (4.9)
LVEF, median (IQR), %	44 (30, 56)	50 (40, 60)	40 (25, 55)	45 (26, 60)	41 (30, 55)	45 (30, 60)	40 (30, 55)	0.071
HFpEF, ≥50%	480 (42.3)	221 (60.7)	259 (33.6)	92 (43.8)	388 (42.0)	268 (45.7)	212 (38.8)	0.022
HFrEF, <50%	654 (57.7)	143 (39.3)	511 (66.4)	118 (56.2)	536 (58.0)	319 (54.3)	335 (61.2)
LA dilation, n (%)								0.039
Mild	383 (33.8)	109 (29.9)	274 (35.6)	59 (28.1)	324 (35.1)	178 (30.3)	205 (37.5)
Moderate or severe	242 (21.3)	89 (24.5)	153 (19.9)	42 (20.0)	200 (21.6)	131 (22.3)	111 (20.3)
RA dilation, n (%)								0.873
Mild	212 (18.7)	57 (15.7)	155 (20.1)	35 (16.7)	177 (19.2)	113 (19.3)	99 (18.1)
Moderate or severe	85 (7.5)	36 (9.9)	49 (6.4)	13 (6.2)	72 (7.8)	43 (7.3)	42 (7.7)
Valvular dysfunction, n (%)
Mitral stenosis								0.715
Mild	11 (1.0)	6 (1.6)	5 (0.6)	1 (0.5)	10 (1.1)	7 (1.2)	4 (0.7)
Moderate or severe	28 (2.5)	22 (6.0)	6 (0.8)	5 (2.4)	23 (2.5)	15 (2.6)	13 (2.4)
Aortic stenosis								0.419
Mild	29 (2.6)	15 (4.1)	14 (1.8)	1 (0.5)	28 (3.0)	17 (2.9)	12 (2.2)
Moderate or severe	40 (3.5)	15 (4.1)	25 (3.2)	2 (1.0)	38 (4.1)	24 (4.1)	16 (2.9)
Mitral regurgitation								0.879
Trivial	129 (11.4)	49 (13.5)	80 (10.4)	29 (13.8)	100 (10.8)	66 (11.2)	63 (11.5)
Mild	472 (41.6)	161 (44.2)	311 (40.4)	91 (43.3)	381 (41.2)	242 (41.2)	230 (42.0)
Moderate or severe	110 (9.7)	34 (9.3)	76 (9.9)	18 (8.6)	92 (10.0)	61 (10.4)	49 (9.0)
Aortic regurgitation								0.213
Trivial	441 (38.9)	136 (37.4)	305 (39.6)	101 (48.1)	340 (36.8)	211 (35.9)	230 (42.0)
Mild	193 (17.0)	83 (22.8)	110 (14.3)	26 (12.4)	167 (18.1)	105 (17.9)	88 (16.1)
Moderate or severe	40 (3.5)	10 (2.7)	30 (3.9)	6 (2.9)	34 (3.7)	21 (3.6)	19 (3.5)
Tricuspid regurgitation								0.679
Trivial	128 (11.3)	32 (8.8)	96 (12.5)	25 (11.9)	103 (11.1)	66 (11.2)	62 (11.3)
Mild	385 (34.0)	141 (38.7)	244 (31.7)	70 (33.3)	315 (34.1)	207 (35.3)	178 (32.5)
Moderate or severe	71 (6.3)	39 (10.7)	32 (4.2)	8 (3.8)	63 (6.8)	33 (5.6)	38 (6.9)

Data are median (IQR) or n (%) as stated.

LV, left ventricular; IVSd, inter-ventricular septal dimension at end-diastole; LVPWd, LV posterior wall thickness at end-diastole; RWT, relative wall thickness; LVM, LV mass; BSA, body surface area; LVEF, LV ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFmrEF, heart failure with mid-range preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LA, left atrium; RA, right atrium.

^aFor these variables, missing values are reported as the proportion of records with incomplete entries, and analyses are conducted using available data, excluding missing observations.

