Abstract
Aims
This study aimed to characterize a contemporary population with subtypes of incident or prevalent heart failure (HF) based on reduced (HFrEF), mildly reduced, or preserved (HFpEF) left ventricular ejection fraction (LVEF) and to assess how outcomes, healthcare, treatments, and healthcare costs vary between each subtype of incident HF.
Methods and results
Using Swedish data from the CardioRenal and Metabolic disease Heart Failure (CaReMe HF) study, updated to cover a more recent time period, this population‐based study characterized patients from Stockholm County, Sweden, with incident HF (patients with a first HF diagnosis between 1 January 2015 and 31 December 2019) or prevalent HF (patients with a first HF diagnosis before 1 January 2020). Patients with incident HF had LVEF measured by echocardiography within ±90 days of their first HF diagnosis, and patients with prevalent HF within 5 years prior to the index date. The 13 375 patients with prevalent HF (39.2% women, mean age 73.9 years) had multiple comorbidities (cardiovascular diseases, chronic kidney disease, diabetes, and cancer). These were already highly prevalent at the time of the first HF diagnosis in the 8042 patients with incident HF (40.5% women, mean age 72.3 years). Patients with incident HFpEF received less specialist HF care at outpatient secondary care facilities following their first HF diagnosis than those with incident HFrEF. Patients with HFrEF had higher risks of complications and exerted a higher burden, in terms of care for and costs of HF, on the healthcare system.
Conclusions
This study of contemporary patients with incident HF demonstrates that those with HFpEF and HFrEF differ considerably in terms of clinical presentation, prognosis, and care, highlighting a potential to improve HF outcomes.
Keywords: Heart failure, Left ventricular ejection fraction, HFpEF, HFmrEF, HFrEF
Introduction
The global prevalence of heart failure is expected to increase beyond the 64 million people that it currently affects. 1 Subsequently, its burden to healthcare systems is also expected to increase, exceeding the proportion (1–2%) of the annual healthcare budget that is currently allocated towards the condition. 2 , 3
The characteristics and prognosis of patients with heart failure change over time with the development of new treatments and updated guidelines. 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 Therefore, there is a continuous need to characterize the population with heart failure, to identify evolving risks and weaknesses in the management of the condition. However, such information is not easily accessible, often lacking granularity, such as information on left ventricular ejection fraction (LVEF), N‐terminal pro‐brain natriuretic peptide (NT‐proBNP), or blood pressures, if based on administrative registries, 13 , 14 or lacking external validity if based on highly selected cohorts. 15 , 16 We recently made a contribution to addressing this need with the CardioRenal and Metabolic disease Heart Failure (CaReMe HF) study, which characterized a contemporary, multinational population with prevalent heart failure.
Using Swedish data from the CaReMe HF study, composed of medical records from routine practice across all levels of healthcare (primary, secondary, and tertiary care) and updated to cover a more recent time period, we now aimed to compile a uniquely comprehensive set of information characterizing a contemporary population with subtypes of incident or prevalent heart failure based on reduced (HFrEF), mildly reduced (HFmrEF), or preserved (HFpEF) LVEF. Additionally, we aimed to describe how adverse outcomes, patterns of healthcare utilization, pharmacological heart failure treatments, and healthcare costs vary between each subtype of incident heart failure, to explore if there are groups of patients with heart failure that are disproportionally affected by the condition or underserved in current healthcare settings.
Methods
Data sources
The study was set in Stockholm County, Sweden. Sweden has a comprehensive, nationwide public healthcare system that each resident can access with a minor co‐payment for healthcare visits, hospitalizations, and medications. 17 Residents have a unique personal identification number (person‐ID), 18 which is mandatory for all administrative purposes, including any contact with the healthcare system and filling of drug prescriptions, thus providing a basis for complete population‐wide medical history.
Data were extracted from the CELOSIA database, 2 the details of which can be viewed in Supporting Information, Methods S1 . Briefly, it includes data from the Swedish Prescribed Drug Register, the Cause of Death Register, and the National Patient Register and from several data sources with extensive regional coverage, including the electronic health records of Stockholm County. A complete list of the Anatomical Therapeutic Chemical (ATC) codes, 19 International Classification of Diseases (ICD)‐10 codes, 20 procedure codes, 21 and laboratory values used in the CELOSIA database is available in Supporting Information, Table S1 .
Individual patient‐level data from the national registers and regional data sources were linked using the person‐ID. The linked pseudonymized database was managed separately by Sence Research AB, Uppsala, Sweden. This study conforms with the principles outlined in the Declaration of Helsinki and was approved by the Swedish Ethical Review Authority (Approval Numbers 2020‐03850 and 2020‐06716). Given the nature of the study, informed consent was not required.
Study sample
This population‐based study assessed two cohorts aged ≥18 years from Stockholm County, Sweden: one cohort with incident heart failure and one with prevalent heart failure. Patients could be included in both cohorts.
Incident cohort
The cohort with incident heart failure included all patients who had their first diagnosis of heart failure between 1 January 2015 and 31 December 2019 and who could be followed for at least 1 day. The date of discharge for the first diagnosed heart failure event was considered the date of the first heart failure diagnosis. Patients needed to have had their LVEF measured at least once by echocardiography within ±90 days of their first heart failure diagnosis and had to be residing in Stockholm County uninterrupted for the 5 years prior to and at the time of their diagnosis.
Prevalent cohort
The cohort with prevalent heart failure included all patients who had their first diagnosis of heart failure before 1 January 2020 and who were still alive on that index date. Patients needed to have had their LVEF measured at least once by echocardiography within the 5 years prior to the index date. However, that measurement had to have been performed after their first diagnosis of heart failure. Patients also needed to have been residing in Stockholm County uninterrupted during that same 5 year period and on the index date.
Patients with incident or prevalent heart failure were identified using ICD‐10 codes (I50, I11.0, I13.0, and I13.2) recorded in any position in primary care, outpatient secondary care, and/or inpatient care. The LVEF value closest to the first diagnosis of heart failure was used to define subtypes of incident heart failure. The most recent LVEF value within the 5 years prior to 1 January 2020 was used to define subtypes of prevalent heart failure. Values of ≤40%, 41–49%, and ≥50% represented HFrEF, HFmrEF, and HFpEF, respectively.
