Introduction

Related to improved revascularization techniques and materials, as well as increasing supply with invasive cardiac devices (such as implantable cardioverter defibrillators (ICD) or resynchronization therapies) and cardiac pharmacotherapies, overall survival rates in patients with heart failure (HF) are steadily improving [1,2,3,4]. As a result of improved treatment strategies for patients with HF, characteristics of patients with HF have consistently changed during the past decades, leading into a higher proportion of older patients with increased burden with cardiac and non-cardiac comorbidities [5, 6]. Chronic kidney disease (CKD) was demonstrated to be an independent risk factor of HF progression and sudden cardiac death (SCD), whereas more than half of patients with CKD die due to cardiovascular diseases [7,8,9]. However, the prevalence and prognostic impact of CKD may vary among different HF entities. For instance, the presence of CKD revealed higher prognostic accuracy in patients with HF with reduced (HFrEF) than in HF with preserved left ventricular ejection fraction (LVEF) (HFpEF) within the “Swedish Heart Failure Registry.”

Following the introduction of HF with mildly reduced ejection fraction (HFmrEF) — which are characterized by a LVEF 41–49% as an independent category of HF patients, a scientific gap was created with limited data concerning the effect of comorbidities and treatment strategies for these HF patients [10, 11]. HFmrEF was shown to share features with both HFrEF (i.e., higher rate of males, increased prevalence of CAD) and HFpEF (i.e., higher cardiovascular risk profiles) [12,13,14]. However, data concerning the prevalence and prognostic impact of CKD in patients with HFmrEF, specifically with regard to the stage and etiology of CKD, remains limited [12, 15].

Therefore, the present study aims to [1] investigate the prognostic impact of different stages of CKD [2], the impact of different CKD etiologies, and [3] the effect of CKD stratified by important subgroups of patients with HFmrEF, using a large dataset of consecutive patients hospitalized with HFmrEF from 2016 to 2022.

Methods

Study patients, design, and data collection

For the present study, all consecutive patients hospitalized with HFmrEF at one university medical centre were included from January 2016 to December 2022 [16]. Using the electronic hospital information system, all relevant clinical data related to the index event were documented, such as baseline characteristics, vital signs on admission, prior medical history, prior medical treatment, length of index hospital and intensive care unit (ICU) stay, laboratory values, data derived from all non-invasive or invasive cardiac diagnostics and device therapies, such as echocardiographic data, coronary angiography, data being derived from prior or newly implanted cardiac devices, and pharmacotherapies at discharge.

The present study is derived from the “Heart Failure With Mildly Reduced Ejection Fraction Registry” (HARMER), representing a retrospective single-center registry including consecutive patients with HFmrEF hospitalized at the University Medical Center Mannheim (UMM), Germany (clinicaltrials.gov identifier: NCT05603390). The registry was carried out according to the principles of the declaration of Helsinki and was approved by the medical ethics committee II of the Medical Faculty Mannheim, University of Heidelberg, Germany (ethical approval code: 2022–818).

Inclusion and exclusion criteria

All consecutive patients with ≥ 18 years of age hospitalized with HFmrEF at one institution were included. The diagnosis of HFmrEF was determined according to the “2021 ESC Guidelines for the diagnosis and treatment of acute and chronic HF” [10]. Accordingly, all patients with LVEF 41–49% with symptoms and/or signs of HF were included. The presence of elevated amino-terminal prohormone of brain natriuretic peptide (NT-proBNP) levels and other evidence of structural heart disease were considered to make the diagnosis more likely but were not mandatory for diagnosis of HFmrEF. Transthoracic echocardiography was performed by cardiologists during routine clinical in accordance with current European guidelines [17, 18]. Patients without measurement of the estimated glomerular filtration rate (eGFR) were excluded for the present study.

Risk stratification

For the present study, risk stratification was performed according to the presence and severity of CKD. CKD was defined as abnormalities of kidney function with implication for health accompanied with an estimated glomerular filtration rate (eGFR) < 60 ml/min/1.73 m2 (i.e., according to kidney disease improving global outcome KDIGO stages 3a–5) and a duration > 3 months. Therefore, a minimum of two eGFR measurements separated by at least 3 months was required for the diagnosis of CKD [19,20,21]. All available eGFR values prior to (ambulatory or hospitalized patients) and during index hospitalization were evaluated with regard to the definition of CKD. In patients admitted with acute decompensated HF, eGFR levels were considered following recompensation to decrease the risk to misclassify patients with acute kidney injury.

