Introduction

WASJ

World Academy of Sciences Journal

2632-2900 2632-2919

D.A. Spandidos

WASJ-8-3-00458

10.3892/wasj.2026.458

Articles

Fetuin-A and dephosphorylated uncarboxylated matrix Gla protein levels in patients with established coronary heart disease undergoing hemodialysis

Divani

Maria

Katsanaki

Aikaterini

Tziastoudi

Maria

Makri

Panagiota

Poulianiti

Christina

Lykotsetas

Evangelos

Balatsouka

Andriani

Stefanidis

Ioannis

Eleftheriadis

Theodoros

Department of Nephrology, Faculty of Medicine, University of Thessaly, 41110 Larissa, Greece

Correspondence to: Professor Theodoros Eleftheriadis, Department of Nephrology, Faculty of Medicine, University of Thessaly, Biopolis, Mezourlo Hill, 41110 Larissa, Greece teleftheriadis@yahoo.com

May-Jun2026

27032026

8 3

10 01 2026 16 03 2026

2026

This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ) and either DOI or URL of the article must be cited.

Fetuin-A and matrix Gla protein (MGP) protect against vascular calcification. It has been suggested that low fetuin-A levels or high levels of inactivated, dephosphorylated, uncarboxylated MGP (dp-ucMGP) may be associated with poor cardiovascular outcomes. The present study evaluated the potential of fetuin-A and dp-ucMGP levels as markers of coronary heart disease (CHD) in patients undergoing hemodialysis (HD). A total of 126 patients undergoing HD (40 patients with established CHD) and 24 healthy subjects were enrolled in the study. Serum fetuin-A and dp-ucMGP levels were measured using ELISA. In the patients undergoing HD, fetuin-A levels were significantly lower, whereas dp-ucMGP levels were significantly higher than in the healthy subjects. In patients undergoing HD, fetuin-A levels did not differ between those with or without CHD, diabetes mellitus, a C-reactive protein (CRP) level >1 mg/dl, or heart failure. Dephosphorylated, uncarboxylated MGP was lower in patients with CHD, diabetes mellitus, a CRP level >1 mg/dl and heart failure with a reduced ejection fraction. Dephosphorylated, uncarboxylated MGP levels were shown to perform moderately well as an indicator of CHD-free patients (area under the curve, 0.659). An increase in dp-ucMGP by 1 ng/ml was found to decrease the risk of CHD by 23.4%. On the whole, the present study demonstrates that, in patients with HD, fetuin-A levels are not associated with CHD. Conversely, elevated dp-ucMGP levels indicate CHD-free patients, although its accuracy is moderate.

fetuin-A matrix Gla protein hemodialysis coronary heart disease

Funding: No funding was received.

Introduction

Cardiovascular diseases (CVD) are the leading cause of mortality among patients undergoing hemodialysis (HD) (1), with the cardiovascular (CV) mortality rate being 10-20-fold higher than that in the general population following stratification for age, race and sex (2). In the general population, vascular calcification is associated with increased CV mortality (3). Compared to the general population, vascular calcification is prevalent in patients undergoing HD and is associated with a higher risk of CV-related mortality (4-6).

Fetuin-A and matrix Gla protein (MGP) are two proteins that provide a defense against extraosseous calcification. Fetuin-A is produced in the liver and functions as a calcification inhibitor by forming soluble colloidal spheres of fetuin-A, calcium and phosphate, preventing the crystallization of hydroxyapatite (7). Fetuin-A-deficient mice on a mineral/vitamin D-rich diet develop calcification in small vessels of the heart, lungs, kidneys and skin (8). Matrix Gla protein is secreted by chondrocytes and vascular smooth muscle cells. It binds to newly formed hydroxyapatite crystals, hindering their accumulation within the arterial wall. Additionally, MGP inhibits the binding of bone morphogenetic protein-2 to its receptor, preventing vascular smooth muscle cells from differentiating into osteoblasts (9). MGP-deficient mice undergo spontaneous and extensive calcification of cartilage and blood vessels, leading to death within 2 months of birth due to blood vessel rupture (10). For MGP to become active, it must undergo the carboxylation of certain γ-glutamate residues, followed by the phosphorylation of specific serine residues, in a process that depends on vitamin K. Consequently, dephosphorylated uncarboxylated MGP (dp-ucMGP) represents its inactive form and elevated levels of dp-ucMGP indicate a reduced capacity to inhibit vascular calcification (11,12). In line with the increased incidence of vascular calcification in patients undergoing HD, serum fetuin-A levels are lower in this population (13,14), whereas dp-ucMGP levels are higher (15,16). However, the data do not univocally support a role for serum fetuin-A or dp-ucMGP levels as markers of CV risk or established CVD.

In a community-based survey, low fetuin-A levels were shown to be associated with aortic arch calcification (17). Another study detected a trend towards lower levels in patients with peripheral artery disease (18). In another study, among community-dwelling individuals, lower fetuin-A levels were independently associated with greater severity of coronary artery calcification, but not with peripheral artery disease or carotid intima-media thickness (19). Conversely, in the Multi-Ethnic Study of Atherosclerosis, fetuin-A was not associated with the risk of CVD (20). In another cohort of patients who underwent carotid or lower extremity endarterectomy, fetuin-A was not associated with the coronary artery calcification score (21). In a previous study, an analysis of seven prospective studies found that fetuin-A genetic variants associated with reduced fetuin-A levels were not related to the risk of coronary heart disease (CHD) (22). Additionally, in another study involving 1,049 patients with CHD, fetuin-A levels failed to predict secondary CHD events over a 6-year follow-up period (23). Notably, in another study, fetuin-A was found to be higher in patients with CHD than in those without (24).

A similar discrepancy is observed in studies involving patients undergoing HD. Studies have shown that lower fetuin-A levels are associated with a higher CV mortality (13,14,25-27), an increased carotid intima-media thickness (13,26,27), elevated carotid-to-femoral pulse wave velocity (28), higher coronary artery calcification scores (25) and stroke (29). Conversely, in one study involving 104 patients undergoing HD, fetuin-A levels did not differ between those with or without coronary artery or abdominal aortic calcification (30). In another cohort of 220 patients undergoing HD, serum fetuin-A was not associated with carotid intima-media thickness, the degrees of carotid atherosclerotic plaques or calcification, left ventricular ejection fraction, prevalent CVD and all-cause or CV mortality (31). Similarly, in another cohort of 143 patients undergoing HD, no difference in fetuin-A levels was observed between those with a carotid intima-media thickness ≥0.8 mm, plaque or stenosis ≥50% and those without (32).