Characterization of patients with incident or prevalent heart failure
The LVEF values used to define subtypes of heart failure were searched for in electronic health records using different keywords. Values were extracted in both structured numerical format and unstructured free‐text format. Measurements in unstructured format were manually curated. Numerical measurements were favoured over categorical definitions, and calculated measurements over eyeballing. Exact numeric or semi‐quantitative LVEF values were extracted where available.
Socio‐economic status was based on the most recent data available in the Mosaic system (a geodemographic classification of households) within the 10 years prior to or on the day of the first diagnosis of heart failure or the index date in patients with incident heart failure or prevalent heart failure, respectively. Diagnoses in any position in primary care, outpatient secondary care, and/or inpatient care prior to and on the day of the first diagnosis of heart failure or the index date were used to identify prevalent comorbidities in the respective cohorts. Diagnoses of cancer were only searched for in the 10 years up until or on the day of the first diagnosis of heart failure or the index date. A complete list of the comorbidities of interest and the ICD‐10 codes used to identify them is available in Supporting Information, Table S2 .
In addition to using ICD‐10 codes, laboratory measurements available in electronic health records were used to identify the prevalence of chronic kidney disease in each cohort (detailed in Supporting Information, Methods S2 ). 18 Similarly, additional methods were used to define the type of diabetes (1 or 2) in patients with the condition (detailed in Supporting Information, Methods S3 ). Laboratory measurements were also used to establish clinical profiles of each cohort. The most recent, single measurement within the year prior to or on the day of the first heart failure diagnosis or the index date was used, with exception for potassium values, where the maximum value within the 5 years prior to or on the day of the first heart failure diagnosis or the index date was used. Medication use in patients with incident heart failure was based on data available 1 year prior to and on the day of their first heart failure diagnosis. Medication use in patients with prevalent heart failure was based on data available in the year prior to the index date. Medications of interest and the ATC codes used to identify them are listed in Supporting Information, Table S3 .
Adverse outcomes during 1 year follow‐up in patients with incident heart failure
Follow‐up could be performed up until the end of December 2020. Adverse outcomes (all‐cause, cardiovascular, and renal mortality; and hospitalizations with a primary diagnosis of heart failure, chronic kidney disease, myocardial infarction, stroke, or peripheral artery disease) were monitored for the year following discharge from the first diagnosis of heart failure in patients with incident heart failure. Only primary diagnoses in inpatient care or the primary cause of death were used. The ICD‐10 codes used to identify cause of death are listed in Supporting Information, Table S4 . The ICD‐10 codes used to identify non‐fatal adverse outcomes of interest during follow‐up were the same as those used to identify prevalent comorbidities. However, ICD‐10 codes for myocardial infarction and stroke were limited to I21–I22 and I60–I63, respectively.
Adverse outcomes were presented as the 1 year event rate per 100 person‐years. The Poison regression model fit was used to estimate the average 1 year event rate adjusted for age, sex, and LVEF. The model was fitted with a spline with three knots, one for age and two for interactions (age, sex, and LVEF; and sex and LVEF). All analyses were performed using R (Version 3.6.0).
Patterns of healthcare in patients with incident heart failure
Outpatient secondary care was defined as any recorded visit at a hospital or other secondary care facility. Primary care was defined as all other outpatient care. Frequencies of visits to outpatient secondary care facilities and primary care facilities for treatment of heart failure (indicated by visits with an ICD‐10 code, I50, diagnosed in any position) were tallied for patients with incident heart failure during each year with complete follow‐up after the first diagnosis of heart failure. Each patient's healthcare utilization was followed for a maximum period of 5 years.
Healthcare costs for patients with incident heart failure
The average yearly costs of primary care, outpatient secondary care, and inpatient care visits by patients with subtypes of incident heart failure over the 2 years immediately following, and inclusive of, their first diagnosis of heart failure were calculated separately for each non‐mortality outcome of interest and where outcomes could be combined [cardiovascular and renal disease (any diagnosis of heart failure, chronic kidney disease, myocardial infarction, stroke, and/or peripheral artery disease), cardiorenal disease (heart failure and/or chronic kidney disease), and atherosclerotic cardiovascular disease (myocardial infarction, stroke, and/or peripheral artery disease)]. Diagnoses in any position were used. Costs were expressed in US dollars (USD), using an exchange rate of 1 SEK to 0.0968 USD. The average yearly costs were calculated using the following formula:
Results
Of 29 938 patients with incident heart failure and 41 357 patients with prevalent heart failure, 8042 (27%) and 13 375 (32%), respectively, had available LVEF data and, thus, could be included in the analyses. The patients with available LVEF data were somewhat younger, mostly men, had higher NT‐proBNP levels, and more often had prevalent causes of heart failure including ischaemic heart disease, myocardial infarction, and valvular disease (Supporting Information, Table S5 ). The characteristics of the contemporary cohort with prevalent heart failure are presented in Supporting Information, Table S6 and are described in Supporting Information, Results S1 . The results herein focus on the cohort with incident heart failure.
Characteristics of patients with incident heart failure
Comorbidities
The average age of the 8042 patients with incident heart failure was 72 years, and 41% were women (Table 1 ). Systolic blood pressure was on average 138 mmHg, NT‐proBNP was 5000 ng/L, and one‐third of patients had laboratory‐confirmed chronic kidney disease [estimated glomerular filtration rate (eGFR) was on average 60 mL/min/1.73 m2, and urine albumin–creatinine ratio 45 g/mol]. However, only one‐sixth of patients had a diagnosis of chronic kidney disease. Seventy per cent of the cohort were diagnosed with hypertension, two out of five patients had ischaemic heart disease, nearly half had atrial fibrillation, and a quarter had diabetes. Substantial proportions of the cohort had records of filling prescriptions for renin‐angiotensin system inhibitors (RASis; 53%), beta‐blockers (55%), statins (37%), and loop diuretics (27%) within the year prior to their first diagnosis of heart failure.
Table 1.
