Endocrine DOI 10.1007/s12020-015-0673-z
ORIGINAL ARTICLE
Calcimimetic and vitamin D analog use in hemodialyzed patients is associated with increased levels of vitamin K dependent proteins Maria Fusaro1 • Sandro Giannini2 • Maurizio Gallieni3 • Marianna Noale1 • Giovanni Tripepi4 • Maurizio Rossini5 • Piergiorgio Messa6 • Paolo Rigotti7 • Tecla Pati8 • Francesco Barbisoni9 • Antonio Piccoli10 • Andrea Aghi10 • Marianna Alessi10 • Luciana Bonfante10 • Fabrizio Fabris2 • Sabina Zambon2 Stefania Sella2 • Giorgio Iervasi11 • Mario Plebani12
•
Received: 9 April 2015 / Accepted: 20 June 2015 Ó Springer Science+Business Media New York 2015
Abstract Matrix Gla protein (MGP) and bone Gla protein (BGP) are two vitamin K-dependent proteins (VKDPs) involved in the regulation of vascular calcification (VC). We carried out a secondary analysis of the VIKI study to evaluate associations between drug consumption and VKDP levels in 387 hemodialyzed patients. The VIKI study assessed the prevalence of vitamin K deficiency in hemodialysis patients. We evaluated drug consumption, determined BGP and MGP levels, and verified the presence of any vertebral fractures (VF) and VC by spine radiographs. Total BGP levels were twice as high with calcimimetics versus no calcimimetics (290 vs. 158.5 mcg/L, p \ 0.0001) and 69 % higher with vitamin D analogs (268 vs. 159 mcg/L, p \ 0.0001). Total MGP was 19 % higher with calcimimetics (21.5 vs. 18.1 mcg/L, p = 0.04) and 54 % higher with calcium acetate (27.9 vs. 18.1 mcg/L,
p = 0.003); no difference was found with vitamin D analogs (21.1 vs. 18.3 mcg/L, p = 0.43). Median Total BGP level was 29 % lower in patients with C1 VF (151 vs. 213 mcg/L, p = 0.0091) and 36 % lower in patients with VC (164 vs. 262.1 mcg/L, p = 0.0003). In non-survivors, median BGP and MGP were lower, but only for MGP this difference reached the statistical significance (152 vs. 191 mcg/L, p = 0.20 and 15.0 vs. 19.7 mcg/L, p = 0.02, respectively). Pending studies on vitamin K supplementation, calcimimetics, and vitamin D analogs may play a role in preserving vitamin K-dependent protein activity, thus contributing to bone and vascular health in CKD patients.
& Maria Fusaro [email protected]
7
Kidney – Pancreas Transplant Unit, Department of Surgical, Oncological and Gastroenterological Sciences, University of Padua, Padua, Italy
National Research Council (CNR), Neuroscience Institute, Via Giustiniani, 2, 35128 Padua, Italy
8
Nephrology and Dialysis Unit, Hospital of Rovigo, Rovigo, Italy
Department of Medicine, Clinica Medica 1, University of Padua, Padua, Italy
9
Nephrology and Dialysis Unit, Hospital of Lodi, Lodi, Italy
10
Nephrology Unit, University of Padua, Padua, Italy
11
CNR Institute of Clinical Physiology, Pisa, Italy
12
Laboratory Medicine Unit, Department of Medicine, University of Padua, Padua, Italy
1
2
3
Nephrology and Dialysis Unit, San Carlo Borromeo Hospital, Department of Biomedical and Clinical Sciences ‘‘L. Sacco’’, University of Milan, Milan, Italy
4
CNR-IFC Pathophysiology of Renal Diseases and Hypertension Unit, Reggio Calabria, Italy
5
Rheumatology Unit, Department of Medicine, University of Verona, Verona, Italy
6
Fondazione Ca’ Granda-IRCCS Ospedale Maggiore Policlinico, Milan, Italy
Keywords Matrix mineralization Biochemical markers of bone turnover CKD Vitamin K proteins Vascular calcification
123
Endocrine
Introduction
Materials and methods
Vascular calcification (VC) is an important risk factor for cardiovascular morbidity and mortality in chronic kidney disease (CKD) [1]. The prevalence of calcification of large arteries has been found to be as high as 60–80 % in hemodialysis patients [2, 3]. It has recently been ascertained that VC is an actively regulated process and there is a general consensus that the extent of VC is the outcome of the balance between factors inducing calcification, such as high serum phosphorus, and inhibiting factors, such as the vitamin K-dependent proteins [1, 4]. The role of these proteins in both bone mineralization and VC has been demonstrated by gene knockout experiments: matrix Gla protein (MGP) gene knockout mice develop pathological fractures due to severe osteoporosis and severe medial vascular calcifications leading to aortic rupture and death early in life [5]. Also another vitamin K-dependent protein, namely bone Gla protein (BGP), also called osteocalcin, has proved to be involved both in the regulation of VC and in bone mineralization: BGP knockout mice develop early hyperostosis [6] and a recent mortality trial has shown that it is associated with reduced aortic VC [7]. Interestingly, a significant association between VC in medium caliber arteries and the rate of prevalent vertebral fractures has been reported in haemodialysis patients [8]. In the same study, both severe VC and vertebral fractures were positively associated with mortality, with a threefold and fivefold increase in relative risk, respectively, in women [8]. Secondary hyperparathyroidism (HPT) is a common complication of CKD. Calcitriol is an effective treatment for lowering parathyroid hormone (PTH), but it may cause detrimentally elevated levels of serum calcium and phosphate. Moreover, at high doses, it activates specific vitamin D receptors (VDR) expressed in vascular smooth muscle cells (VSMCs) and promotes VC both in vitro and in vivo [9, 10]. The development of vitamin D analogs has resulted in therapy that not only prevents excessive increases in calcium and phosphorus levels, but also VC, both in vitro and in vivo [9, 10]. Calcimimetics bind to calcium-sensing receptors, increasing their sensitivity to extracellular calcium and thereby interfering with signals stimulating PTH synthesis and secretion. Unlike vitamin D sterols, they do not induce elevated calcium and phosphorus levels and have proved to be able to prevent VC [11]. We performed a secondary analysis of the VIKI study, an observational study designed to assess the prevalence of vitamin K deficiency in hemodialysis patients, to investigate the effects of ongoing treatment for mineral and bone disorders on BGP and MGP levels.
