Introduction
Heart failure (HF) is a leading cardiovascular problem worldwide and increasing evidence demonstrates that advanced glycation end products (AGEs) play a pivotal role in the development and progression of the disease (see for a review [
1]). AGEs are generated non-enzymatically. Following their interaction with the receptors for AGEs (RAGE), a series of events leading to vascular and myocardial damage are elicited, resulting in diastolic and systolic dysfunction. Some of the molecular mechanisms involved in these events include oxidative stress, increased inflammation, enhanced extracellular matrix accumulation, and renal damage. Consequently, AGEs and soluble RAGE (sRAGE) levels are related to the severity [
2,
3] and the bad prognosis [
3‐
5] of HF, and even to the development of post-infarction HF [
6].
The association between sRAGE and incident HF has been recently examined in a prospective study (The Atherosclerosis Risk in Communities Study) concluding that lower circulating levels of sRAGE are independently associated with the development of HF in a community-based population [
7]. However, little is known about the evolution of sRAGE and AGEs levels during HF progression, so prospective studies are needed for this aim. Some data suggest AGEs accumulate during cardiovascular disease progression, mainly in the context of renal failure [
8]. It has also been reported that the amount of AGEs in cardiomyocytes increases significantly in both diabetes and HF [
9].
Evolution and progression of HF is also related with obesity and body composition. Extensive evidences demonstrate the adverse effects of obesity on central and peripheral hemodynamics, as well as on cardiac structure and function [
10]. This explains the positive relationship between body mass index (BMI) and the risk of incident HF [
11] or cardiovascular disease [
12]. However, despite the adverse effects of obesity on cardiac structure and function, numerous studies have suggested that obese patients with HF have a better prognosis than non-obese patients, considering obesity as a category of BMI [
13], a high percentage of body fat [
14], or a high waist circumference [
15]. This is the so-called “obesity paradox”. These data have suggested a possible protective role of body fat on HF that can be related with the fact that cardiac cachexia is related with cardiac dysfunction. All together suggests that even the weight loss recommendation usually given to obese patients with HF should be taken with care and new studies of body composition in patients with HF should be made.
AGEs could participate in the complex relationship between obesity and the outcome in patients with HF [
16]. Although the underlying molecular mechanisms are far from clear some possibilities have been suggested. On one hand, the combined effects in obesity of enhanced food consumption, low-energy expenditure, hyperglycaemia, hyperlipidaemia and increased oxidative stress may boost the formation of AGEs. Moreover, AGEs could be formed during the maturation of adipocytes [
17]. AGE-RAGE mediated activation may stimulate inflammatory signalling in adipose tissue, resulting in dysregulation of adipokines and contributing to the development of obesity-related complications [
17]. On the other hand, AGEs are possible targets of the protective role of body fat. Gaens et al. [
18] suggested that lower levels of circulating AGE (measured as carboxymethyl lysine) in obesity were associated with their trapping into the adipose tissue resulting in lower levels of circulating AGEs. This is in accordance with central obesity being characterized by chronic low grade inflammation and low levels of circulating AGEs [
19]. In subjects on hemodialysis, those who naturally accumulate AGEs, the changes in skin autofluorescence (a measurement of AGEs) over 1 year period were related to BMI [
20]. Accordingly, sRAGE levels, which in some conditions could act as a decoy of circulating AGEs, are inversely related with BMI in severely obese subjects [
21], in young adults [
22] and in subjects with metabolic syndrome, both adults [
23] and adolescents [
24].
Taking all these questions into account, our main objective was to analyze, for the first time, the role of AGEs and sRAGE levels in the progression of HF after an episode of acute HF decompensation in relation with other metabolic parameters like body composition, body fat distribution and nutritional status.
Discussion
In this work, the evolution of AGEs and sRAGE levels during the 6 months following an acute HF episode of decompensation is studied for the first time. AGEs and sRAGE levels continuously increased in this period, but the increase of AGEs was mainly observed in those patients with incident HF. This relation with the evolution of HF is a novel finding of this work. It was also confirmed that sRAGE levels are predictive biomarkers of bad HF prognosis independent of age, gender, body mass index and other risk factors, in a 1-year follow-up period. Interestingly, high AGE1m levels after the acute decompensation are also indicative of a worse clinical outcome in terms of cardiovascular death and HF readmission. Finally, AGEs and sRAGE levels showed a direct association with clinical malnutrition (CONUT score), but an inverse relation with BMI or the percentage of body fat.