For the present study, risk stratification was performed according to different KDIGO stages, etiology of CKD, and in important pre-specified subgroups (i.e., ischemic vs. non-ischemic cardiomyopathy, as well as stratified by patients with stable (i.e., prior HFmrEF), deteriorated (i.e., prior HFpEF), and improved HF (i.e., prior HFrEF)).

Measurement of creatinine and eGFR

During index hospitalization, creatinine determination was carried out predominantly from lithium heparinate plasma. This assay is a modification of the Jaffé method with blank correction and axis segment adjustment. This blank correction is used to minimize interferences with bilirubin. The assay was carried out on a clinical chemistry analyzer (Atellica CH 930, Siemens Healthineers, Erlangen, Germany). eGFR was estimated by using the CKD-EPI formula. This is more accurate compared to the MDRD formula in estimating the GFR in the threshold region of beginning renal insufficiency [22, 23].

Study endpoints

The primary endpoint was all-cause mortality at 30 months (median follow-up). Secondary endpoints comprised in-hospital all-cause mortality, all-cause mortality at 12 months, rehospitalization for worsening HF, cardiac rehospitalization, acute myocardial infarction (AMI), stroke, coronary revascularization, major adverse cardiac and cerebrovascular events (MACCE), and changes in LVEF and NT-pro BNP levels during follow-up. All-cause mortality was documented using the electronic hospital information system and by directly contacting state resident registration offices (“bureau of mortality statistics”). Cardiac rehospitalization was defined as rehospitalization due to a primary cardiac condition, including worsening HF, AMI, coronary revascularization, and symptomatic atrial or ventricular arrhythmias. MACCE was defined as a composite of all-cause mortality, coronary revascularization, non-fatal AMI, and non-fatal stroke. All echocardiographic examinations of inpatients and outpatients treated at our institution were assessed and LVEF, and NT-pro BNP levels were documented for the intervals of 0–6 months, 6–12 months, 12–18 months, 18–24 months, and 24–30 months during routine follow-up to further investigate the prognostic role of CKD on LVEF and NT-pro BNP levels during follow-up.

Statistical methods

Quantitative data is presented as mean ± standard error of the mean (SEM), median and interquartile range (IQR), and ranges depending on the distribution of the data. They were compared using the Student’s t test for normally distributed data or the Mann–Whitney U test for nonparametric data. Deviations from a Gaussian distribution were tested by the Kolmogorov–Smirnov test. Qualitative data is presented as absolute and relative frequencies and were compared using the chi-square test or the Fisher’s exact test, as appropriate. Kaplan–Meier analyses were performed stratified by the severity of CKD, as well as to investigate the etiology of CKD and the prognostic value of CKD in pre-specified subgroups. Univariable hazard ratios (HR) were given together with 95% confidence intervals. The prognostic impact of the severity of CKD was investigated within multivariable Cox regression models using the “forward selection” option.

The results of all statistical tests were considered significant at p ≤ 0.05. SPSS (Version 28, IBM, Armonk, NY) was used for statistics.

Results

Study population

A total of 2228 consecutive patients with HFmrEF were hospitalized at our institution from 2016 to 2022. Forty-four patients lost to follow-up, and 29 patients with no measurement of eGFR on admission were excluded (Supplemental Fig. 1; Flow chart). The final study cohort comprised 2155 patients with HFmrEF with a total prevalence of CKD of 31%. Among patients with CKD, most patients presented with KDIGO stage 3b (32.8%), followed by 3a (30.1%), whereas only 11.9% of patients had KDIGO stage 5. Renal replacement therapy was performed in all patients with CKD and KDIGO stage 5. Patients with CKD were older (median age 81 vs. 72 years; p = 0.001) and had higher rates of prior CAD (54.0% vs. 35.2%; p = 0.001), prior AMI (30.9% vs. 20.8%; p = 0.036), as well as higher prevalence of congestive HF (52.2% vs. 25.3%; p = 0.001) (Table 1). In line with this, patients with CKD had higher rates of cardiovascular risk factors including arterial hypertension (89.2% vs. 72.8%; p = 0.001), diabetes mellitus (47.5% vs. 31.5%; p = 0.001), and hyperlipidemia (34.7% vs. 28.3%; p = 0.001). Ischemic cardiomyopathy was the most common HF etiology in both groups (60.1% vs. 56.7%), followed by hypertensive cardiomyopathy (8.0% vs. 8.0%) and primary non-ischemic cardiomyopathies (7.3% vs. 6.7%; p = 0.062). Patients with CKD had lower hemoglobin levels (median 10.9 vs. 13.0 g/dl; p = 0.001) and lower platelet count (median 213 vs. 233 × 109/l; p = 0.001). Finally, the prescription rates of beta-blockers (81.5% vs. 75.7%; p = 0.001), aldosterone antagonists (16.5% vs. 12.9%; p = 0.031), and loop diuretics (76.3% vs. 35.9%; p = 0.001) were higher in patients with CKD.

