Relationship between Fetuin-A and vascular or valvular calcification in hemodialysis patients

Methods: The study was conducted on 122 MHD patients. Serum levels of calcium, phosphorus, parathormone, and Fetuin-A were tested. Intima-media thickness (IMT) ≥0.8 mm and the presence of stenosis>50% or plaques in carotid doppler sonography were considered as vascular calcifi cation. Calcifi cation of cardiac valves or mitral annular calcifi cation in twodimensional echocardiography, were considered as cardiac valvular calcifi cation. The presence of any or both of the two conditions was defi ned as cardiovascular calcifi cation (CVC).


Introduction
Cardiovascular diseases account for 50% of mortality in patients with end-stage kidney disease (ESKD) [1]. The prevalence of these diseases among hemodialysis patients is about 10-30 times higher than that among general population of the same age and sex [2]. The high prevalence of cardiovascular events in ESKD patients is associated with traditional risk factors including hypertension, dyslipidemia, diabetes mellitus, and smoking. However, all cases of mortality in these patients are not explained by these factors, suggesting the presence of some other risk factors such as cardiovascular calci ication (CVC) [3,4]. that in non-dialysis patients. In addition, the prevalence of cardiac valvular calci ication in dialysis patients is more than 50% [7]. It has been shown that vascular calci ication is an independent predictor of morbidity and mortality in ESKD patients [8].
Vascular calci ication is not only caused by passive precipitation of calciumphosphate in vascular wall but is also due to an active process caused by induction of the expression of bone-associated genes in vascular smooh muscle cells (VSMC) [9,10]. A number of calci ication inhibitors such as Fetuin-A, Matrix Gla protein (MGP) and Osteoprotegerin (OPG) are able to prevent soft tissue calci ication [10].
Vascular calci ication may happen in two main types of intimal calci ication, as a part of diffuse atherosclerosis, and medial calci ication, which is generally focal in distribution, unrelated to atherosclerotic risk factors, and seen in younger hemodialysis patients. Pathogenesis may be genetic, mineral metabolism-related, or non-mineral metabolism related [10].
Fetuin-A or a2-Heremans-Schmid glycoprotein is produced in the liver and acts as a strong inhibitor of calcium -phosphate deposition [10]. Fetuin-A can regulate several of the key cellular events that lead to VSMC calci ication, including apoptosis, vesicle calci ication, and phagocytosis [11]. It interacts directly with matrix vesicle release and also forms stable colloidal spheres with calcium and phosphorus, so-called "calciprotein particles" which inhibit hydroxyapatite precipitation and hence may modulate vascular calci ication processes locally and at early stages [10] In patients on maintenance hemodialysis (MHD) the level of Fetuin-A is lower than that in healthy population and various studies have found an association between low levels of Fetuin-A and coronary artery calci ication, the presence of carotid plaques, valvular calci ication, and mortality in ESKD patients [12][13][14][15]. However, some studies have failed to ind a direct relationship between the reduction in serum level of Fetuin-A and calci ication in chronic kidney disease (CKD) or hemodialysis patients [16][17][18]. It has also been reported that Fetuin-A level is inversely associated with increased risk of cardiovascular mortality in hemodialysis patients [19]. However, this has not been proved in a number of other studies [18][19][20]. Thus, apparently further studies are required to investigate the complex relationship between Fetuin-A level and CVC in dialysis patients.
The majority of previous studies investigating the role of Fetuin-A in CVC have been focused on coronary artery calci ication and few studies have investigated the relationship between Fetuin-A and valvular calci ication and carotid intima media thickness (IMT) [21][22][23]. Also, the results of such studies have been controversial. Accordingly, the aim of this study was to examine the relationship between Fetuin-A levels of a cohort of MHD patients with cardiac valvular and vascular calci ication.

