Advertisement

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

Open AccessPublished:June 19, 2022DOI:https://doi.org/10.1016/j.atherosclerosis.2022.06.1017

      Highlights

      • Dp-ucMGP is a marker of vitamin K insufficiency resulting in impaired inhibition of vascular calcification.
      • Dp-ucMGP was associated with future cardiovascular disease events and all-cause mortality.
      • Dp-ucMGP was associated with vascular deaths and non-vascular/non-cancer deaths.
      • Dp-ucMGP was not associated with myocardial infarction and sudden cardiac deaths.
      • Clinical trials are warranted to examine whether vitamin K substitution results in improved health outcomes.

      Abstract

      Background and aims

      Matrix Gla protein (MGP), a vitamin K-dependent protein, is a potent inhibitor of vascular calcification. Desphospho-uncarboxylated MGP (dp-ucMGP), a marker of vitamin K insufficiency, has been shown to predict cardiovascular disease (CVD) and all-cause mortality in high-risk populations. Whether the increased risk associated with dp-ucMGP also applies to the general, and especially, the elderly population has not yet been fully elucidated.

      Methods and results

      Plasma dp-ucMGP was measured in 684 individuals aged 50–89 years of the prospective population-based Bruneck Study (baseline evaluation in 2000). Baseline median dp-ucMGP was 478.4 (IQR 335.0–635.2) pmol/L. Over a median follow-up of 15.5 years, 163 CVD events occurred and 235 participants died. Age-/sex-adjusted hazard ratios (HRs) per 1-SD higher level of loge transformed dp-ucMGP were 1.30 (95%CI: 1.09–1.55; p=0.004) for incident CVD and 1.36 (95%CI: 1.17–1.57; p<0.001) for all-cause mortality. After multivariable adjustment, the associations remained significant with HRs of 1.23 (95%CI: 1.02–1.47, p=0.029) for CVD and 1.40 (95%CI: 1.20–1.64; p<0.001) for all-cause mortality. The associations remained virtually unchanged after additional adjustment for dietary quality as measured with the Alternative Healthy Eating Index. We found no association of dp-ucMGP with myocardial infarction and sudden cardiac deaths, but a strong association with other vascular deaths and non-vascular/non-cancer deaths.

      Conclusions

      This study shows a significant association of plasma dp-ucMGP with incident CVD and a significant and even stronger association with all-cause mortality. Clinical trials are needed to investigate whether vitamin K substitution results in improved health outcomes.

      Graphical abstract

      Keywords

      1. Introduction

      Vascular calcification is a highly prevalent phenotype and an independent predictor of cardiovascular disease (CVD) and mortality [
      • Rennenberg R.J.M.W.
      • Kessels A.G.H.
      • Schurgers L.J.
      • van Engelshoven J.M.A.
      • de Leeuw P.W.
      • Kroon A.A.
      Vascular calcifications as a marker of increased cardiovascular risk: a meta-analysis.
      ]. It is a biologically active process, tightly regulated via complex molecular signaling pathways [
      • Sallam T.
      • Cheng H.
      • Demer L.L.
      • Tintut Y.
      Regulatory circuits controlling vascular cell calcification.
      ]. Matrix Gla protein (MGP) is the most powerful inhibitor of vascular calcification in humans, primarily expressed by vascular smooth muscle cells in the arterial wall [
      • Roumeliotis S.
      • Dounousi E.
      • Eleftheriadis T.
      • Liakopoulos V.
      Association of the inactive circulating matrix gla protein with vitamin K intake, calcification, mortality, and cardiovascular disease: a review.
      ]. MGP activity depends on two post-translational modifications: vitamin K-dependent γ-glutamate carboxylation and serine phosphorylation. While γ-carboxylation is crucial for MGP calcification inhibitory properties, phosphorylation has been suggested to be involved in regulating its secretion into the extracellular environment. Nonphosphorylated- (desphospho-) uncarboxylated MGP (dp-ucMGP) is the inactive isoform of MGP and a marker of poor vitamin K status, resulting in impaired inhibition of vascular calcification. A recent meta-analysis [
      • Chen H.-G.
      • Sheng L.-T.
      • Zhang Y.-B.
      • Cao A.-L.
      • Lai Y.-W.
      • Kunutsor S.K.
      • Jiang L.
      • Pan A.
      Association of vitamin K with cardiovascular events and all-cause mortality: a systematic review and meta-analysis.
      ] summarized data from cohort studies including primarily high-risk patients and reported that higher dp-ucMGP levels were associated with CVD and all-cause mortality. However, studies from the general population [
      • Dalmeijer G.W.
      • van der Schouw Y.T.
      • Magdeleyns E.J.
      • Vermeer C.
      • Verschuren W.M.M.
      • Boer J.M.A.
      • Beulens J.W.J.
      Circulating desphospho-uncarboxylated matrix γ-carboxyglutamate protein and the risk of coronary heart disease and stroke.
      ,
      • Liu Y.-P.
      • Gu Y.-M.
      • Thijs L.
      • Knapen M.H.J.
      • Salvi E.
      • Citterio L.
      • Petit T.
      • Carpini S.D.
      • Zhang Z.
      • Jacobs L.
      • Jin Y.
      • Barlassina C.
      • Manunta P.
      • Kuznetsova T.
      • Verhamme P.
      • Struijker-Boudier H.A.
      • Cusi D.
      • Vermeer C.
      • Staessen J.A.
      Inactive matrix gla protein is causally related to adverse health outcomes.
      ,
      • van den Heuvel E.G.H.M.
      • van Schoor N.M.
      • Lips P.
      • Magdeleyns E.J.P.
      • Deeg D.J.H.
      • Vermeer C.
      • den Heijer M.
      Circulating uncarboxylated matrix Gla protein, a marker of vitamin K status, as a risk factor of cardiovascular disease.
      ,
      • Riphagen I.J.
      • Keyzer C.A.
      • Drummen N.E.A.
      • De Borst M.H.
      • Beulens J.W.J.
      • Gansevoort R.T.
      • Geleijnse J.M.
      • Muskiet F.A.J.
      • Navis G.
      • Visser S.T.
      • Vermeer C.
      • Kema I.P.
      • Bakker S.J.L.
      Prevalence and effects of functional vitamin K insufficiency: the PREVEND study.
      ,
      • Shea M.K.
      • Booth S.L.
      • Weiner D.E.
      • Brinkley T.E.
      • Kanaya A.M.
      • Murphy R.A.
      • Simonsick E.M.
      • Wassel C.L.
      • Vermeer C.
      • Kritchevsky S.B.
      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).
      ] published so far yielded inconsistent results. Since these studies mainly included middle-aged individuals and since both vitamin K deficiency and vascular calcification are more pronounced with advancing age, more data from the elderly population are needed.
      Our objective was to assess the association of dp-ucMGP levels with incident CVD and all-cause and cause-specific mortality using data from the prospective population-based Bruneck Study.

