INTRODUCTION

Patients with end-stage heart failure (HF) constitute an etiologically and functionally heterogeneous group, which makes it difficult to optimize their management and assess their outcomes [1, 2]. Physicians routinely use biomarkers reflecting pathophysiological pathways involved in HF development and progression for diagnosis, risk stratification, and management of HF patients [3, 4]. Candidate biomarkers for risk stratification are still being investigated because their type and prognostic value have changed over time as a result of medical progress, as reflected by modifications in the guidelines for HF management. The most useful stratification tools in daily clinical practice are those that explore the main disease pathways, such as activation of neurohormonal systems, inflammation, oxidative stress, dysfunctional calcium handling, impaired energy utilization, mitochondrial dysfunction, cardiac fibrosis, and matrix remodeling, and impaired endothelial function [5, 6]. They should also be accurately and precisely measurable, their results should be easily interpretable and quickly available, their biological variation and reference limits should be defined, and possible sources of error should be well known [5–7].

A promising new HF biomarker is growth differentiation factor-15 (GDF-15), which has not yet been adopted in the guidelines or routine clinical care for HF patients [6, 7]. GDF-15 is emerging as a biomarker that integrates information from cardiac and extracardiac disease pathways and is linked to the incidence, progression, and prognosis of HF. Given that end-stage HF affects multiple organ systems, GDF-15 may provide adequate prognostic information in this population [3, 6, 7].

As a member of the transforming growth factor-beta superfamily, GDF-15 is a newly studied heart-derived endocrine hormone that regulates inflammatory response, tissue repair, and cardiac remodeling [3, 6, 7]. The effects of GDF-15 on the heart seem to be linked to its spectrum of autocrine/paracrine properties, including anti-inflammatory, antifibrotic, antihypertrophic, antioxidant, and antiapoptotic effects [5–7]. GDF-15 is expressed and secreted by numerous cell types, including smooth and endothelial muscle cells and cardiomyocytes, in response to cellular stress, inflammation, and mitochondrial dysfunction, i.e. processes closely related to HF [6]. GDF-15 is also involved in coordinating endothelial function by regulating vasoconstriction and relaxation and the release of nitric oxide [7]. Given the associations of biomarkers with the complex pathogenesis of HF, we aimed to determine the potential risk factors associated with the one-year mortality rate in patients with end-stage HF, with particular emphasis on GDF-15.

METHODS

We prospectively analyzed 404 consecutive patients with end-stage HF who were hospitalized in the cardiology department and underwent heart transplantation evaluation between 2018 and 2022. The flowchart of the study design with the inclusion and exclusion criteria is presented in Figure 1.

Figure 1. Flow chart of the study design for the inclusion and exclusion criteria

Abbreviations: HF, heart failure; INTERMACS, Interagency Registry for Mechanically Assisted Circulatory Support; NYHA, New York Heart Association

At the time of enrollment, a panel of laboratory tests, an ergospirometric exercise test, echocardiography, and right heart catheterization were performed on all patients. Blood samples of peripheral venous blood were drawn after 12 hours of fasting at the time of enrollment. The hematologic parameters were measured with automated blood cell counters (Sysmex XS1000i and XE2100, Sysmex Corporation, Kobe, Japan), which had intra-assay and interassay coefficients of variation of 5% and 4.5%, respectively. The biochemical parameters of the blood samples were determined with a COBAS Integra 800 analyzer (Roche Instrument Center AG, Rotkreuz, Switzerland). Fibrinogen levels were measured using a modified Clauss method (Dade Behring) in a Sysmex CA-6000 automated coagulation analyzer. The plasma high-sensitivity C-reactive protein (hs-CRP) concentration was determined with a COBAS Integra 70 analyzer (Roche Diagnostics, Ltd., Mannheim, Germany). The detection limit of hs-CRP was 0.0175 mg/dl. Plasma N-terminal pro-B-type natriuretic peptide (NT-proBNP) was measured by a commercially available kit from Roche Diagnostics (Mannheim, Germany) on an Elecsys 2010 analyzer with an analytical sensitivity of <5 pg/ml. In addition, 10 ml of peripheral blood was collected to measure the serum GDF-15 concentration. Human GDF-15 was measured by sandwich enzyme-linked immunosorbent assay (ELISA) with a commercially available kit (Human GDF-15 ELISA, SunRedBio Technology Co., Ltd., Shanghai, China). This ELISA was performed using a BioTek Elx50 reader (BioTek Instruments Inc., Tecan Group, Switzerland). The concentration of GDF-15 was expressed as pg/ml.