This is a sub-study of the VIKI Dialysis study [12]. Briefly, the VIKI study was an observational study performed in 387 adult patients of both genders who had been on dialysis for more than 1 year at 18 centers in Italy between November 2008 and November 2009. It was designed primarily to assess the prevalence of deficiency of vitamin K1 and vitamin K2 in dialysis patients. One of its secondary objectives was to assess the impact of vitamin K status on vitamin K-dependent proteins, namely MGP and BGP and their association with VF and VC. Concomitant treatment, including mineral and bone disorders treatment, i.e. oral calcitriol, vitamin D analogs, calcimimetics, and phosphate-binding drugs, was amongst the information collected. Two 5 mL samples of venous blood were drawn when the patients were fasting for the determination, before dialysis session, of routine bone biochemistry (total alkaline phosphatase, PTH, 25-OH Vitamin D), total osteocalcin (Bone Gla Protein, BGP), undercarboxylated osteocalcin (ucBGP), total matrix Gla protein (MGP), and undercarboxylated MGP. A radiograph of the thoracic and lumbar regions of the spinal column (D5 to L4) in the latero-lateral view with the patient in the lateral recumbent position was obtained. A vertebral fracture was considered to be present when the height of the vertebral body was reduced by at least 20 % (4 mm), according to Genant [13, 14]. Vascular calcifications were quantified by measuring the length of calcific deposits along the abdominal aortic wall (mild 0.1–5 cm, moderate 5.1–10 cm and severe [10 cm). The presence of calcifications of the iliac arteries was evaluated through the same radiograph (mild 0.1–3 cm, moderate 3.1–5 cm, and severe [5 cm) [15]. The vital status of patients was checked through a follow-up telephone interview of 2.7 ± 0.5 years [12].
123
Laboratory determination Parathyroid hormone (PTH) The method for quantitative determination of PTH in serum was the automated LIAISONÒ N-TactÒ PTH Assay 310910 (DiaSorin Inc., Stillwater MN, USA), which is a direct, two-site, sandwich-type chemiluminescence immunoassay (CLIA) carried out on the LIAISONÒ (DiaSorin Inc., Stillwater MN, USA) instrument. The analytical sensitivity is 1 pg/mL, and the intra-assay and inter-assay CVs were 3.7–6.3 and 3.5–5.3 %, respectively.