Similarly, for fetuin-A, a discrepancy relative to dp-ucMGP has been observed. In a population-based study with a 15.5-year follow-up period, higher dp-ucMGP levels were associated with an increased incidence of CVD and all-cause mortality (33). That study did not find a significant link between dp-ucMGP levels and the risk of myocardial infarction or sudden cardiac death. Nonetheless, elevated levels of dp-ucMGP were associated with an increased risk of other CVDs (33). One possible explanation for this is that myocardial infarction often results from the rupture of non-calcified atherosclerotic plaques (34). In another cohort of patients who experienced myocardial infarction, coronary revascularization, or first ischemic stroke, those in the highest quartile of dp-ucMGP had a higher risk of all-cause and CV 5-year mortality (35). In the Multi-Ethnic Study of Atherosclerosis, a positive association was detected between dp-ucMGP and the incidence and progression of coronary arteries and thoracic aorta calcification (36). Interestingly, in patients with acute coronary syndrome, those with non-ST-elevation myocardial infarction (NSTEMI) had higher levels of dp-ucMGP than those with ST-elevation myocardial infarction (STEMI), suggesting a potential link between dp-ucMGP and coronary artery calcification burden and more stable atherosclerotic plaques (37). Conversely, in the Health, Aging, and Body Composition Study, no association was detected between dp-ucMGP and incident CVD (38). Mendelian randomization studies have demonstrated that lower genetically predicted dp-ucMGP levels are associated with a reduced risk of CHD (39,40). Notably, in the Multi-Ethnic Study of Atherosclerosis, elevated dp-ucMGP was associated with an increased risk of incident CVD, CHD and all-cause mortality, but only in the youngest age quartile (41).

A similar discrepancy has been observed in studies involving patients undergoing HD. In these patients, high dp-uc MGP levels have been linked to increased vascular calcification (15,16,42). In a previous study, in a cohort of 141 patients with end-stage kidney disease (ESKD) undergoing living donor kidney transplantation, elevated dp-ucMGP levels were associated with the coronary artery calcification score and medial vascular calcification in the epigastric artery (43). However, in another cohort of 493 patients with ESKD from the same center with a median follow-up period of 42 months, elevated dp-ucMGP levels were associated with increased all-cause mortality, but they were not identified as an independent risk factor of coronary artery or aortic valve calcification (44). Conversely, another study demonstrated that, among 391 patients on incident HD, neither dp-ucMGP nor fetuin-A were associated with coronary artery calcification, pulse wave velocity or mortality risk (45). Similarly, among participants in the Chronic Renal Insufficiency Cohort with mild-to-moderate chronic kidney disease, dp-ucMGP was not associated with either coronary artery calcification score or pulse wave velocity (46). Moreover, in the same cohort, elevated dp-ucMGP levels were associated with increased all-cause mortality, but not with more atherosclerotic CV events (47). Of note, in a study on 198 patients undergoing HD, low dp-ucMGP levels were associated with a higher risk of all-cause mortality and CV mortality (48).

In this complex landscape, the present study investigated the association between fetuin-A or dp-ucMGP and established CHD in patients undergoing HD. Additionally, the present study explored other confounding factors that could impact these associations, underscoring the need for a better understanding of these interactions to improve patient outcomes.

Patients and methods <sec> <title>Patients

A total of 126 patients undergoing HD participated in the study. The mean age of the patients was 65.94±11.85 years, comprising 89 males and 37 females. The cause of ESKD was diabetic nephropathy (n=36), primary glomerulonephritis (n=24), hypertension (n=21), autosomal dominant polycystic kidney disease (n=11), secondary focal segmental glomerulosclerosis (n=6), cardiorenal syndrome (n=4), vasculitis (n=4), obstructive nephropathy (n=2), analgesic nephropathy (n=1) and unknown causes (n=17).

In total, 50 patients had diabetes mellitus, and 40 had a history of CHD. CHD was ascertained by coronary artery angiography performed for angina symptoms or following a myocardial infarction. A total of 110 patients underwent recent transthoracic echocardiography. Among these, 38 patients were identified with heart failure with preserved ejection fraction (HFpEF), while 26 patients were diagnosed with heart failure with reduced ejection fraction (HFrEF). Statin therapy was administered to 82 patients, and the majority received antihypertensive medications. The attending nephrologists determined the need for phosphate binders, vitamin D analogues and calcimimetics. All patients had received HD for at least 6 months prior to inclusion. Each patient underwent 4-h HD sessions three times per week, using polysulfone dialyzers and bicarbonate-based dialysate with calcium concentrations of 1.25 or 1.5 mmol/l.

Patients with active infection, autoimmune disease, malignancy, liver pathology, or who have received cytotoxic, immunosuppressive, or corticosteroid therapy within the previous 6 months were excluded from the study. Patients receiving vitamin K antagonists were also excluded. Further clinical and demographic features are presented in Table I.

A control group consisting of 24 healthy individuals (mean age, 64.42±6.49 years; 15 males and 9 females) was included following a review of medical records and a physical examination.

Written informed consent was obtained from each participant, and the study protocol received approval from the Ethics Committee of the Faculty of Medicine, University of Thessaly, Larissa, Greece (approval no. 558/10-2-2017).

Patient samples and analyses

Blood samples were drawn at the onset of the second hemodialysis session of the week, and serum was preserved at -80˚C.

Serum fetuin-A concentrations were measured using the Human Fetuin-A ELISA kit (cat. no. CSB-E12882h, Cusabio Technology, Wuhan, China), which has a sensitivity of 3.9 ng/ml. Serum dp-ucMGP levels were assessed using the human dp-ucMGP ELISA kit (cat. no. EH4755, Wuhan Fine Biotech Co., Ltd.), which has a sensitivity of 46.875 pg/ml. All other parameters were part of routine laboratory assessments performed concurrently with serum collection for the measurements of the factors described above.

Statistical analysis

Statistical analysis was performed using the IBM SPSS Statistics version 29 (IBM Corp.). Since the one-sample Kolmogorov-Smirnov test revealed that serum fetuin-A and dp-ucMGP levels were not normally distributed, non-parametric methods were applied. Specifically, for group comparisons, the Mann-Whitney U test was performed, with data presented as the median (interquartile range). Furthermore, Spearman's rank correlation coefficient was performed to evaluate correlations between continuous variables, while the Chi-squared test was used to examined associations between categorical variables. In addition, receiver operating characteristic (ROC) curve analysis was performed, and the optimal cut-off was identified by maximizing Youden's index. Finally, binary logistic regression was used to assess the independent effects of variables on outcomes. A value of P<0.05 was considered to indicate a statistically significant difference.

Results

Serum fetuin-A and dp-ucMGP levels in patients undergoing HD and healthy volunteers. Serum fetuin-A levels were significantly lower in patients undergoing HD than in healthy volunteers [1.54 (0.76-2.13) vs. 2.22 (1.423-3.980) mg/ml, respectively; P=0.02, Mann-Whitney U test] (Fig. 1A).

