Mortality, Outcomes, Costs, and Use of Medicines Following a First Heart Failure Hospitalization

.

H eart failure (HF) is common, and its prevalence is expected to rise with aging populations and improved diagnosis. 1,2HF is associated with an impaired quality of life, poor outcomes, and places a substantial economic burden on health care systems. 3eatment strategies to improve prognosis vary depending on left ventricular ejection fraction (LVEF).For patients with heart failure with preserved ejection fraction (HFpEF), guideline-recommended treatments are currently limited.In contrast, for patients with chronic HF with reduced ejection fraction (HFrEF), initiation of several guideline-directed medical therapies (GDMTs) reduces hospitalizations for heart failure (hHFs) and mortality and is costeffective, with effects beginning to appear within a few days or weeks of initiation. 4,5][8][9][10] Failure to implement GDMT has an adverse effect on outcomes. 11,12OLUTION HF (Utilization of Dapagliflozin and Other Guideline Directed Medical Therapies in Heart Failure Patients: A Multinational Observational Study Based on Secondary Data) is a multinational, observational study that provides insights into the management of patients after discharge from an hHF. 6The present analysis provides a contemporary description of rehospitalization rates, hospital health care costs, use of key therapies, and mortality after a first hHF in Japan, Sweden, the United Kingdom, and the United States (Central Illustration).

METHODS
STUDY DESIGN.EVOLUTION HF 6 is a multinational, observational, longitudinal cohort study that uses data extracted from well-established electronic health records or claims data sources in Japan (hospital sourced), Sweden (hospital sourced and national registries), the United Kingdom (primary care sourced), and the United States (hospital sourced and prescription claims data) (Supplemental Methods, Supplemental Figure 1).
STUDY POPULATIONS AND STUDY PERIODS.Adults were included if they had a first-ever registered inpatient hHF during the study period of 2018 through 2022 (Supplemental Table 1).Patients with new-onset hHF were included to increase the validity of an HF diagnosis and the likelihood that the investigated medicines were initiated for HF treatment (rather than for other indications).To ensure that only patients with new-onset HF were included, patients were excluded if they had any prior HF diagnosis during all the available periods for each database and were required to have records in the 12 months prior to the event.Patients with a prior type 1 diabetes diagnosis were also excluded.HF was defined by the following International Classification of Diseases-10th Revision diagnosis codes in all countries: I50, I11.0, I13.0, and I13.2 (Supplemental Table 2).
COHORTS.Two cohorts were created within each country: one (cohort 1) to study clinical outcomes and hospital health care costs, and another (cohort 2) to describe contemporary use of HF medicines after hHF discharge (Figure 1).For both cohorts, the index date (start of follow-up) was defined as the date of discharge from a first registered inpatient hHF, and follow-up periods were defined as the time from the index date to the data-extraction date, date of death, or 12 months after discharge, whichever came first.
Data were also censored 14 days after the last registered activity (ie, a dispensed medicine) in the database to avoid including patients who had been lost to follow-up.

Bozkurt et al
Outcomes, Costs, and Treatment Following Incident hHF  BASELINE CHARACTERISTICS.For patients in both cohorts, baseline data were extracted for at least 12 months before each study period start date to allow a 12-month "lookback" period from the index hHF admission for all patients (baseline periods).Patient characteristics were described prior to the first hHF admission, and included demographics, comorbidities, measurements of blood pressure, and estimated glomerular filtration rate (Supplemental Tables 2   and 3).Use of medicines was based on at least 1 filled prescription during the year prior to the first hHF admission (Supplemental Table 4).
OUTCOMES.Outcomes for each cohort were assessed from first hHF discharge date during the respective follow-up periods for each country.a Concurrent use of a renin-angiotensin system inhibitor, beta-blocker, mineralocorticoid receptor antagonist, and sodium-glucose cotransporter-2 inhibitor.
hypertensive CKD, dialysis, glomerular diseases, renal tubulointerstitial disease, or other), 13 myocardial infarction, stroke, or peripheral artery disease, and cardiovascular and all-cause mortality during the 12 months after index (Supplemental Table 2).The index hHF was not included when counting outcome events.