Characteristics of patients with incident heart failure at time of first diagnosis of heart failure
| All patients | HFrEF | HFmrEF | HFpEF | |
|---|---|---|---|---|
| N (%) | 8042 | 3407 (42.4%) | 1737 (21.6%) | 2898 (36.0%) |
| Age, years, mean (SD) | 72.3 (13.3) | 70.2 (13.7) | 71.4 (13.3) | 75.2 (12.3) |
| Female, n (%) | 3261 (40.5%) | 1162 (34.1%) | 651 (37.5%) | 1448 (50.0%) |
| Level of healthcare for first heart failure diagnosis and diagnosis position, N (%) | ||||
| Primary care | 617 (7.7%) | 147 (4.3%) | 121 (7.0%) | 349 (12.0%) |
| Outpatient care, 1st position | 1281 (15.9%) | 459 (13.5%) | 292 (16.8%) | 530 (18.3%) |
| Outpatient care, not 1st | 827 (10.3%) | 304 (8.9%) | 263 (15.1%) | 260 (9.0%) |
| Inpatient care, 1st position | 2370 (29.5%) | 1136 (33.3%) | 388 (22.3%) | 846 (29.2%) |
| Inpatient care, not 1st | 2947 (36.6%) | 1361 (39.9%) | 673 (38.7%) | 913 (31.5%) |
| Socio‐economic status, N (%) | ||||
| High | 1620 (20.2%) | 675 (19.8%) | 361 (20.8%) | 584 (20.2%) |
| Middle | 3320 (41.3%) | 1396 (41.0%) | 729 (42.0%) | 1195 (41.3%) |
| Low | 3093 (38.5%) | 1330 (39.1%) | 646 (37.2%) | 1117 (38.6%) |
| Missing | 9 (0.1%) | 6 (0.2%) | ≤5 | ≤5 |
| Comorbidities, N (%) | ||||
| Ischaemic heart disease | 3167 (39.4%) | 1346 (39.5%) | 812 (46.7%) | 1009 (34.8%) |
|
ASCVD |
4247 (52.8%) | 1753 (51.5%) | 997 (57.4%) | 1497 (51.7%) |
| Myocardial infarction | 2407 (29.9%) | 1119 (32.8%) | 662 (38.1%) | 626 (21.6%) |
| Stroke | 1305 (16.2%) | 499 (14.6%) | 275 (15.8%) | 531 (18.3%) |
| Peripheral artery disease | 1003 (12.5%) | 389 (11.4%) | 179 (10.3%) | 435 (15.0%) |
| Atrial fibrillation | 3715 (46.2%) | 1468 (43.1%) | 761 (43.8%) | 1486 (51.3%) |
| Hypertension | 5695 (70.8%) | 2128 (62.5%) | 1234 (71.0%) | 2333 (80.5%) |
| Valvular disease | 1466 (18.2%) | 446 (13.1%) | 266 (15.3%) | 754 (26.0%) |
| COPD | 1088 (13.5%) | 374 (11.0%) | 231 (13.3%) | 483 (16.7%) |
| Pulmonary hypertension | 139 (1.7%) | 17 (0.5%) | 25 (1.4%) | 97 (3.3%) |
| Myocarditis | 497 (6.2%) | 171 (5.0%) | 74 (4.3%) | 252 (8.7%) |
| Congenital heart disease | 107 (1.3%) | 41 (1.2%) | 27 (1.6%) | 39 (1.3%) |
| Sarcoidosis | 40 (0.5%) | 17 (0.5%) | 9 (0.5%) | 14 (0.5%) |
| Amyloidosis | 30 (0.4%) | 7 (0.2%) | ≤5 | 18 (0.6%) |
| Type 1 diabetes | 126 (1.6%) | 48 (1.4%) | 30 (1.7%) | 48 (1.7%) |
| Type 2 diabetes | 1947 (24.2%) | 792 (23.2%) | 389 (22.4%) | 766 (26.4%) |
| Any cancer | 2049 (25.5%) | 776 (22.8%) | 449 (25.8%) | 824 (28.4%) |
| CKD diagnosis or laboratory, N (%) | ||||
| Diagnosis | 1352 (16.8%) | 489 (14.4%) | 259 (14.9%) | 604 (20.8%) |
| KDIGO‐confirmed CKD by eGFR | 2683 (33.4%) | 918 (26.9%) | 526 (30.3%) | 1239 (42.8%) |
| CKD diagnosis or KDIGO‐CKD | 3035 (37.7%) | 1067 (31.3%) | 590 (34.0%) | 1378 (47.6%) |
| Laboratory measurements, mean (SD) | ||||
| NT‐proBNP, ng/L | 4996.2 (6966.8) | 6639.9 (8072.0) | 4385.9 (6637.3) | 3597.3 (5352.6) |
| SBP, mmHg | 138.3 (25.1) | 136.0 (24.6) | 138.8 (25.0) | 140.8 (25.4) |
| Body mass index, kg/m2 | 27.2 (5.7) | 26.9 (5.6) | 27.2 (5.7) | 27.6 (5.8) |
| HbA1c, mmol/mol | 46.3 (14.5) | 47.0 (15.7) | 45.4 (14.0) | 46.3 (13.6) |
| eGFR, mL/min/1.73 m2 | 59.9 (20.6) | 61.5 (20.4) | 61.9 (20.6) | 56.9 (20.4) |
| UACR, g/mol | 45.4 (110.1) | 47.6 (106.6) | 37.5 (101.7) | 47.7 (116.9) |
| C‐reactive protein, mg/L | 25.6 (49.3) | 25.4 (47.8) | 24.3 (48.5) | 26.6 (51.4) |
| Haemoglobin, g/L | 130.8 (20.4) | 133.8 (20.1) | 131.6 (20.1) | 126.6 (20.2) |
| LDL cholesterol, mmol/L | 2.5 (1.1) | 2.6 (1.0) | 2.5 (1.1) | 2.5 (1.1) |
| Total cholesterol, mmol/L | 4.5 (1.2) | 4.5 (1.2) | 4.4 (1.3) | 4.5 (1.3) |
| Potassium, mmol/L | 4.1 (0.5) | 4.2 (0.5) | 4.1 (0.5) | 4.1 (0.5) |
| Maximum potassium, mmol/L | 4.6 (0.7) | 4.6 (0.6) | 4.6 (0.6) | 4.7 (0.7) |
| Medication use, N (%) | ||||
| RASis | 4293 (53.4%) | 1556 (45.7%) | 962 (55.4%) | 1775 (61.2%) |
| Beta‐blockers | 4421 (55.0%) | 1503 (44.1%) | 1006 (57.9%) | 1912 (66.0%) |
| MRAs | 445 (5.5%) | 178 (5.2%) | 66 (3.8%) | 201 (6.9%) |
| Statins | 2990 (37.2%) | 1069 (31.4%) | 668 (38.5%) | 1253 (43.2%) |
| SGLT2is | 51 (0.6%) | 23 (0.7%) | 9 (0.5%) | 19 (0.7%) |
| Loop diuretics | 2151 (26.7%) | 685 (20.1%) | 418 (24.1%) | 1048 (36.2%) |
| Digitalis | 298 (3.7%) | 102 (3.0%) | 51 (2.9%) | 145 (5.0%) |
ASCVD, atherosclerotic cardiovascular disease; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; HbA1c, haemoglobin A1c; KDIGO, Kidney Disease: Improving Global Outcomes; LDL, low‐density lipoprotein; MRAs, mineralocorticoid receptor antagonists; NT‐proBNP, N‐terminal pro‐brain natriuretic peptide; RASis, renin‐angiotensin system inhibitors; SBP, systolic blood pressure; SD, standard deviation; SGLT2is, sodium‐glucose cotransporter‐2 inhibitors; UACR, urinary albumin–creatinine ratio.