On the contrary, compared to the healthy subjects, the serum dp-ucMGP level was significantly higher in patients undergoing HD [1334.35 (920.8-1449.6) vs. 3892.38 (2958.9-6983.4) pg/ml, respectively, P<0.001, Mann-Whitney U test] (Fig. 1B)

<sec> <title>Serum fetuin-A levels in patients undergoing HD

Serum fetuin-A levels did not differ between patients undergoing HD with or without established CHD [1.62 (0.74-1.93) vs. 1.47 (0.77-2.15) mg/ml, respectively, P=0.987, Mann-Whitney U test] (Fig. 2A). Thus, ROC curve analysis revealed that in patients undergoing HD, fetuin-A is not a reliable marker of CHD [area under the curve (AUC), 0.499; 95% confidence interval (CI), 0.390-0.608; P=0.987] (Fig. 2B).

Notably, among the factors depicted in Table I, serum fetuin-A levels exhibited positive correlations with serum dp-ucMGP levels (Rho=0.311, P<0.001) (Fig. 3A), and phosphorus (Rho=0.184, P=0.042) (Fig. 3B). No correlations were detected between fetuin-A and the nutritional markers, body mass index (BMI; Rho=0.072, P=0.426) and serum albumin (Rho=-0.001, P=0.992) (data not shown). Fetuin-A was also not correlated with the urea reduction ratio (Rho=0.082, P=0.364) (data not shown). Although a positive correlation was detected with the white blood cell count (Rho=0.280, P=0.002) (Fig. 3C) and neutrophils (Rho=0.226, P=0.012) (Fig. 3D), no correlation was detected with serum C-reactive protein (CRP; Rho=0.037, P=0.680) (data not shown). In addition, using a serum CRP cut-off of 1 mg/dl to detect inflammation did not reveal that inflammation affects serum fetuin-A levels. Its levels were 1.62 (0.78-2.02) mg/ml in those with a CRP level <1 mg/ml, and 1.39 (0.75-2.28) mg/ml in those with a CRP level >1mg/dl (P=0.658, Mann-Whitney U test) (Table II).

Sex did not affect fetuin-A levels, which were 1.54 (0.73-2.21) mg/ml in males and 1.52 (1.05-2.12) mg/ml in females (P=0.783, Mann-Whitney U test). Also, fetuin-A levels did not differ in patients with or without diabetes mellitus [1.55 (0.75-2.42) vs. 1.47 (0.76-2.04) mg/ml, respectively, P=0.625, Mann-Whitney U test], HFpEF [1.55 (1.06-2.15) vs. 1.68 (0.73-2.80) mg/ml, respectively, P=0.921, Mann-Whitney U test], or HFrEF [1.30 (0.72-1.85) vs. 1.67 (0.96-2.15) mg/ml, respectively, P=0.325, Mann-Whitney U test] (Table II).

Serum dp-ucMGP levels in patients undergoing HD

Serum dp-ucMGP levels were significantly lower in patients undergoing HD with established CHD. They were 3,112.62 (2,487.32-4,221.92) pg/ml in those with CHD and 4,704.96 (3,135.95-7,207.95) pg/ml in those without CHD (P=0.004, Mann-Whitney U test) (Fig. 4A). ROC curve analysis revealed an AUC of 0.659 (95% CI, 0.560-0.757; P=0.004) (Fig. 4B). Below the optimal cut-off point of 4,241 pg/ml, serum dp-ucMGP exhibited a specificity for detecting CHD of 80.00% (95% CI, 65.24-89.50%) and a sensitivity of 55.81% (95% CI, 45.29-65.84%). Thus, serum dp-ucMGP may be characterized as a moderately significant marker of CHD in patients undergoing HD.

From the factors depicted in Table I, serum dp-ucMGP levels were correlated negatively with age (Rho=-0.302, P<0.001) (Fig. 5A), and positively with BMI (Rho=0.230, P=0.010) (Fig. 5B), creatinine (Rho-0.187, P=0.037) (Fig. 5C), calcium (Rho=0.398, P<0.001) (Fig. 5D), intact parathyroid hormone (iPTH; Rho=0.192, P=0.031) (Fig. 5E) and fetuin-A (Rho=0.311, P<0.001) (Fig. 5F). Notably, dp-ucMGP was not correlated with serum albumin (Rho=0.139, P=0.121) or the urea reduction ratio (Rho=-0.074, P=0.409) (data not shown). Although no correlation was detected with CRP (Rho=-0.153, P=0.087), when a CRP cut-off point of 1 mg/dl was set for defining inflammation, those with inflammation had lower dp-ucMGP levels [3,631.09 (2,492.57-6,218.71) pg/ml] than those without inflammation [4,684.97 (3,181.79-7,197.20) pg/ml] (P=0.032, Mann-Whitney U test) (Table III).

Sex did not affect serum dp-ucMGP levels. They were 3,766.63 (3,056.58-6,910.13) pg/ml in males and 4,451.02 (2,094.04-6,983.42) pg/ml in females (P=0.750, Mann-Whitney U test). However, dp-ucMGP levels were lower in patients with diabetes mellitus. They were 3,512.70 (2,490.82-4,943.77) pg/ml in those with diabetes mellitus and 4,736.58 (3,088.39-7,377.55) pg/ml in those without diabetes mellitus (P=0.006, Mann-Whitney U test). The levels of dp-ucMGP did not differ in those with HFpEF [4,451.02 (3,054.16-7,207.95) pg/ml] or without HFpEF [3,810.37 (3,328.87-7,840.34) pg/ml] (P=0.571, Mann-Whitney U test). However, they were significantly lower in patients with HFrEF [3,094.49 (2,474.14-4,357.54) pg/ml] than in those without HFrEF [4,008.10 (3,076.50-7,707.95) pg/ml)] (P=0.028, Mann-Whitney U test) (Table III). Notably, CHD was associated with HFrEF (Chi-squared test, 35.267, P<0.001) (Table IV), indicating that, in many cases of HFrEF, CHD may be responsible.

Of note, the male sex was associated with CHD (Chi-squared test, 15.955, P<0.001), as was diabetes mellitus (Chi-squared test, 10.107, P=0.001) (Table IV). As regards the other variables that were associated with serum dp-ucMGP levels, age (P=0.182, Mann-Whitney U test), BMI (P=0.317, Mann-Whitney U test), calcium (P=0.825, Mann-Whitney U test), iPTH (P=0.992, Mann-Whitney U test), creatinine (P=0.067, Mann-Whitney U test) and CRP (P=0.902, Mann-Whitney U test) did not differ between those with CHD and those without CHD (Table V). When the CRP cut-off of 1 mg/dl was set for defining inflammation, no association was detected between inflammation and CHD (Chi-squared test, 0.147, P=0.702) (Table IV).

