Acute Heart Failure developed as worsening of Chronic Heart Failure is associated with increased mortality compared to de novo cases

Acute heart failure (AHF) emerges either de novo or from worsening of chronic heart failure (CHF). The aim of the present study was to evaluate the association between worsening of CHF and mortality in AHF patients. Out of 152 included AHF patients, 47 (30.9%) were de novo AHF patients and 105 (69%) were AHF patients with worsening of CHF. The proportion dying in hospital (19.0% vs. 4.3%, p = 0.023) and within 3 months after hospitalization (36.6% vs. 6.7%, p < 0.001) was significantly higher in AHF patients with worsening of CHF. Logistic regression analyses also showed a significant positive association of AHF emerging as worsening of CHF with hospital mortality [odds ratio (OR) and 95% confidence interval (CI): 5.29 (1.46–34.10), p = 0.029] and 3-month mortality [8.09 (2.70–35.03), p = 0.001]. While the association with hospital mortality was no longer significant after adjusting for comorbidities and clinical as well as laboratory parameters known to be associated with mortality in heart failure patients, the association with 3-month mortality remained significant. We conclude that compared to de novo AHF, AHF evolved from worsening of CHF is a more severe condition and is associated with increased mortality.

Considering the persistent hemodynamic and metabolic burden imposed on CHF patients we hypothesized that the clinical and laboratory characteristics are worse and the mortality rate higher in AHF patients with worsening of CHF compared to de novo AHF patients. Therefore, we compared the clinical and laboratory characteristics of patients with worsening of CHF with those of de novo AHF patients and evaluated the association between worsening of CHF and mortality in AHF patients.

Results
Clinical characteristics. The patients' baseline characteristics have been described elsewhere 9 and are shown in Table 1. Out of 152 included patients, 105 (69.1%) were patients with worsening of CHF and 47 (30.9%) were de novo AHF patients. As seen in Table 1, both groups were similar regarding age, sex, smoking status, body mass index (BMI) and New York Heart Association (NYHA) functional classification. Compared to de novo AHF patients, patients with worsening of CHF had a significantly lower mean arterial pressure (MAP) and heart rate ( Table 1), whereas respiratory rate and the incidence of dyspnoea and ortopnoea were similar in both groups (Supplementary Table 1). Venous congestion, a consequence of right-sided HF, was more frequent in patients with worsening of CHF; this was exemplified by a significantly higher incidence of jugular venous distension (JVD), enlarged liver, and peripheral edema ( Table 1). The incidence of ascites was higher, albeit not statistically significantly in patients with worsening of CHF compared to de novo AHF patients (Table 1). Furthermore, systolic pulmonary artery pressure (SPAP) was significantly higher in patients with worsening of CHF compared to de novo AHF patients, whereas ejection fraction (EF) ( Table 1), the incidence of aortic insufficiency, aortic stenosis, and dilatation of right or left ventricle did not differ significantly between the groups (Supplementary Table 1).
Regarding comorbidities, the incidence of chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), and cardiomyopathy (CM) were significantly higher, and the incidence of acute coronary syndrome (ACS) was significantly lower in AHF patients with worsening of CHF compared to de novo AHF patients (Supplementary Table 2). In contrast, the incidence of hyperlipidemia, hypercholesterolemia, hypertension, type 2 diabetes mellitus (T2DM) as well as coronary and peripheral atherosclerosis was similar in both groups.
Laboratory parameters. The markers of kidney function urea and creatinine, were significantly higher and glomerular filtration rate (GFR) significantly lower in patients with worsening of CHF compared to de novo AHF patients (Table 2). While the N-terminal pro-Brain Natriuretic Peptide (NT-proBNP) and interleukin-6 (IL-6) levels were significantly higher, the levels of alanine aminotransferase (ALT), total cholesterol, low-density lipoprotein (LDL) cholesterol, and high-density lipoprotein (HDL) cholesterol, as well as leukocytes, and platelets were significantly lower in patients with worsening of CHF compared to de novo AHF patients ( Table 2). Levels of aspartate transaminase (AST), C-reactive protein (CRP), triglycerides and erythrocytes did not differ significantly between groups (Table 2). While the levels of total serum protein and albumin, as well as glucose, sodium, and potassium were not significantly different between groups, the chloride levels were significantly lower in patients with worsening of CHF compared to de novo AHF patients (Supplementary Table 3).   Mortality. Significantly more patients with AHF caused by worsening of CHF died in the hospital as well as within 3 months after hospitalization compared to de novo AHF patients (Table 3). In line with this, the univariable logistic regression analyses showed a significant positive association of worsening of CHF with hospital and 3-month mortality (Table 4). While this association did not remain significant for in-hospital mortality after adjusting for age, sex, BMI, NT-proBNP, MAP, GFR, and either urea, IL-6, and LDL cholesterol, or COPD, CKD, CM, ACS, and NYHA, the adjustment did not qualitatively change the results for 3-month mortality (Table 4).