Characteristics at the time of the first heart failure diagnosis in all patients with incident heart failure and available left ventricular ejection fraction data, and that cohort stratified by subtypes of heart failure based on reduced (HFrEF), mildly reduced (HFmrEF), or preserved (HFpEF) left ventricular ejection fraction.
Type of heart failure
Of the 8042 patients with incident heart failure, 42.4% had HFrEF, 21.6% had HFmrEF, and 36.0% had HFpEF. Patients with HFpEF were 5 years older (75.2) than those with HFrEF (70.2 years), and half of all patients with HFpEF were women, as compared with 34% of those with HFrEF. Socio‐economic status was similar across the LVEF‐based groups. A lower proportion of patients with HFpEF had prevalent ischaemic heart disease than those with HFrEF. Patients with HFpEF also had, on average, lower NT‐proBNP and slightly higher systolic blood pressure. Diagnoses of hypertension, pulmonary hypertension, and atrial fibrillation were more common in HFpEF than in HFrEF. Valvular disease was twice as common in HFpEF as in HFrEF. Myocarditis, chronic obstructive pulmonary disease, and cancer were all more common in HFpEF than in HFrEF. Treatment with RASis, beta‐blockers, statins, and loop diuretics was more common in patients with HFpEF than in those with HFrEF.
Healthcare level at first diagnosis of heart failure
More patients with incident heart failure received their first diagnosis of heart failure in inpatient care (66%) than in outpatient secondary care (26%) or primary care (8%). Of patients who received their first diagnosis of heart failure in inpatient care, most had HFrEF (47%), while the majority of those diagnosed in primary care had HFpEF (57%). Most patients diagnosed in outpatient secondary care had either HFpEF (37%) or HFrEF (36%).
Healthcare and treatment during follow‐up in patients with incident heart failure
In patients with HFrEF, HFmrEF, or HFpEF, the proportion who received specialist care for heart failure at outpatient secondary care facilities was highest in the year immediately following the first heart failure diagnosis (Figure 1 ). During that first year, a smaller proportion of patients with HFpEF received outpatient secondary care than those with HFmrEF or HFrEF. While utilization of outpatient secondary care decreased thereafter for all LVEF‐based groups, the reduction was far more pronounced in patients with HFpEF. Indeed, only 26% of patients with HFpEF still received outpatient secondary care in the second year, compared with 34% and 56% of patients with HFmrEF and HFrEF, respectively. The proportion of patients who received primary care was also highest in the year immediately following the first heart failure diagnosis. However, the decrease in primary care during the years following was far less substantial than that in outpatient secondary care. Additionally, the primary care visits followed a relatively similar pattern between the LVEF‐based groups.
Figure 1.
The frequency of visits to (A) outpatient secondary care facilities for specialist heart failure (HF) treatment or (B) primary care facilities, each year during the first 5 years following the first diagnosis of HF in patients with incident HF. Patients with incident HF are stratified by subtypes of HF based on reduced (HFrEF), mildly reduced (HFmrEF), or preserved (HFpEF) left ventricular ejection fraction. [Correction added on 10 May 2024, after first online publication: Figure 1 image has been corrected in this version.]
High proportions of patients with HFrEF received specialist care for heart failure at outpatient secondary care facilities during the 2 years following their first diagnosis, regardless of the level of care in which that first diagnosis was made (Supporting Information, Figure S1 ). The proportions of patients who received outpatient secondary care over the same 2 year period were lower in those with HFmrEF or HFpEF, especially in those who received their first heart failure diagnosis in primary care. High proportions of those patients continued to receive heart failure treatment in primary care settings. In that context, patients who received their first diagnosis of heart failure in outpatient secondary care more often received subsequent, outpatient secondary care. However, high proportions of patients with HFpEF still received treatment in primary care even if their first diagnosis of heart failure was made in outpatient secondary care.
In an analysis of patients who had not filled a prescription for a pharmacological heart failure treatment [RASis (including angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, and angiotensin receptor/neprilysin inhibitors), beta‐blockers, mineralocorticoid receptor antagonists (MRAs), sodium‐glucose cotransporter‐2 inhibitors (SGLT2is), and/or loop diuretics] prior to their first heart failure diagnosis, prescriptions for RASis and beta‐blockers were filled more often than any other drug classes of interest in the 2 years thereafter (Supporting Information, Figure S2 ). Larger proportions of the patients who received their first diagnosis of heart failure in inpatient care received drug treatment during follow‐up than those whose first diagnosis was in outpatient secondary or primary care. A lesser proportion of the patients with HFpEF began any of the drug treatments of interest within the 2 years following their first diagnosis of heart failure than those with HFrEF. Initiation of SGLT2i treatment was low during follow‐up compared with other drug classes in all LVEF‐based subtypes of incident heart failure.