Materials and Methods
This cross-sectional study was conducted in hemodialysis ward of Hasheminejad kidney Center in Tehran in 2013. Inclusion criteria were being over 18 years old and undergoing MHD for at least 3 months. Exclusion criteria were a recent history of trauma, acute coronary syndrome, acute infection, and acute hepatitis which could have caused an increase in serum in lammatory markers. In addition, lack of consent to participate in the study was considered as an exclusion criterion. After explaining the objectives of the study to 179 MHD patients, 122 patients consented to participate in the study. The study received an ethical approval from the ethics committee of Hasheminejad Kidney center.
The demographic data (age, gender), cause of ESKD, duration of dialysis (dialysis vintage) and monthly laboratory results were extracted from the software of our dialysis ward, i.e., Hemodialysis Data Processor Software (HDPS), AIP Company, 2014, Tehran. Laboratory tests included serum hemoglobin, creatinine), blood urea nitrogen, sodium, potassium, calcium, phosphorus, parathyroid hormone (iPTH), vitamin D, iron, total iron binding capacity, ferritin, lipid pro ile (cholesterol, triglycerides, low density lipoprotein and high density lipoprotein), albumin, uric acid, and highly sensitive C reactive protein (hsCRP). Fetuin-A was measured by using ZB-1386-H9648 kit through ELISA test.
Vascular calci ication was examined though performing doppler sonography of the right and left carotid arteries. Carotid intima-media thickness (IMT) was measured at three points of the artery (upper, middle and lower) on each side and the calculated mean value was considered as IMT value. An IMT≥ 0.8mm on either side was considered as increased IMT [24]. A stenosis of more than 50%, the presence of any hypoechoic or calci ied plaque, or increased IMT were considered as vascular calci ication. Finally, using two-dimensional echocardiography, calci ication of the mitral or aortic valves and mitral annular calci ication (MAC) were determined and the presence of any of these was considered as valvular calci ication. The presence of any vascular and/or valvular calci ication was considered as CVC. The laboratory indings were compared between the patients with and without vascular calci ication, valvular calci ication, and CVC and the predictors of CVC were determined.
For statistical analysis IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp was used. Frequency and percentage were used to describe the qualitative variables and mean and standard deviation (SD) or median (interquartile range), were used to describe the quantitative variables, where appropriate. To compare quantitative and qualitative variables between the two groups with and without calci ication chi-square test and independent t test or Mann-Whitney U test were used, respectively. In addition, logistic regression was used to determine the predictors of CVC. P value less than 0.05 was considered as signi icance.
Tables 4 to 6 compare the characteristics of the patients with and without vascular calci ication, valvular calci ication and CVC, respectively. As compared with the group without calci ication, the patients with any type of calci ication were older, were more affected by diabetes and had longer dialysis vintage and lower levels of serum creatinine. In addition, patients with vascular calci ication also had signi icantly higher level of hemoglobin and lower level of serum uric acid. The patients with valvular calci ication and CVC had also signi icantly lower level of calcium. There was no difference in mean Fetuin-A or the percentage of patients with reduced Fetuin-A in any type of calci ication. There was also no difference between the levels of hsCRP, albumin and ferritin, assumed as indicators of in lammation, as well as serum phosphorus, iPTH, Vitamin D and lipids between the 2 groups.
In order to determine the predictors of CVC we used logistic regression analysis. Accordingly, the presence or absence of calci ications was considered as the outcome variable and the variables of age, dialysis vintage, presence or absence of diabetes, and the results of laboratory tests were considered as predictors. The predictors were entered into the model via Forward Conditional method. Table 7 shows the inal predictors (Nagelkerke R Square = 0.539). Increased risk of calci ication was seen with a chance of 1.1 times for each year of increase in age [odds ratio (OR) = 1.1, CI 95% = 1.1 -1.2], a chance of 2.8 times for every unit of increase in calcium (OR = 2.8, CI 95% = 1.1 -7.6), and a chance of 7.4 times with existence of diabetes (OR = 7.4, CI 95% = 1.1 to 47.4).