Among the variables that may causally influence dp-ucMGP levels, only diabetes mellitus exhibited a causal association with CHD. Binary logistic regression analysis with CHD as the outcome variable and 1,000 pg (1 ng) increments of dp-ucMGP as the predictor variable revealed that for every dp-ucMGP increase of 1,000 pg/ml, the risk for CHD decreased by 23.4% [odds ratio (OR), 0.776; 95% CI, 0.649-0.928; P=0.005] (Table VI). Binary logistic regression analysis with CHD as the outcome variable and 1,000 pg increments of dp-ucMGP and diabetes mellitus as predictor variables was performed using the enter method. The overall model was statistically significant (Chi-squared test, 15.850; P<0.001), indicating that the predictors reliably distinguished between individuals with and without CHD. The model explained 16.6% of the variance (Nagelkerke R²=0.166) and correctly classified 73.8% of cases. Diabetes mellitus almost triplicates the possibility for CHD (OR, 2.909; 95% CI, 1.295-6.534; P=0.01), and for every dp-ucMGP increase of 1,000 pg/ml, the risk for CHD decreased by 18.8% (OR, 0.812; 95% CI, 0.677-0.973; P=0.024). Thus, dp-ucMGP levels predicted CHD independently of diabetes mellitus (Table VI).

Discussion

Since CVD is the primary cause of mortality in patients undergoing HD, evaluating CVD pathogenesis and identifying potential markers for its early detection is critical. The present study investigated whether fetuin-A, a vascular calcification inhibitor, and dp-ucMGP, an inactive vascular calcification inhibitor, are associated with established CHD in a cohort of stable patients undergoing HD.

As previous studies have shown (13,14), serum fetuin-A levels were lower in HD patients than in healthy volunteers. However, its levels did not differ in patients undergoing HD with or without CHD. The data on the role of fetuin-A as a risk factor or a marker of established CVD are mixed. Some studies in the general population have shown that low fetuin-A levels are associated with an increased risk of CV or established CVD (17-19), other studies did not find an association (20-23), and at least one study demonstrated that fetuin-A was higher in patients with CHD than in those without (24). Studies involving patients undergoing HD have also yielded mixed outcomes. Some studies support the role of fetuin-A as a risk factor or marker of CVD (13,14,25-29), while others do not (30-32). According to the present study, serum fetuin-A cannot serve as a marker of established CHD in patients undergoing HD. This result does not entirely contradict studies suggesting that fetuin-A levels can indicate the risk of CVD. Numerous of those studies focused on subclinical atherosclerosis or vascular calcification (13,25-28,30,32), or they assessed incident CVD rather than established CVD in general or CHD specifically (13,14,25-27).

In the present study, serum fetuin-A levels correlated with dp-ucMGP, phosphorus, white blood cell count and neutrophil levels, but not with CRP. Additionally, factors such as sex, diabetes mellitus, heart failure and inflammation, as defined by a CRP cut-off, did not affect fetuin-A levels. Of note, fetuin-A is considered a negative inflammatory reactant (13,14,26-29). Notably, not all studies involving patients undergoing HD have shown a negative association with CRP (31,49). As regards diabetes mellitus, while some studies have reported an association with decreased fetuin-A levels (29,31), others have not (26). Considering the lack of an association between fetuin-A levels and heart failure in the present study, in the patients included herein, heart failure was strongly associated with CHD, implying that, in the majority of cases, CHD may be the cause of heart failure.

As in previous studies (15,16), serum dp-ucMGP levels were higher in patients undergoing HD than in healthy volunteers. In the present study, in patients undergoing HD, dp-ucMGP levels were significantly lower in those with established CHD than in those without CHD. ROC curve analysis revealed an AUC of 0.659 for identifying CHD-free patients, indicating serum dp-ucMGP as a moderately significant marker of CHD in patients undergoing HD.

The data on the role of high dp-ucMGP as a risk factor or a marker of established CVD are mixed. In the general population, some studies have identified high dp-ucMGP levels as a risk factor for CVD or a marker of established CVD (33-37,41). By contrast, others have not found such an association (38), whereas, particularly for CHD, others have supported the opposite (39,40). However, even in studies that have supported an association between high dp-ucMGP levels and CVD, several interesting points emerge. In one of these, no link was detected between dp-ucMGP levels and the risk of incident myocardial infarction or sudden cardiac death (33). In another study, patients with NSTEMI had higher dp-ucMGP levels than those with STEMI (37). The results of the last two studies suggest that high dp-ucMGP levels are associated with greater coronary artery calcification and, consequently, a lower risk of atherosclerotic plaque rupture (34). Lastly, one study demonstrated that elevated dp-ucMGP levels were associated with an increased risk of incident CVD, CHD and all-cause mortality, but only in the youngest age quartile (45-to-53-year-olds) (41), i.e., in a younger age group than the patients undergoing HD in the present study.

In studies involving patients undergoing HD, the findings on the role of high dp-ucMGP as a CVD risk factor or marker of established CVD are also inconsistent. Some studies have shown such an association (15,16,42-44), while others have failed to detect it (45,46); notably, in one study, low dp-ucMGP levels were linked to increased all-cause [hazard ratio (HR), 1.71; 95% CI, 0.92-3.17] and CV mortality (HR, 1.83; 95% CI, 0.9-3.7) (48). Of note, in one study involving patients with mild to moderate chronic kidney disease, elevated dp-ucMGP levels were associated with increased all-cause mortality, but not with more atherosclerotic CV events (47). Finally, in support of the results of the present study, studies have failed to confirm CV benefits despite the potential of vitamin K supplementation to reduce dp-ucMGP levels in patients with ESKD (49).

In the present study, serum dp-ucMGP levels were negatively correlated with age and positively correlated with BMI, creatinine, calcium, iPTH and fetuin-A. Although no correlation was detected with CRP, when a CRP cut-off of 1 mg/dl was used to define inflammation, those with inflammation had lower dp-ucMGP levels than those without. Studies involving patients undergoing HD conclude with inconsistent results regarding the association between dp-ucMGP levels and CRP. Some have detected a positive association (42,44,45), while others have found no association (50). Sex did not affect serum dp-ucMGP levels. However, dp-ucMGP levels were lower in patients with diabetes mellitus. Studies on patients undergoing HD conclude with varying results regarding the association between dp-ucMGP levels and diabetes mellitus. Some have detected a positive association (16,51), while others have found no association (43,44), or a negative association (38,46). Finally, dp-uc MGP levels were lower in patients with HFrEF. However, the strong association between HFrEF and CHD suggests that CHD may have caused HFrEF in the patients in the present study.