Adverse outcomes in patients with incident heart failure
The rate of adverse outcomes increased with age (Figure 2 ). When evaluating specific causes for adverse outcomes, a substantial proportion of adverse events throughout the year following the first diagnosis of heart failure was composed of hospitalizations due to heart failure or atherosclerotic cardiovascular disease (Supporting Information, Table S7 ). Death rates due to heart failure, chronic kidney disease, or atherosclerotic cardiovascular disease were low (Supporting Information, Figure S3 ), and there were no notable differences in adverse outcomes between men and women (Supporting Information, Figure S4 ). When only hospitalizations and deaths due to heart failure were considered, the prognosis for patients with HFrEF was worse than those with HFmrEF or HFpEF (Figure 3 ).
Figure 2.
Adverse outcomes throughout the year following the first diagnosis of heart failure (HF) in patients with incident HF. Patients with incident HF are stratified by sex and subtypes of HF based on reduced (HFrEF), mildly reduced (HFmrEF), or preserved (HFpEF) left ventricular ejection fraction. Adverse outcomes are presented as the 1 year event rate per 100 person‐years and are adjusted for age.
Figure 3.
Kaplan–Meier curves presenting the proportion of patients with incident heart failure (HF) impacted by adverse outcomes throughout the year following their first diagnosis of HF. Adverse outcomes included all‐cause hospitalizations and deaths, and all hospitalizations and deaths due to HF, chronic kidney disease (CKD), and atherosclerotic cardiovascular disease (ASCVD). Patients with incident HF are stratified by age groups and subtypes of HF based on reduced (HFrEF), mildly reduced (HFmrEF), or preserved (HFpEF) left ventricular ejection fraction.
Healthcare costs in patients with incident heart failure
The average cumulative cost of all healthcare during the year following the first diagnosis of heart failure in patients with incident heart failure exceeded 15 000 USD per patient (Supporting Information, Figure S5 ). At least half of these healthcare costs were due to care for heart failure. When all healthcare costs were considered, there was not a clear pattern, indicating that one LVEF‐based group was more costly than any other groups (Supporting Information, Figure S6 ). However, when only costs due to healthcare for heart failure were considered, the proportion of costs that could be attributed to care for patients with HFrEF was higher than that attributable to those with HFmrEF or HFpEF, across all age groups (Supporting Information, Figure S7 ). Further, the proportion of costs stemming from care for heart failure that could be attributed to patients (of any LVEF‐based group) who had their first diagnosis of heart failure in inpatient care was higher than that attributable to those whose first diagnosis was in outpatient secondary or primary care (Supporting Information, Figure S8 ).
Discussion
Using comprehensive data from routine clinical practice across all levels of healthcare in Stockholm County, Sweden, this study provides a unique description of the demographics, prevalent comorbidities, clinical profiles, and treatments of contemporary patients with subtypes of incident or prevalent heart failure based on HFrEF, HFmrEF, and HFpEF. This study also provides novel insights into the healthcare journey of patients with incident heart failure following their first diagnosis of heart failure, demonstrating that patients with HFpEF receive less follow‐up specialist care for heart failure in outpatient secondary care settings than those with HFrEF and that a large proportion of patients with HFpEF or HFmrEF, who are first diagnosed with heart failure in primary care, will more often continue to receive their subsequent healthcare in the primary care setting. Further, patients with HFpEF receive less heart failure drug treatments following their first diagnosis of heart failure than those with HFrEF. However, patients with HFrEF are at a higher risk of complications following their first diagnosis of heart failure than those with HFpEF, more often requiring follow‐up care for heart failure. Subsequently, more money is being spent on patients with HFrEF for heart failure healthcare than those with HFmrEF or HFpEF.
A large proportion of contemporary patients in the cohort with prevalent heart failure has multiple comorbidities (hypertension, atrial fibrillation, ischaemic heart disease, and myocardial infarction, adding to the burden of other prevalent comorbidities including chronic kidney disease, type 2 diabetes, and cancer). Although these comorbidities affected a slightly smaller proportion of the patients in the cohort with incident heart failure, prevalence was already evident at the time of the first heart failure diagnosis, when patients were on average 72 years old. Consistent with the CaReMe HF study, 2 this research indicates that 50% of the population with incident heart failure might have stage III–V chronic kidney disease. However, despite the availability of laboratory measurements indicating that chronic kidney disease had developed, only half of that population with chronic kidney disease had an ICD‐reported diagnosis for the condition.
Preserved and mildly reduced heart failure was more common in both incident and prevalent heart failure. Indeed, 58% of patients with incident heart failure had an LVEF > 40% (HFpEF, 36%; and HFmrEF, 22%) at the time of their first heart failure diagnosis. The proportion with an LVEF > 40% was even higher in patients with prevalent heart failure (71%; HFpEF, 49%; and HFmrEF, 22%) mainly due to an increase in the proportion of patients with an LVEF > 50% (36% in incident heart failure compared with 49% in prevalent heart failure). Given that patients with HFrEF had a much worse prognosis following their first heart failure diagnosis, part of this transition is likely due to high mortality in patients with HFrEF. At the time of the first heart failure diagnosis, atrial fibrillation was predictably more common and ischaemic heart disease less common in patients with HFpEF. The observation that myocarditis was less common in patients with incident HFrEF is novel. Other clinical characteristics were similar between the LVEF‐based groups, except for differences in NT‐proBNP, systolic blood pressure, and eGFR. Similarities between patients with HFpEF and those with HFrEF may be explained by high variability in those traits due to high underlying variance in LVEF. Socio‐economic status and other comorbidities were surprisingly similar between the LVEF‐based groups in both the incident and prevalent cohorts.