Discussion
According to the indings of this study, 87% of our MHD patients had either vascular or valvular, i.e. cardiovascular calci ication. Comparison of different characteristics between patients with and without calci ications showed that patients with calci ication were more likely to be older, have longer dialysis vintage, and have diabetes mellitus. They also had lower serum creatinine and higher calcium levels. However, there was no signi icant in Fetuin-A level between the patients with or without calci ication. The results of logistic regression analysis showed that age, serum calcium level, and diabetes mellitus were the most important predictors of calci ication in these patients.
Calci ication of the aortic and mitral valves is a common inding in MHD patients with a prevalence of four to ive times higher than that in the general population [25][26]. In the study by Ribeiro et al., the prevalence of mitral and aortic calci ication were 44.5% and 52% in patients with CKD-5D, and were higher than those with normal renal function (10% and 4.3%, respectively) [26]. Raggi et al. reported a prevalence of 45% and 34%, for mitral and aortic valve calci ications in patients with ESRD, as detected by CT scan. In addition, the prevalence of simultaneous involvement of both valves was 21% [25]. In the study by Rao et al, the prevalence of MAC in dialysis patients was 35% [27]. Sayarlioglu et al. studied 129 hemodialysis patients and their results showed a 23.3% prevalence of mitral valve calci ication, a 21.7% prevalence of aortic valve calci ication and the overall 33.3% prevalence of valvular abnormalities was [28]. In our study, the prevalence of mitral calci ication, aortic calci ication, and MAC, were 57%, 54%, and 55% respectively, and the overall valvular calci ication was 73%, which is higher than the rates reported by other studies. This higher prevalence might be attributed to several factors such as differences in the study populations, especially regarding mean age and dialysis vintage, different de initions and diagnostic methods for detection of CVC, the type of phosphate binders used and dialysate calcium concentration. We have been using Calcium carbonate as the main phosphate binder in the country till about 6 years ago, when sevelamer was covered by insurance companies and this could have been an explanation for high rate of vascular calci ication. On the other hand, the concentration of dialysate calcium is low at 2.5 mmol/L in our country and may be bene icial regarding causing less calci ication. Further studies on cumulative doses of calcium-based vs. non-calcium-based phosphate binders may help to determine the effect of oral calcium supplements and phosphate binders on vascular calci ication.
According to previous studies, dialysis patients with vascular calci ication are older and have longer duration of dialysis [29][30]. Our study also con irmed that patients with any type of calci ication were older and had a longer dialysis vintage, though dialysis vintage lost its effect in our regression analysis. These indings suggests that age, as a major atherosclerotic risk factor plays an important role in calci ication of dialysis patients and also longer exposure to the uremic milieu has a detrimental role in provocation of vascular calci ication.
Factors such as diabetes mellitus, disorders of metabolism of calcium and phosphorus including high serum calcium level, high calcium-phosphate product, and in lammation have been introduced as important predictors of CVC in these patients [26,28,[31][32][33][34]. In our study hsCRP, ferritin and (low) albumin, assumed as markers of in lammation, had no correlation with vascular calci ication. It is possible that more speci ic markers of in lammation such as pre-albumin and Il6 would have shown the correlation between vascular calci ication and in lammation. On the other hand, higher levels of calcium was a predictor of CV in both univariate and logistic regression analysis. Other studies have mentioned the relationship between high calcium intake, high calcium-phosphorus product or high serum calcium levels with vascular calci ication in MHD patients (31,33,35,36).
Although there is some evidence indicating an association between low levels of Fetuin-A and coronary artery calci ication, the incidence of cardiovascular events, and mortality, not all studies prove such an association [37][38][39]. Hermans et al. reported that Fetuin-A does not act as an independent predictor of vascular stiffness in dialysis patients, after adjusting for age, gender, mean arterial pressure and diabetes mellitus [16]. Also Jung et al. showed that, the inverse relationship between Fetuin-A and coronary artery calci ication is lost after adjusting for age [40]. Moreover, Pencak et al. did not ind any signi icant difference in Fetuin-A level and between hemodialysis patients and healthy subjects, despite the apparent difference in coronary artery calci ication score between the two groups [41].
There are a limited number of studies on the relationship between Fetuin-A and carotid IMT or valvular calci ication. However, as mentioned in coronary artery calci ication, the indings are controversial. Odamaki et al. reported an inverse relationship between Fetuin-A and carotid IMT [21]. Caglar et al. reported the same inverse relationship in non-diabetic patients with CKD stages 1 to 5; however, they did not observe this relationship in the multivariate analysis [42]. Schlieper et al. found no relationship between Fetuin-A and IMT in hemodialysis patients and Hermans et al. also noted that the relationship between Fetuin-A disappears after adjusting for age [43]. According to Ziolkowska et al., low level of Fetuin-A in children and adults with ESKD is an independent predictor of higher carotid IMT; however, this was not true for patients with CKD stages 2-4 [44]. In the study by Pertosa et al, the baseline level of Fetuin-A was inversely and independently associated with carotid IMT measured two years later [45]. Moreover, Pateinakis et al. reported an inverse relationship in 81 chronic hemodialysis patients; however, the observed relationship disappeared after adjusting for age [46].
There are also controversial indings and reports on the relationship between Fetuin-A level and the valvular calci ication in hemodialysis patients. Hermans et al failed to prove the role of Fetuin-A as an independent risk factor for aortic stiffness [16]. Lee et al. reported that the Fetuin-A level was lower in patients with valvular calci ication [37]. However, Mann et al. who conducted a study on patients with ESKD in India failed to ind a relationship between Fetuin-A and aortic calci ication [47].
Considering the controversial results of the studies on the relationship between Fetuin-A level and valvular or vascular calci ication in hemodialysis patients, results of our study, which showed no signi icant relationship between CVC and Fetuin-A level, are not much unexpected. The differences in methodology and study populations with different characteristics can partially explain the observed controversies and differences. It has also been stated that serum Fetuin-A in CKD patients is more observed as Fetuin-mineral complex (FMC), and not as free Fetuin-A [48]. FMC is a combination of Fetuin-A, ibrinogen, ibronectin 1, and calcium. In fact, with reducing GFR level, the FMC level, and not the level of free Fetuin-A, increases. This is associated with higher levels of coronary artery calci ication score in dialysis patients [48]. So it seems that, based on the current knowledge, circulating serum Fetuin-A has a poor relationship with CVC [49]. In addition, some other factors such as obesity, insulin resistance, metabolic syndrome, nonalcoholic fatty liver, lifestyle, smoking, adynamic bone diseases, and the use of vitamin D metabolites can affect the Fetuin-A level [50]. Moreover, the presence of different coding genes of Fetuin-A can affect its level in different populations [51].

Conclusion
The indings of our study showed that a high percentage of patients undergoing chronic hemodialysis (87%) had vascular or valvular calci ication and age, calcium level, and diabetes were identi ied as the most important predictors of calci ication in these patients. The Fetuin-A level was not different between patients with and without CVC and was not recognized as a predictor.
It is recommended to conduct longitudinal studies to examine the evolution of CVC and its' risk factors over time and to study the level of Fetuin-mineral complex, which may be a better representative of Fetuin effect on soft tissues.