Differences in the Biomarker Profile of De Novo Acute Heart Failure versus Decompensation of Chronic Heart Failure

The perception of acute heart failure (AHF) as a single entity is increasingly outdated, as distinct patient profiles can be discerned. Key heart failure (HF) studies have previously highlighted the difference in both the course and prognosis of de novo AHF and acute decompensated chronic HF (ADHF). Accordingly, distinct AHF profiles with differing underlying pathophysiologies of disease progression can be shown. We compared a range of selected biomarkers in order to better describe the profile of de novo AHF and ADHF, including the inter alia—serum lactate, bilirubin, matrix metallopeptidase 9 (MMP-9), follistatin, intercellular adhesion molecule 1 (ICAM-1), lipocalin and galectin-3. The study comprised 248 AHF patients (de novo = 104), who were followed up for one year. The biomarker data of the de novo AHF and ADHF profiles was then compared in order to link biomarkers to their prognosis. Our study demonstrated that, although there are similarities between each patient profile, key biomarker differences do exist—predominantly in terms of NTproBNP, serum lactate, bilirubin, ICAM-1, follistatin, ferritin and sTfR (soluble transferrin receptor). ADHF tended to have compromised organ function and higher risks of both one-year mortality and composite endpoint (one-year mortality or rehospitalization for heart failure) hazard ratios (HR) (95% CI): 3.4 (1.8–6.3) and 2.8 (1.6–4.6), respectively, both p < 0.0001. Among the biomarkers of interest: sTfR HR (95% CI): 1.4 (1.04–1.8), NGAL(log) (neutrophil gelatinase-associated lipocalin) HR (95% CI): 2.0 (1.3–3.1) and GDF-15(log) (growth/differentiation factor-15) HR (95% CI): 4.0 (1.2–13.0) significantly impacted the one-year survival, all p < 0.05.


Introduction
Acute heart failure (AHF) is a multifaceted clinical condition with the potential to be life-threatening and requiring urgent treatment [1]. Although AHF is perceived as a single entity, it includes heterogenic populations of patients; thus, distinct patient profiles can be recognized [2]. The differences can be shown in several aspects of clinical presentation, such as: renal function, natriuretic response, volume status or neurohormonal activity [3][4][5][6][7][8]. We speculate that the pathophysiology underlying patients at various stages of the natural course of heart disease is different, at least in terms of the severity of some phenomena. Furthermore, the mechanisms that drive both the decompensation itself and poor outcomes may vary. This observation has critical practical implications, as the analyses of several key AHF clinical trials have shown differences in both the clinical course and prognosis of de novo and acute decompensated chronic heart failure (ADHF) [9][10][11].
We aim to provide a comprehensive picture of several selected biomarkers and compare them between patients experiencing their first episode of AHF (de novo AHF) with those who had heart failure diagnosed prior to hospital admission (ADHF).

Study Population
This was a single-center, observational study that was undertaken in the Centre of Heart Diseases, 4th Military Hospital, Wroclaw, Poland between January 2016 and September 2017. AHF was defined according to the European Society of Cardiology (ESC) guidelines criteria, and events were classified as new onset without a prior history of heart failure (de novo AHF) and ADHF [12,13]. All adult patients (≥18 years) hospitalized with AHF as the primary cause of hospitalization treated with intravenous furosemide at admission and those willing to participate (by signing an informed consent form) were enrolled. Exclusion criteria included the following: cardiogenic shock, diagnosis of acute coronary syndrome, known severe liver disease, end-stage renal disease requiring renal replacement therapy and evidence of infection. Patients were treated in accordance with the attending physicians' recommendations of the ESC guidelines.

Study Design
After admission to hospital, detailed information concerning patient demographics (including history of heart failure), clinical history, comorbidities, previous therapies and physical findings, were collected. Dyspnea was measured with the use of a self-reported 10-point Likert scale at the time of hospital admission, where 0 meant 'lack of dyspnea' and 10 points meant 'dyspnea of the worst severity/maximal dyspnea'. Clinical assessments alongside venous blood were obtained at baseline and the following subsequent time points: day 1 and day 2 and discharge. The samples were also collected, centrifuged and frozen (at −70 • C) for additional prespecified analyses. A current literature review was undertaken, and biomarkers were selected on the basis of their relation to the pertinent heart failure (HF) pathophysiological pathways. The selected biomarkers were involved in inflammation (CRP, IL-6, IL-22 and WBC); liver function (AST, ALT and bilirubin); perfusion and congestion (NTproBNP and lactate); iron status (Fe, total iron binding capacity, sTfR and ferritin) and cardiac remodeling (MMP-3, follistatin, selectin, lipocalin, PF4, myostatin, ICAM-1, GDF-15 and galectin-3). The properties of each marker and evidence in HF have been described elsewhere [14][15][16][17][18][19][20][21]; a brief description of each biomarker can be found in the Supplementary Materials. Information regarding the biomarkers of interest was not available to the treating physicians, and neither clinical nor therapeutic decisions were based on these results.