Among the variables that affected dp-ucMGP levels, only diabetes mellitus was associated with CHD. For every dp-ucMGP increase of 1 ng/ml, the risk of CHD decreased by 23.4%. Still, following adjustment for diabetes mellitus, for every dp-ucMGP increase of 1 ng/ml, the risk for CHD decreased by 18.8%.

The present study has certain limitations. First, further prospective studies involving a larger patient population are necessary to verify the findings. The modest discriminatory capacity of dp-ucMGP for detecting CHD (AUC, 0.659) further underscores the need for additional research. In addition, the present study relied on the medical records of patients to determine the presence of CHD. As a consequence, some cases of asymptomatic CHD may have been overlooked. Nevertheless, regular thrice-weekly attendance at the HD unit enables early detection of CHD, and patients undergo periodic cardiological assessments and transthoracic echocardiography. Moreover, the results do not entirely contradict studies suggesting that fetuin-A or dp-ucMGP levels can indicate the risk of CVD. A number of those studies focused on subclinical atherosclerosis or vascular calcification (15,16,42-45), or they assessed incident CVD rather than established CHD (43-45,47,48). Finally, we assessed fetuin-A and dp-ucMGP at a single time point to determine whether they are associated with established CHD. However, interpreting the results is challenging as the patients who survive to that time point are not a random sample. Patients with more calcified arteries may have an improved prognosis, since more calcified atherosclerotic plaques are less prone to rupture (34). Thus, patients with CHD with higher dp-ucMGP levels or lower fetuin-A levels, and possibly more calcified coronary arteries, are more likely to be alive and were included in the assessment. This survival imbalance may influence the observed association between fetuin-A or dp-ucMGP and CHD. In other words, dp-ucMGP may be a negative marker of prevalent CHD and, concurrently, a risk factor for incident CHD. Following-up on the patient cohort for incident CHD will help clarify this concept.

In conclusion, the present study demonstrates that, in patients undergoing HD, fetuin-A levels are not associated with established CHD. Conversely, elevated dp-ucMGP levels indicate CHD-free patients, though its accuracy is moderate.

Acknowledgements

Not applicable.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

TE designed the study. MD and PM obtained patient samples and clinical data. AK performed the ELISA measurements. TE, MD, AK, MT, PM, CP, EL, AB and IS analyzed and interpreted the results. T.E. wrote the draft manuscript; T.E., M.D., and M.T. edited the manuscript. TE, AK and MD confirm the authenticity of all the raw data. All authors drafted the manuscript, critically revised the manuscript, and agreed to be fully accountable for ensuring the integrity and accuracy of the work, and have read and approved the final manuscript.

Ethics approval and consent to participate

The present study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the University of Thessaly, Faculty of Medicine (Approval no. 558/10-2-2017). Informed consent was obtained from all subjects involved in the study.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References 1

Jankowski

Floege

Fliser

Böhm

Marx

Cardiovascular disease in chronic kidney disease

Circulation143115711722021

Foley

Parfrey

Sarnak

Clinical epidemiology of cardiovascular disease in chronic renal disease

Am J Kidney Dis32S112S1191998

9820470

10.1053/ajkd.1998.v32.pm9820470

Allison

Hsi

Wassel

Morgan

Wright

Criqui

Calcified atherosclerosis in different vascular beds and the risk of mortality

Arteriosclerosis Thromb Vasc Biol321401462012

22034514

10.1161/ATVBAHA.111.235234

Goodman

Goldin

Kuizon

Yoon

Gales

Sider

Wang

Chung

Emerick

Greaser

Coronary-artery calcification in young adults with end-stage renal disease who are undergoing dialysis

New Engl J Med342147814832000

10816185

10.1056/NEJM200005183422003

Liabeuf

Desjardins

Diouf

Temmar

Renard

Choukroun

Massy

The addition of vascular calcification scores to traditional risk factors improves cardiovascular risk assessment in patients with chronic kidney disease

PLoS One10e01317072015

26181592

10.1371/journal.pone.0131707

Hutcheson

Goettsch

Cardiovascular calcification heterogeneity in chronic kidney disease

Circul Res13299310122023

37053279

10.1161/CIRCRESAHA.123.321760

Heiss

DuChesne

Denecke

Grötzinger

Yamamoto

Renné

Jahnen-Dechent

Structural basis of calcification inhibition by alpha 2-HS glycoprotein/fetuin-A. Formation of colloidal calciprotein particles

J Biol Chem27813333133412003

12556469

10.1074/jbc.M210868200

Schäfer

Heiss

Schwarz

Westenfeld

Ketteler

Floege

Muller-Esterl

Schinke

Jahnen-Dechent

The serum protein alpha 2-Heremans-Schmid glycoprotein/fetuin-A is a systemically acting inhibitor of ectopic calcification

J Clin Invest1123573662003

12897203

10.1172/JCI17202

Iyemere

Proudfoot

Weissberg

Shanahan

Vascular smooth muscle cell phenotypic plasticity and the regulation of vascular calcification

J Intern Med2601922102006

16918817

10.1111/j.1365-2796.2006.01692.x

Luo

Ducy

McKee

Pinero

Loyer

Behringer

Karsenty

Spontaneous calcification of arteries and cartilage in mice lacking matrix GLA protein

Nature38678811997

9052783

10.1038/386078a0

Schurgers

Cranenburg

ECM

Vermeer

Matrix Gla-protein: The calcification inhibitor in need of vitamin K

Thromb Haemost1005936032017

18841280

Kumric

Borovac

Kurir

Martinovic

Separovic

Baric

Bozic

Role of matrix gla protein in the complex network of coronary artery disease: A comprehensive review

Life (Basel)117372021

34440481

10.3390/life11080737

Stenvinkel

Wang

Qureshi

Axelsson

Pecoits-Filho

Gao

Barany

Lindholm

Jogestrand

Heimbürger

Low fetuin-A levels are associated with cardiovascular death: Impact of variations in the gene encoding fetuin

Kidney Int67238323922005

15882283

10.1111/j.1523-1755.2005.00345.x

Ketteler

Bongartz

Westenfeld

Wildberger

Mahnken

Böhm

Metzger

Wanner

Jahnen-Dechent

Floege

Association of low fetuin-A (AHSG) concentrations in serum with cardiovascular mortality in patients on dialysis: A cross-sectional study

Lancet3618278332003

12642050

10.1016/S0140-6736(03)12710-9

Schurgers

Barreto

Liabeuf

Renard

Magdeleyns

Vermeer

Choukroun

Massy

The circulating inactive form of matrix Gla protein is a surrogate marker for vascular calcification in chronic kidney disease