H o s p i t a l h e a l t h c a r e c o s t s ( c o h o r t 1 )
. Costs for planned and unplanned inpatient and outpatient hospital visits associated with any diagnosis, separately, of HF, CKD, myocardial infarction, stroke, and peripheral artery disease were cumulatively summarized for the 12 months following the index date for each patient. 1,14,15The index hHF was not included.
For this analysis, multiple diagnoses could be registered for a given hospitalization.For detailed methods, see Supplemental Methods.
Use of HF medicines (cohort 2).HF medicines were defined as those represented by GDMT (ie, those  recommendations for the treatment of HFpEF. 4,5Information on LVEF was not available for the majority of patients included in this analysis.As such, we cannot report on whether prescribing adhered to guidelines, but rather only on which medicines were prescribed and when. STATISTICAL ANALYSIS.Continuous variables were reported using median (IQR).Categorical variables were reported as absolute frequency and percentage.
All analyses of event rates are descriptive, and no formal between-group comparisons were made.The event rates are described separately by country, and DerSimonian and Laird random-effects meta-analysis was used when pooling data, taking heterogeneity between countries into consideration. 16Tau was used to describe heterogeneity, corresponding with the estimated SD of the underlying data across countries.
Cumulative percentages of patients using HF medicines after index were calculated using the Kaplan-Meier method.Event rates were calculated as events per 100 patient-years based on time to first event.
CLINICAL OUTCOMES (COHORT 1).Event rates were generally similar across all countries (Table 2).The most frequent event was all-cause rehospitalization (96.8 events per 100 patient-years), followed by allcause death (28.4 events per 100 patient-years-data from Sweden and the United Kingdom only), cardiovascular death (16.2 events per 100 patient-yearsdata from Sweden and the United Kingdom only), rehospitalizations for HF (13.6 events per 100 patientyears), all-cause in-hospital death (12.9 events per 100 patient-years), and rehospitalizations for CKD (4.5 events per 100 patient-years).Rehospitalizations for atherosclerotic cardiovascular diseases were less frequent than rehospitalizations for HF and CKD (myocardial infarction: 2.0 events per 100 patientyears; stroke: 3.0 events per 100 patient-years; and peripheral artery disease: 0.9 events per 100 patientyears).In-hospital all-cause mortality was similar for all 4 countries (Table 2).Rates per 100 patient-years for all-cause rehospitalizations were 99.  6).
HOSPITAL HEALTH CARE COSTS (COHORT 1).All countries had hospital health care cost data available, and the cost levels varied between countries.
Following hHF, hospital health care costs for cardiorenal events (HF or CKD) were consistently higher than those for atherosclerotic events (myocardial infarction or stroke), across all countries (Figure 2).

USE OF HF MEDICINES (COHORT 2)
. The initiation of RAS inhibitors and beta-blockers within 3 months of first hHF discharge was approximately 60% to 80% across countries and index quarters, and remained Outcomes, Costs, and Treatment Following Incident hHF relatively unchanged from first to last relevant calendar quarters (Figure 3, Supplemental Figure 2).

MRA use at 3 months was approximately 20% to 40%
in the first index quarter and changed little over time, except in Sweden, in which a slight increase was noted in the final quarter.Sacubitril/valsartan use was low (<10% at 3 months), with small changes over time.Use of SGLT2 inhibitors was 2% to 11% at 3 months after discharge in the first quarter, increasing to 8% to 35% in the last quarter for Japan, Sweden, and the United States (no last index quarter data were available for the United Kingdom).
Use of 2 or more concurrent HF medicines was approximately 40% to 80% at 3 months across the countries and showed a slight increase from the first to last quarter (Figure 4, Supplemental Figure 3).The use of 3 or more concurrent HF medicines was approximately 10% to 30% at 3 months across the countries, with a notable increase from the first to last quarter.For patients enrolled in the last index quarter, the concurrent use of 4 GDMTs at 3 months after hHF discharge was 10% in Japan, 21% in Sweden, 2% in the United Kingdom (first quarter results only), and 3% in the United States (Supplemental Figure 3).

