Introduction

Guideline-directed medical therapy (GDMT) has been shown to improve survival, reduce the risk of heart failure (HF) hospitalization, and reduce symptoms in patients with HF with reduced ejection fraction (HFrEF)1,2,3,4,5,6. Initiation and uptitration of GDMT for all patients with HFrEF is recommended by class I indications in the 2021 European Society of Cardiology (ESC) HF guidelines and 2023 focused update7,8. In patients with HF with mildly reduced ejection fraction (HFmrEF), implementation of GDMT is also recommended weakly (class IIb) except for a sodium-glucose cotransporter 2 (SGLT2) inhibitor (dapagliflozin or empagliflozin)7,8. Nevertheless, it has been reported that the rates of initiation of GDMT for patients with HFrEF or HFmrEF were low (20–40%)9,10.

Several factors can be considered as potential barriers for GDMT implementation, including patient factors, social factors, and physician factors11. Recently, some studies have been focused on the reasons why GDMT was not implemented for patients with HFrEF12,13, but the detailed reasons of non-implementation remained unclear in patients admitted for acute heart failure. In Japan, the length of hospital stay is much longer than that in other countries14. Thus, it makes it possible to initiate GDMT during the HF hospitalization, allowing smooth uptitration in the subsequent outpatient setting. This study aimed to clarify the prevalence, clinical significance, and the barriers of GDMT implementation at discharge using the multicenter observational cohort study of Japanese patients with acute heart failure hospitalization.

Methods

Study population

The Kyoto Congestive Heart Failure (KCHF) registry is a physician-initiated, prospective, observational, multicenter cohort study that enrolled the consecutive patients hospitalized for acute heart failure for the first time between October 1, 2014, and March 31, 2016, at 19 secondary and tertiary hospitals in Japan. The details of the KCHF registry have been described previously15,16. Briefly, we consecutively enrolled the patients with acute decompensated heart failure defined by the modified Framingham criteria who were admitted to the participating centers and who received treatment for heart failure involving intravenous drugs within 24 h after hospital presentation. We included patients with HFmrEF as well as those with HFrEF, because some studies reported that the prognostic impact of medication for patients with HFmrEF was more comparable to those with HFrEF, but not to those with HFpEF17,18. The study protocol was approved by the ethics committee of Kyoto University Hospital (approval number: E2311) and other participating centers (Supplementary Appendix 1). The institutional review boards of Kyoto University Hospital and each participating center waived the need for written informed consent, as this study met the conditions outlined in the ‘Ethical Guidelines for Medical and Health Care Research Involving Human Subjects’. All the data were anonymized prior to the analysis. The investigation conformed to the principles outlined in the Declaration of Helsinki.

Definitions

We defined the use of angiotensin-converting enzyme inhibitors (ACE-I)/angiotensin receptor blockers (ARB), beta-blockers (BB), or mineralocorticoid receptor antagonists (MRA) as GDMT. The use of each medication was defined as prescription at discharge from the index hospitalization. We did not include SGLT2 inhibitors in GDMT, because they were not approved for heart failure during the enrollment period of this study in Japan. The study patients were classified according to the number of GDMT classes including ACEI or ARB, BB, and MRA at discharge (i.e. GDMT = 0, 1, 2, and 3). Moreover, patients with GDMT = 3 were classified as full-GDMT group, whereas those with GDMT = 0, 1, and 2 were classified as no full-GDMT group. The detailed definitions of other baseline patient characteristics were provided in Supplementary Appendix 2.

Outcomes

The primary outcome measure in the current study was a composite of all-cause death or HF hospitalization at 1 year. The secondary outcome measures were individual components of the primary outcome measure and cardiovascular death at 1 year. Death was regarded as cardiovascular in origin unless obvious non-cardiovascular causes could be identified. Cardiovascular death included HF-related death, sudden death, stroke-related death, and death from other cardiovascular causes. Sudden death was defined as unexplained death in a previously stable patient. Stroke included either ischemic or hemorrhagic stroke that required hospitalization with symptoms lasting more than 24 h. HF hospitalization was defined as hospitalization due to worsening of HF, requiring intravenous drug therapy. The detailed definitions of other clinical outcome measures were described previously16. All the endpoint events were adjudicated by a clinical event committee. Laboratory data were obtained at the time of admission.