Study Outcomes
The clinical endpoints of the study were: 1.
Composite endpoint of one-year mortality and rehospitalization for heart failure.

Clinical Follow-Up
Discharged patients were tracked at the heart failure clinic for at least one year. Information regarding rehospitalizations and survival status was obtained either directly from patients or their relatives (telephone contact), from the HF clinic database, from the hospital system or from the national citizen registry by the investigators, who were blinded to the biomarker results. No patient was lost to follow-up.

Statistical Analysis
Continuous variables with a normal distribution were described using means ± standard deviation, variables with skewed distribution were described by medians with (upper and lower quartiles) and categorized variables were given as numbers and percentages. The statistical significance of the differences between groups was assessed using the t-test, Mann-Whitney U test or Kruskal-Wallis test. The normality of the distributions for these variables was evaluated by three different statistical tests: the Kolmogorov-Smirnov test, Lilliefors test and the W Shapiro-Wilk test. The Cox proportional hazards models were used to calculate the hazard ratio (HR) with the corresponding 95% confidence interval (95% CI) for all-cause mortality. Multivariable analyses were adjusted for: age, ejection fraction, systolic blood pressure at admission, hemoglobin, NTproBNP and blood urea nitrogen. Kaplan-Meier survival curves were constructed to demonstrate the survival. p < 0.05 was considered statistically significant. Statistical analyses were performed using STATISTICA 12 (StatSoft Polska Sp. z o.o., Krakow, Poland).

Comparison of Clinical and Basic Laboratory Characteristics
Patients with ADHF presented significantly lower LVEF 35 ± 1 compared to 40 ± 13% amongst the de novo AHF patients, p < 0.05. Higher levels of systolic and diastolic blood pressure on admission were observed in the de novo AHF patients (145 ± 33 vs. 126 ± 27 mmHg and 85 ± 16 vs. 75 ± 15 mmHg, all p < 0.001,respectively), alongside a higher average heart rate (95 ± 25 vs. 87 ± 23 bpm, p = 0.007) in comparison to ADHF. There was no statistically relevant difference in the length of hospitalization and reported dyspnea between both groups. Lower levels of sodium were observed amongst ADHF patients (138 ± 5 vs. 140 ± 4 mmol/L, p < 0.01) ( Table 1).

Comparison of Comorbidity and Risk Factors
Of the comorbidity and risk factors we observed in the study population, CAD was significantly more prevalent in ADHF patients (69.4%) than in the de novo group (38.5%), p < 0.01. Conversely, hypertension was more frequent in the de novo group (82.7% vs. 77.8%), albeit not significantly, p = 0.34 (Table 1).

Comparison of Selected Biomarkers-De Novo AHF and ADHF
No significant difference between selected biomarkers concerning inflammation (including CRP, IL-6 and IL-12) could be observed between the de novo AHF and ADHF groups (Table 3). Of the liver function tests we analyzed, the total bilirubin at admission was significantly higher in the ADHF group (1.6 ± 1.6 mg/dL) in comparison to the de novo AHF group (1.2 ± 0.8 mg/dL, p = 0.011). The majority of biomarkers associated with remodeling did not demonstrate significant differences between the two patient profiles. One exception here was ICAM-1, which was significantly lower in the de novo group 367.5 ± 152.4 vs. 416.7 ± 195.8 ng/mL, p = 0.035. Perfusion and congestion markers in patients with de novo AHF and ADHF differed-mean (±SD) lactate concentrations at discharge were 1.9 ± 0.6 and 2.1 ± 1.1 mmol/L. The median NTproBNPs at admission were 5108 (2593-11,579) and 5797 (3829-11,920) pg/mL, p = 0.06, and the median NTproBNPs at discharge were 2743 (1531.5-4798.5) and 3601 (2084-7284) pg/mL, p < 0.01, respectively.

Comparison of Selected Biomarkers between Patients Who Experienced an Event (Death or Heart Failure Rehospitalization, Whichever Occurred First) and the One-Year Event-Free Group
Lower lactate levels at day 1 were observed in the event-free group 2.0 ± 0.6 vs. 2.3 ± 1.7 mmol/L, p = 0.021; likewise, the NTproBNP levels for the event-free group were lower at admission: 5080 (2944-9101) vs. 6312 (4083-13,944) pg/mL, p < 0.004 and at discharge: 2636 (1499-4802) vs. 4318 (2616-8210) pg/mL, p < 0.0005. NGAL differed significantly between profiles where we observed lower levels in the event-free group: 77.9 ± 51.8 in comparison to 95.7 ± 55.6 ng/mL, p = 0.039 in death or rehospitalized patients. Concerning the iron status, sTfR was lower in the event-free group compared to the death or rehospitalization group: 1.9 ± 0.7 and 2.2 ± 0.9 mg/L, p = 0.006, respectively (Table 4).