      2. Patients and methods

      2.1 Bruneck Study population

      The Bruneck Study is a prospective population-based survey. The objective and methods of this cohort have been published previously [
      • Kiechl S.
      • Schett G.
      • Schwaiger J.
      • Seppi K.
      • Eder P.
      • Egger G.
      • Santer P.
      • Mayr A.
      • Xu Q.
      • Willeit J.
      Soluble receptor activator of nuclear factor-kappa B ligand and risk for cardiovascular disease.
      ,
      • Stegemann C.
      • Pechlaner R.
      • Willeit P.
      • Langley S.R.
      • Mangino M.
      • Mayr U.
      • Menni C.
      • Moayyeri A.
      • Santer P.
      • Rungger G.
      • Spector T.D.
      • Willeit J.
      • Kiechl S.
      • Mayr M.
      Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck study.
      ,
      • Willeit P.
      • Kiechl S.
      • Kronenberg F.
      • Witztum J.L.
      • Santer P.
      • Mayr M.
      • Xu Q.
      • Mayr A.
      • Willeit J.
      • Tsimikas S.
      Discrimination and net reclassification of cardiovascular risk with lipoprotein(a): prospective 15-year outcomes in the Bruneck Study.
      ,
      • Kiechl S.
      • Wittmann J.
      • Giaccari A.
      • Knoflach M.
      • Willeit P.
      • Bozec A.
      • Moschen A.R.
      • Muscogiuri G.
      • Sorice G.P.
      • Kireva T.
      • Summerer M.
      • Wirtz S.
      • Luther J.
      • Mielenz D.
      • Billmeier U.
      • Egger G.
      • Mayr A.
      • Oberhollenzer F.
      • Kronenberg F.
      • Orthofer M.
      • Penninger J.M.
      • Meigs J.B.
      • Bonora E.
      • Tilg H.
      • Willeit J.
      • Schett G.
      Blockade of receptor activator of nuclear factor-κB (RANKL) signaling improves hepatic insulin resistance and prevents development of diabetes mellitus.
      ]. Briefly, in 1990, the study population was recruited as a sex- and age-stratified random sample of all residents aged 40–79 years living in the city of Bruneck (125 women and 125 men in each decade of age; n = 1000; all of Caucasian origin). Participants were re-examined every five years from 1990 to 2010, and most recently in 2016. The present analysis focused on the observation period from 2000 to 2016, with the study visit in 2000 serving as baseline. The study was conducted according to the guidelines of the Declaration of Helsinki. The study protocol was approved by the Ethics Committees of Verona and Bolzano, and all participants gave their written informed consent before entering the study [
      • Kiechl S.
      • Schett G.
      • Schwaiger J.
      • Seppi K.
      • Eder P.
      • Egger G.
      • Santer P.
      • Mayr A.
      • Xu Q.
      • Willeit J.
      Soluble receptor activator of nuclear factor-kappa B ligand and risk for cardiovascular disease.
      ,
      • Stegemann C.
      • Pechlaner R.
      • Willeit P.
      • Langley S.R.
      • Mangino M.
      • Mayr U.
      • Menni C.
      • Moayyeri A.
      • Santer P.
      • Rungger G.
      • Spector T.D.
      • Willeit J.
      • Kiechl S.
      • Mayr M.
      Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck study.
      ,
      • Willeit P.
      • Kiechl S.
      • Kronenberg F.
      • Witztum J.L.
      • Santer P.
      • Mayr M.
      • Xu Q.
      • Mayr A.
      • Willeit J.
      • Tsimikas S.
      Discrimination and net reclassification of cardiovascular risk with lipoprotein(a): prospective 15-year outcomes in the Bruneck Study.
      ,
      • Kiechl S.
      • Wittmann J.
      • Giaccari A.
      • Knoflach M.
      • Willeit P.
      • Bozec A.
      • Moschen A.R.
      • Muscogiuri G.
      • Sorice G.P.
      • Kireva T.
      • Summerer M.
      • Wirtz S.
      • Luther J.
      • Mielenz D.
      • Billmeier U.
      • Egger G.
      • Mayr A.
      • Oberhollenzer F.
      • Kronenberg F.
      • Orthofer M.
      • Penninger J.M.
      • Meigs J.B.
      • Bonora E.
      • Tilg H.
      • Willeit J.
      • Schett G.
      Blockade of receptor activator of nuclear factor-κB (RANKL) signaling improves hepatic insulin resistance and prevents development of diabetes mellitus.
      ].