Discussion
The present study clearly shows that AHF developed from CHF represents a more severe condition that is associated with higher mortality when compared to de novo AHF. Compared to de novo AHF patients, AHF patients with worsening of CHF presented with markedly higher concentrations of NT-proBNP, a marker of heart congestion and elevated left-sided filling pressure 10 , indicating a more pronounced left ventricular dysfunction, accompanied by a lower MAP and higher SPAP and in turn, more severe right-sided HF with consequently more severe venous congestion.
Interestingly, despite a higher incidence of JVD, enlarged liver and peripheral edema, the body weight in AHF patients with worsening of CHF was significantly lower than in de novo AHF patients. This finding most probably reflects a more pronounced impairment of the intestinal absorption of nutrients due to a more severe intestinal edema, as well as a more severe chronic inflammation, the postulated common link between CHF, wasting syndrome and hypoalbuminemia 8,11,12 . In line with this, our AHF patients with worsening of CHF had lower serum levels of albumin (albeit not statistically significant) and significantly higher concentrations of IL-6 when compared to de novo AHF patients.
It is generally accepted that the augmented translocation of bacterial endotoxins from the edematous intestine into the circulation 13 as well as insufficient binding and neutralization of endotoxines by circulating lipoproteins, whose levels are decreased in HF 14,15 , underlie the persistent inflammatory response in HF 16 . Additionally, the pro-inflammatory activation of venous endothelial cells by circumferential stretch due to congestion contributes to the chronic inflammation in HF 17 . Accordingly, lower serum levels of LDL cholesterol and HDL cholesterol together with a more severe congestion, in patients with worsening of CHF compared to de novo AHF patients may well explain the more pronounced inflammatory response, reflected by higher IL-6 levels.   The augmented inflammatory response, which is known to promote the formation and clearance of leukocyteplatelet complexes 18,19 , is most likely responsible for the lower leukocyte and platelet levels in our AHF patients with worsening of CHF.
It is well established that decreased cholesterol and lipoprotein serum levels are strongly and independently associated with increased mortality in HF 8,14 . In contrast to total, LDL, and HDL cholesterol, which were significantly lower in patients with worsening of CHF, triglyceride levels were similar in both studied groups.
Serum levels of lipids and lipoproteins are determined by their intestinal absorption, hepatic synthesis, and removal from the circulation by various catabolic routes. Accordingly, modulation of these physiological processes by the underlying AHF pathophysiology, such as congestion, which was more severe in AHF patients with worsening of CHF, may explain the different levels of total, LDL, and HDL cholesterol in the studied groups. Similar triglyceride levels in the two studied groups may indicate either that the triglyceride synthesis is not affected by the underlying AHF pathophysiology or that the decreased triglyceride synthesis is counterbalanced by a likewise decreased triglyceride catabolism. The latter possibility is likely, because lipoprotein lipase, a serum enzyme responsible for triglyceride degradation, is decreased by inflammation 20 , which indeed was more severe in AHF patients with worsening of CHF than in de novo AHF patients.
Renal dysfunction, also a strong predictor of mortality in HF 21,22 , was more pronounced in our AHF patients with worsening of CHF than in de novo AHF patients. This was reflected by a markedly decreased GFR and concomitantly increased serum levels of urea and creatinine. The reduction in GFR may be due to several factors including kidney hypoperfusion 23 , renal co-morbidities 24 , renal venous congestion 25,26 , or inflammatory cytokines such as IL-6 27 . Accordingly, a higher incidence of CKD, together with a more severe venous congestion and a more pronounced inflammatory response, as well as higher serum levels of NT-proBNP, an established predictor of worsening renal function in CHF 28 , may well explain the more severe impairment of renal function in our AHF patients with worsening of CHF compared to de novo AHF patients.