DISCUSSION
This is the largest and most contemporary study of patients following an incident hHF.Despite differences in study designs and patient population definitions, the characteristics of patients prior to an incident hHF in our study showed similarities with other studies (Supplemental Discussion).We found Event rates for hospitalizations with a main diagnosis of heart failure, chronic kidney disease, myocardial infarction, stroke, peripheral artery disease, all-cause death, and cardiovascular death for patients in Japan, Sweden, the United Kingdom, and the United States, excluding the index hHF (cohort 1; N ¼ 263,525).a Based on the first event during the 12 months following the index hHF discharge (excluding the index hHF).b Random-effects models were used to calculate pooled values; the heterogeneity measure tau corresponds to the estimated SD of the underlying data.c Data on deaths outside of hospital not available.d Cause of death registries not available.
Abbreviations as in Table 1.Following discharge from a first hHF, rehospitalization and mortality were higher in this real-world cohort than in observational studies of patients with acute hHF. 17,18Although speculative, this may partly be because the EVOLUTION HF population exclusively included cases of incident hHF.This population, prior to hospitalization, was undiagnosed and therefore not treated optimally.Indeed, most new cases of HF are diagnosed during a hospital admission.These patients may have rapid progression of disease and severe comorbid conditions (such as respiratory infections or acute coronary syndrome) that precipitated the event.It is to be expected that The pooled mortality of 28% within the first year was high compared with other studies, 17,18 and the results from Sweden and the United Kingdom suggest that more than half of the deaths occur outside of hospital following a first hHF.The high risks of rehospitalization and death observed following a first   HOSPITAL HEALTH CARE COSTS.The cumulative cost analyses account for repeated events during follow-up, rather than for only first events.These demonstrated that, over a 12-month period, hospital health care costs increased quite rapidly, mainly driven by hospitalizations involving HF and CKD.
Although we studied patients with a first hHF (a substantial proportion of whom are likely to have had heart failure with mildly reduced ejection fraction [HFmrEF] and HFpEF), it is possible that more rapid initiation of GDMTs may reduce future hospitalization events and may delay the progression of CKD. 20,21 MEDICINE UPTAKE FOLLOWING A FIRST hHF.When assessing the initiation of individual GDMT medications, large proportions of patients were using a RAS inhibitor or beta-blocker immediately after hHF discharge.This probably reflects the widespread use of these therapies for comorbid conditions prior to admission as well as historical conventions about the sequence of medicine initiation.However, in the first year after discharge, few patients were initiated on sacubitril/valsartan SGLT2 inhibitors reduce the combined risk of hHF or cardiovascular death in patients with HFrEF, HFmrEF, and HFpEF.[24][25][26][27] Most patients encountered in clinical practice are eligible for treatment with SGLT2 inhibitors according to trial and label criteria.28 It is to be hoped that SGLT2 inhibitor use for the treatment of HF will increase with greater awareness and successful implementation strategies.In an analysis from SwedeHF (the Swedish Heart Failure Registry), after the HF benefits of GDMTs vs usual care following hHF, including patients with a broad range of LVEF, and underlines the feasibility of early and rapid optimization of multiple medical therapies in HF. 30 Among patients hospitalized with HF, a substantial proportion are likely to have HFrEF or HFmrEF.1,2 The uptake of all 4 GDMTs may have been higher in patients with HFrEF.We recognize that guidelines do not include recommendations for the use of beta-blockers in patients with HFpEF (these may even worsen outcomes and symptoms in some patients) 31,32 and that the class of recommendations are lower for ARNIs, MRAs, or ARBs than SGLT2 inhibitors.4,5 Nevertheless, many patients will receive RAS inhibitors and beta-blockers for indications other than HF.There are few trial data to support use of all 4 medicines in new-onset hospitalized HF, and no data for HFpEF.However, we assessed initiation for up to 1 year postdischarge from a first hHF and about half of such patients would be expected to have HFrEF. 2 In addition, patients with HFmrEF have similar indications to those with HFrEF, and therefore the proportion of patients not eligible for all 4 medicines is probably lower than 50%.Therefore, it is not unreasonable to expect greater use of all 4 medicines than observed to date.
There are a number of potential barriers to optimal GDMT use, such as patients' willingness to take medicines; real or anticipated side effects and tolerability; 33,34 age, frailty, and comorbidities; 7,8,35 health care professional inertia; the cost to patients; a lack of reimbursement and payer coverage; a lack of optimized multidisciplinary health care structures for GDMT optimization and follow-up; and a lack of strategic care plans and staff to implement them, which creates uncertainty about who will implement therapy and when.Initiatives to improve GDMT uptake and subsequently patient outcomes might include better patient and clinician education and understanding of HF and its management; 11,36 assurance of access to appropriate care and followup; 35,37,38  Stockholm, Sweden.E-mail: gianluigi.savarese@ki.se.