Temporal trends in incidence rates of young HF

After age- and sex-standardisation, the IRs of young HF increased by 20%, from 97 in 2014 to 116 in 2023 per 100,000 people (IRR 1.20, 95%CI 1.13–1.27) (Fig. 1). The increasing trend was consistent in the 45–65 years category (82 in 2014 and 100 in 2023 per 100,000 people, IRR 1.22, 95%CI 1.14–1.31) but not in the <45 years category (15 in 2014 and 16 in 2023 per 100,000 people, IRR 1.06, 95%CI 0.91–1.24) (Fig. 1B). The total IRs for men rose from 76 in 2014 to 92 in 2023 per 100,000 people, showing a 21% increase (IRR 1.21, 95% CI 1.12–1.30) (Fig. 1B). Specifically, the IRs for men aged 45–65 were 65 and 80 per 100,000 people in 2014 and 2023, respectively. For men aged <45, the IRs were 11 and 12 per 100,000 people in 2014 and 2023, respectively (Fig. 1A). The total IRs for women also increased, from 21 in 2014 to 24 in 2023 per 100,000 people, indicating a 15% rise (IRR 1.15, 95%CI 1.03–1.29) (Fig. 1B). In women aged 45–65 years, the IRs were 17 and 20 in 2014 and 2023, respectively, while remaining at 4 per 100,000 people for those aged <45 years (Fig. 1A).

Fig. 1.

Overall and standardised incidence rates of young heart failure, 2014 through 2023. (A) Absolute number of cases and standardised incidence rates of young heart failure by age and sex; (B) Overall and category-specific incidence rate ratios of young heart failure in 2023 versus 2014 (based on mid-year census estimates).

Temporal trends in comorbidities and shift patterns of risk factors

The average number of comorbidities at or before index HF was 2.04 (1.52) and declined over time, from 2.26 (1.64) in 2014 to 1.75 (1.52) in 2023 (difference adjusted for age and sex −0.51, 95% CI -0.61 to −0.41) (Fig. 2A). Overall, 57% had two or more comorbidities, decreasing from 63% in 2014 to 47% in 2023. The 20 most prevalent comorbidities among young HF patients were illustrated in Fig. 2B.

Fig. 2.

Temporal trends in comorbidities among young heart failure patients, 2014 through 2023. (A) Number of comorbidities (out of 20 major conditions) affecting young heart failure patients over time. (B) Cumulative percentage of patients affected by individual comorbidities over time. COPD, chronic obstructive pulmonary disease; SCA, sudden cardiac arrest.

After multivariable adjustment, young HF patients in 2019–2023 were more likely to be obese (OR 1.68, 95%CI 1.55–1.82), recipients of CSSA (OR 1.61, 95%CI 1.44–1.80), diagnosed with cardiomyopathy (OR 1.55, 95%CI 1.42–1.68), and aged 45–65 years (OR 1.21, 95%CI 1.11–1.32) compared to those in 2014–2018 (Fig. 3). In contrast, young HF patients in 2014–2018 were frequently associated with alcohol consumption and demonstrated higher likelihoods of T2DM, AF, valvular heart disease, history of SCA, and CKD compared to those in 2019–2023. Subgroup analyses stratified by age and sex demonstrated temporal patterns consistent with those observed in the overall cohort (Fig. S2).

Fig. 3.

Forest plot depicting shifting associations between past (2014–2018) and current (2019–2023) young heart failure in multivariable analysis. CSSA, comprehensive social security assistance; AF, atrial fibrillation; SCA, sudden cardiac arrest; COPD, chronic obstructive pulmonary disease.

Temporal trends in adverse outcomes

The incidence of all-cause mortality within one year declined from 2014 to 2018 to 2019–2023 (15.6 [14.8, 16.5] vs. 14.6 [13.8, 15.5] events per 100 person-years), driven primarily by a reduction in cardiovascular mortality (9.32 [8.69, 10.0] vs. 8.35 [7.75, 9.01]), while non-cardiovascular mortality remained unchanged (Table 3; Fig. 4). Significant declines were also observed in HF hospitalisation, cardiovascular mortality or HF hospitalisation, MACE, and sudden death. After adjustment for age, sex, obesity, CSSA beneficiaries, and cardiomyopathy, hazard ratios for one-year mortality and secondary outcomes consistently decreased over time (Table S4).

Table 3.

Temporal trends in the incidence of one-year outcomes.