Clin J Am Soc Nephrol55685752010

20133489

10.2215/CJN.07081009

Fain

Kapuku

Paulson

Williams

Raed

Dong

Knapen

MHJ

Vermeer

Pollock

Inactive matrix gla protein, arterial stiffness, and endothelial function in african american hemodialysis patients

Am J Hypertension317357412018

29635270

10.1093/ajh/hpy049

Lin

Shi

Lin

Yang

Liao

Chen

Tsai

Lin

The impact of fetuin-A on predicting aortic arch calcification: Secondary analysis of a community-based survey

Front Cardiovasc Med1114154382024

39040998

10.3389/fcvm.2024.1415438

Jirak

Mirna

Wernly

Paar

Thieme

Betge

Franz

Hoppe

Lauten

Kammler

Analysis of novel cardiovascular biomarkers in patients with peripheral artery disease

Minerva Med1094434502018

29652038

10.23736/S0026-4806.18.05628-8

Barrett-Connor

Wassel

Cummins

Bergstrom

Daniels

Laughlin

The associations of fetuin-A with subclinical cardiovascular disease in community-dwelling persons

J Am Coll Cardiol58237223792011

22115642

10.1016/j.jacc.2011.08.035

Aroner

St-Jules

Mukamal

Katz

Shlipak

Criqui

Kestenbaum

Siscovick

de Boer

Jenny

Fetuin-A, glycemic status, and risk of cardiovascular disease: The multi-ethnic study of atherosclerosis

Atherosclerosis2482242292016

27038419

10.1016/j.atherosclerosis.2016.03.029

Cahalane

Barrett

Ross

Mulvihill

JJE

Purtill

Selvarajah

O'Brien

Kavanagh

Moloneye

Egan

On the association between circulating biomarkers and atherosclerotic calcification in a cohort of arterial disease participants

Nutr Metab Cardiovasc Dis31153315412021

33810961

10.1016/j.numecd.2021.02.005

Laugsand

Bartz

Djousse

Kizer

Tracy

Dehghan

Rexrode

Lopez

Rimm

Fetuin-A and risk of coronary heart disease: A Mendelian randomization analysis and a pooled analysis of AHSG genetic variants in 7 prospective studies

Atherosclerosis24344522015

26343871

10.1016/j.atherosclerosis.2015.08.031

Roos

von Eynatten

Heemann

Rothenbacher

Brenner

Breitling

Serum fetuin-A, cardiovascular risk factors, and six-year follow-up outcome in patients with coronary heart disease

Am J Cardiol105166616722010

20538112

10.1016/j.amjcard.2010.01.342

Akin

Celik

Altun

Ayca

Diker

Satılmıs

Sahin

Relationship of fibroblast growth factor 23 and fetuin-A to coronary atherosclerosis

J Diabetes Complications295505552015

25772252

10.1016/j.jdiacomp.2015.02.013

Mohamed

MRM

Hassan

Alakkad

Othman

Setouhi

Issa

The relationship of fetuin-A with coronary calcification, carotid atherosclerosis, and mortality risk in non-dialysis chronic kidney disease

J Lipid Atheroscler131942112024

38826181

10.12997/jla.2024.13.2.194

Pertosa

Simone

Ciccone

Porreca

Zaza

Dalfino

Memoli

Procino

Bonomini

Sirolli

Serum Fetuin A in hemodialysis: A link between derangement of calcium-phosphorus homeostasis and progression of atherosclerosis?

Am J Kidney Dis534674742009

19167798

10.1053/j.ajkd.2008.10.046

Hermans

MMH

Brandenburg

Ketteler

Kooman

van der Sande

Boeschoten

Leunissen

Krediet

Dekker

Netherlands cooperative study on the adequacy of Dialysis (NECOSAD)

Association of serum fetuin-A levels with mortality in dialysis patients

Kidney Int722022072007

17342178

10.1038/sj.ki.5002178

Pateinakis

Papagianni

Douma

Efstratiadis

Memmos

Associations of fetuin-A and osteoprotegerin with arterial stiffness and early atherosclerosis in chronic hemodialysis patients

BMC Nephrol141222013

23758931

10.1186/1471-2369-14-122

Chen

Chiu

Hsu

Pai

Yang

Peng

Low serum fetuin A levels and incident stroke in patients with maintenance haemodialysis

Eur J Clin Invest433873962013

23419133

10.1111/eci.12057

Pencak

Czerwieńska

Ficek

Wyskida

Kujawa-Szewieczek

Olszanecka-Glinianowicz

Więcek

Chudek

Calcification of coronary arteries and abdominal aorta in relation to traditional and novel risk factors of atherosclerosis in hemodialysis patients

BMC Nephrol14102013

23317172

10.1186/1471-2369-14-10

Collado

Coll

Nicolau

Azqueta

Pons

Cruzado

de la Torre

Deulofeu

Mojal

Pascual

Cases

Serum osteoprotegerin in prevalent hemodialysis patients: Associations with mortality, atherosclerosis and cardiac function

BMC Nephrol182902017

28882110

10.1186/s12882-017-0701-8

Ossareh

Rayatnia

Vahedi

Jafari

Zebarjadi

Association of serum fetuin-A with vascular calcification in hemodialysis patients and its' impact on 3-year mortality

Iran J Kidney Dis145005092020

33277456

Willeit

Santer

Tschiderer

Pechlaner

Vermeer

Willeit

Kiechl

Association of desphospho-uncarboxylated matrix gla protein with incident cardiovascular disease and all-cause mortality: Results from the prospective Bruneck Study

Atherosclerosis35320272022

35764030

10.1016/j.atherosclerosis.2022.06.1017

Virmani

Burke

Farb

Kolodgie

Pathology of the unstable plaque

Prog Cardiovasc Dis443493562002

12024333

10.1053/pcad.2002.122475

Mayer

Seidlerová

Bruthans

Filipovský

Timoracká

Vaněk

Cerná

Wohlfahrt

Cífková

Theuwissen

Vermeer

Desphospho-uncarboxylated matrix Gla-protein is associated with mortality risk in patients with chronic stable vascular disease

Atherosclerosis2351621682014

24835435

10.1016/j.atherosclerosis.2014.04.027

Berlot

Shea

Tracy

Budoff

Kim

Naveed

Booth

Kizer

Bortnick

Matrix Gla protein and the long-term incidence and progression of coronary artery and aortic calcification in the multi-ethnic study of atherosclerosis

Atherosclerosis3921175052024

38527383

10.1016/j.atherosclerosis.2024.117505

Bilalic

Kurir

Kumric

Borovac

Matetic

Supe-Domic

Bozic

Circulating levels of dephosphorylated-uncarboxylated matrix gla protein in patients with acute coronary syndrome