      2.2 Outcome definition and assessment

      The CVD endpoint comprised all incident cases of transient ischemic attack, fatal and non-fatal ischemic stroke, myocardial infarction and revascularization procedures (angioplasty and surgery), as well as sudden cardiac deaths and death from rupture of aortic aneurysms. Ischemic stroke and transient ischemic attack were classified according to the criteria of the National Survey of Stroke [
      • Walker A.E.
      • Robins M.
      • Weinfeld F.D.
      The national Survey of stroke. Clinical findings.
      ]. Presence of myocardial infarction was determined by World Health Organization criteria (definite disease status) [
      • World Health Organization
      Ischaemic Heart Disease Registers: Report of the Fifth Working Group (Including a Second Revision of the Operating Protocol).
      ]. Vascular death consisted of CVD mortality, death from haemorrhagic stroke, heart failure, pulmonary embolism and other CVD deaths. Other vascular deaths were vascular deaths excluding myocardial infarction and sudden cardiac death. Non-vascular/non-cancer deaths included death from trauma, violence, infections, mental, neurological, liver, digestive, renal, respiratory and endocrine diseases, as well as deaths from unknown causes.
      CVD events and causes of deaths were determined by careful review of general practitioners' medical records, databases from the Bruneck hospital, death certificates and autopsy reports, and the thorough clinical and laboratory examinations performed as part of the study protocol. A major strength of the Bruneck Study is that virtually all participants of this study were referred to the local hospital and that the network existing between the hospital and general practitioners allowed for the retrieval of full medical information of all individuals. Follow-up was 100% complete for the CVD and all-cause mortality endpoints.