Table 1 Baseline characteristics
Full size table

Correlations of eGFR with clinical and laboratory data

During index hospitalization, eGFR correlated with hemoglobin (r = 0.340; p = 0.001), platelet count (r = 0.105; p = 0.001), total cholesterol (r = 0.147; p = 0.001), and LDL cholesterol (r = 0.195; p = 0.001), whereas an inverse correlation with age (r =  − 0.361; p = 0.001), potassium (r =  − 0.177; p = 0.001) and NT-proBNP (r =  − 0.141; p = 0.001) was observed (Supplemental Table 1).

Prognostic impact of different stages of CKD in patients with HFmrEF

During a median follow-up of 30 months, the primary endpoint all-cause mortality occurred in 49.4% with CKD and in 23.3% without (Table 2: endpoints). Even milder stages of CKD (i.e., KDIGO stage 3a) were already associated with a higher risk of 30-month all-cause mortality compared to patients without CKD (HR = 1.242; 95% CI 1.147–1.346; p = 0.001) (Fig. 1; left panel). In line with this, patients with CKD and KDIGO stage 3b were associated with a higher risk of long-term all-cause mortality compared to patients with KDIGO stage 3a (HR = 1.833; 95% CI 1.041–3.229; p = 0.036). However, the long-term prognosis did not differ in patients with CKD and KDIGO stage 5 compared to patients with stage 4 (HR = 0.886; 95% CI 0.616–1.275; p = 0.515) or stage 3b (HR = 1.032; 95% CI 0.815–1.307; p = 0.791) Furthermore, the highest risk of HF-related rehospitalization was observed in patients with CKD and KDIGO stages 3b and 4 (log rank p ≤ 0.015), whereas patients with CKD and KDIGO stage 5 had a lower risk of HF-related rehospitalization compared to patients with KDIGO stage 4 (HR = 0.440; 95% CI 0.228–0.849; p = 0.014) and 3b (HR = 0.596; 95% CI 0.388–0.915; p = 0.018) (Fig. 1; right panel).

Table 2 Follow-up data, primary and secondary endpoints
Full size table
Fig. 1
figure 1

Kaplan–Meier analyses demonstrating the prognostic impact of different KDGIO stages on the primary endpoint all-cause mortality at 30 months (left panel), as well as on the risk of HF-related rehospitalization (right panel) within the entire study cohort

Full size image

After multivariable adjustment, already patients with KGIDO stage 3a had a higher risk of 30-month all-cause mortality compared to patients without CKD (HR = 1.331; 95% CI 1.016–1.744; p = 0.038) (Table 3). However, the risk of HF-related was specifically increased in patients with CKD and KDIGO stage 3b (HR = 2.058; 95% CI 1.444–2.933; p = 0.001) and 4 (HR = 1.912; 95% CI 1.274–2.869; p = 0.002), but not in patients with CKD and KDIGO stage 5 after multivariable adjustment (Table 3).

Table 3 Multivariable Cox regression analyses with regard to 30-month all-cause mortality and heart failure–related rehospitalization stratified by the severity of CKD
Full size table

During follow-up, patients with CKD had a lower LVEF at 6 months (median 45 vs. 48; p = 0.001), at 12 months (median 45 vs. 48; p = 0.025), and at 24 months (median 45 vs. 49; p = 0.030) as compared to patients without CKD (Fig. 2; left panel). In line with this, NT-pro BNP levels were higher in patients with CKD during index hospitalization (median 1957 vs. 1441; p = 0.001), as well as at 6 months (median 1993 vs. 1226; p = 0.001), 12 months (median 1499 vs. 1126 p = 0.022) and 30 months (median 2337 vs. 452; p = 0.001) thereafter (Fig. 2; right panel).

Fig. 2
figure 2

Changes in LVEF (left panel) and NT-pro BNP levels (right panel) among patients with and without CKD during 30 months of follow-up. Data is presented as median and interquartile range (IQR)

Full size image

Prognostic impact of the etiology of CKD in HFmrEF

In patients with HFmrEF, most patients had CKD related to arterial hypertension (38.4%), followed by diabetes mellitus (21.7%), whereas only 4.9% had CKD related to glomerulonephritis (Fig. 3). In patients with advanced stages of CKD (i.e., KDIGO stage 5), diabetes mellitus was the most common etiology of CKD in 31.3% of patients. However, the etiology of CKD was not associated with the risk of 30-month all-cause mortality (log rank p ≥ 0.347) and HF-related rehospitalization (log rank p ≥ 0.149) when comparing the prognosis of patients with the four most common etiologies of CKD (Fig. 4).