Time period	Cases/No.	Events per 100 PY (95%CI)
All-cause mortality
2014–2018	1324/9427	15.6 (14.8, 16.5)
2019–2023	1201/10110	14.6 (13.8, 15.5)
CV mortality
2014–2018	791/9427	9.32 (8.69, 10.0)
2019–2023	686/10110	8.35 (7.75, 9.01)
non-CV mortality
2014–2018	533/9427	6.28 (5.76, 6.84)
2019–2023	515/10110	6.27 (5.75, 6.84)
HF hospitalisation
2014–2018	3655/9427	62.3 (60.4, 64.3)
2019–2023	3345/10110	55.8 (54.0, 57.6)
CV mortality or HF hospitalisation
2014–2018	4206/9427	71.7 (69.6, 73.8)
2019–2023	3841/10110	64.0 (62.1, 66.0)
MACE
2014–2018	1963/9427	26.2 (25.1, 27.4)
2019–2023	1616/10110	21.6 (20.6, 22.7)
Sudden death
2014–2018	399/9427	4.70 (4.26, 5.19)
2019–2023	254/10110	3.09 (2.73, 3.50)

CV, cardiovascular; MACE, major adverse cardiovascular events.

Fig. 4.

Kaplan–Meier curves forone-yearoutcome risks. (A) All-cause mortality (Log-rank test: χ² = 7.6, df = 1, p = 0.006); (B) CV mortality (Log-rank test: χ² = 10.1, df = 1, p = 0.001); (C) non-CV mortality (Log-rank test: χ² = 0.4, df = 1, p = 0.53); (D) HF hospitalisation (Log-rank test: χ² = 33.5, df = 1, p < 0.001); (E) CV mortality or HF hospitalisation (Log-rank test: χ² = 39.9, df = 1, p < 0.001); (F) MACE (Log-rank test: χ² = 55.2, df = 1, p < 0.001); (G) Sudden death (Log-rank test: χ² = 34.3, df = 1, p < 0.001). CV, cardiovascular; MACE, major adverse cardiovascular events.

Subgroup analyses demonstrated a consistent decline in one-year all-cause mortality across the study period, with the exception of individuals with T2DM, stroke, or CKD, in whom one-year all-cause mortality showed an upward trend (Table S5). Sensitivity analyses excluding 515 patients with a history of COVID-19 and 335 who developed COVID-19 during follow-up yielded results consistent with the primary analysis, indicating that the observed outcome trends were robust (Table S6). Among older adults aged 65–90 years (n = 69,379; Table S7), exclusion of 11,171 individuals with COVID-19 indicated a decrease in one-year mortality among those aged 65–75 years, whereas individuals aged 75–90 years exhibited lower survival. Both age groups demonstrated declining temporal trends in the incidence of cardiovascular adverse events. Further analysis identified a shift in the leading cause of mortality (Fig. S3), from sudden death at 31% in 2014 to IHD at 17% in 2023, with IHD marginally exceeding sudden death, which accounted for 16% in 2023.

Discussion

In the territory-wide cohort study, we identified a substantial increase in the incidence of young HF between 2014 and 2023, predominantly among men aged 45–65. Compared to young HF patients in 2014–2018, those in 2019–2023 demonstrated a lower comorbidity burden, similar patterns of cardiac remodeling and valvulopathy, a rising subtype of HFrEF, and higher GDMT use. Notably, shifting risk profiles were observed with increasing rates of obesity, CSSA beneficiaries, cardiomyopathy, and individuals aged 45–65 years. Despite these temporal changes, one-year mortality and cardiovascular adverse events reduced over the study period. The present study advances prior research by 1) being the first large-scale, multicenter, territory-wide investigation targeting young HF patients with standardised characterisation and systematic monitoring; 2) providing new perspectives on echocardiographic traits and shifting patterns of risk factors; 3) incorporating the most recent recommendations for GDMT.

Epidemiological studies have previously indicated an increasing trend in the incidence of young HF,5, 6, 7, 8, 9 consistent with our observations in Hong Kong, which suggest a worldwide tendency.²⁸ Notably, our data revealed a 20% increase in young HF incidence between 2014 and 2023, exceeding the 10% rise reported among French individuals aged 18–50 years from 2013 to 2018, and the 1.5% annual increase among New Zealanders aged 20–49 years between 2006 and 2018.^5,8 Interestingly, both the French data and ours highlighted an age-dependent susceptibility,⁵ with a greater burden of young HF observed with advancing age (<45 vs. 45–65 years: 16% vs. 22% in France). In contrast, divergent findings were observed in a Swedish cohort, where individuals <45 years experienced more than a 40% increase in HF incidence from 1991 to 2006, compared to a 16% increase among those aged 45–55 years and a decrease in those above 55 years.⁶ Although the risks differ between the <45 years and 45–65 years age groups across studies, these variations may reflect disproportionate influences of geographical, ethnic, and lifestyle-specific disparities across distinct young age categories.28, 29, 30 Nonetheless, the overall population of young HF is rising globally, posing a concerning trend.