      2.3 Clinical examinations and biochemical measurements

      At each study visit, participants were evaluated with regards to their medical history, life-style behaviors, socio-economic status and vascular risk factors using validated standard procedures [
      • Kiechl S.
      • Schett G.
      • Schwaiger J.
      • Seppi K.
      • Eder P.
      • Egger G.
      • Santer P.
      • Mayr A.
      • Xu Q.
      • Willeit J.
      Soluble receptor activator of nuclear factor-kappa B ligand and risk for cardiovascular disease.
      ,
      • Stegemann C.
      • Pechlaner R.
      • Willeit P.
      • Langley S.R.
      • Mangino M.
      • Mayr U.
      • Menni C.
      • Moayyeri A.
      • Santer P.
      • Rungger G.
      • Spector T.D.
      • Willeit J.
      • Kiechl S.
      • Mayr M.
      Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck study.
      ,
      • Willeit P.
      • Kiechl S.
      • Kronenberg F.
      • Witztum J.L.
      • Santer P.
      • Mayr M.
      • Xu Q.
      • Mayr A.
      • Willeit J.
      • Tsimikas S.
      Discrimination and net reclassification of cardiovascular risk with lipoprotein(a): prospective 15-year outcomes in the Bruneck Study.
      ,
      • Kiechl S.
      • Wittmann J.
      • Giaccari A.
      • Knoflach M.
      • Willeit P.
      • Bozec A.
      • Moschen A.R.
      • Muscogiuri G.
      • Sorice G.P.
      • Kireva T.
      • Summerer M.
      • Wirtz S.
      • Luther J.
      • Mielenz D.
      • Billmeier U.
      • Egger G.
      • Mayr A.
      • Oberhollenzer F.
      • Kronenberg F.
      • Orthofer M.
      • Penninger J.M.
      • Meigs J.B.
      • Bonora E.
      • Tilg H.
      • Willeit J.
      • Schett G.
      Blockade of receptor activator of nuclear factor-κB (RANKL) signaling improves hepatic insulin resistance and prevents development of diabetes mellitus.
      ]. Body mass index was calculated as weight divided by height squared (kg/m2). Waist-to-hip ratio was determined by dividing the circumferences of waist and hip (cm/cm) as measured using a plastic tape meter at the level of the umbilicus and of the greater trochanters. Lifetime smoking was assessed as pack-years. Diabetes was defined based on ADA criteria [
      Report of the expert committee on the diagnosis and classification of diabetes mellitus.
      ]. Hypertension was defined as blood pressure ≥140/90 mmHg or the use of anti-hypertensive drugs. Social status (low versus high) was determined by collecting information on the educational level of the participant and the occupational status of the person with the highest income in the household. A low social status was assumed if the participant had <12 years of education and/or the average monthly income of the person with the highest income in the household was <$2000. Physical activity was quantified using the validated Baeke Score [
      • Baecke J.A.
      • Burema J.
      • Frijters J.E.
      A short questionnaire for the measurement of habitual physical activity in epidemiological studies.
      ]. An overall physical activity score was calculated by averaging scores for work (3 categories: 1 = sedentary, 2 = moderate, 3 = heavy) and sports/leisure activities (3 categories: 0, ≤2, or >2 h per week). Symptoms suggestive of heart failure were classified using the New York Heart Association (NYHA) classification [
      • Remme W.J.
      • Swedberg K.
      Task force for the diagnosis and treatment of chronic heart failure, European society of cardiology, guidelines for the diagnosis and treatment of chronic heart failure.
      ]. Glomerular filtration rate was estimated using the CKD-EPI equation [
      • Levey A.S.
      • Stevens L.A.
      • Schmid C.H.
      • Zhang Y.L.
      • Castro A.F.
      • Feldman H.I.
      • Kusek J.W.
      • Eggers P.
      • Van Lente F.
      • Greene T.
      • Coresh J.
      CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration), A new equation to estimate glomerular filtration rate.
      ]. Alternate Healthy Eating Index (AHEI), a measure of diet quality, significantly associated with the risk of major chronic diseases in numerous studies, was calculated as described previously [
      • McCullough M.L.
      • Feskanich D.
      • Stampfer M.J.
      • Giovannucci E.L.
      • Rimm E.B.
      • Hu F.B.
      • Spiegelman D.
      • Hunter D.J.
      • Colditz G.A.
      • Willett W.C.
      Diet quality and major chronic disease risk in men and women: moving toward improved dietary guidance.
      ]. We did not consider the ‘duration of multivitamin use component’ because multivitamin supplementation was almost absent in our cohort. Venous blood samples were drawn after an overnight fast and 12 h of abstinence from smoking. Laboratory parameters were measured by standard methods as described previously [
      • Kiechl S.
      • Schett G.
      • Schwaiger J.
      • Seppi K.
      • Eder P.
      • Egger G.
      • Santer P.
      • Mayr A.
      • Xu Q.
      • Willeit J.
      Soluble receptor activator of nuclear factor-kappa B ligand and risk for cardiovascular disease.
      ,
      • Stegemann C.
      • Pechlaner R.
      • Willeit P.
      • Langley S.R.
      • Mangino M.
      • Mayr U.
      • Menni C.
      • Moayyeri A.
      • Santer P.
      • Rungger G.
      • Spector T.D.
      • Willeit J.
      • Kiechl S.
      • Mayr M.
      Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck study.
      ,
      • Willeit P.
      • Kiechl S.
      • Kronenberg F.
      • Witztum J.L.
      • Santer P.
      • Mayr M.
      • Xu Q.
      • Mayr A.
      • Willeit J.
      • Tsimikas S.
      Discrimination and net reclassification of cardiovascular risk with lipoprotein(a): prospective 15-year outcomes in the Bruneck Study.
      ,
      • Kiechl S.
      • Wittmann J.
      • Giaccari A.
      • Knoflach M.
      • Willeit P.
      • Bozec A.
      • Moschen A.R.
      • Muscogiuri G.
      • Sorice G.P.
      • Kireva T.
      • Summerer M.
      • Wirtz S.
      • Luther J.
      • Mielenz D.
      • Billmeier U.
      • Egger G.
      • Mayr A.
      • Oberhollenzer F.
      • Kronenberg F.
      • Orthofer M.
      • Penninger J.M.
      • Meigs J.B.
      • Bonora E.
      • Tilg H.
      • Willeit J.
      • Schett G.
      Blockade of receptor activator of nuclear factor-κB (RANKL) signaling improves hepatic insulin resistance and prevents development of diabetes mellitus.
      ]. Plasma levels of dp-ucMGP were measured in duplicates using the inaKtif MGP iSYS kit, which is a fully automated dual-antibody test based on the sandwich ELISA developed by VitaK, Maastricht University, The Netherlands [
      • Cranenburg E.C.M.
      • Koos R.
      • Schurgers L.J.
      • Magdeleyns E.J.
      • Schoonbrood T.H.M.
      • Landewé R.B.
      • Brandenburg V.M.
      • Bekers O.
      • Vermeer C.
      Characterisation and potential diagnostic value of circulating matrix Gla protein (MGP) species.
      ]. Intra-assay variation coefficients were 3.1% for the upper limit of the normal range and 5.4% for the lower limit of the normal range. Inter-assay variation coefficients were 6.9% for the upper limit of the normal range and 13.6% for the lower limit of the normal range [
      • Pivin E.
      • Ponte B.
      • Pruijm M.
      • Ackermann D.
      • Guessous I.
      • Ehret G.
      • Liu Y.-P.
      • Drummen N.E.A.
      • Knapen M.H.J.
      • Pechere-Bertschi A.
      • Paccaud F.
      • Mohaupt M.
      • Vermeer C.
      • Staessen J.A.
      • Vogt B.
      • Martin P.-Y.
      • Burnier M.
      • Bochud M.
      Inactive matrix gla-protein is associated with arterial stiffness in an adult population-based study.
      ].

      2.4 Statistical analysis

      Continuous variables were reported as mean ± SD or medians (interquartile range) and dichotomous variables as counts (percentages). Wald test using linear regression or likelihood ratio test (LRT) using logistic regression were used to assess the associations of baseline variables with quartile groups of dp-ucMGP. Continuous variables that were not approximately normally distributed were loge transformed before entered into models.
      Within-person variability of dp-ucMGP levels over 5 and 10 years were quantified using Spearman's and Pearson correlation coefficients in participants without intake of vitamin K antagonists.
      For the CVD endpoint, person-years of follow-up for each participant were accrued from the 2000 baseline until diagnosis of CVD, death or April 1, 2016, whichever came first. Participants who experienced an event were censored with respect to subsequent follow-up i.e. only the first CVD event was considered in the analysis. For the all-cause mortality endpoint, participants accrued follow-up time from baseline until death or April 1, 2016. Cox proportional hazard models were used to estimate HRs and 95%CIs per 1-SD higher loge transformed dp-ucMGP as well as across quartiles of dp-ucMGP levels. 1-SD of loge transformed dp-ucMGP equaled 0.518 pmol/L (corresponding to a 1.67-fold change in dp-ucMGP). The multivariable adjustment included age, sex, smoking (as loge transformed pack years), hypertension, body mass index, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, loge transformed triglycerides level, physical activity, diabetes, social status and estimated glomerular filtration rate. In sensitivity analyses, we excluded participants who (a) had experienced a CVD event prior to study baseline, (b) were taking vitamin K antagonists at baseline, (c) additionally adjusted for the AHEI, (d) additionally adjusted for various medications showing a significant association with dp-ucMGP and, thus, representing potential confounding factors, and (e) used repeat measurements of dp-ucMGP over time. For repeat measurements of dp-ucMGP, we used a long-term average of dp-ucMGP (“usual-levels”) estimated with regression calibration including data from 2000, 2005 and 2010 [
      • Studies Collaboration Fibrinogen
      Correcting for multivariate measurement error by regression calibration in meta-analyses of epidemiological studies.
      ]. In subgroup analyses, we investigated potential modifications by age, sex (male versus female), heart failure (no versus NYHA class ≥ I), smoking (ex/never versus current), history of hypertension (no versus yes), history of diabetes (no versus yes), estimated glomerular filtration rate, and body mass index by including the variables into the models together with an interaction term with loge transformed dp-ucMGP and adjusted for age and sex, if appropriate. For purpose of presentation, we categorized continuous variables into relevant groups. P values for interaction are reported for continuous variables. The proportional hazard assumptions were tested for all variables using Schoenfeld residuals and by graphical inspection, and were satisfied.
      Statistical analyses were performed with Stata 15.1 MP software package. All statistical tests were 2-sided, and p ≤ 0.05 was considered as statistically significant.