Fig. 3
figure 3

Distribution of the etiologies of CKD in patients with HFmrEF and concomitant CKD, as well as stratified by KDIGO stages

Full size image
Fig. 4
figure 4

Kaplan–Meier analyses demonstrating the prognostic impact of different CKD etiologies on the primary endpoint all-cause mortality at 30 months (left panel), as well as on the risk of HF-related rehospitalization (right panel) in patients with CKD

Full size image

Effect of CKD in pre-specified subgroups

Even when stratified by patients with ischemic and non-ischemic cardiomyopathy, patients with CKD and KDIGO stages 4 and 5 were associated with the highest risk of 30-month all-cause mortality, especially in patients with ischemic cardiomyopathy (stage 4 vs. 3b: log rank p = 0.010) (Supplemental Fig. 2). Interestingly, even patients with CKD and KDIGO stage 3a were associated with a higher risk of 30-month all-cause mortality compared to patients without CKD, which was observed in patients with ischemic (KDIGO stage 3a vs. non-CKD: HR = 1.273; 95% CI 1.138–1.424; p = 0.001) and non-ischemic cardiomyopathy (KDIGO stage 3a vs. non-CKD: HR = 1.211; 95% CI 1.080–1.358; p = 0.001). Furthermore, LVEF was lower in patients with CKD compared to patients without at 6 and 18 months of follow-up in patients with ischemic cardiomyopathy, whereas NT-proBNP levels were higher during index hospitalization and 6 months thereafter (Supplemental Fig. 3).

When stratified by prior LVEF, patients with CKD and KDIGO stage 3a were specifically associated with a higher risk of all-cause mortality in patients with stable HF (i.e., prior HFmrEF) (HR = 1.358; 95% CI 1.012–1.822; p = 0.041) and deteriorated HF (i.e., prior HFpEF) (HR = 1.164; 95% CI 0.981–1.381; p = 0.081, statistical trend), whereas this association was not observed in patients with improved HF (i.e., prior HFrEF) (HR = 1.009; 95% CI 0.615–1.656; p = 0.971) (Supplemental Fig. 4). Finally, LVEF and NT-proBNP levels did not differ in patients with and without CKD in the presence of stable, deteriorated, or improved HF (Supplemental Fig. 5).

Discussion

The present study investigates the prognostic impact of the severity and etiology of CKD in patients with HFmrEF using a large retrospective registry from 2016 to 2022. The data suggests that even milder stages of CKD (i.e., KDIGO stage 3a) were already associated with an increased risk of 30-month all-cause mortality compared to patients without CKD. This association was observed independently of the presence or absence of ischemic cardiomyopathy and confirmed after multivariable adjustment. In contrast, the long-term prognosis did not differ in patients with CKD and KDIGO stages 3b, 4, and 5. In contrast, patients with KDIGO stages 3b and 4 had the highest risk of HF-related rehospitalization at 30 months. Finally, the long-term prognosis was not affected by the etiology of CKD.

Related to ongoing demographic changes, the overall number of patients with HF suffering from cardiac and non-cardiac comorbidities is steadily increasing [6]. The prognostic value of CKD among patients with HFrEF, HFpEF, and HFmrEF was investigated within a large-scale sub-study of the “Swedish Heart Failure Registry.” In their study, the authors demonstrated an increased risk of all-cause mortality at 1 and at 5 years in patients with CKD, whereas this association was weaker in patients with HFpEF than in patients with HFrEF and HFmrEF [24]. However, in their study, the definition of CKD was based on a single creatinine and eGFR measurement leading to an overall increased prevalence of CKD (i.e., 45–56%). The overall higher rate of CKD compared to the contemporary study is in line with data from the BIOSTAT-CHF study [25]. This may be related to the fact that the definition of CKD required at least 2 eGFR measurements with a duration > 3 months for the present study. This may minimize the chance of including patients with acute kidney injury. In the present study, patients with CKD were accompanied by a risk of all-cause mortality of 34.0% at 1 year and 49.4% at 3 years, which was higher compared to previous studies in patients with HFmrEF (i.e., 23% at 1 year in the Swedish Heart Failure Registry) [24] and even higher compared to contemporary HFrEF registries (12.7% at 1 year in the Chang Gung Research Database) [26]. The mortality rates in the present study may be related to the all-comers setting of the present study and the absence of exclusion criteria, as well as the lower probability of including patients with acute kidney injury related to serial eGFR measurements. From this perspective, previous studies suggest an even higher risk of all-cause mortality in patients with CKD as compared to acute kidney injury [27].