Endocrine
25-OH vitamin D For quantitative determination of total 25-OH vitamin D (both D2 and D3 form) in serum, we used the automated LIAISONÒ 25 OH Vitamin D TOTAL Assay 310600, which is a direct competitive chemiluminescence immunoassay (CLIA) executed on the LIAISONÒ (DiaSorin Inc., Stillwater MN, USA) instrument. The analytical sensitivity is \10 nmol/L, and the intra-assay coefficients of variation (CV) have been found to range between 2.9 and 5.5 %, while the inter-assay CV is 6.3–12.9 %. Total matrix GLA protein (MGP) The quantitative determination of MGP was performed using the Human MGP—Matrix Gla Protein Kit (Biomedica Medizinprodukte GmbH & Co KG, Wien, A). It is a manual competitive ELISA method designed to detect MGP in serum. The analytical sensitivity is 0.3 nmol/L, and the intra-assay and inter-assay coefficients of variation (CVs) are 5–6 and 7–9 %, respectively. Undercarboxylated MGP (ucMGP) The measurement of the total undercarboxylated Matrix GLA Protein was performed by VitaK using a competitive ELISA, as described previously [16]. The analytical sensitivity is 21 nmol/L, and the intra-assay and inter-assay CVs have been found to be 8.9 and 11.4 %, respectively. Total osteocalcin (BGP) The method for quantitative determination of total osteocalcin in serum was the automated LIAISONÒ Osteocalcin Assay 310950 (DiaSorin Inc., Stillwater MN, USA), which is a direct, two-site, sandwich-type chemiluminescence immunoassay (CLIA) executed on the LIAISONÒ (DiaSorin Inc., Stillwater MN, USA) instrument. The analytical sensitivity is\0.3 ng/mL, and the intra-assay CV is 3–8 %, while the inter-assay CV is 4–9 %. Undercarboxylated osteocalcin (ucBGP) For quantitative determination of the undercarboxylated form, we used Glu-osteocalcin EIA Kit MK118 (Takara Bio Inc., Otsu, Shiga, Japan), a manual solid-phase EIA based on a sandwich method that utilizes two mouse monoclonal anti-ucBGP antibodies to detect ucBGP by a two-step procedure. One of the mouse monoclonal antiundercarboxylated BGPs is immobilized onto the microtiter plate and blocked against non-specific binding. Samples are added to each well and incubated. The second step is to wash the plate and to add the second anti-BGP labeled
with peroxidase (POD). The reaction between POD and substrate (H2O2 and 3,30 , 5,50 tetramethylbenzidine) results in color development with intensities proportional to the amount of ucBGP present in the samples. The analytical sensitivity is 0.25 ng/mL, and the intra-assay and interassay CVs are 4.4–6.7 and 5.7–9.9 %, respectively. Statistical analysis The data are shown as mean ± standard deviation (SD) for quantitative variables or median and interquartile range (IQ) for not normal or strongly asymmetric variables, and frequency percentages for all discrete variables. The comparisons between BGP and MGP medians (including their undercarboxylated forms) for subjects on mineral and bone disorders treatment (oral calcitriol or not, vitamin D analogs or not, calcimimetics or not, and phosphate-binding drugs or not) for subjects with or without VF, with or without aortic VC, with or without iliac VC, and for subjects alive or dead at the telephone interview follow-up, were performed using the Wilcoxon sum-rank test. Four multiple regression models were defined, with outcomes BGP and MGP (total and undercarboxylated); each outcome underwent logarithmic transformation. Demographic data (gender, age), renal failure history, lifestyle (smoking status, alcohol consumption), medical history (cardio- and cerebrovascular disease, diabetes mellitus, malabsorption syndrome and liver disease), BMI, results from routine biochemical examinations, and mineral and bone disorders treatment (oral calcitriol, vitamin D analogs, calcimimetics, and phosphate-binding drugs) were considered independent variables. A stepwise selection procedure was adopted, with p value to entry 0.10 and a p value to stay 0.15. Further regression models were defined considering as possible independent variables combinations of drugs: calcimimetics and vitamin D analogs for BGP outcomes (total and undercarboxylated), calcimimetics and calcium acetate binding for MGP (total and undercaboxylated). Data were expressed as unstandardized regression coefficients (parameter estimate) and p value. All statistical analyses were performed using SAS statistical package (version 9.3, SAS, Cary, NC).
Results Data were collected in 387 hemodialyzed patients. Patients were mostly male (n = 244, 63 %). Their mean age was 64.2 ± 14.1 years and they had been on dialysis on average for 4 years, mainly bicarbonate dialysis (n = 189, 48.7 %) or hemofiltration (n = 102, 26.4 %). Many of
123
Endocrine Table 1 Mineral and bone disorders treatments Phosphate binders Aluminum, n (%) Calcium carbonate, n (%) Calcium acetate, n (%) Sevelamer, n (%)
96 (24.8) 132 (34.1) 21 (5.4) 163 (42.1)
Lanthanum, n (%)
56 (14.5)
Oral calcitriol, n (%)
177 (45.7)
Intravenous calcitriol, n (%)
12 (3.1)
Intravenous vitamin D analogs, n (%)
77 (19.9)
Calcimimetics, n (%)
75 (19.4)