Statistical analysis

Categorical variables were presented as numbers and percentages. Continuous variables were presented as mean with standard deviation or median with interquartile range (IQR) according to their distributions. Categorical variables were compared using the Chi-square test or Fisher’s exact test. Continuous variables were compared using the Student’s t-test or Wilcoxon rank-sum test based on their distributions. To determine the factors associated with full-GDMT at discharge, multivariable logistic regression models were developed to identify the clinical characteristics, laboratory data, and echocardiography data that were independently associated with full-GDMT using 23 potential candidate variables which were clinically important or extracted with P value < 0.05 by univariate analysis from the variables listed in Supplementary Table 1. Continuous variables were dichotomized using clinically meaningful reference values or median values, as shown in Table 1. The results were expressed as the odds ratios (ORs) with their 95% CI.

Table 1 Patient characteristics in the 4 groups classified based on the number of GDMT classes at discharge.
Full size table

The date of discharge from the index hospitalization was regarded as time 0 for clinical follow-up. The cumulative incidence was estimated using the Kaplan–Meier method with intergroup differences assessed by the log-rank test. Multivariable Cox proportional hazard models were used to estimate the hazard ratios (HRs) and their 95% confidence intervals (CI) of patients with GDMT = 0, 1, and 2, respectively, relative to those with GDMT = 3 for the primary and secondary outcome measures. To adjust for the confounders, we incorporated the 22 clinically relevant risk-adjusting variables listed in Table 1 in consistent with the previous reports19,20. We also performed the subgroup analysis stratified by LVEF, i.e. HFrEF and HFmrEF. We estimated the interactions between the subgroup of LVEF and the effect of the number of GDMT classes at discharge on clinical outcomes. Additionally, we conducted a sensitivity analysis using propensity score (PS) matching analysis. A logistic regression model was developed to make the propensity score for full-GDMT with clinically relevant 31 variables (Supplementary Table 1). To create PS-matched cohorts, the patients in full-GDMT group were matched to those in no full-GDMT group on the PSs by the nearest neighbor matching technique with a caliper of width equal to 0.05 of the standard deviation of the PS (Supplementary Fig. 2). P-values were two-tailed, and statistical significance was set at P < 0.05. All data were analyzed using JMP version 18.0.2 software (SAS Institute Inc, Cary, NC, USA).

Results

Patient characteristics

Among 4056 patients enrolled in the KCHF registry, we excluded 271 patients who died during the index hospitalization, 11 patients without baseline left ventricular EF (LVEF) data, and 57 patients without follow-up data after discharge. In the remaining 3717 patients, we further excluded 1631 patients with HF with preserved EF (HFpEF). Finally, the current study population consisted of 2086 patients with HFrEF or HFmrEF, who were discharged alive. There were 181 (8.7%), 508 (24.4%), 791 (37.9%), and 606 (29.1%) patients with GDMT = 0, 1, 2, and 3, respectively (Fig. 1). Regarding the patient characteristics, as the number of GDMT classes increased, patients were younger, and more often men, and had a higher body mass index, a lower prevalence of valvular heart disease, prior HF hospitalization, prior PCI or CABG, dementia, and a higher prevalence of cardiomyopathy, current working, current smoker, ambulatory status, and living alone (Table 1). Patients with bradycardia, anemia, hypoalbuminemia, hyperkalemia, and impaired renal function at admission less often had greater number of GDMT classes. Patients with GDMT = 3 more often had HFrEF and less often had moderate or severe TR or MR.

Fig. 1
figure 1

Study flowchart. The study patients were classified according to the number of GDMT classes including ACEI or ARB, BB, and MRA at discharge (i.e. GDMT = 0, 1, 2, and 3).