Besides their deleterious effects on the kidneys, inflammatory cytokines such as IL-6 exert a direct detrimental effect on the myocardial structure and function by promoting oxidative stress and endothelial dysfunction, thus contributing to the progression of HF 29 . Therefore, a more pronounced inflammatory state, reflected by higher IL-6 levels, seems to be an important underlying pathophysiology responsible for worse clinical status of our AHF patients with worsening of CHF compared to de novo AHF patients.
The above-mentioned differences between the groups regarding the clinical characteristics observed in our AHF patients with worsening of CHF translated into a higher mortality rate, whereby the impact on 3-month mortality was more pronounced than the impact on hospital mortality. This is in line with two recent reports showing a strong association of the chronicity of HF with 6-month and 3-year mortality, but only a weak association with 1-month mortality 30,31 . In the present study, the strong association of worsening of CHF with 3-month mortality was weakened, but remained significant upon adjustment for comorbidities and established risk factors in HF 8,32,33 . This implies that other features inherent to this group of patients underlie their increased mortality compared to de novo AHF patients.
There are several limitations to our present study: The design precludes conclusions on the pathophysiological mechanisms in terms of cause and consequence, or temporal development and onset of signs, and symptoms associated with worsened status and increased mortality. Furthermore, our data do not facilitate the examination whether the duration of HF chronicity affects mortality rates. Finally, the limited number of participants in this monocentric study influences the statistical power of our analyses. Therefore, further large studies are needed to confirm our results.
Based on our results, we conclude that AHF developed from CHF is a more severe condition than de novo AHF and is associated with higher mortality rates.  Laboratory procedures. The collection of blood samples, standard laboratory methods, and the determination of the NT-proBNP concentration have been described in our previous reports on our AHF cohort 34,37 . The levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were determined with the Abbott Architect c8000 (Chicago, IL, USA).

Statistical analyses.
Categorical data are shown as absolute and relative frequencies, whereas continuous data are summarized as median and range (minimum to maximum) due to the skewed distribution of many of the laboratory parameters. Patients with de novo AHF and with worsening of CHF were compared by Mann-Whitney U test or Fisher's exact test. The impact of worsening of CHF as compared to de novo AHF on hospital and 3-month mortality was assessed by logistic regression analysis. In addition to a univariable model, we also adjusted for age, sex, BMI, NT-proBNP, GFR, MAP, urea, IL-6, and LDL cholesterol, as well as for age, sex, BMI, NT-proBNP, MAP, GFR, COPD, CKD, CM, ACS, and NYHA, the comorbidities as well as clinical and laboratory parameters known to be associated with mortality in HF patients. We checked the variance inflation factor to prevent multi-collinearity among the covariates. Results are presented as odds ratios (OR) with the respective 95% confidence interval (CI). All analyses are exploratory and a p-value of <0.05 was considered significant. R version 3.4.2. was used for all statistical analyses.

Data Availability
All data generated or analysed during this study are included in this manuscript and accompanied Supplementary materia.