them had cardiovascular disease, especially arterial hypertension (n = 304, 78.6 %) and/or peripheral artery disease (n = 134, 34.6 %); some had a history of cardiovascular events (myocardial infarction n = 73, 18.9 %, cerebrovascular event n = 41, 10.6 %). In addition, 64 subjects (16.5 %) reported angina and 85 (22 %) had diabetes mellitus. More than half of the patients had at least one VF (55.3 %) and 40 (10.4 %) had experienced a fracture after starting dialysis. Regarding mineral and bone disorder treatment, the most common were oral calcitriol (45.7 %) and sevelamer (42.1 %). Similar proportions of patients were on treatment with vitamin D analogs (19.9 %) or with calcimimetics (19.4 %), whereas only 5.4 % were on treatment with calcium acetate (Table 1). At univariate analysis, median total BGP levels were significantly increased in patients treated only with calcimimetics or vitamin D analogs. Total BGP levels were almost twice as high in patients on calcimimetic Fig. 1 Median total BGP levels were significantly higher in patients on calcimimetic treatment than those who were not (290 vs. 158.5 mcg/L, p \ 0.0001). A significant difference was also seen in patients on vitamin D analog treatment versus those who were not (268 vs. 159 mcg/L, p \ 0.0001). Univariate analysis
123
treatment compared to those who were not (290 vs. 158.5 mcg/L, p \ 0.0001) and 69 % higher in patients on vitamin D analog treatment than in those who were not (268 vs. 159 mcg/L, p \ 0.0001) (Fig. 1). Furthermore, univariate analysis highlighted total BGP levels of 147 mcg/L among patients on calcium carbonate treatment versus 204 among patients who were not, (p = 0.0008) and BGP levels of 210 mcg/L among patients on Sevelamer treatment versus 152 among patients who were not, (p = 0.0022). Total BPG levels were not significantly different with the remaining mineral and bone disorders treatment (Table 2). Median total MGP levels were higher in patients treated with calcimimetics than in those who were not (21.5 vs. 18.1 mcg/L, p = 0.041). The same was true for calcium acetate (27.9 vs. 18.1 mcg/L, p = 0.0030) (Fig. 2) and aluminum phosphate binders (22.1 vs. 16.9 mcg/L, p = 0.044). Median total MGP and undercarboxylated MGP (ucMGP) levels were lower in patients treated with calcium carbonate than in those who were not (15 treated vs. 20.1 mcg/L, p = 0.031 and 438 vs. 634 mcg/L, p = 0.001, respectively). Sevelamer use was associated with reduced total MGP levels: 16.4 versus 20.3 mcg/L, p = 0.0368 (Fig. 2). Median total MGP was not significantly different among patients on the other kinds of mineral and bone disorder treatments (Table 2). The multiple regression model with log-transformed total BGP values showed that the geometric mean of total BGP was 37.7 % higher in subjects treated with calcimimetics than in those who were not (p = 0.0002) and 20.7 % higher in subjects treated with vitamin D analogs than in those who were not (p = 0.0228). The multiple
Endocrine Table 2 Associations of treatments for CKD-MBD and serum levels of BGP and MGP Total BGP, mcg/L median (IQ)
ucBGP, mgc/L median (IQ)
Total MGP, nmol/L median (IQ)
ucMGP, nmol/L median (IQ)
Treated with
200.5 (96.4, 448.6)
11.8 (7.7, 22.1)
22.1 (15.2, 32.5)
583.5 (253.5, 1107.5)
Not treated with
170.0 (94.8, 280)
10.8 (4.3, 17)
16.9 (12, 29.4)
569 (296.9, 908)
p value
0.1287
0.0855
0.044
0.5973
Treated with
147 (82.6, 233.5)
10.9 (4.5, 19.4)
15 (12, 29.5)
438 (255.1, 791.2)
Not treated with p value
204 (106, 362) 0.0008
11.1 (5, 17) 0.7351
20.1 (13.3, 31.9) 0.0311
634 (315, 1071) 0.0010
Treated with
179 (83.1, 248.6)
10.9 (6.4, 15.6)
27.9 (20, 44.5)
791 (446, 1062)
Not treated with
183 (97.2, 319)
11 (4.6, 17.2)
18.1 (12.4, 30.1)
559.3 (288, 922)
p value
0.4870
0.9185
0.0030
0.3366
Treated with
210 (117, 363.2)
11.9 (6.3, 17)
16.4 (12, 27.6)
547 (296.9, 920)
Not treated with
152 (63.7, 276)
10.6 (4.2, 18.7)
20.3 (13.6, 33.1)
584 (266, 943.5)
p value
0.0022
0.2926
0.0368
0.9930
Treated with
199.5 (106, 415.3)
10.1 (4.6, 15.9)
17.9 (13.8, 35.9)
613.9 (360.4, 1050)
Not treated with
180 (93.5, 290)
11.1 (4.6, 17.4)
18.9 (12.6, 30)
557 (264, 918)
p value
0.1978
0.7389
0.3533
0.1825
Treated with Not treated with
184 (91.3, 347) 175.5 (102, 297)
11.2 (6, 16.2) 10.9 (4, 19)
17.8 (13, 27.6) 20.4 (12.6, 32.1)
585 (286.8, 921) 549.4 (288, 933)
p value
0.8449
0.8136
0.1916
0.7350
Treated with
237.5 (153.5, 464.2)
13.6 (9.1, 17.6)
19.9 (13.6, 24.6)
719 (258, 861.5)
Not treated with
179 (94, 312.4)
11 (4.6, 17.2)
18.7 (12.7, 30.9)
563.4 (288, 942)
p value
0.1267
0.2503
0.9801
0.9071
Treated with
268 (146, 459)
15.2 (8.4, 43.2)
21.1 (12.4, 34.7)
457 (237, 876)
Not treated with
159 (86.1, 275)
10.3 (4, 16.2)
18.3 (13, 29.8)
584 (305, 942)
p value
ORIGINAL ARTICLE
Calcimimetic and vitamin D analog use in hemodialyzed patients is associated with increased levels of vitamin K dependent proteins Maria Fusaro1 • Sandro Giannini2 • Maurizio Gallieni3 • Marianna Noale1 • Giovanni Tripepi4 • Maurizio Rossini5 • Piergiorgio Messa6 • Paolo Rigotti7 • Tecla Pati8 • Francesco Barbisoni9 • Antonio Piccoli10 • Andrea Aghi10 • Marianna Alessi10 • Luciana Bonfante10 • Fabrizio Fabris2 • Sabina Zambon2 Stefania Sella2 • Giorgio Iervasi11 • Mario Plebani12
•
Received: 9 April 2015 / Accepted: 20 June 2015 Ó Springer Science+Business Media New York 2015