Full size image

Factors associated with full GDMT at discharge

To investigate the factors associated with full GDMT (the number of GDMT classes = 3) at discharge, 2086 patients in the current study were divided into the full-GDMT group (N = 606 (29.1%)) and the no full-GDMT group (the number of GDMT classes at discharge = 0, 1, 2: N = 1480 (70.9%)). The patient characteristics of 2 groups were summarized in Supplementary Table 1. The proportion of patients in the full-GDMT group and the percentages of prescription for each medication between participating sites were shown in Supplementary Fig. 1. In the multivariable logistic regression analysis, current smoker (OR: 1.36, 95% CI: 1.02–1.81, P = 0.03), ambulatory status (OR: 1.54, 95% CI: 1.07–2.21, P = 0.02) and HFrEF (OR: 1.65, 95% CI: 1.29–2.12, P < 0.001) were positively associated with full GDMT at discharge, whereas age ≥ 80 years old (OR: 0.64, 95% CI: 0.50–0.83, P < 0.001), acute coronary syndrome (OR: 0.42, 95% CI: 0.25–0.70, P < 0.001), anemia at admission (OR: 0.67, 95% CI: 0.53–0.85, P < 0.001), and eGFR < 30 mL/min/1.73m2 (OR: 0.29, 95% CI: 0.20–0.43, P < 0.001) were negatively associated with full GDMT at discharge (Table 2).

Table 2 Multivariable logistic regression analysis for the factors associated with full GDMT at discharge.
Full size table

Clinical outcomes

The median follow-up duration was 462 (IQR: 354–645) days, with an 88.0% follow-up rate at 1 year. The cumulative 1-year incidence of the primary outcome measure (a composite of all-cause death or HF hospitalization) was significantly lower with increasing number of GDMT classes at discharge (GDMT = 0, 56.3%; GDMT = 1, 40.7%; GDMT = 2, 31.9%; GDMT = 3, 25.1%, P < 0.001) (Fig. 2A). The cumulative 1-year incidences of all-cause death, HF hospitalization, and cardiovascular death were also significantly lower with increasing number of GDMT classes at discharge (Fig. 2B and D). After adjusting the confounders, the excess risk of patients with GDMT = 0 or 1, but not GDMT = 2 relative to those of GDMT = 3 remained significant for the primary outcome and for all-cause death (Table 3). The excess risk of those of GDMT = 0 relative to those of GDMT = 3 also remained significant for HF hospitalization and for cardiovascular death (Table 3). There were no significant interactions between the subgroup of LVEF and the effect of the number of GDMT classes at discharge for the primary and secondary outcome measures (Fig. 3). Additionally, in PS-matched cohorts, the excess risk of those of GDMT = 0 relative to those of GDMT = 3 was significant for the primary outcome, all-cause death, and cardiovascular death (Supplementary Fig. 3, Supplementary Table 2).

Fig. 2
figure 2

Kaplan-Meier curves for the primary and secondary outcome measures according to the number of GDMT classes at discharge. (A) the primary outcome measure, (B) all-cause death, (C) HF hospitalization, and (D) cardiovascular death. The primary outcome measure was a composite of all-cause death or HF hospitalization. Unadjusted HR and 95% CI were presented in the figure.

Full size image
Table 3 Clinical outcomes at 1 year.
Full size table
Fig. 3
figure 3

Effects of HFrEF or HFmrEF on clinical outcomes. HR hazard ratio, CI confidence interval, GDMT guideline-directed medical therapy, HF heart failure.

Full size image

Discussion

The main findings of the current study were as follows: (1) The proportion of patients who achieved full GDMT at discharge was 29.1%; (2) The factors positively associated with full GDMT at discharge were current smoker, ambulatory status, and HFrEF, whereas age ≥ 80 years, acute coronary syndrome, anemia, and eGFR < 30 mL/min/1.73m2 were negatively associated with full GDMT at discharge; (3) The increase in the number of GDMT classes at discharge was associated with a lower risk of a composite of all-cause death or HF hospitalization.

GDMT, including ACE-I/ARB or angiotensin receptor-neprilysin inhibitor (ARNI), BB, MRA, and SGLT2 inhibitor (dapagliflozin or empagliflozin) has been demonstrated to improve outcomes in patients with HFrEF. In patients with acute heart failure, initiation and rapid uptitration of GDMT was associated with lower all-cause death or HF hospitalization6,21,22. In this study, we included the patients with HFmrEF as well. Also with regard to patients with HFmrEF, candesartan was reported to be effective for reduction of cardiovascular death or HF hospitalization (HR 0.76, 95% CI: 0.61–0.96, P = 0.02)17, and the prognostic impacts of medication in patients with HFmrEF were more comparable to those with HFrEF, but not to those with HFpEF18. In addition, considering that LVEF of 50% is lower than the lower boundary of the normal LVEF range23, it is reasonable that GDMT is implemented for patients with HFmrEF, as for those with HFrEF in clinical practice. This study found that in acute heart failure patients with HFrEF or HFmrEF, the greater number of GDMT classes at discharge was associated with reduced risk of clinical outcomes at 1 year, even after multivariable adjustment for confounders. Besides, the trend of the association was consistent without significant interaction, regardless of HFrEF or HFmrEF. These results supported the premise that GDMT should be implemented for patients with HFrEF or HFmrEF at discharge of acute heart failure hospitalization with the aim of achieving full GDMT. In addition, the patients with GDMT = 0 still exhibited higher risk for clinical outcomes relative to those with GDMT = 3 in propensity score-matched cohorts. Patients discharged with GDMT = 0 are uncommon in clinical practice, constituting less than 10% of the total cohort. Considering, for example, especially low prescription rate of loop diuretics of 66.3%, this population might have negative reasons preventing GDMT implementation, such as difficulty in taking regular oral medication or highly impaired renal function.