When investigating the effect of different stages of CKD, the “Get With The Guidelines-Heart Failure” (GWTG-HF) suggested an increased risk of all-cause mortality in patients with HFmrEF and eGFR ≥ 60 mL/min/1.73 m2, whereas higher eGFR values were not associated with mortality; however, only minor part of the patients suffered from HFmrEF [28]. In line, data from the “RECOLFACA” registry revealed advanced stages of CKD (i.e., eGFR < 30 mL/min/1.73 m2) were associated with increased risk of all-cause mortality compared to patients with eGFR ≥ 60 mL/min/1.73 m2; however, the risk of all-cause mortality was not affected in patients with moderate CKD (i.e., GFR 30–59 mL/min/1.73 m2) [29]. The present study suggests that even milder stages of CKD were independently associated with an increased risk of 30-month all-cause mortality in consecutive patients with HFmrEF, which was confirmed using multivariable Cox regression analyses.

Of note, the risk of HF-related rehospitalization at 30 months was lower in patients with CKD and KDIGO stage 5 compared to patients with stages 3b and 4 within the present study. This may be associated with improved volume management in patients with renal replacement therapy, as well as the closer clinical follow-up in patients with end-stage renal disease. The “Continuous Ultrafiltration for Congestive Heart Failure” trial investigated the prognostic impact of ultrafiltration as compared to diuretic treatment in 56 patients with congestive HF. Although the risk of all-cause mortality at 1 year did not significantly differ in both groups, patients undergoing ultrafiltration had improved freedom from rehospitalization for HF and lower NT-proBNP levels compared to patients treated with diuretics [30]. In line with this, the “Aquapheresis versus Intravenous Diuretics and Hospitalization for Heart Failure” (AVOID-HF) trial demonstrated a longer time until the first HF-related rehospitalization at 90 days in patients undergoing ultrafiltration as compared to diuretic treatment only in 224 patients hospitalized for HF [31]. This may be in line with findings from the present study, suggesting improved freedom from HF-related rehospitalization in patients with CKD and KDIGO stage 5, whereas all patients required renal replacement therapy which may be in line with improved volume management as compared to diuretic treatment only. Finally, only a minor part of the study population in the present study was treated with sodium-glucose cotransporter 2 inhibitors (SGLT2i), which were shown to decrease the risk of cardiovascular mortality and HF-related rehospitalization in HFmrEF [32]. The low prescription rates of SGLT2i may be related to their recent upgraded level of evidence in 2023 and are supported by their low use in contemporary HFmrEF studies [11]. From this perspective, the “EMPEROR-Preserved” demonstrated the superiority of treatment with SGLT2i versus placebo, irrespective of the use and daily dose of concomitant diuretic treatment [33]. Furthermore, treatment with SGLT2i was associated with improved outcomes independent of CKD and KDIGO categories [34]. Further real-life studies are however needed to investigate the effect of CKD in HFmrEF following their inclusion into daily clinical practice.

Study limitations

The study has several limitations. For the present study, at least two eGFR measurements separated by at least 3 months were required for the diagnosis of CKD in accordance with current international guidelines. Defining CKD based on the assessment of at least two eGFR measurements may, however, result in delayed or missed identification of patients with CKD in the absence of regular testing episodes within the present registry related to the consecutive inclusion of patients with HFmrEF. Furthermore, the inclusion of two eGFR values may lead to a survivor bias for fulfilling the criteria of eGFR according to the present guidelines [19]. In addition, patients with recurrent episodes of acute kidney injury may be misdiagnosed into the group of CKD. To decrease the chance of misclassification, eGFR values were considered following recompensation in patients with acute decompensated HF and multiple eGFR measurements during index hospitalization (> 95% patients with multiple eGFR measurements during index hospitalization within the present study). The presence of albuminuria was not assessed for the present study, although albuminuria was demonstrated to increase the risk of all-cause mortality among patients with CKD; therefore, the prevalence and prognosis of patients with early stages of CKD were beyond the scope of the present study [9]. HF-related and cardiac rehospitalization was assessed at our institution only. Finally, causes of death beyond during index hospitalization were not available for the present study.

Conclusions

CKD represents one of the most common non-cardiovascular risk factors being present in 31% of patients with HFmrEF. Even milder stages of CKD were already independently associated with impaired long-term prognosis.