      3. Results

      3.1 Baseline plasma dp-ucMGP levels and their within-person variability over time

      Median (interquartile range) dp-ucMGP was 478.4 (335.0–635.2) pmol/L in the study population at baseline. Repeated measurements of dp-ucMGP were performed in 2005 and 2010. Distribution of dp-ucMGP levels in individuals without current intake of vitamin K antagonists and 5- and 10-year within-person correlations are depicted in Fig. 1. In individuals not taking vitamin K antagonists, median dp-ucMGP was 470.9 (333.1–625.1) at baseline, 553.1 (416.7–716.6) in 2005, and 600.4 (421.7–787.0) pmol/L in 2010. Correlations between levels of dp-ucMGP assessed in individuals in 2000 and 2005 as well as 2000 and 2010 were high (Spearman's correlation coefficient, r = 0.74, 95%CI: 0.69–0.77; and r = 0.61, 95%CI: 0.54–0.66, respectively).
      Fig. 1
      Fig. 1Distribution of dp-ucMGP at baseline in 2005 and in 2010 (A) and correlations between dp-ucMGP at baseline and during follow-up (B and C) in individuals without intake of vitamin K antagonists.
      (A) Violin plots show box plots together with kernel densities. (B) Scatter plots of dp-ucMGP levels at baseline and in 2005 (n = 525) together with a regression line (solid line) and 95% confidence interval (shaded area). (C) Scatter plots of dp-ucMGP levels at baseline and in 2010 (n = 445) together with a regression line (solid line) and 95% confidence interval (shaded area). For the purpose of presentation, dp-ucMGP levels higher than 2500 pmol/L have been omitted in (A) and are indicated as “x” (B) and (C). Three individuals included in the graphs had missing values on vitamin K antagonist intake in 2005 (n = 2) or 2010 (n = 1). rp indicates Pearson and rs Spearman's correlation coefficient. CI, confidence interval.