In the current study, the prescription rate for each medication at discharge for patients with HFrEF or HFmrEF was 61.9% for ACEI/ARB, 75.9% for BB, and 49.5% for MRA. This was not as high as in previous studies including many outpatients or the latest clinical trial in Japan10,12,24, due that our registry enrolled patients with acute heart failure and more than 60% of whom were hospitalized for de novo heart failure without prior HF hospitalization. In addition, 29.1% of all patients achieved full GDMT at discharge. In this study, we divided total cohort into 2 groups: full-GDMT group and no full-GDMT group. This is because, particularly in the setting of long hospital stay in Japan, the strategy is often adopted in which GDMT is initiated during hospitalization and then uptitrated in outpatient follow-up, and initiation of full GDMT at discharge is considered to be important.

There are various potential barriers for GDMT initiation and uptitration. The first is patient factors such as age, frailty, vital signs, laboratory data, patient adherence, or side effects. The second is social factors such as costs, insurance, or accessibility to healthcare. The third is physician factors, such as inadequate understanding of the guidelines, fear of side effects, or clinical inertia11. In this study, patients with acute coronary syndrome were less likely to be associated with full GDMT, which was thought to be due to intolerance or prioritization of other medications such as antiplatelets or statins. Age, smoking status, and ambulatory status, which were likely to affect patients’ activity or adherence, were also associated with full GDMT. The fact that current smoker was positively associated with full GDMT might seem counterintuitive, but it was considered that the adjustment was not perfect and residual confounding might have resulted in current smokers having more favorable background compared to non-smokers. Moreover, the fact that there was a variation in the proportion of patients achieving full GDMT across the participating sites might reflect social factors in each region or hospital, as well as physician factors such as clinical inertia or difference in the quality of clinical practice. In summary, this study suggests that it might be good to recognize factors associated with full GDMT and attempt to implement GDMT during acute heart failure hospitalization for patients with HFrEF or HFmrEF if possible.

The current study has several limitations. First, this was an observational study and the patients not achieving full GDMT were thought to be more likely to have serious comorbidities, who would be excluded from randomized controlled trial. Accordingly, the extent to which GDMT was associated with a reduced risk of clinical outcomes in this study might have been overestimated. However, comparison between full GDMT and absence of full GDMT could not be performed in terms of ethics. Thus, we adjusted for 22 clinical conceivable factors in the multivariable logistic regression analysis to minimize the confounders. Second, we did not assess the dose and uptitration of the medication. However, as the current study was the data for acute heart failure hospitalization, our analysis focused on the initiation of the medication at discharge rather than uptitration. Third, we have no data about ARNI and SGLT2 inhibitor. These drugs were not yet approved for heart failure in Japan during the enrollment period of the KCHF registry between 2014 and 2016. While ARNI and SGLT2 inhibitor contribute to improve clinical outcomes in heart failure, the efficacy was proved not to be significantly different, regardless of background medication24,25. Therefore, the clinical significance of the number of GDMT in this study remains undiminished, but future studies are needed to align with contemporary clinical practice including ARNI and SGLT2 inhibitor. Fourth, LVEF was derived solely from the initial data on admission. We did not analyze on the time course of LVEF during follow-up.

Conclusion

The greater number of GDMT classes at discharge was associated with a lower risk for all-cause death or HF hospitalization in patients with HFrEF or HFmrEF. However, we should be cautious about residual confounding, because serious morbidities were negatively associated with full GDMT.