Information on survival data during long-term follow-up was obtained from the official registry of the National Health Fund or during protocol visits at our institution. The endpoint was defined as all-cause mortality within 12 months of the index hospitalization. The Medical University of Silesia’s local Institutional Review Board approved the study protocol, and all patients provided informed consent (specific ethics codes — KNW/0022/KB1/53/18, PCN/0022/KB1/69/I/19, PCN/0022/KB/159/20, PCN/CBN/0052/KB1/20/II/21/22). The study was performed in accordance with the ethical standards of the 1964 Declaration of Helsinki and its later amendments.

Statistical analysis

Statistical analysis was performed with SAS software version 9.4. Continuous variables were expressed as the mean (standard deviation) or, if not normally distributed, median (first–third quartiles). Differences in continuous variables were compared using Student’s t-test for normally distributed variables and the Mann–Whitney U test for nonnormally distributed variables. Categorical variables were expressed as numbers and percentages and were compared by the χ2 test. Cox’s univariable proportional hazard analysis was used to select potential independent predictive factors of death for inclusion in the multivariable analysis. The examined covariates included age, male sex, ischemic etiology of HF, history of arterial hypertension and dyslipidemia, body mass index, maximal oxygen uptake, cardiac index, and echocardiography parameters (left atrium, left ventricular end-diastolic diameter, right ventricular, and left ventricular ejection fraction), as well as laboratory parameters (bilirubin, albumin, glomerular filtration rate, gamma-glutamyl transpeptidase [GGTP], alkaline phosphatase [ALP], hs-CRP, fibrinogen, NT-proBNP, and uric acid). The relationships between explanatory variables were calculated by the Spearman rank correlation coefficient, and multicollinearity was evaluated using tolerance and variance inflation factors. Since several covariates were highly correlated (e.g., the correlation was 0.74 for GDF-15 with dyslipidemia, 0.51 for ALP with GGTP, and 0.41 for NT-proBNP with ALP), those covariates that provided a better fit for the model were selected for further analysis. Univariable predictors of death with a P-value of 0.05 or less, which did not correlate significantly, were entered into the multivariable Cox proportional hazards model. Schoenfeld residuals were used to check the proportional hazards assumption. The proportional hazards assumption was not met, so weighted Cox regression analysis was conducted using the SAS macro language [8, 9]. The results were presented as hazard ratios with 95% confidence intervals. A P-value of <0.05 was considered significant.

RESULTS

The final study group consisted of 315 patients with end-stage HF who underwent heart transplantation evaluation classified into New York Heart Association (NYHA) functional classes III and IV (79.7% and 20.3%, respectively) and profiles 4 to 6 according to the Interagency Registry for Mechanically Assisted Circulatory Support classification. During the one-year follow-up, the mortality rate was 31% (n = 97). The demographic and clinical characteristics of the patients in the whole population and in the survival and nonsurvival groups are presented in Table 1. The multivariable Cox proportional hazards analysis demonstrated that an elevated GDF-15 concentration, along with the routinely used laboratory parameters fibrinogen, bilirubin, GGTP, N-terminal pro-B-type natriuretic peptide, and high-sensitivity C-reactive protein, was strongly associated with one-year mortality in patients with end-stage HF. The univariable and multivariable Cox proportional hazards analyses for the prediction of one-year mortality are shown in Table 2.

DISCUSSION

Our single-center prospective study showed that higher GDF-15, hs-CRP, fibrinogen, GGTP, bilirubin, and NT-proBNP were associated with one-year outcomes in patients with end-stage HF. From a pathophysiological point of view, the most interesting finding of our study was the strong association between higher GDF-15 levels and worse prognosis. Our results highlight the continuing importance of elevated GDF-15 levels in the risk stratification of patients with end-stage HF treated according to the current HF guidelines.

GDF-15 is a pleiotropic protein indicating several pathological processes associated with HF, but its mechanism of action is still unclear [6, 7, 10]. GDF-15 expression is controlled by various transcriptional pathways and is present nonpathologically only in placental and prostate tissue while in other tissues, its expression is very low. Its expression increases strongly in response to stress stimuli that activate several signaling pathways. Cardiac myocytes produce and secrete GDF-15 in response to injury, such as inflammation, ischemia, pressure, and volume overload, possibly through proinflammatory cytokines and oxidative stress-dependent signaling pathways [6, 7, 10–12]. Additionally, GDF-15 plays an important role in the development of endothelial dysfunction, a key feature in the onset and progression of HF [11]. GDF-15 directly modulates vascular contraction and relaxation responses in the endothelium by reducing nitric oxide release [11]. GDF-15 is not specific to cardiac myocytes. Other cellular sources of GDF-15 include macrophages, vascular smooth muscle cells, endothelial cells, and adipocytes in organs such as the kidney and liver [6, 7, 10, 11]. Thus, GDF-15, a noncardiospecific marker, may indicate both myocardial stress and related systemic abnormalities. This allows GDF-15 to integrate information from cardiac and extracardiac disease pathways that are linked to multiorgan dysfunction in the advanced stages of the disease. Elevated GDF-15 may reflect additional inflammatory and oxidative stress in HF beyond hemodynamic wall stress [13].