One of the most striking findings in our investigation was the declining trend in the prevalence of major comorbidities, including hypertension, T2DM, and IHD from 2014 to 2023. This unprecedented observation points to an evolving paradigm in standardised therapeutics for young individuals, aligning with previous studies that focused on trends in earlier time periods rather than recent years.^6,31 For instance, a decline in the prevalence of T2DM and IHD among young HF patients under 45 years was observed in a Swedish cohort from 1985 to 2006.⁶ Similarly, a decreasing trend in hypertension and diabetes (type I and II) was reported in the French national hospitalisation database among young HF patients (aged 18–50) from 2013 to 2018.⁵ Although the general population has experienced a growing prevalence of hypertension over the past two decades,¹² recent evidence suggests that timely detection and management could reduce the complication rate of HF by 30% in young individuals.³² Besides, decreasing trends in HF complication rates were identified within young patients with diabetes (aged 45–65) between 1995 and 2015, utilising data from the US Centers for Disease Control and Prevention (CDC) National Health Interview Survey.³³ This finding aligns with ours, which may be attributed to the general decrease in diabetes prevalence at a young age in Hong Kong,³⁴ and the rising use of metformin and SGLT2 inhibitors among this population, as evidenced by the Hong Kong Diabetes Surveillance Database (HKDSD) from 2002 to 2019.³⁴ The established cardioprotective benefits of these pharmacotherapies, particularly in HF,^35,36 further support decreasing number of T2DM cases in our study. The observed increasing trend of AF in the Swedish data from 1985 to 2006 in young HF⁶ and in cohort from the community surveillance component of the ARIC study from 2005 to 2014 that included patients ≥55 years of age,³¹ however contrasts with our study, which found a decreasing number of AF cases among young HF patients. Our recent cohort contributes contemporary insights, incorporating advancements in effective HF treatments. It further confirms that the current population-attributable estimate for HF associated with AF is relatively low, 2.6% across all age groups and below 5% primarily for incident HFpEF.^1,37 This likely reflects the compensatory benefits of effective treatments rather than the direct impact of the increasing AF prevalence on HF incidence.

Beyond the declining number of comorbidities, the patterns of risk factors signify a remarkable transformation. Obesity, known for its causal association with early development of HF,^11,38 nearly doubled among young HF patients in 2019–2023 compared to those in 2014–2018. While this rising trend parallels findings from Denmark,⁷ the prevalence in our cohort was twice that reported in both Denmark⁷ and Sweden,¹⁴ underscoring a substantial obesity burden within this population. The underlying mechanisms involving cardiometabolic disturbances, elevated inflammation, and premature cardiac structural abnormalities have been extensively described,^39,40 which calls for immediate efforts to enhance the understanding and management of this emerging epidemic. Another novel finding was that CSSA beneficiaries, representing the lowest socioeconomic stratum in Hong Kong, have shown increasing trends and substantial wealth inequality compared to the wealthiest individuals over the past decade. While the worsening wealth inequality in Hong Kong was primarily noted in the elderly,⁴¹ the increasing financial vulnerability among young individuals demands equal attention. A recent study revealed that individuals with lower socioeconomic status were predisposed to develop HF at a younger age compared to their higher socioeconomic status counterparts.¹³ Our study extended this discovery, revealing an alarming uptrend in the development of HF among individuals with lower socioeconomic status, particularly among younger populations. Cardiomyopathy, commonly associated with a higher risk of HF in young individuals, showed an increasing trend in our analysis, with the largest contribution attributed to the growing prevalence of DCM, consistent with observations from Sweden.^6,14 In Sweden, Barasa et al. reported a comparable prevalence of cardiomyopathy in patients under 55 years compared to our study (14.4% vs. 15.9%), with DCM driving this increasing trend. Later observations in 2020 by Basic et al.,¹⁴ further supported this, reporting a prevalence of 27.2% for DCM and 2.0% HCM. Remarkably, a higher prevalence of DCM was noted in patients under 50 years in CHARM study² and MAGGIC meta-analysis,⁴² (50% and 48.5%, respectively), suggesting significant regional variations. While most patients with HCM do not develop symptomatic HF,⁴³ this explains the limited number of HCM cases in our cohort. Aging is another important contributing factor to HF. In our analysis, the <45 years age group remained consistent over time, probably due to its small proportion, as observed in the French study (<45% vs. 45–70 years: 8.7% vs 91.3%),⁵ which may have impeded significant temporal changes within this demographic. Conversely, the 45–65 years age group demonstrated an increasing trend after multivariable adjustments, despite the unfavorable results observed in univariable analysis. This finding, while not fully understood, could potentially be explained by the decreasing incidence of congenital heart disease due to advancements in therapeutic interventions. Furthermore, numerous risk factors remain unassessed, including substance abuse (e.g., cocaine and anabolic steroids), iron deficiency, myocardial inflammation, neuromuscular diseases, psychological stress, sleep disorders, and human immunodeficiency virus related conditions.^12,44 These factors, common in young populations, warrant further investigation in future research.