PERSPECTIVES COMPETENCY IN PATIENT CARE AND PROCEDURAL
SKILLS: Following a first hHF, there are high rates of rehospitalization and death, as well as high costs.The 4 foundational pharmacological classes of medicines for HFrEF had low implementation rates in broad, generalizable HF populations in Japan, Sweden, the United Kingdom, and the United States.This may be partially explained by weaker indications and recommendations for these treatments for HFmrEF and HFpEF and also de novo and acute HF.Nevertheless, outcomes may be improved with better use of 4 GDMT medications after hHF in appropriate patients.
TRANSLATIONAL OUTLOOK: There are many barriers to the rapid initiation of GDMTs after a first hHF.These may include perceptions regarding the safety and efficacy of GDMTs in the acute setting, a lack of strategic care plans and a lack of staff with the expertise and experience to implement them, negative patient and clinician perceptions about particular GDMTs, and challenges posed by a lack of access to optimal HF care.This study highlights the need to address these barriers to optimal care in order to improve patient outcomes and to use health care resources wisely.This will require effective planning, organization and implementation, and transparent auditing.

-2 0 2 3 :
---C o h o r t 1 ( c l i n i c a l o u t c o m e s a n d c o s t s d u r i n g 2 0 1 8 -2 0 2 0 ) .This cohort was created to allow sufficient follow-up while avoiding the potential influence of the COVID-19 pandemic on the results (Figure 1).Patients were included if the date of discharge from their index hHF took place between January 1, 2018, and January 1, 2020.Patient characteristics, 1-year readmission rates, and hospital health care costs are reported for this cohort.However, postdischarge mortality data were available only from Sweden and the United Kingdom.

C o h o r t 2 ( c o n t e m p o r a r y u s e o f m e d i c i n e s d u r i n g 2
In this cohort, use of HF medicines following discharge from a first hHF within each country is reported for the years 2020 through 2022 (Figure1).The study periods for the individual countries within this cohort covered the first full month after approval of the SGLT2 inhibitor dapagliflozin for HFrEF in each country (Japan: December 2020; Sweden: December 2020; United Kingdom: January 2021; United States: June 2020), up to the latest available update of the databases.Because cohort 1 and cohort 2 were independent, the association between GDMT use and outcomes was not studied.

C l i n
i c a l o u t c o m e s ( c o h o r t 1 ) .The clinical outcomes studied were rehospitalizations with a primary diagnosis of HF, chronic kidney disease (CKD) (including diagnoses of acute kidney failure, unspecified kidney failure, diabetic kidney disease, CENTRAL ILLUSTRATION Risks, Costs, and HF Medicine Use After a First hHF Study Population Follow-Up Period Aim Post-hHF discharge description of: • What is the risk of rehospitalization and death in the year after a first hHF?• What diseases drive hospital health care cost during this period?• How has HF medicine use changed during the period 2020-2022?et al.J Am Coll Cardiol HF. 2023;-(-):---.
with a strong, evidence-based recommendation): RAS inhibitors (angiotensin-converting enzyme inhibitors and angiotensin receptor blockers [ARBs]), the ARNI sacubitril/valsartan, beta-blockers, MRAs, and SGLT2 inhibitors (dapagliflozin and empagliflozin).Kaplan-Meier estimates were reported for time to initiation and time to the concurrent use of 2 or more medicines, 3 or more medicines, and 4 medicines in the 12 months following index hHF discharge.All HF medicine prescriptions filled in the month prior to the index hHF and during follow-up were analyzed.GDMT initiation prior to the index hHF may have occurred for indications other than HF, and such medicines may be continued after the index hHF without the need to refill the prescription.Therefore, GDMTs initiated in the month prior to or during the index hHF were counted as use at index and had day 0 as the day of initiation, with day 1 being the date of discharge (index date).