      3.2 Baseline characteristics and their associations with dp-ucMGP

      Baseline characteristics of the whole study population and according to quartiles of plasma dp-ucMGP levels are presented in Table 1. The study population comprised a total of 684 individuals with a mean age of 66.1 ± 10.2 (SD) years, 354 (51.8%) women and 49 (7.2%) individuals with a prior CVD event. Across increasing quartiles of dp-ucMGP levels, more individuals had hypertension and were on vitamin K antagonists. Age, body mass index, systolic blood pressure, NT-proBNP, high-sensitivity CRP, and intake of calcium-channel blockers, diuretics and digitalis glycosides increased, while renal function, AHEI scores and intake of beta-blockers and statins decreased across dp-ucMGP quartiles.
      Table 1Baseline characteristics of the whole Bruneck Study cohort and according to quartiles of plasma dp-ucMGP.
      CharacteristicsAll participants (n = 684)Quartiles of dp-ucMGP
      Q1<335.4 pmol/LQ2335.4–478.3 pmol/LQ3>478.3–634.0 pmol/LQ4>634.0 pmol/Lp value
      Age- and sex-adjusted p values for trend were derived from Wald test using linear regression or likelihood ratio test using logistic regression.
      Demographic variables
       Age, years66.1 ± 10.260.5 ± 8.363.6 ± 8.066.8 ± 10.073.5 ± 9.5<0.001
      Sex-adjusted p value for age.
       Female sex, n (%)354 (51.8)81 (47.4)89 (52.1)93 (54.4)91 (53.2)0.807
      Age-adjusted p value for sex.
       Low social status, n (%)404 (59.1)85 (49.7)92 (53.8)107 (62.6)120 (70.2)0.388
      Anthropometric parameters
       Body mass index, kg/m225.4 ± 4.024.5 ± 3.825.4 ± 3.826.0 ± 4.225.7 ± 4.1<0.001
       Waist-to-hip ratio, cm0.92 ± 0.070.91 ± 0.070.92 ± 0.080.92 ± 0.080.93 ± 0.070.399
      Life-style variables
       Smoking, pack years0 (0, 24.9)10 (0, 25.5)0 (0, 23.3)0 (0, 26.8)0 (0, 26.0)0.061
       Physical activity, score2.35 ± 0.782.61 ± 0.802.41 ± 0.792.31 ± 0.772.08 ± 0.680.236
       Hypertension, n (%)384 (56.1)74 (43.3)83 (48.5)105 (61.4)122 (71.4)0.011
       Systolic blood pressure, mmHg139.8 ± 18.9133.8 ± 16.8138.6 ± 19.0141.2 ± 19.0145.8 ± 18.70.019
       Diastolic blood pressure, mmHg84.0 ± 8.383.1 ± 8.684.1 ± 7.684.4 ± 8.784.5 ± 8.30.221
       Diabetes mellitus, n (%)78 (11.4)17 (9.9)19 (11.1)24 (14.0)18 (10.5)0.063
       Alternative Healthy Eating Index
      Data on 25-hydroxyvitamin D was available for 679 participants; and on NT-proBNP and Alternative Healthy Eating Index for 682 participants.
      35.3 ± 7.337.1 ± 6.835.7 ± 6.934.7 ± 7.733.7 ± 7.30.010
      Laboratory parameters
       Total cholesterol, mmol/L6.05 ± 1.095.97 ± 0.946.09 ± 1.036.12 ± 1.176.00 ± 1.200.639
       HDL-cholesterol, mmol/L1.48 ± 0.391.46 ± 0.401.51 ± 0.371.52 ± 0.381.45 ± 0.410.243
       LDL-cholesterol, mmol/L3.86 ± 0.953.82 ± 0.813.89 ± 0.893.89 ± 1.023.82 ± 1.060.873
       Triglycerides, mmol/L1.36 (0.98, 1.83)1.30 (0.97, 1.80)1.30 (0.94, 1.83)1.36 (0.94, 1.81)1.44 (1.04, 1.89)0.151
       Fasting glucose, mmol/L5.33 (5.00, 5.83)5.22 (4.94, 5.66)5.38 (5.00, 5.83)5.27 (4.94, 5.83)5.49 (5.11, 5.88)0.228
       High-sensitivity CRP, nmol/L17.33 (8.67, 37.81)13.71 (7.43, 28.29)14.57 (7.24, 29.24)21.71 (10.76, 51.14)22.67 (10.95, 52.29)0.004
       Osteoprotegerin, pmol/L3.82 (3.29, 4.53)3.55 (3.15, 4.21)3.60 (3.20, 4.33)3.83 (3.23, 4.48)4.26 (3.71, 5.15)0.507
       eGFR (ml/min/1.73 m2)82.0 ± 14.786.6 ± 12.684.8 ± 13.081.7 ± 14.675.1 ± 15.8<0.001
       25-hydroxyvitamin D, ng/ml
      Data on 25-hydroxyvitamin D was available for 679 participants; and on NT-proBNP and Alternative Healthy Eating Index for 682 participants.
      34.1 (25.9, 43.8)37.7 (29.8, 47.4)36.2 (27.9, 46.7)33.1 (24.2, 43.2)28.5 (22.3, 37.8)0.077
       NT-proBNP, pg/ml
      Data on 25-hydroxyvitamin D was available for 679 participants; and on NT-proBNP and Alternative Healthy Eating Index for 682 participants.
      3.21 (2.49, 4.14)2.64 (2.09, 3.43)2.97 (2.34, 3.82)3.37 (2.63, 4.19)3.90 (3.00, 4.96)0.005
      Medications
       ACEi or ARBs, n (%)80 (11.7)15 (8.8)18 (10.5)18 (10.5)29 (17.0)0.857
       Beta-blockers, n (%)60 (8.8)22 (12.9)14 (8.2)12 (7.0)12 (7.0)0.001
       Calcium-channel blockers, n (%)82 (12.0)11 (6.4)13 (7.6)18 (10.5)40 (23.4)0.046
       Diuretics, n (%)103 (15.1)13 (7.6)15 (8.8)29 (17.0)46 (26.9)0.030
       Digitalis glycosides, n (%)46 (6.7)1 (0.6)2 (1.2)10 (5.9)33 (19.3)<0.001
       Statins, n (%)63 (9.2)22 (12.9)13 (7.6)13 (7.6)15 (8.8)0.017
       Vitamin K antagonists, n (%)20 (2.9)1 (0.6)0 (0)0 (0)19 (11.1)<0.001
       Acetylsalicylic-acid, n (%)82 (12.0)12 (7.0)15 (8.8)23 (13.5)32 (18.7)0.965
       Oral antidiabetic agents, n (%)29 (4.2)6 (3.5)7 (4.1)8 (4.7)8 (4.7)0.103
       Antibiotics
      Intake of antibiotics within the last 3 months.
      , n (%)
      44 (6.4)7 (4.1)15 (8.8)13 (7.6)9 (5.3)0.460
      Clinical conditions
       History of CVD, n (%)49 (7.2)10 (5.9)9 (5.3)12 (7.0)18 (10.5)0.248
       Heart failure (NYHA ≥1), n (%)228 (33.3)28 (16.4)39 (22.8)62 (36.3)99 (57.9)0.282
      Data presented are mean ± SD, median (interquartile range) for highly skewed variables, or numbers (%).
      ACEi, angiotensin converting enzyme inhibitors; ARBs, angiotensin receptor blockers; CVD, cardiovascular disease; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MI, myocardial infarction; na, not applicable; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA, New York Heart Association Classes; Q, quartile.
      a Age- and sex-adjusted p values for trend were derived from Wald test using linear regression or likelihood ratio test using logistic regression.
      b Data on 25-hydroxyvitamin D was available for 679 participants; and on NT-proBNP and Alternative Healthy Eating Index for 682 participants.
      c Sex-adjusted p value for age.
      d Age-adjusted p value for sex.
      e Intake of antibiotics within the last 3 months.