GDF-15 might have a limited role in the diagnosis of HF, but it may provide incremental prognostic information on HF morbidity and mortality [11]. Clinical studies have shown that initial and repeat measurements of GDF-15 can predict worse outcomes in different populations of HF patients [7, 10, 11, 14]. However, most of these studies followed earlier guidelines for HF management before modern medical therapy and devices came into use. Anand et al. [10], in a post hoc analysis of patients from the Val-Heft Trial (Valsartan in Heart Failure Trial), reported that baseline circulating GDF-15 was associated with adverse outcomes independent of established clinical, echocardiographic, and biochemical markers. We must emphasize that patients from Val-Hefts were randomized 20 years ago, and the study population differed significantly from ours, especially in terms of pharmacological treatment for HF. Only 33% of patients were treated with beta-blockers, 2% with spironolactone, and none with flozins.

In a meta-analysis by Luo et al. [14], the relationship between GDF-15 and all-cause mortality was studied in more than 6000 stable HF patients randomized between 2007 and 2017 [15]. The authors found that GDF-15 was strongly associated with all-cause mortality among chronic ischemic HF patients, while no such association was found among chronic nonischemic HF patients [15]. Another meta-analysis of 8 studies with over 4000 stable NYHA II–III HF patients randomized before 2017 also indicated an association between elevated GDF-15 and a greater risk of mortality in this population [15]. These results should be interpreted with caution due to the substantial heterogeneity and publication bias among the studies included in the meta-analyses. Some studies reported that both initial and repeat measurements of GDF-15 have greater prognostic utility than NT-proBNP, which is considered the gold-standard biomarker in HF [10, 16].

Other factors associated with the one-year mortality rate in the analyzed group of patients were markers of systemic inflammation, such as hs-CRP and fibrinogen. Hs-CRP is considered the most important single marker of low-grade inflammatory processes in HF. It is synthesized by hepatocytes in response to cytokines, and it is part of the immune response [16]. Fibrinogen, a nonspecific acute-phase reactant, is also a component of the inflammatory cascade increasingly implicated in the pathogenesis of HF [17]. There is strong evidence that inflammation contributes to the pathogenesis, progression, and complications of HF through various pathways [18]. Proinflammatory cytokines, which are components of the immune response that mediate inflammation and arise from the gastrointestinal tract, spleen, and adipose tissue, induce low-grade inflammation that promotes ventricular systolic and diastolic dysfunction and contributes to endothelial dysfunction, cardiomyocyte apoptosis, cardiac fibrosis, and cardiac remodeling [19].

There is growing evidence that one of the important causes of inflammation in HF patients is translocation of bacterial endotoxin through edematous bowel walls. The activation of angiotensin II and mineralocorticoid receptors may cause vascular inflammation by promoting the production of reactive oxygen species and adhesion molecules and the polarization of macrophages toward a proinflammatory phenotype [19, 20]. The hs-CRP level is associated with cardiac decompensation, myocardial tissue and other organ damage, low cardiac output, and venous congestion, as well as the presence of muscle wasting, anemia, renal dysfunction, and other common comorbidities in HF [21]. There is also evidence to suggest that CRP itself is toxic to the myocardium, suggesting a mechanism by which inflammation may continue to damage the heart [17, 22].