Molecules2611082021

33669806

10.3390/molecules26041108

Shea

Booth

Weiner

Brinkley

Kanaya

Murphy

Simonsick

Wassel

Vermeer

Kritchevsky

Health ABC Study

Circulating vitamin K Is inversely associated with incident cardiovascular disease risk among those treated for hypertension in the health, aging, and body composition study (Health ABC)

J Nutr1478888952017

28356433

10.3945/jn.117.249375

Liu

Thijs

Knapen

Salvi

Citterio

Petit

Carpini

Zhang

Jacobs

Inactive matrix gla protein is causally related to adverse health outcomes

Hypertension654634702015

25421980

10.1161/HYPERTENSIONAHA.114.04494

Zwakenberg

Burgess

Sluijs

Weiderpass

Beulens

JWJ

van der Schouw

Circulating phylloquinone, inactive Matrix Gla protein and coronary heart disease risk: A two-sample mendelian randomization study

Clin Nutr39113111362020

31103344

10.1016/j.clnu.2019.04.024

Berlot

Shea

Tracy

Budoff

Kim

Naveed

Booth

Kizer

Bortnick

Inactive matrix gla protein and cardiovascular outcomes: The multi-ethnic study of atherosclerosis

J Am Heart Assoc14e0364592025

39968786

10.1161/JAHA.124.036459

Delanaye

Krzesinski

Warling

Moonen

Smelten

Médart

Pottel

Cavalier

Dephosphorylated-uncarboxylated Matrix Gla protein concentration is predictive of vitamin K status and is correlated with vascular calcification in a cohort of hemodialysis patients

BMC Nephrol151452014

25190488

10.1186/1471-2369-15-145

Jaminon

AMG

Dai

Qureshi

Evenepoel

Ripsweden

Söderberg

Witasp

Olauson

Schurgers

Stenvinkel

Matrix Gla protein is an independent predictor of both intimal and medial vascular calcification in chronic kidney disease

Sci Rep1065862020

32313061

10.1038/s41598-020-63013-8

Dai

Erlandsson

Jaminon

AMG

Qureshi

Ripsweden

Brismar

Witasp

Heimbürger

Jørgensen

Functional vitamin K insufficiency, vascular calcification and mortality in advanced chronic kidney disease: A cohort study

PLoS One16e02476232021

33626087

10.1371/journal.pone.0247623

Fitzpatrick

Kim

Sozio

Jaar

Estrella

Monroy-Trujillo

Parekh

Calcification biomarkers, subclinical vascular disease, and mortality among multiethnic dialysis patients

Kidney Int Rep5172917372020

33102965

10.1016/j.ekir.2020.07.033

Shea

Wang

Barger

Weiner

Townsend

Feldman

Rosas

Chen

Flack

Association of vitamin K status with arterial calcification and stiffness in chronic kidney disease: The chronic renal insufficiency cohort

Curr Dev Nutr71000082023

37181121

10.1016/j.cdnut.2022.100008

Shea

Barger

Booth

Wang

Feldman

Townsend

Chen

Flack

Jaar

Vitamin K status, all-cause mortality, and cardiovascular disease in adults with chronic kidney disease: The chronic renal insufficiency cohort

Am J Clin Nutr1159419482022

34788785

10.1093/ajcn/nqab375

Schlieper

Westenfeld

Krüger

Cranenburg

Magdeleyns

Brandenburg

Djuric

Damjanovic

Ketteler

Vermeer

Circulating nonphosphorylated carboxylated matrix gla protein predicts survival in ESRD

J Am Soc Nephrol223873952011

21289218

10.1681/ASN.2010040339

Gluba-Brzózka

Michalska-Kasiczak

Franczyk-Skóra

Nocuń

Banach

Rysz

Markers of increased cardiovascular risk in patients with chronic kidney disease

Lipids Health Dis131352014

25145866

10.1186/1476-511X-13-135

Mao

Huang

Zhou

Correlation between circulating dephosphorylated uncarboxylated matrix Gla protein and vascular calcification in peritoneal dialysis patients

Int J Artif Organs478858932024

39482813

10.1177/03913988241293980

Nyvad

Christensen

Andersen

Reinhard

Nørgaard

Madsen

Nielsen

Thomsen

Jensen

Peters

Buus

PIVKA-II but not dp-ucMGP is associated with aortic calcification in chronic kidney disease

BMC Nephrol254262024

39604863

10.1186/s12882-024-03876-5

Figure 1

Comparison of serum fetuin-A and dp-ucMGP levels between healthy individuals and patients undergoing HD. (A) Serum fetuin-A levels were significantly lower in patients undergoing HD compared to healthy subjects. (B) Conversely, dp-ucMGP levels were significantly higher in patients undergoing HD compared to healthy subjects. The line inside the box indicates the median. The box represents the IQR, and the whiskers extend to 1.5 x IQR above and below the upper and lower quartiles, respectively. Points outside the whiskers are outliers. ^*P<0.05, vs. healthy subjects. dp-ucMGP, uncarboxylated matrix Gla protein; HD, hemodialysis; IQR, interquartile range.

Figure 2

Serum fetuin-A levels in patients undergoing HD with or without CHD. (A) Serum fetuin-A levels did not differ significantly between patients undergoing HD with or without CHD. (B) ROC curve analysis revealed an area under the curve of 0.499 (P>0.05). HD, hemodialysis; CHD, coronary heart disease.

Figure 3

Scatterplots of statistically significant correlations between fetuin-A and the evaluated factors. Serum fetuin-A levels correlated significantly with (A) dp-ucMGP (Rho=0.311, P<0.001), (B) phosphorus (Rho=0.184, P=0.042), (C) white blood cell count (Rho=0.280, P=0.002), and (D) neutrophils (Rho=0.226, P=0.012). dp-ucMGP, uncarboxylated matrix Gla protein.

Figure 4

Serum dp-ucMGP levels in patients undergoing HD with or without CHD. (A) Serum dp-ucMGP levels were significantly lower in patients undergoing HD with CHD compared to those without CHD. ^*P<0.05, vs. those with no CHD. (B) dp-ucMGP levels can identify patients who are free of CHD, although its accuracy is moderate, as indicated by an area under the curve of 0.659 (P<0.05) in ROC curve analysis. dp-ucMGP, uncarboxylated matrix Gla protein; HD, hemodialysis; CHD, coronary heart disease.

Figure 5

Scatterplots of statistically significant correlations between dp-ucMGP and the evaluated factors. Serum dp-ucMGP levels significantly correlated with (A) age (Rho=-0.302, P<0.001), (B) body mass index (Rho=0.230, P=0.010), (C) creatinine (Rho=0.187, P=0.037), (D) calcium (Rho=0.398, P<0.001), (E) parathyroid hormone (Rho=0.192, P=0.031), and (F) fetuin-A (Rho=0.311, P<0.001). dp-ucMGP, uncarboxylated matrix Gla protein.