Differences in follow‐up and treatment after first heart failure diagnosis
Evaluating patterns of healthcare utilization in patients with incident heart failure, this study highlighted substantial differences in the frequency of care for patients following their first diagnosis of heart failure, which might mirror national and local treatment guidelines for heart failure. Patients with HFpEF received less specialist care for heart failure at secondary outpatient care facilities than those with HFrEF throughout the 5 years following their first heart failure diagnosis. It appears that the level of healthcare at which the first diagnosis of heart failure is made carries significant importance for patients with HFmrEF and that those diagnosed in outpatient secondary care are more likely than those diagnosed in primary care to receive outpatient secondary care for any follow‐up care. While a similar trend was observed in those with HFpEF, it was less overt. Further, lesser proportions of patients with HFpEF receive outpatient secondary care after their first diagnosis, regardless of the level of healthcare at which that first diagnosis was made, and they remain at that care level until their LVEF deteriorates. The differences in the level of follow‐up care might reflect a lack of disease‐modifying treatment options for patients with HFmrEF and HFpEF during the study period. As new treatments and interventions for these LVEF‐based subtypes of heart failure become available, in particular for patients with HFpEF, healthcare utilization patterns might change.
The differences in care between patients with HFpEF and those with HFrEF extend to pharmacological treatments for heart failure following the first diagnosis of heart failure. Indeed, in line with contemporary guidelines for the pharmacological treatment of heart failure of this cohort with incident heart failure, patients with HFpEF received less pharmacological heart failure treatment. 22 Higher proportions of the patients with HFrEF were treated with RASis, MRAs, and/or loop diuretics than those with HFpEF, and only a very low proportion was treated with SGLT2is. Given that SGLT2is are a very recent addition to guidelines for the pharmacological treatment of heart failure, 6 , 7 this was to be expected in this cohort. Collectively, these findings highlight a potential to improve outcomes, addressing the condition before it deteriorates and, subsequently, easing the burden that heart failure places on healthcare systems.
Prognosis and healthcare utilization by left ventricular ejection fraction
Patients with incident heart failure and HFpEF or HFmrEF were generally less at risk of severe complications than those with HFrEF. Indeed, patients with HFrEF had higher rates of hospitalization for heart failure following their first heart failure diagnosis, the most common adverse outcome during follow‐up. This burden of care for patients with HFrEF translated to a financial burden, where a large proportion of costs for heart failure healthcare following the first diagnosis of heart failure could be attributed to subsequent hospitalizations for heart failure in patients with HFrEF.
Consistent with previous research involving patients with prevalent heart failure and HFrEF, HFmrEF, or HFpEF, 23 this study further demonstrated that patients with prevalent heart failure and HFmrEF have characteristics that are similar to those with HFpEF or HFrEF. Similar to that study's cohort with prevalent heart failure, this present study indicates that the prognosis of patients with incident heart failure and HFmrEF is more like that of patients with HFpEF than those with HFrEF. Additionally, patterns of healthcare utilization in patients with HFmrEF more resembled that in patients with HFpEF than those with HFrEF, possibly reflecting a similar historical limitation in the availability of pharmacological heart failure treatments for patients with HFmrEF and those with HFpEF. The similarities in characteristics, healthcare journey, and prognosis between patients with HFmrEF and those with HFrEF or HFpEF only further emphasize the need to question if the classification, HFmrEF, has any clinical value or if it is of academic interest only. 24
Strengths and limitations
This study had access to health registries and electronic health records detailing most incidences of healthcare and medication use by residents in one of Sweden's largest regions, representing ~23% of the national population. In Sweden, each resident has a unique person‐ID consistently that is used across all levels of public and private healthcare. This identification number makes it possible to extract complete data for all healthcare use, dispensed drugs, and death from nationwide registries with full coverage. Moreover, the Swedish Cause of Death Register has complete coverage of all Swedish residents and is based on mandatory death reporting without exceptions. Despite its strength in its capacity to identify patients relevant to the study aims, to characterize the subsequent study population, and to document a patient's journey through healthcare, several limitations must be acknowledged.
Only 27% of the patients from the cohort with incident heart failure and 32% of the patients from the cohort with prevalent heart failure had available LVEF data, limiting the identification of HFrEF, HFmrEF, and HFpEF across the study sample and the subsequent analyses of the LVEF‐based groups by potential introduction of selection bias. Low extraction rates of LVEF data were mainly due to the methods used in this study. While LVEF values can be detected in different modules in the electronic health records based on several keywords, a large proportion of LVEF values in Stockholm County was recorded under a specific keyword that was not included in the search protocol. Indeed, inclusion of this keyword in future studies of Stockholm County's electronic health records would capture a larger proportion of the available LVEF values.
While the methods used in this study limited its capacity to extract available LVEF values, low availability of LVEF data is also a possible reflection of low access to and a long waiting list for LVEF evaluation by echocardiography in the Stockholm area, where those with elevated NT‐proBNP levels, hospital admissions, and referrals by cardiologists are prioritized. Although large samples could still be formed, it is unknown if the characteristics of the LVEF‐based groups would change by the inclusion of the unanalysed patients. Further, heart failure often goes undetected or undiagnosed for a period of time. Therefore, the proportions of patients with incident or prevalent heart failure who could not be included in the analyses may be larger.
Given that LVEF measurements could have been performed up to 5 years prior to the index date in patients with prevalent heart failure, it must be acknowledged that each LVEF‐based group may include patients whose LVEF could have improved or worsened beyond the thresholds defining their heart failure subtype between the time that the classifying LVEF measurement was taken and other measurements describing the cohort characteristics were performed.
This study did not assess if patients from a specific socio‐economic background or ethnicity were disproportionally affected by heart failure or underserved in healthcare. Additionally, the study population was derived from a single county in Sweden. While a large proportion of the Swedish population resides within this county and, thus, the study's findings may be generalized to the wider Swedish population, it is unknown whether the finding can be generalized to most other nations where the healthcare systems can differ considerably.