FIGURE 1
FIGURE 1 Study Design and Timelines

FIGURE 2
FIGURE 2 Health Care Costs 12 Months After a First hHF Discharge

FIGURE 3
FIGURE 3 HF Medicine Use After a First hHF Discharge

FIGURE 4
FIGURE 4 Concurrent HF Medicine Use After a First hHF Discharge .

J
A C C : H E A R T F A I L U R E V O L .-, NO. -, 2023 Bozkurt et al -2 0 2 3 : ---Outcomes, Costs, and Treatment Following Incident hHF hHF indicate an urgent need for early initiation of multiple concurrent GDMTs.

SGLT2 inhibitors became known
from the cardiovascular outcomes trials, SGLT2 inhibitor use was noted to increase rapidly in patients with type 2 diabetes and HF. 29 High clinical risk has been associated with suboptimal GDMT use by other studies.The recent randomized STRONG-HF (Safety, Tolerability and Efficacy of Rapid Optimization, Helped by NT-proBNP testinG, of Heart Failure Therapies) study demonstrated the benefits of rapid initiation of 3 concurrent

J
A C C : H E A R T F A I L U R E V O L .

TABLE 1
Characteristics of 263,525 Patients Prior to a First hHF Values are median (IQR) or n (%).The table shows characteristics of patients discharged from a first hHF between January 1, 2018, and January 1, 2020 (cohort 1; N ¼ 263,525).a Maximum duration was capped at 100 days; patients with index hospitalization longer than 100 days were excluded as outliers.b Includes angina pectoris, unstable angina, myocardial infarction, and percutaneous coronary intervention/coronary artery bypass grafting.c Laboratory measurements represent the last registered value in the year prior to a first hHF.d Systolic blood pressure not measured at admission.eGFR ¼ estimated glomerular filtration rate; hHF ¼ hospitalization for heart failure; MRA ¼ mineralocorticoid receptor antagonist; NA ¼ not available; RAS ¼ reninangiotensin system; SGLT2 ¼ sodium-glucose cotransporter-2.

TABLE 2
Rehospitalization Risk and Mortality After Discharge From a First hHF Event Rates (Events per 100 Patient-Years) a Dr Lund has received grant support, lecture fees, and/or consulting fees from Abbott Laboratories, Alleviant, AstraZeneca, Bayer, Boehringer Ingelheim, Boston Scientific, Edwards Lifesciences, Lexicon Pharmaceuticals, Medscape, Merck Sharp & Dohme, MyoKardia, Novartis, OrionPharma, Pharmacosmos, Sanofi, and Vifor.Dr Kishi has reported that he has no relationships relevant to the contents of this paper to disclose.
EVOLUTION HF is funded by AstraZeneca.Dr Cleland was supported by a British Heart Foundation Centre of Research Excellence (grant ADDRESS FOR CORRESPONDENCE: Dr Gianluigi Savarese, Karolinska Institutet, Department of Medicine, Solna, Karolinska University Hospital, D1: 04, 171 76 -, NO. -, 2023 Bozkurt et al -2 0 2 3 : ---Outcomes, Costs, and Treatment Following Incident hHF 37. Kapelios CJ, Canepa M, Benson L, et al.Noncardiology vs. cardiology care of patients with heart failure and reduced ejection fraction is associated with lower use of guideline-based care and higher mortality: observations from the Swedish Heart Failure Registry.Int J Cardiol.2021;343:63-72.38.Schrage B, Lund LH, Benson L, et al.Predictors of primary prevention implantable cardioverter-defibrillator use in heart failure with reduced ejection fraction: impact of the predicted risk of sudden cardiac death and all-cause mortality.Eur J Heart Fail.2022;24: 1212-1222.39.Ghazi L, Yamamoto Y, Riello RJ, et al.Electronic alerts to improve heart failure therapy in outpatient practice: a cluster randomized trial.J Am Coll Cardiol.2022;79: A C C : H E A R T F A I L U R E V O L .-, NO. -, 2023 Bozkurt et al -2 0 2 3 : ---Outcomes, Costs, and Treatment Following Incident hHF J