Table I

Patient characteristics.

	No. of patients	Mean	SD
Age (years)	126	65.94	11.85
Duration (months)	126	56.21	50.04
Males/females	89/37
Diabetes mellitus (yes/no)	50/76
Coronary heart disease (yes/no)	40/86
HFrEF/HFpEF/No HF	26/38/46
Systolic blood pressure (mmHg)	126	132.12	21.41
Diastolic blood pressure (mmHg)	126	65.57	13.32
White blood cell count (c/µl)	126	6,884.81	2,286.36
Neutrophils (c/µl)	126	4,589.67	1,726.59
Lymphocytes (c/µl)	126	1,681.77	606.01
Hemoglobin (g/dl)	126	11.77	0.84
Platelets (c/µl)	126	203.99	58.97
Creatinine (mg/dl)	126	6.38	2.04
Urea (mg/dl)	126	130.35	27.48
Urea reduction rate (%)	126	67.53	7.60
Residual diuresis (ml)	126	321.43	422.30
Body mass index (Kg/m²)	126	26.90	5.71
Albumin (g/dl)	126	3.63	0.10
Cholesterol (mg/dl)	126	130.93	42.42
Triglycerides (mg/dl)	126	130.33	68.77
Calcium (mg/dl)	126	9.12	0.46
Phosphorus (mg/dl)	126	5.34	1.01
Parathyroid hormone (pg/ml)	126	347.73	287.29
Alkaline phosphatase (U/l)	126	205.80	105.30
SGOT (U/l)	126	12.88	7.31
SGPT (U/l)	126	11.72	9.16
Ferritin (ng/ml)	126	153.47	161.76
TSAT (%)	126	19.01	12.12
CRP (mg/dl)	126	1.24	0.87
CRP >1 mg/dl (yes/no)	63/63
Fetuin-A (mg/ml)	126	1.73	1.22
Dp-uc-matrix Gla protein (pg/ml)	126	4,743.50	2,556.75

HFrEF, heart failure with reduced ejection fraction; HFpEF, heart failure with preserved ejection fraction; HF, heart failure; SGOT, serum glutamic-oxaloacetic transaminase; SGPT, serum glutamic-pyruvic transaminase; TSAT, transferrin saturation; CRP, C-reactive protein.

Table II

Differences in serum fetuin-A (mg/ml) between different groups.

	No. of patients	Yes	No	P-value
Male sex	90	1.54 (0.73-2.21)	1.52 (1.05-2.12)	0.783
Diabetes mellitus	50	1.55 (0.75-2.42)	1.47 (0.76-2.04)	0.625
Coronary heart disease	40	1.62 (0.74-1.93)	1.47 (0.77-2.15)	0.987
HFpEF	38	1.55 (1.06-2.15)	1.68 (0.73-2.80)	0.921
HFrEF	26	1.30 (0.72-1.85)	1.67 (0.96-2.15)	0.325
CRP >1 mg/dl	63	1.39 (0.75-2.28)	1.62 (0.78-2.02)	0.658

Values correspond to median (IQR) analyzed using the Mann-Whitney U test. HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; CRP, C-reactive protein.

Table III

Differences in serum dp-ucMGP (pg/ml) between different groups.

	No. of patients	Yes	No	P-value
Male sex	90	3,766.63 (3,056.58-6,910.13)	4,451.02 (2,094.04-6,983.42)	0.750
Diabetes mellitus	50	3,512.70 (2,490.82-4,943.77)	4,736.58 (3,088.39-7,377.55)	0.006
Coronary heart disease	40	3,112.62 (2,487.32-4,221.92)	4,704.96 (3,135.95-7,207.95)	0.004
HFpEF	38	4,451.02 (3,054.16-7,207.95)	3,810.37 (3,328.87-7,840.34)	0.571
HFrEF	26	3,094.49 (2,474.14-4,357.54)	4,008.10 (3,076.50-7,707.95)	0.028
CRP >1 mg/dl	63	3,631.09 (2,492.57-6,218.71)	4,684.97 (3,181.79-7,197.20)	0.032

Table IV

Associations of CHD with sex, diabetes, HFrEF and CRP.

Variable	Category	CHD (no. of patients	No CHD (no. of patients)	χ²	P-value
Sex	Male	38	52	15.955	<0.001
	Female	2	34
Diabetes mellitus	Yes	24	26	10.107	0.001
	No	16	60
HFrEF	Yes	22	4	35.267	<0.001
	No	18	68
CRP >1 mg/dl	Yes	21	42	0.147	0.702
	No	19	44

HFrEF, heart failure with reduced ejection fraction; CRP, C-reactive protein; CHD, coronary heart disease.

Table V

Differences in factors correlated with dp-ucMGP in patients with or without CHD.

	CHD	No CHD	P-value
Age (years)	71.5 (63-77)	67.0 (59-74)	0.182
Body mass index (kg/m²)	25.30 (22.25-27.80)	25.54 (23.43-29.77)	0.317
Creatinine (mg/dl)	6.0 (3.77-7.03)	6.4 (5.1-7.7)	0.067
Calcium (mg/dl)	9.05 (8.7-10.0)	9.10 (8.8-9.3)	0.825
Parathyroid hormone (pg/ml)	258.7 (199.9-391.9)	267.3 (175.0-455.0)	0.992
Fetuin-A (mg/ml)	1.62 (0.745-1.93)	1.47 (0.773-2.145)	0.987
CRP (mg/dl)	1.02 (0.79-1.61)	0.945 (0.78-1.60)	0.902

Values correspond to median (IQR) analyzed using the Mann-Whitney U test. dp-ucMGP, dephosphorylated-uncarboxylated matrix Gla protein; CRP, C-reactive protein; CHD, coronary heart disease.

Table VI

Logistic regression analysis for CHD with dp-ucMGP alone (model 1) or in combination with diabetes (the only variable affecting both) (model 2).

					95% C.I. for OR
	B	S.E.	Sig.	OR	Lower	Upper
Model 1
dp-ucMGP (ng/ml)	-0.254	0.091	0.005	0.776	0.649	0.928
Model 2
Diabetes mellitus	1.068	0.413	0.010	2.909	1.295	6.534
dp-ucMGP (ng/ml)	-0.208	0.092	0.024	0.812	0.677	0.973

Model 1 fit: Omnibus test of model coefficients, P=0.03; Nagelkerke R²=0.097, χ²=9.055, P=0.03. Model 2 fit: Omnibus test of model coefficients, P<0.001; Nagelkerke R²=0.166, χ²=25.850, P<0.01.