Abstract Matrix Gla protein (MGP) and bone Gla protein (BGP) are two vitamin K-dependent proteins (VKDPs) involved in the regulation of vascular calcification (VC). We carried out a secondary analysis of the VIKI study to evaluate associations between drug consumption and VKDP levels in 387 hemodialyzed patients. The VIKI study assessed the prevalence of vitamin K deficiency in hemodialysis patients. We evaluated drug consumption, determined BGP and MGP levels, and verified the presence of any vertebral fractures (VF) and VC by spine radiographs. Total BGP levels were twice as high with calcimimetics versus no calcimimetics (290 vs. 158.5 mcg/L, p \ 0.0001) and 69 % higher with vitamin D analogs (268 vs. 159 mcg/L, p \ 0.0001). Total MGP was 19 % higher with calcimimetics (21.5 vs. 18.1 mcg/L, p = 0.04) and 54 % higher with calcium acetate (27.9 vs. 18.1 mcg/L,
p = 0.003); no difference was found with vitamin D analogs (21.1 vs. 18.3 mcg/L, p = 0.43). Median Total BGP level was 29 % lower in patients with C1 VF (151 vs. 213 mcg/L, p = 0.0091) and 36 % lower in patients with VC (164 vs. 262.1 mcg/L, p = 0.0003). In non-survivors, median BGP and MGP were lower, but only for MGP this difference reached the statistical significance (152 vs. 191 mcg/L, p = 0.20 and 15.0 vs. 19.7 mcg/L, p = 0.02, respectively). Pending studies on vitamin K supplementation, calcimimetics, and vitamin D analogs may play a role in preserving vitamin K-dependent protein activity, thus contributing to bone and vascular health in CKD patients.
& Maria Fusaro [email protected]
7
Kidney – Pancreas Transplant Unit, Department of Surgical, Oncological and Gastroenterological Sciences, University of Padua, Padua, Italy
National Research Council (CNR), Neuroscience Institute, Via Giustiniani, 2, 35128 Padua, Italy
8
Nephrology and Dialysis Unit, Hospital of Rovigo, Rovigo, Italy
Department of Medicine, Clinica Medica 1, University of Padua, Padua, Italy
9
Nephrology and Dialysis Unit, Hospital of Lodi, Lodi, Italy
10
Nephrology Unit, University of Padua, Padua, Italy
11
CNR Institute of Clinical Physiology, Pisa, Italy
12
Laboratory Medicine Unit, Department of Medicine, University of Padua, Padua, Italy
1
2
3
Nephrology and Dialysis Unit, San Carlo Borromeo Hospital, Department of Biomedical and Clinical Sciences ‘‘L. Sacco’’, University of Milan, Milan, Italy
4
CNR-IFC Pathophysiology of Renal Diseases and Hypertension Unit, Reggio Calabria, Italy
5
Rheumatology Unit, Department of Medicine, University of Verona, Verona, Italy
6
Fondazione Ca’ Granda-IRCCS Ospedale Maggiore Policlinico, Milan, Italy
Keywords Matrix mineralization Biochemical markers of bone turnover CKD Vitamin K proteins Vascular calcification
123
Endocrine
Introduction
Materials and methods
Vascular calcification (VC) is an important risk factor for cardiovascular morbidity and mortality in chronic kidney disease (CKD) [1]. The prevalence of calcification of large arteries has been found to be as high as 60–80 % in hemodialysis patients [2, 3]. It has recently been ascertained that VC is an actively regulated process and there is a general consensus that the extent of VC is the outcome of the balance between factors inducing calcification, such as high serum phosphorus, and inhibiting factors, such as the vitamin K-dependent proteins [1, 4]. The role of these proteins in both bone mineralization and VC has been demonstrated by gene knockout experiments: matrix Gla protein (MGP) gene knockout mice develop pathological fractures due to severe osteoporosis and severe medial vascular calcifications leading to aortic rupture and death early in life [5]. Also another vitamin K-dependent protein, namely bone Gla protein (BGP), also called osteocalcin, has proved to be involved both in the regulation of VC and in bone mineralization: BGP knockout mice develop early hyperostosis [6] and a recent mortality trial has shown that it is associated with reduced aortic VC [7]. Interestingly, a significant association between VC in medium caliber arteries and the rate of prevalent vertebral fractures has been reported in haemodialysis patients [8]. In the same study, both severe VC and vertebral fractures were positively associated with mortality, with a threefold and fivefold increase in relative risk, respectively, in women [8]. Secondary hyperparathyroidism (HPT) is a common complication of CKD. Calcitriol is an effective treatment for lowering parathyroid hormone (PTH), but it may cause detrimentally elevated levels of serum calcium and phosphate. Moreover, at high doses, it activates specific vitamin D receptors (VDR) expressed in vascular smooth muscle cells (VSMCs) and promotes VC both in vitro and in vivo [9, 10]. The development of vitamin D analogs has resulted in therapy that not only prevents excessive increases in calcium and phosphorus levels, but also VC, both in vitro and in vivo [9, 10]. Calcimimetics bind to calcium-sensing receptors, increasing their sensitivity to extracellular calcium and thereby interfering with signals stimulating PTH synthesis and secretion. Unlike vitamin D sterols, they do not induce elevated calcium and phosphorus levels and have proved to be able to prevent VC [11]. We performed a secondary analysis of the VIKI study, an observational study designed to assess the prevalence of vitamin K deficiency in hemodialysis patients, to investigate the effects of ongoing treatment for mineral and bone disorders on BGP and MGP levels.