In turn, fibrinogen is a major coagulation protein in the blood and is an important determinant of thrombogenicity, blood viscosity, and platelet aggregation [23]. Fibrinogen regulates the function of intercellular adhesion molecules that enhance monocyte and leukocyte adhesion to endothelial cells [23, 24]. It also stimulates mononuclear cells to express pro-inflammatory cytokines, which induce nitric oxide-mediated negative inotropic effects and cardiac myo­cyte apoptosis in experimental animal models [18, 23, 24]. Upon binding to endothelial cells, fibrinogen also causes the release of vasoactive mediators, promotes smooth muscle cell proliferation, and modulates endothelial cell permeability and migration [25]. Fibrinogen augments platelet reactivity, aggregation, and degranulation and promotes microthrombus formation [25–27]. The results of platelet hyperreactivity, hypercoagulability from enhanced thrombogenicity, and impaired fibrinolysis predispose a patient to microvascular thrombosis. The association between fibrinogen and myocardial function suggests that the postulated pathogenetic mechanisms of procoagulation and inflammation may contribute to subclinical impairment of myocardial function independently of altered ventricular filling dynamics. It has also been suggested that fibrinogen contributes directly to the onset of arrhythmias, which often lead to sudden cardiac death [27]. The involvement of fibrinogen in arrhythmogenesis has been attributed to pathways associated with the inflammatory cascade and hemodynamic alterations [27].

Our study also demonstrated that two liver function parameters, GGTP and bilirubin, were strongly associated with worse prognosis in patients with end-stage HF. End-stage HF is characterized by a prevalent cholestatic profile pattern characterized by increased bilirubin, GGTP, and ALP, whereas transaminase levels (hepatitis profile) usually remain within the normal range [28]. These abnormalities reflect the pathophysiology of liver function abnormalities in end-stage HF, which is associated with reduced cardiac output and the transmission of elevated central venous pressure to the hepatic venous system, leading to passive hepatic congestion [29, 30]. Another theory suggests that cardiac cirrhosis is not a simple reaction to chronic pressure and stasis but is associated with high-grade vascular occlusion due to intrahepatic thrombosis [29–32]. In individuals with cholestasis secondary to HF, GGTP is increased because of increased synthesis of the enzyme protein, and this increase is usually accompanied by an elevated bilirubin level [33]. Furthermore, GGTP, which is less specific for hepatobiliary injury than bilirubin, may not only reflect disease severity and hemodynamic perturbations but also be associated with systemic inflammation and oxidative stress [28]. In turn, an increase in total bilirubin in HF is also caused by impaired uptake of indirect bilirubin from blood, conjugation in hepatocytes, and secretion of direct bilirubin into bile. Moreover, hemolysis secondary to lung congestion may indirectly contribute to an increase in bilirubin levels [34].

In concordance with other studies, our findings demonstrated that elevated NT-proBNP concentrations were associated with worse prognosis in patients with end-stage HF [35–38]. NT-proBNP is a useful biomarker that supports diagnosing or ruling out HF as well as identifying patients with worsening HF [35–37]. Natriuretic peptides play an important role in HF by counteracting the effects of excessive stimulation of the sympathetic nervous system, renin-angiotensin-aldosterone system, and arginine-vasopressin system [39]. Moreover, natriuretic peptides influence cardiac remodeling through antihypertrophic and antifibrotic effects, including the inhibition of cardiac fibroblast proliferation, macrophage infiltration, and the secretion of proinflammatory factors. In the early stages of HF, natriuretic peptides play a protective role in maintaining homeostasis. However, as cardiac function deteriorates, the effectiveness of natriuretic peptides decreases due to decreased availability, increased removal, enzymatic degradation by neprilysin, or an increase in the proportion of inactive proBNP secreted from the heart [36, 39].

Limitations

Several limitations must be noted. This single-center study was limited to patients with HF of NYHA classes III and IV. The inclusion and exclusion criteria applied may have limi­ted the generalizability of our findings. Furthermore, blood samples were obtained only at the time of enrollment, but serial assessment of parameters during follow-up is required to accurately assess the prognostic value of the analyzed markers. A further limitation of the study is the lack of an independent validation cohort that would support our results. Although the cohort was relatively large for a single-center study, it can be considered small compared with the epidemiological scale of HF.

In conclusion, this study evaluated noninvasive and simple indicators associated with poor prognosis in patients with end-stage HF. Although limited to a single center, our study demonstrated that an elevated GDF-15 concentration, along with routinely used laboratory parameters such as fibrinogen, bilirubin, GGTP, NT-proBNP, and hs-CRP, is strongly associated with one-year mortality in patients with end-stage HF.

Article information

Conflict of interest: None declared.

Funding: This work was supported by an internal grant from the Medical University of Silesia in Katowice (grant no. BNW-1-047/K/3/K to BSJ).

Open access: This article is available in open access under Creative Common Attribution-Non-Commercial-No Derivatives 4.0 International (CC BY-NC-ND 4.0) license, which allows downloading and sharing articles with others as long as they credit the authors and the publisher, but without permission to change them in any way or use them commercially. For commercial use, please contact the journal office at polishheartjournal@ptkardio.pl