      3.3 Analysis of the association of dp-ucMGP with incident cardiovascular disease and all-cause and cause-specific mortality

      Over a median follow-up of 15.5 years, 163 CVD events and 235 deaths were recorded – corresponding to incidence rates of 20.2 and 26.7 per 1000 person-years, respectively. CVD and all-cause mortality incidence rates increased with higher dp-ucMGP quartiles: from 11.6 and 11.3 per 1000 person-years in the first quartile, to 15.2 and 16.7 in the second, 26.0 and 28.1 in the third, and 32.8 and 60.4 per 1000 person-years in the fourth quartile, respectively.
      In Cox proportional hazard models, a 1-SD higher loge transformed dp-ucMGP level was significantly associated with CVD risk in the unadjusted (HR 1.61, 95%CI: 1.36–1.90; p<0.001) and age-/sex-adjusted model (HR 1.30, 95%CI: 1.09–1.55, p=0.004) (Table 2). In the multivariable adjusted model, which included a range of potential confounders, the association remained statistically significant with a HR of 1.23 (95%CI: 1.02–1.47; p=0.029). In sensitivity analyses, a significant association was obtained when the model was restricted to individuals without prior CVD, additionally adjusted for the AHEI, additionally adjusted for potential confounding by various medications, and when long-term averages of repeat measurements of dp-ucMGP over time were used, whereas it lost its significance when we excluded subjects taking vitamin K antagonists (Table 3). In subgroup analyses, we observed no statistically significant effect modification by age, sex, heart failure, smoking, hypertension, diabetes, impairment of renal function, and body-mass index (Fig. 2). In individual disease endpoints, dp-ucMGP showed no significant association with myocardial infarction and sudden cardiac death (age-/sex-adjusted HR 1.10, 95%CI: 0.81–1.49; p=0.546; multivariable adjusted HR 0.98, 95%CI: 0.71–1.35; p=0.886), but showed a consistent significant association with other cardiovascular diseases.
      Table 2Association of dp-ucMGP with incident cardiovascular disease and all-cause mortality in the Bruneck Study.
      Outcome, type of analysisContinuous loge transformed (dp-ucMGP)p valueQ2 (335.4–478.3 pmol/L)
      Q1 (dp-ucMGP level <335.4 pmol/L) was used as reference and included 171 subjects. Number of events in Q1 were 27 for all CVD, 21 for CVD excluding MI and SCD, and 28 for all-cause mortality.
      Q3 (>478.3–634.0 pmol/L)
      Q1 (dp-ucMGP level <335.4 pmol/L) was used as reference and included 171 subjects. Number of events in Q1 were 27 for all CVD, 21 for CVD excluding MI and SCD, and 28 for all-cause mortality.
      Q4 (>634.0 pmol/L)
      Q1 (dp-ucMGP level <335.4 pmol/L) was used as reference and included 171 subjects. Number of events in Q1 were 27 for all CVD, 21 for CVD excluding MI and SCD, and 28 for all-cause mortality.
      p value
      p value for trend.
      CVD
      All CVD
       Unadjusted1.61 (1.36–1.90)<0.0011.31 (0.79–2.17)2.24 (1.41–3.58)2.85 (1.78–4.55)<0.001
       Age-/sex-adjusted1.30 (1.09–1.55)0.0041.14 (0.69–1.89)1.78 (1.11–2.86)1.61 (0.97–2.66)0.021
       Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (no versus yes), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dL), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.23 (1.02–1.47)0.0291.14 (0.68–1.90)1.68 (1.03–2.74)1.47 (0.87–2.49)0.077
       No. of events/subjects163/68434/17151/17151/171
      Excluding MI and SCD
       Unadjusted1.58 (1.31–1.91)<0.0011.24 (0.69–2.22)2.09 (1.22–3.57)2.92 (1.72–4.97)<0.001
       Age-/sex-adjusted1.32 (1.08–1.61)0.0081.10 (0.61–1.97)1.71 (0.99–2.94)1.82 (1.03–3.22)0.014
       Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (no versus yes), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dL), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.24 (1.00–1.52)0.0461.15 (0.64–2.08)1.70 (0.97–2.99)1.75 (0.96–3.18)0.035
       No. of events/subjects123/68425/17137/17140/171
      All-cause mortality
      Unadjusted2.05 (1.79–2.34)<0.0011.49 (0.92–2.42)2.53 (1.62–3.96)5.60 (3.69–8.50)<0.001
      Age-/sex-adjusted1.36 (1.17–1.57)<0.0011.22 (0.75–1.98)1.60 (1.02–2.53)1.97 (1.27–3.05)0.001
      Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (no versus yes), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dL), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.40 (1.20–1.64)<0.0011.36 (0.83–2.22)1.68 (1.05–2.68)2.19 (1.38–3.48)<0.001
      No. of events/subjects235/68440/17161/171106/171
      Hazard ratio (95% confidence interval) according to continuous (1-SD higher) loge transformed dp-ucMGP and dp-ucMGP quartiles (Q) for the outcome all CVD, CVD excluding MI and SCD, and all-cause mortality. Hazard ratios (95% confidence intervals) were derived from Cox regression models.
      CVD, cardiovascular disease; MI, myocardial infarction; No, numbers; SCD, sudden cardiac death.
      a Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (no versus yes), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dL), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      b p value for trend.
      c Q1 (dp-ucMGP level <335.4 pmol/L) was used as reference and included 171 subjects. Number of events in Q1 were 27 for all CVD, 21 for CVD excluding MI and SCD, and 28 for all-cause mortality.
      Table 3Sensitivity analyses of the association between plasma dp-ucMGP and cardiovascular disease and all-cause mortality in the Bruneck Study.
      Outcome, type of analysisNo. of events/subjectsHR per 1-SD higher loge transformed dp-ucMGP level (95% CI)