Conclusions
In this study of contemporary patients in a real‐world clinical setting, the characteristics of patients with incident heart failure differed notably between those with HFpEF and those with HFrEF. Patients with incident HFpEF received less heart failure care in outpatient secondary care settings, and more in primary care, than those with incident HFrEF. Patients with HFrEF had higher risks of complications and exerted a higher burden on the healthcare system, in terms of care for and costs of heart failure, compared with the other phenotypes. These findings demonstrate important differences in the clinical presentation, prognosis, and care between subtypes of heart failure based on LVEF and highlight a potential to improve patient outcomes and ease the burden of heart failure.
Conflict of interest
J.S. reports stock ownership in Symptoms Europe AB and Anagram Kommunikation AB, outside of the submitted work. J.Ä. has received payment/honoraria from AstraZeneca and Novartis and has participated on a data safety monitoring/advisory board for AstraZeneca, Astella, and Boehringer Ingelheim. S.K. has participated as a consultant on advisory board committees and educational activities with AstraZeneca and Novo Nordisk. J.B. and K.E. are employees of AstraZeneca. S.G. is an employee of Sence Research AB (an independent company in epidemiology and outcomes research) that received funding from AstraZeneca for statistical analysis within this research project. T.C. is an employee of and stockowner of Sence Research AB. M.K.S. has participated as a consultant on advisory board committees and educational activities with Amgen, AstraZeneca, Boehringer Ingelheim, Novo Nordisk, and GSK. A.N. has participated as a consultant on advisory board committees and educational activities with AstraZeneca, Novo Nordisk, Boehringer Ingelheim, and Eli Lilly.
Funding
The study was sponsored by AstraZeneca Mölndal.
Supporting information
Methods S1. The CELOSIA database.
Methods S2. Identifying chronic kidney disease.
Methods S3. Defining the type of diabetes mellitus.
Results S1. Characteristics of patients with prevalent heart failure.
Table S1. Inclusion criteria for the CELOSIA database from which the study sample was extracted.
Table S2. International classification of diseases (ICD)‐10 codes used to identify prevalent comorbidities of interest in each cohort.
Table S3. Anatomical Therapeutic Chemical (ATC) codes and procedure codes used to identify medication use of interest in each cohort.
Table S4. International classification of diseases (ICD)‐10 codes used to identify causes of death in each cohort.
Table S5. Characteristics of patients with incident or prevalent heart failure stratified by availability of left ventricular ejection fraction data.
Table S6. Characteristics of patients with prevalent heart failure as of the index date.
Table S7. Adverse outcomes throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S1. The cumulative proportions of patients with incident heart failure that received healthcare for heart failure during the two years following their first diagnosis.
Figure S2. The cumulative proportions of patients with incident heart failure who received pharmacological heart failure treatments for the first time following their first diagnosis of heart failure.
Figure S3. Adverse outcomes throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S4. Adverse outcomes throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S5. The average per patient cumulative cost of all healthcare and healthcare for heart failure throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S6. The average per patient cumulative cost of all healthcare throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S7. The average per patient cumulative cost of healthcare for heart failure throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S8. The average per patient cumulative cost of all healthcare and healthcare for heart failure, at each level of health care, throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Acknowledgements
The authors thank Jordan Loader, PhD, of Sence Research AB, Uppsala, Sweden, for providing medical writing support, which was funded by AstraZeneca, Stockholm, Sweden, in accordance with Good Publication Practice (GPP3) guidelines (http://www.ismpp.org/gpp3). The authors also thank Susanna Jerström from AstraZeneca for valuable comments on the manuscript.
Sundström, J. , Ärnlöv, J. , Karayiannides, S. , Bodegard, J. , Ersmark, K. , Gustafsson, S. , Cars, T. , Svensson, M. K. , and Norhammar, A. (2024) Heart failure outcomes by left ventricular ejection fraction in a contemporary region‐wide patient cohort. ESC Heart Failure, 11: 1377–1388. 10.1002/ehf2.14685.
References
- 1. Savarese G, Lund LH. Global public health burden of heart failure. Card Fail Rev 2017;3:7‐11. doi: 10.15420/cfr.2016:25:2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Norhammar A, Bodegard J, Vanderheyden M, Tangri N, Karasik A, Maggioni AP, et al. Prevalence, outcomes and costs of a contemporary, multinational population with heart failure. Heart 2023;109:548‐556. doi: 10.1136/heartjnl-2022-321702 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Stewart S, Jenkins A, Buchan S, McGuire A, Capewell S, McMurray JJJV. The current cost of heart failure to the National Health Service in the UK. Eur J Heart Fail 2002;4:361‐371. doi: 10.1016/s1388-9842(01)00198-2 [DOI] [PubMed] [Google Scholar]
- 4. Anker SD, Butler J, Filippatos G, Ferreira JP, Bocchi E, Böhm M, et al. Empagliflozin in heart failure with a preserved ejection fraction. N Engl J Med 2021;385:1451‐1461. doi: 10.1056/NEJMoa2107038 [DOI] [PubMed] [Google Scholar]
- 5. Conrad N, Judge A, Tran J, Mohseni H, Hedgecott D, Crespillo AP, et al. Temporal trends and patterns in heart failure incidence: A population‐based study of 4 million individuals. Lancet 2018;391:572‐580. doi: 10.1016/S0140-6736(17)32520-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, et al. 2022 AHA/ACC/HFSA guideline for the management of heart failure: A report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation 2022;145:e895‐e1032. doi: 10.1016/j.jacc.2021.12.012 [DOI] [PubMed] [Google Scholar]
- 7. McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: Developed by the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC) with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J 2021;42:3599‐3726. doi: 10.1002/ejhf.2333 [DOI] [PubMed] [Google Scholar]