This is a sub-study of the VIKI Dialysis study [12]. Briefly, the VIKI study was an observational study performed in 387 adult patients of both genders who had been on dialysis for more than 1 year at 18 centers in Italy between November 2008 and November 2009. It was designed primarily to assess the prevalence of deficiency of vitamin K1 and vitamin K2 in dialysis patients. One of its secondary objectives was to assess the impact of vitamin K status on vitamin K-dependent proteins, namely MGP and BGP and their association with VF and VC. Concomitant treatment, including mineral and bone disorders treatment, i.e. oral calcitriol, vitamin D analogs, calcimimetics, and phosphate-binding drugs, was amongst the information collected. Two 5 mL samples of venous blood were drawn when the patients were fasting for the determination, before dialysis session, of routine bone biochemistry (total alkaline phosphatase, PTH, 25-OH Vitamin D), total osteocalcin (Bone Gla Protein, BGP), undercarboxylated osteocalcin (ucBGP), total matrix Gla protein (MGP), and undercarboxylated MGP. A radiograph of the thoracic and lumbar regions of the spinal column (D5 to L4) in the latero-lateral view with the patient in the lateral recumbent position was obtained. A vertebral fracture was considered to be present when the height of the vertebral body was reduced by at least 20 % (4 mm), according to Genant [13, 14]. Vascular calcifications were quantified by measuring the length of calcific deposits along the abdominal aortic wall (mild 0.1–5 cm, moderate 5.1–10 cm and severe [10 cm). The presence of calcifications of the iliac arteries was evaluated through the same radiograph (mild 0.1–3 cm, moderate 3.1–5 cm, and severe [5 cm) [15]. The vital status of patients was checked through a follow-up telephone interview of 2.7 ± 0.5 years [12].
123
Laboratory determination Parathyroid hormone (PTH) The method for quantitative determination of PTH in serum was the automated LIAISONÒ N-TactÒ PTH Assay 310910 (DiaSorin Inc., Stillwater MN, USA), which is a direct, two-site, sandwich-type chemiluminescence immunoassay (CLIA) carried out on the LIAISONÒ (DiaSorin Inc., Stillwater MN, USA) instrument. The analytical sensitivity is 1 pg/mL, and the intra-assay and inter-assay CVs were 3.7–6.3 and 3.5–5.3 %, respectively.