One-year mortality improved, aligning with previous reports from Sweden, where young adult mortality rates were similarly observed (11.9% vs. 12.9%).⁶ The marked decline in one-year mortality observed in Sweden from 1987 to 2006 was largely attributable to the widespread use of ACE inhibitors and Beta-blockers.⁴⁵ Extending these observations, our contemporary data from 2014 to 2023 demonstrated significant reductions in both one-year mortality and cardiovascular adverse events, likely reflecting the cardioprotective benefits of more recently recommended therapies, including ARNI and SGLT2 inhibitors.⁴⁶ Nevertheless, in adjusted analyses, the observed reduction in one-year mortality was less than 10%, considerably lower than the over 30% decrease reported in the Swedish cohort, suggesting that the coverage of GDMT remains suboptimal.⁴⁷ Greater efforts are needed to improve adherence to and persistence with GDMT in this demographic. Additionally, the lack of significant improvement in non-cardiovascular mortality in our study, as reported by Tao et al. in the American cohort,⁴⁸ may reinforce the notion that young HF patients are inadequately treated with GDMT, or it could indicate limited efficacy of contemporary GDMT use in addressing non-cardiac risks within this population. Moreover, as our analysis did not account for cancer therapies administered prior to or at the time of HF diagnosis, we cannot exclude the potential impact of oncologic treatments on outcomes in this cohort. In total, these results highlight the need for further strategies and therapies to better enhance survival outcomes among young HF patients.

Clinical perspectives

This population-based study has important implications for young HF care delivery. The observed 20% increase in HF incidence within this demographic, coupled with only modest improvements in one-year mortality, points to a growing clinical challenge and increasing demands on healthcare systems. These temporal trends may be partially attributable to the growing prevalence of obesity and cardiomyopathy, an increasing number of CSSA beneficiaries, and a greater proportion of individuals aged 45–65 years, which necessitates targeted risk factor surveillance and the implementation of effective early intervention strategies. Ensuring equitable access to healthcare resources is crucial, particularly in view of the rising proportion of young patients presenting with HFrEF at diagnosis. Encouragingly, reductions in adverse cardiovascular outcomes, including HF hospitalisation, cardiovascular mortality or HF hospitalisation, MACE, and sudden death were observed, conveying a critical and optimistic message for young HF patients.

Limitations

This study has several limitations. First, utilising diagnostic codes poses a risk of underestimation, especially in the young population, where cases lacking substantial medical care could remain uncoded. Nonetheless, we discerned an upsurge in the incidence of young HF, surpassing the findings of previous studies. The absence of natriuretic peptides and NYHA functional class in the CDARS did not hinder the detection of young HF; however, we acknowledge that their incorporation could have offered supplementary value to this study. Second, as BMI is not systematically recorded in CDARS and may be missing for some patients, this approach may result in underestimation or misclassification of obesity status. The pervasive issues of under-diagnosed and under-coded smoking, alcohol use, depression are widely recognised in administrative data, which may lead to residual confounding in the assessment of these comorbidities. Third, echocardiographic data were available only from Queen Mary Hospital, located in a relatively affluent district of Hong Kong, raising the possibility of selection bias in analyses restricted to this subset. Differences in baseline characteristics between patients with and without echocardiographic data (Table S8) limit the generalisability and reproducibility of observed temporal trends in cardiac remodeling and valvulopathy.