Basal levels of AGEs were related to the severity of the HF in terms of NYHA functional classes at the time of hospital admission, but not in relation with preserved or reduced ejection fraction. On the contrary, sRAGE did not show this relationship with HF severity. These results are different than those previously observed in patients with chronic HF, where sRAGE levels correlated with NYHA class [
2,
5]. However, it should be taken into account that in the present study we are analyzing patients in the moment of the acute decompensation by HF, most of them in class III and IV, and that 64% of these patients are cases of incident HF. In the previously mentioned studies the patients presented chronic and stable HF with a more homogeneous distribution over the NYHA classes. In this sense, the influence of the inflammatory status of each patient in the moment of sampling could affect AGEs or sRAGE levels. However, the levels measured in this work for PAI-1 and resistin, two cytokines that have been related with the sub-clinical inflammatory status [
31,
32], were no different regarding the cutoff point of sRAGE
0, and they were contradictory regarding the cutoff point of AGE
1m: whereas resistin was higher in the group of AGE
1m > 40 a.u., PAI-1 was lower.
Moreover, we did not observe differences in sRAGE levels between patients with preserved or reduced LVEF as Willensen et al. [
33] did, but we considered reduced LVEF as <50% whereas they used <40%. A possible explanation for these differences could arise from one of the novel findings of this work: AGEs and sRAGE levels increase during the development of HF. At least during 6 months following an acute decompensation by HF, both biomarkers continuously increased their values, being this event more evident for AGEs in the case of incident HF. Therefore, the levels of these biomarkers seem to be influenced by the stage of the HF. Our data show that AGEs levels increase noticeably after an acute episode of incident HF, whereas sRAGE seems to react in the same way, but with a slower rate, which is why the increase is better observed in patients with previous HF. This means that AGEs and sRAGE levels are not only biomarkers of the severity of HF, but also of its progression. This is an important point since the rest of the biomarkers measured in the study, all related with diabetes and obesity, showed the opposite relationship with acute HF. They all presented a marked reduction from their elevated levels during hospitalization, to lower levels during the months following the discharge.
AGE-RAGE axis seems to be related to the nutritional status of patients since CONUT scores indicating risk of malnutrition were directly associated with AGE and sRAGE levels. This is an interesting and new finding, which is in accordance with the inverse relationship observed between AGE and sRAGE levels with respect to parameters of obesity such as BMI or BF percentage. We previously commented that sRAGE levels were inversely related to BMI in severely obese subjects [
21], in young adults [
22] and in subjects with metabolic syndrome either adults [
23] or adolescents [
24]. More recently, the inverse relationship between sRAGE and BMI was observed in healthy women [
34]. Soluble RAGE levels were also inversely associated with fat mass, and in patients with prediabetes [
35], where it also was negatively correlated with body weight and waist and hip circumferences. As we will discuss below, high sRAGE levels are indicators of bad prognosis in the population under study. Therefore, these findings support the hypothesis of the so-called “obese paradox”, by which obesity and particularly body fat, have shown to be protective in the context of HF [
14]. In this case, high BMI and BF correlate with lower sRAGE levels, predicting a better outcome than in patients with higher sRAGE levels and lower BMI and BF.
The relationship between AGEs and obesity seems to be more complex. According to previous data [
17], AGEs accumulation in adipose tissue could explain our findings about the inverse relationship between AGEs levels and BMI, which agrees with previous results [
18,
19], but disagrees with Amin et al. [
36], that found a positive relation between AGEs, measured as carboxymethyl lysine, and BMI. Although the mechanisms are controversial, AGEs can be absorbed from diet. In this sense, whereas low-AGE diets can increase insulin sensitivity [
37], high-AGE diets are related with increased fat intake and higher risk of abdominal obesity [
38]. However, metabolic syndrome can make the difference: serum AGEs levels and AGEs consumption from diet are higher in obese people with metabolic syndrome than in obese without metabolic syndrome [
39]. Once in the body, RAGE mediates AGE accumulation in adipose tissue [
18], and RAGE activation can induce the production of inflammatory mediators in adipocytes that could ultimately lead to dysregulation of adipokines in obesity [
17].