Differences in One-Year Outcomes between De Novo AHF and ADHF
There were 66 (31.3%) deaths and 88 (41.7%) patients who experienced the composite study endpoint (death or rehospitalization for HF, whichever occurred first) within one year of observation. In the multivariable analyses, the ADHF group had a significantly higher risk of both endpoints, hazard ratios (HRs) (95% CI): 3.4 (1.8-6.3) and 2.8 (1.6-4.6), respectively, both p < 0.0001. The Kaplan-Meier curves for both endpoints are presented in Figure 1a

Discussion
The principal finding of our study is that de novo AHF and ADHF patients may present themselves with numerous comparable clinical characteristics, leading to the categorization of AHF as a single entity. However, several variances in biomarker profiles can be recognized in AHF, which may be linked to differentiation in the pathophysiology of the syndrome at different stages of the disease and prognosis [22]. Surprisingly, a few of the examined biomarkers did not reveal significant differences, which may actually be related to the fact that a number of changes may be more quantitative than qualitative during the course of the disease.
There were numerous similarities in the clinical and laboratory characteristics of de novo AHF and ADHF patients. In our study, age and dyspnea were equable between both groups, as were the parameters of hemoglobin, white blood cell and platelet counts. Interestingly, a previous meta-analysis has indicated that a lower age of de novo AHF patients is likely; however, we did not observe such an association [23]. The inflammatory biomarkers, like C-reactive protein, IL-6 and IL-22, were also comparable between both groups, alongside several remodeling markers. However, the ADHF patients were more likely to have a reduced LVEF and a greater likelihood of ischemic heart disease, whereas higher blood pressure was more prevalent in de novo AHF patients.
In terms of comorbidity, CAD was significantly more frequently observed in the ADHF group, whereas de novo patients were more likely to be hypertensive. ADHF patients were more likely to have been administered either an ACEI/ARB, beta-blocker or MRA prior to administration, and this trend continued during hospitalization.
We also showed that the organ functions differed between the two patient profiles, with ADHF likely to have worse organ functions in general. This observation was not surprising, as multiorgan dysfunction is a well-recognized marker of poor outcome in AHF [7,[24][25][26][27][28]. It has been reported previously that elevated bilirubin is an ominous sign