- 8. McDonald M, Virani S, Chan M, Ducharme A, Ezekowitz JA, Giannetti N, et al. CCS/CHFS heart failure guidelines update: Defining a new pharmacologic standard of care for heart failure with reduced ejection fraction. Can J Cardiol 2021;37:531‐546. doi: 10.1016/j.cjca.2021.01.017 [DOI] [PubMed] [Google Scholar]
- 9. McMurray JJV, Packer M, Desai AS, Gong J, Lefkowitz MP, Rizkala AR, et al. Angiotensin‐neprilysin inhibition versus enalapril in heart failure. N Engl J Med 2014;371:993‐1004. doi: 10.1056/NEJMoa1409077 [DOI] [PubMed] [Google Scholar]
- 10. McMurray JJV, Solomon SD, Inzucchi SE, et al. Dapagliflozin in patients with heart failure and reduced ejection fraction. N Engl J Med 2019;381:1995‐2008. doi: 10.1056/NEJMoa1911303 [DOI] [PubMed] [Google Scholar]
- 11. Packer M, Anker SD, Butler J, Filippatos G, Pocock SJ, Carson P, et al. Cardiovascular and renal outcomes with empagliflozin in heart failure. N Engl J Med 2020;383:1413‐1424. doi: 10.1056/NEJMoa2022190 [DOI] [PubMed] [Google Scholar]
- 12. Solomon SD, McMurray JJV, Claggett B, et al. Dapagliflozin in heart failure with mildly reduced or preserved ejection fraction. N Engl J Med 2022;387:1089‐1098. doi: 10.1056/NEJMoa2206286 [DOI] [PubMed] [Google Scholar]
- 13. Arulmurugananthavadivel A, Holt A, Parveen S, Lamberts M, Gislason GH, Torp‐Pedersen C, et al. Importance of diagnostic setting in determining mortality in patients with new‐onset heart failure: Temporal trends in Denmark from 1997 to 2017. Eur Heart J Qual Care Clin Outc 2022;8:750‐760. doi: 10.1093/ehjqcco/qcab073 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bottle A, Kim D, Aylin P, Cowie MR, Majeed A, Hayhoe B. Routes to diagnosis of heart failure: Observational study using linked data in England. Heart 2018;104:600‐605. doi: 10.1136/heartjnl-2017-312183 [DOI] [PubMed] [Google Scholar]
- 15. Borlaug BA, Redfield MM. Diastolic and systolic heart failure are distinct phenotypes within the heart failure spectrum. Circulation 2011;123:2006‐2014. doi: 10.1161/CIRCULATIONAHA.110.954388 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Chioncel O, Lainscak M, Seferovic PM, Anker SD, Crespo‐Leiro MG, Harjola VP, et al. Epidemiology and one‐year outcomes in patients with chronic heart failure and preserved, mid‐range and reduced ejection fraction: An analysis of the ESC Heart Failure Long‐Term Registry. Eur J Heart Fail 2017;19:1574‐1585. doi: 10.1002/ejhf.813 [DOI] [PubMed] [Google Scholar]
- 17. Anell A. The public‐private pendulum—Patient choice and equity in Sweden. N Engl J Med 2015;372:1‐4. doi: 10.1056/NEJMp1411430 [DOI] [PubMed] [Google Scholar]
- 18. Ludvigsson JF, Otterblad‐Olausson P, Pettersson BU, Ekbom A. The Swedish personal identity number: Possibilities and pitfalls in healthcare and medical research. Eur J Epidemiol 2009;24:659‐667. doi: 10.1007/s10654-009-9350-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. World Health Organization . WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD Index 2017. Available from: https://www.whocc.no/. Accessed 20 June 2023. [Google Scholar]
- 20. World Health Organization . International Classification of Disease and Related Health Problems. Available from: http://www.who.int/classifications/icd/en/. Accessed 20 June 2023. [Google Scholar]
- 21. Nordic Medico‐Statistical Committee . NOMESCO Classification of Surgical Procedures. 2010. Available from: https://norden.diva‐portal.org/smash/get/diva2:970547/FULLTEXT01.pdf. Accessed 20 June 2023. [Google Scholar]
- 22. Ponikowski P, Voors AA, Anker SD, Bueno H, Cleland JGF, Coats AJS, et al. 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure: The Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC)developed with the special contribution of the Heart Failure Association (HFA) of the ESC. Eur Heart J 2016;37:2129‐2200. doi: 10.1093/eurheartj/ehw128 [DOI] [PubMed] [Google Scholar]
- 23. Savarese G, Stolfo D, Sinagra G, Lund LH. Heart failure with mid‐range or mildly reduced ejection fraction. Nat Rev Cardiol 2022;19:100‐116. doi: 10.1038/s41569-021-00605-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Johansson I, Dahlström U, Edner M, Näsman P, Rydén L, Norhammar A. Type 2 diabetes and heart failure: Characteristics and prognosis in preserved, mid‐range and reduced ventricular function. Diab Vasc Dis Res 2018;15:494‐503. doi: 10.1177/1479164118794619 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Methods S1. The CELOSIA database.
Methods S2. Identifying chronic kidney disease.
Methods S3. Defining the type of diabetes mellitus.
Results S1. Characteristics of patients with prevalent heart failure.
Table S1. Inclusion criteria for the CELOSIA database from which the study sample was extracted.
Table S2. International classification of diseases (ICD)‐10 codes used to identify prevalent comorbidities of interest in each cohort.
Table S3. Anatomical Therapeutic Chemical (ATC) codes and procedure codes used to identify medication use of interest in each cohort.
Table S4. International classification of diseases (ICD)‐10 codes used to identify causes of death in each cohort.
Table S5. Characteristics of patients with incident or prevalent heart failure stratified by availability of left ventricular ejection fraction data.
Table S6. Characteristics of patients with prevalent heart failure as of the index date.
Table S7. Adverse outcomes throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S1. The cumulative proportions of patients with incident heart failure that received healthcare for heart failure during the two years following their first diagnosis.
Figure S2. The cumulative proportions of patients with incident heart failure who received pharmacological heart failure treatments for the first time following their first diagnosis of heart failure.
Figure S3. Adverse outcomes throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S4. Adverse outcomes throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S5. The average per patient cumulative cost of all healthcare and healthcare for heart failure throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S6. The average per patient cumulative cost of all healthcare throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S7. The average per patient cumulative cost of healthcare for heart failure throughout the year following the first diagnosis of heart failure in patients with incident heart failure.
Figure S8. The average per patient cumulative cost of all healthcare and healthcare for heart failure, at each level of health care, throughout the year following the first diagnosis of heart failure in patients with incident heart failure.