Endocrine
25-OH vitamin D For quantitative determination of total 25-OH vitamin D (both D2 and D3 form) in serum, we used the automated LIAISONÒ 25 OH Vitamin D TOTAL Assay 310600, which is a direct competitive chemiluminescence immunoassay (CLIA) executed on the LIAISONÒ (DiaSorin Inc., Stillwater MN, USA) instrument. The analytical sensitivity is \10 nmol/L, and the intra-assay coefficients of variation (CV) have been found to range between 2.9 and 5.5 %, while the inter-assay CV is 6.3–12.9 %. Total matrix GLA protein (MGP) The quantitative determination of MGP was performed using the Human MGP—Matrix Gla Protein Kit (Biomedica Medizinprodukte GmbH & Co KG, Wien, A). It is a manual competitive ELISA method designed to detect MGP in serum. The analytical sensitivity is 0.3 nmol/L, and the intra-assay and inter-assay coefficients of variation (CVs) are 5–6 and 7–9 %, respectively. Undercarboxylated MGP (ucMGP) The measurement of the total undercarboxylated Matrix GLA Protein was performed by VitaK using a competitive ELISA, as described previously [16]. The analytical sensitivity is 21 nmol/L, and the intra-assay and inter-assay CVs have been found to be 8.9 and 11.4 %, respectively. Total osteocalcin (BGP) The method for quantitative determination of total osteocalcin in serum was the automated LIAISONÒ Osteocalcin Assay 310950 (DiaSorin Inc., Stillwater MN, USA), which is a direct, two-site, sandwich-type chemiluminescence immunoassay (CLIA) executed on the LIAISONÒ (DiaSorin Inc., Stillwater MN, USA) instrument. The analytical sensitivity is\0.3 ng/mL, and the intra-assay CV is 3–8 %, while the inter-assay CV is 4–9 %. Undercarboxylated osteocalcin (ucBGP) For quantitative determination of the undercarboxylated form, we used Glu-osteocalcin EIA Kit MK118 (Takara Bio Inc., Otsu, Shiga, Japan), a manual solid-phase EIA based on a sandwich method that utilizes two mouse monoclonal anti-ucBGP antibodies to detect ucBGP by a two-step procedure. One of the mouse monoclonal antiundercarboxylated BGPs is immobilized onto the microtiter plate and blocked against non-specific binding. Samples are added to each well and incubated. The second step is to wash the plate and to add the second anti-BGP labeled
with peroxidase (POD). The reaction between POD and substrate (H2O2 and 3,30 , 5,50 tetramethylbenzidine) results in color development with intensities proportional to the amount of ucBGP present in the samples. The analytical sensitivity is 0.25 ng/mL, and the intra-assay and interassay CVs are 4.4–6.7 and 5.7–9.9 %, respectively. Statistical analysis The data are shown as mean ± standard deviation (SD) for quantitative variables or median and interquartile range (IQ) for not normal or strongly asymmetric variables, and frequency percentages for all discrete variables. The comparisons between BGP and MGP medians (including their undercarboxylated forms) for subjects on mineral and bone disorders treatment (oral calcitriol or not, vitamin D analogs or not, calcimimetics or not, and phosphate-binding drugs or not) for subjects with or without VF, with or without aortic VC, with or without iliac VC, and for subjects alive or dead at the telephone interview follow-up, were performed using the Wilcoxon sum-rank test. Four multiple regression models were defined, with outcomes BGP and MGP (total and undercarboxylated); each outcome underwent logarithmic transformation. Demographic data (gender, age), renal failure history, lifestyle (smoking status, alcohol consumption), medical history (cardio- and cerebrovascular disease, diabetes mellitus, malabsorption syndrome and liver disease), BMI, results from routine biochemical examinations, and mineral and bone disorders treatment (oral calcitriol, vitamin D analogs, calcimimetics, and phosphate-binding drugs) were considered independent variables. A stepwise selection procedure was adopted, with p value to entry 0.10 and a p value to stay 0.15. Further regression models were defined considering as possible independent variables combinations of drugs: calcimimetics and vitamin D analogs for BGP outcomes (total and undercarboxylated), calcimimetics and calcium acetate binding for MGP (total and undercaboxylated). Data were expressed as unstandardized regression coefficients (parameter estimate) and p value. All statistical analyses were performed using SAS statistical package (version 9.3, SAS, Cary, NC).
Results Data were collected in 387 hemodialyzed patients. Patients were mostly male (n = 244, 63 %). Their mean age was 64.2 ± 14.1 years and they had been on dialysis on average for 4 years, mainly bicarbonate dialysis (n = 189, 48.7 %) or hemofiltration (n = 102, 26.4 %). Many of
123
Endocrine Table 1 Mineral and bone disorders treatments Phosphate binders Aluminum, n (%) Calcium carbonate, n (%) Calcium acetate, n (%) Sevelamer, n (%)
96 (24.8) 132 (34.1) 21 (5.4) 163 (42.1)
Lanthanum, n (%)