Discussion
The principal finding of our study is that de novo AHF and ADHF patients may present themselves with numerous comparable clinical characteristics, leading to the categorization of AHF as a single entity. However, several variances in biomarker profiles can be recognized in AHF, which may be linked to differentiation in the pathophysiology of the syndrome at different stages of the disease and prognosis [22]. Surprisingly, a few of the examined biomarkers did not reveal significant differences, which may actually be related to the fact that a number of changes may be more quantitative than qualitative during the course of the disease.
There were numerous similarities in the clinical and laboratory characteristics of de novo AHF and ADHF patients. In our study, age and dyspnea were equable between both groups, as were the parameters of hemoglobin, white blood cell and platelet counts. Interestingly, a previous meta-analysis has indicated that a lower age of de novo AHF patients is likely; however, we did not observe such an association [23]. The inflammatory biomarkers, like C-reactive protein, IL-6 and IL-22, were also comparable between both groups, alongside several remodeling markers. However, the ADHF patients were more likely to have a reduced LVEF and a greater likelihood of ischemic heart disease, whereas higher blood pressure was more prevalent in de novo AHF patients.
In terms of comorbidity, CAD was significantly more frequently observed in the ADHF group, whereas de novo patients were more likely to be hypertensive. ADHF patients were more likely to have been administered either an ACEI/ARB, beta-blocker or MRA prior to administration, and this trend continued during hospitalization.
We also showed that the organ functions differed between the two patient profiles, with ADHF likely to have worse organ functions in general. This observation was not surprising, as multiorgan dysfunction is a well-recognized marker of poor outcome in AHF [7,[24][25][26][27][28]. It has been reported previously that elevated bilirubin is an ominous sign in heart failure [26,[29][30][31][32]. In our study, we observed significantly higher values of total bilirubin at admission in ADHF patients. On the other hand, in terms of kidney function, the ADHF profile tended only to have higher levels of creatinine and higher blood urea nitrogen, yet serum sodium was at more favorable levels in de novo AHF.
Our recent paper showed that NTproBNP is a correlate of lactate, an organic acid produced during anerobic cell metabolism and a marker of perfusion mirroring the energetic challenges AHF presents to the body [33]. We also showed that elevated lactate at the admission of AHF patients without over hypoperfusion is a marker of a poor outcome [34]. The lactate levels of the ADHF group at all time points (admission, day 1 and discharge) were numerically higher than that of the de novo AHF patients, albeit only at discharge was the difference statistically significant. In order to better understand whether differences in the biomarkers between our two groups had an influence upon the subsequent course of the disease, we subsequently divided the patient cohort into two groups: those who went event-free through follow-up and those who experienced either death or HF rehospitalization. Lactate and NTproBNP were clearly associated with a poor outcome in that comparison.
A link between lactate and iron levels in HF has also recently been suggested, which is perhaps expected given the well-described role iron plays in cell energy production [35]. Soluble transferrin receptor (sTfR) concentrations are relative to the cellular iron demand, and raised levels are indicative of iron deficiency [36][37][38]. In our study, the ADHF group had significantly higher sTfR when compared to the de novo group, which may link iron metabolism with the natural progression of the disease.
Of the several remodeling markers we examined, we were surprised to see that only values for Intercellular Adhesion Molecule 1 (ICAM-1) presented a significant difference, as these values were increased in the ADHF group. ICAM-1 belongs to the immunoglobulin superfamily and enables the binding of leukocytes to endothelial cells [39]. Studies concerning cardiac remodeling have identified ICAM-1 as a critical component of leukocyte tissue infiltration in the left ventricle [14,39]; thus, our ICAM-1 results were expected, considering the ADHF group had a significantly impaired LVEF. Furthermore, we observed the opposite trend of higher follistatin levels in the de novo AHF patients. Follistatin is known to play a role in the inflammatory cascade, and as far as we are aware, no such studies exist concerning the relationship of follistatin and ICAM-1 in AHF. We were surprised to see that the remodeling markers between both AHF groups were generally alike. Future research is needed to evaluate the potential relationship between follistatin and ICAM-1; additionally, a wider evaluation of the remodeling biomarkers, such as copeptin, Monocyte Chemotactic Protein-3 and ST-2, and their roles in the patient profiles of de novo AHF and ADHF are needed.
Interestingly, when we evaluated the patient data in terms of those who went eventfree or experienced our composite endpoint (death or rehospitalization), only lipocalin/ NGAL was significantly different, with increased serum concentrations in the death or rehospitalization group. The ADHF group had significantly worse outcomes over one year when compared to the de novo group. This may be the result of compromised organ function and a worse baseline status (particularly, liver, kidney and left ventricular function) in the first group. Importantly, several of the biomarkers examined did influence the outcome in our population, such as: lactate, sTfR, Lipocalin/NGAL and GDF-15. Galectin-3 impacted upon the in-hospital mortality in our population. This marker has already been shown to impact the prognosis of ambulatory chronic HF patients [40]. On the other hand, we were surprised to see that some molecules (with a high theoretical/pathophysiological potential to be linked with disease progression and outcome) did not reveal any influence upon the patient outcomes [41][42][43][44].
This study was not without its limitations. Firstly, it was a single-center post hoc analysis with a limited sample size drawn only from the Polish resident population. Secondly, the analyzed biomarkers represent an area of possible selection bias, and finally, as with many similar studies, male patients were overrepresented. Moreover, the study was calculated and powered to reveal differences in many (i.e., bilirubin, lactate, NTproBNP, sTfR, ferritin, ICAM-1, IL-22, follistatin and MMP-9) but not all markers of interest; thus, one needs to take into consideration that the lack of the differences in some cases may actually be a result of study underpower. Confirmation of the differentiation between de novo AHF and ADHF is needed via multi-center studies and a larger number of patients.

Conclusions
Numerous similarities existed between the profiles of de novo AHF and ADHF. Age and dyspnea between both groups were comparable, alongside the inflammatory and nearly all remodeling biomarkers (with the exception of ICAM-1). However, only superficially is AHF a single entity, as during the natural course of AHF, the pathophysiology underlying patients may differ, and therefore, the distinct biomarker profiles of de novo AHF and ADHF can be discerned. In particular, organ function biomarkers (bilirubin, creatinine and blood urea) congestion and perfusion (NTproBNP and lactate) presented more unfavorably in ADHF alongside higher rates of CAD, resulting in a lower probability of survival. De novo AHF, an early stage of the disease, is more hypertensive in profile, with fewer complicating factors. Increased lipocalin/NGAL was associated with death or rehospitalization during follow-up.