Conclusions

In summary, the incidence of HF increased substantially among young adults, who had fewer comorbidities, rising HFrEF subtype, and shifting risk factors between 2014 and 2023. Despite increased utilisation of GDMT, improvements in one-year survival were modest. These findings highlight the urgent need for targeted strategies to address the growing incidence of HF and its underlying determinants in this population.

Contributors

Dr. KH Yiu and Ms. WL Gu had full access to all data in the study and had responsibility for the integrity of the data and the accuracy of the data analysis. The final decision to submit the manuscript for publication was made by KH Yiu. Concept and design: KH Yiu, C.S.P Lam, A Pandey; Acquisition of data: WL Gu, JY Huang, JN Zhang; Statistical analysis: WL Gu, TH-K Teng, W Ouwerkerk; Repeated analysis: C Lawson, C Chandramouli, WT Tay; Drafting of the manuscript: WL Gu, TH-K Teng; Interpretation of data: R Guo, HC Xuan, YH Chan; Critical revision of the manuscript for important intellectual content: All authors; Administrative, technical, or material support: KH Yiu, J Tromp; Supervision: KH Yiu, C.S.P Lam.

Data sharing statement

Data can be obtained upon reasonable request by contacting Dr. Yiu Kai-Hang.

Editor note

The Lancet Group takes a neutral position with respect to territorial claims in published maps and institutional affiliations.

Declaration of interests

KH Yiu has received research support from Novartis and AztraZeneca; is supported by the Sun Chieh Yeh Heart Foundation Fund, General Research Fund (GRF) under Research Grant Council (RGC) of University Grant Committee (UGC), Theme-based Research Scheme (TRS) under UGC and Health and Medical Research Fund (HMRF) under Hong Kong Government Health Bureau.

C.S.P Lam is supported by a Clinician Scientist Award from the National Medical Research Council of Singapore; has Received research support from NovoNordisk and Roche Diagnostics; has Served as consultant or on the Advisory Board/ Steering Committee/ Executive Committee for Alleviant Medical, Allysta Pharma, AnaCardio AB, Applied Therapeutics, AstraZeneca, Bayer, Biopeutics, Boehringer Ingelheim, Boston Scientific, Bristol Myers Squibb, CardioRenal, Eli Lilly, Impulse Dynamics, Intellia Therapeutics, Ionis Pharmaceutical, Janssen Research & Development LLC, Medscape/WebMD Global LLC, Merck, Novartis, Novo Nordisk, Prosciento Inc, Quidel Corporation, Radcliffe Group Ltd., Recardio Inc, ReCor Medical, Roche Diagnostics, Sanofi, Siemens Healthcare Diagnostics and Us2.ai; and serves as co-founder & non-executive director of Us2.ai. Patents issued or pending: Patent pending: PCT/SG2016/050217 (Method for diagnosis and prognosis of chronic heart failure); US Patent No. 10,702, 247 (Automated clinical workflow that recognizes and analyses 2-dimensional and Doppler echo images for cardiac measurements and the diagnosis, prediction and prognosis of heart disease).

A Pandey has received research support from the National Institute on Minority Health and Disparities (R01MD017529), the National Institute of Heart, Lung, and Blood Institute (R21HL169708), American Heart Association, Ultromics, Anumana, and Roche Diagnostics; serves as a consultant for and/or received honoraria outside of the present study as an advisor/consultant for Northwestern University, Tricog Health Inc, Lilly USA, Rivus, Cytokinetics, Roche Diagnostics, Sarfez Therapeutics, Edwards Lifesciences, Merck, Bayer, Novo Nordisk, Alleviant, Axon Therapies, Kilele Health, Acorai, Ultromics, Kardigan, Novartis, Idorsia Pharma, and Science37; also served as consultant for Palomarin Inc. with stocks compensation.

C Lawson is supported by the National Institute for Health Research (NIHR) Leicester Biomedical Research Centre (BRC) and the BHF Centre for Excellence.

J Tromp is supported by the National University of Singapore Start-up grant, the tier 1 grant from the Ministry of Education and the CS-IRG New Investigator Grant from the National Medical Research Council; has received research support from AstraZeneca and consulting or speaker fees from Daiichi-Sankyo, Boehringer Ingelheim, Roche diagnostics and Us2.ai. Patent holder of US-10702247-B2. Stock options in Us2.ai.

W Ouwerkerk has consulted for Us2.ai. Patent holder of US-10702247-B2. Stock options in Us2.ai.

The remaining authors declare nothing.