56 (14.5)
Oral calcitriol, n (%)
177 (45.7)
Intravenous calcitriol, n (%)
12 (3.1)
Intravenous vitamin D analogs, n (%)
77 (19.9)
Calcimimetics, n (%)
75 (19.4)
them had cardiovascular disease, especially arterial hypertension (n = 304, 78.6 %) and/or peripheral artery disease (n = 134, 34.6 %); some had a history of cardiovascular events (myocardial infarction n = 73, 18.9 %, cerebrovascular event n = 41, 10.6 %). In addition, 64 subjects (16.5 %) reported angina and 85 (22 %) had diabetes mellitus. More than half of the patients had at least one VF (55.3 %) and 40 (10.4 %) had experienced a fracture after starting dialysis. Regarding mineral and bone disorder treatment, the most common were oral calcitriol (45.7 %) and sevelamer (42.1 %). Similar proportions of patients were on treatment with vitamin D analogs (19.9 %) or with calcimimetics (19.4 %), whereas only 5.4 % were on treatment with calcium acetate (Table 1). At univariate analysis, median total BGP levels were significantly increased in patients treated only with calcimimetics or vitamin D analogs. Total BGP levels were almost twice as high in patients on calcimimetic Fig. 1 Median total BGP levels were significantly higher in patients on calcimimetic treatment than those who were not (290 vs. 158.5 mcg/L, p \ 0.0001). A significant difference was also seen in patients on vitamin D analog treatment versus those who were not (268 vs. 159 mcg/L, p \ 0.0001). Univariate analysis
123
treatment compared to those who were not (290 vs. 158.5 mcg/L, p \ 0.0001) and 69 % higher in patients on vitamin D analog treatment than in those who were not (268 vs. 159 mcg/L, p \ 0.0001) (Fig. 1). Furthermore, univariate analysis highlighted total BGP levels of 147 mcg/L among patients on calcium carbonate treatment versus 204 among patients who were not, (p = 0.0008) and BGP levels of 210 mcg/L among patients on Sevelamer treatment versus 152 among patients who were not, (p = 0.0022). Total BPG levels were not significantly different with the remaining mineral and bone disorders treatment (Table 2). Median total MGP levels were higher in patients treated with calcimimetics than in those who were not (21.5 vs. 18.1 mcg/L, p = 0.041). The same was true for calcium acetate (27.9 vs. 18.1 mcg/L, p = 0.0030) (Fig. 2) and aluminum phosphate binders (22.1 vs. 16.9 mcg/L, p = 0.044). Median total MGP and undercarboxylated MGP (ucMGP) levels were lower in patients treated with calcium carbonate than in those who were not (15 treated vs. 20.1 mcg/L, p = 0.031 and 438 vs. 634 mcg/L, p = 0.001, respectively). Sevelamer use was associated with reduced total MGP levels: 16.4 versus 20.3 mcg/L, p = 0.0368 (Fig. 2). Median total MGP was not significantly different among patients on the other kinds of mineral and bone disorder treatments (Table 2). The multiple regression model with log-transformed total BGP values showed that the geometric mean of total BGP was 37.7 % higher in subjects treated with calcimimetics than in those who were not (p = 0.0002) and 20.7 % higher in subjects treated with vitamin D analogs than in those who were not (p = 0.0228). The multiple
Endocrine Table 2 Associations of treatments for CKD-MBD and serum levels of BGP and MGP Total BGP, mcg/L median (IQ)
ucBGP, mgc/L median (IQ)
Total MGP, nmol/L median (IQ)
ucMGP, nmol/L median (IQ)
Treated with
200.5 (96.4, 448.6)
11.8 (7.7, 22.1)
22.1 (15.2, 32.5)
583.5 (253.5, 1107.5)
Not treated with
170.0 (94.8, 280)
10.8 (4.3, 17)
16.9 (12, 29.4)
569 (296.9, 908)
p value
0.1287
0.0855
0.044
0.5973
Treated with
147 (82.6, 233.5)
10.9 (4.5, 19.4)
15 (12, 29.5)
438 (255.1, 791.2)
Not treated with p value
204 (106, 362) 0.0008
11.1 (5, 17) 0.7351
20.1 (13.3, 31.9) 0.0311
634 (315, 1071) 0.0010
Treated with
179 (83.1, 248.6)
10.9 (6.4, 15.6)
27.9 (20, 44.5)
791 (446, 1062)
Not treated with
183 (97.2, 319)
11 (4.6, 17.2)
18.1 (12.4, 30.1)
559.3 (288, 922)
p value
0.4870
0.9185
0.0030
0.3366
Treated with
210 (117, 363.2)
11.9 (6.3, 17)
16.4 (12, 27.6)
547 (296.9, 920)
Not treated with
152 (63.7, 276)
10.6 (4.2, 18.7)
20.3 (13.6, 33.1)
584 (266, 943.5)
p value
0.0022
0.2926
0.0368
0.9930
Treated with
199.5 (106, 415.3)
10.1 (4.6, 15.9)
17.9 (13.8, 35.9)
613.9 (360.4, 1050)
Not treated with
180 (93.5, 290)
11.1 (4.6, 17.4)
18.9 (12.6, 30)
557 (264, 918)
p value
0.1978
0.7389
0.3533
0.1825
Treated with Not treated with
184 (91.3, 347) 175.5 (102, 297)
11.2 (6, 16.2) 10.9 (4, 19)
17.8 (13, 27.6) 20.4 (12.6, 32.1)
585 (286.8, 921) 549.4 (288, 933)
p value
0.8449
0.8136
0.1916
0.7350
Treated with
237.5 (153.5, 464.2)
13.6 (9.1, 17.6)
19.9 (13.6, 24.6)
719 (258, 861.5)
Not treated with
179 (94, 312.4)
11 (4.6, 17.2)
18.7 (12.7, 30.9)
563.4 (288, 942)
p value
0.1267
0.2503
0.9801
0.9071
Treated with
268 (146, 459)
15.2 (8.4, 43.2)
21.1 (12.4, 34.7)
457 (237, 876)
Not treated with
159 (86.1, 275)
10.3 (4, 16.2)
18.3 (13, 29.8)
584 (305, 942)
p value