Present data confirm previous findings about the predictive value of AGE-RAGE axis for worse prognosis in HF for AGEs [
3,
6] and sRAGE [
4,
5] levels. Calculated cut-off values for AGE
1m and sRAGE
0 served to predict death and HF readmission in the univariate analyses. The best regression model included hyperlipidemia, NT-proBNP levels and more than 40 a.u. of AGE
1m as predictors of death or HF readmission after acute HF, adjusted by age, gender, renal function, diabetes mellitus and LMI. These results partially agree with those recently reported by Willensen et al. [
33] where AGE predicted hospitalization for HF and the combination of mortality and HF hospitalization whereas sRAGE did not predict events with statistical signification. Apart from the direct molecular damage of the activity of AGE-RAGE axis in cells, tissues and organs of the cardiovascular system [
1], high AGE levels in plasma severely affect kidney function. This damage is always reflected by the positive correlation between AGE and sRAGE levels and parameters indicating bad renal function. Renal impairment is common in HF, and strongly associated with poor outcome [
40], so this could be one reason for the bad prognostic value of AGE and sRAGE in this study.
Another possible reason is the relation with energetic metabolism and obesity in our population. We have previously commented the relationship between AGE and sRAGE levels with malnutrition. Moreover, high AGEs levels were related to higher concentrations of c-peptide and resistin in plasma and to lower concentration of PAI-1, suggesting some relation with the modulation of insulin secretion and insulin resistance for AGEs. PAI-1 can respond to increased inflammation and fibrosis, but also to obesity or metabolic syndrome [
41]. Intriguingly, low levels of insulin were observed in patients with higher levels of sRAGE. The group of patients with events showed higher levels of glucagon and GLP-1, suggesting an enhanced gluconeogenesis and aside from this, a possible implication of increased energy expenditure and decrease of adipose tissue expansion in this group of patients [
42].
Finally, although do not directly supported by our data, we cannot forgot that AGEs have been demonstrated to participate in the structural modification and functional alteration of the extracellular matrix proteins in vessels and myocardium, as well as in the intracellular signals mediated by calcium in cardiovascular cells [
1,
43]. All these mechanisms would also contribute to explain the pathophysiological role of AGE/RAGE axis on HF.
Conclusions
As the conclusion of the work, both AGEs and sRAGE seem to be fine biomarkers of bad prognosis for HF. However, the main novelty is related with the evolution of AGE/RAGE axis in relation with HF. AGEs and sRAGE levels increase at the beginning of the pathology, being useful for following HF progression. Since this increase seems to be related with clinical malnutrition and negatively correlated with body fat quantity, these parameters could serve to modulate their levels in a trend to improve HF clinical outcome.
Limitations
The main limitation of our study is the number of patients included. This reduces the potency of the statistical analysis. Further research in large populations is needed to confirm our findings. Secondly, our study was made only in patients after an acute episode of HF, so, the results can be different in patients with chronic and established HF. Thirdly, the culprit mechanism of the acute episode of HF is unknown in most of the patients, so differences between patients with regard to this could not be explored. Fourthly, sampling during the follow-up occurred at only three time points, hence it is not possible to know what happens in the periods in between. Fifthly, AGE measurement was done for a group of modifications, as explained, not for all the types of AGEs or for only a specific AGE, so, 1) the results cannot be extrapolated to all the AGEs and 2) the specific effect of one single AGE of the type of AGEs measured cannot be extracted from the whole.
Authors’ contributions
BPD, AFT, DBT contributed to sample acquisition and processing and biochemical analyses, RGF, RMA, AVR contributed to patients’ inclusion, medical examination and follow-up, AICP, MCC, FFC contributed to the analysis of body composition by DEXA, SE, EA and JRGJ contributed to the data analysis and conception of the work. All listed authors have contributed substantially to the work, participated in the writing of the manuscript. All authors read and approved the final manuscript.