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (no vs yes), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dL), physical activity (score), diabetes (no vs yes), social status (low vs high), and estimated glomerular filtration rate (ml/min).
      p value
      CVD
      Restricted to subjects without prior CVD137/6351.32 (1.08–1.61)0.007
      Restricted to subjects without vitamin K antagonist intake155/6641.20 (0.98–1.48)0.078
      Additionally adjusted for Alternative Healthy Eating Index163/6821.22 (1.02–1.47)0.031
      Additionally adjusted for medications
      Beta-blockers, calcium-channel blockers, diuretics, digitalis glycosides, statins.
      163/6841.21 (1.00–1.45)0.044
      Using repeat measurements of dp-ucMGP over time163/6841.41 (1.04–1.92)0.029
      All-cause mortality
      Restricted to subjects without prior CVD200/6351.42 (1.20–1.68)<0.001
      Restricted to subjects without vitamin K antagonist intake220/6641.36 (1.14–1.62)0.001
      Additionally adjusted for Alternative Healthy Eating Index163/6821.40 (1.19–1.64)<0.001
      Additionally adjusted for medications
      Beta-blockers, calcium-channel blockers, diuretics, digitalis glycosides, statins.
      163/6841.36 (1.16–1.60)<0.001
      Using repeat measurements of dp-ucMGP over time235/6841.77 (1.36–2.31)<0.001
      Hazard ratios (HRs) and 95% confidence intervals (95%CI) were derived from Cox regression models and calculated for a 1-SD higher loge transformed dp-uc MGP level.
      CVD, cardiovascular disease; No, numbers.
      a Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (no vs yes), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dL), physical activity (score), diabetes (no vs yes), social status (low vs high), and estimated glomerular filtration rate (ml/min).
      b Beta-blockers, calcium-channel blockers, diuretics, digitalis glycosides, statins.
      Fig. 2
      Fig. 2Hazard ratios and 95%CIs per 1-SD higher concentration of dp-ucMGP for incident cardiovascular disease (A) and all-cause mortality (B) across relevant subgroups.
      ap value for interaction. For the variables age, eGFR, and BMI, p values are reported when using continuous variables. Models were additionally adjusted for age and sex, if appropriate. BMI, body mass index; CI, confidence interval; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; HR, hazard ratio; SD, standard deviation.
      Dp-ucMGP was significantly and even stronger associated with all-cause mortality (Table 2). The HR for all-cause mortality associated with a 1-SD higher loge transformed level of dp-ucMGP was 2.05 (95%CI: 1.79–2.34; p<0.001) in the unadjusted model and attenuated to 1.36 (95%CI: 1.17–1.57; p<0.001) in the age-/sex-adjusted, and to 1.40 (95%CI: 1.20–1.64; p<0.001) in the multivariable adjusted model. There was a significant continuous increase in the HRs for all-cause death across quartiles of dp-ucMGP in the unadjusted, age-/sex-adjusted and multivariable model (all p values for trends ≤0.001). In sensitivity analyses, the associations remained virtually unchanged after exclusion of individuals with pre-existing CVD, additionally adjusted for the AHEI, additionally adjusted for potential confounding by various medications, and when excluding vitamin K antagonist users. The hazard ratio for CVD appeared to be stronger when considering long-term averages of dp-ucMGP levels over time (Table 3). Findings were robust across several relevant subgroups (Fig. 2).
      When examining the association with cause-specific deaths, a 1-SD higher value of loge transformed dp-ucMGP was not statistically significantly associated with vascular (HR 1.24, 95%CI: 0.96–1.59; p=0.094) and cancer death (HR 1.32, 95%CI: 0.98–1.78; p=0.065), whereas a significant increased risk was observed for non-vascular/non-cancer death (HR 1.74, 95%CI: 1.32–2.31; p<0.001) in the multivariable adjusted model (Table 4). Notably, fatal myocardial infarction and sudden cardiac death were not associated with dp-ucMGP (age-/sex-adjusted HR 1.11, 95%CI: 0.75–1.65; p=0.605; multivariable adjusted HR 0.95, 95%CI: 0.62–1.46; p=0.820), however, other vascular deaths showed a significant association (age-/sex-adjusted HR 1.45, 95%CI: 1.09–1.92; p=0.010; multivariable adjusted HR 1.42, 95%CI: 1.04–1.92; p=0.025).
      Table 4The association between plasma dp-ucMGP and cause-specific deaths in the Bruneck Study.
      Outcome, No. of (events/subjects), type of analysisHR per 1-SD higher loge transformed dp-ucMGP level (95%CI)p value
      Vascular death (95/684)
      Unadjusted2.12 (1.72–2.62)<0.001
      Age-/sex-adjusted1.33 (1.05–1.67)0.016
      Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (0 versus 1), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dl), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.24 (0.96–1.59)0.094
      Vascular death excluding MI and SCD (64/684)
      Unadjusted2.38 (1.85–3.07)<0.001
      Age-/sex-adjusted1.45 (1.09–1.92)0.010
      Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (0 versus 1), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dl), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.42 (1.04–1.92)0.025
      Cancer death (64/684)
      Unadjusted1.37 (1.05–1.79)0.020
      Age-/sex-adjusted1.13 (0.86–1.48)0.394
      Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (0 versus 1), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dl), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.32 (0.98–1.78)0.065
      Non-vascular/non-cancer death (76/684)
      Unadjusted2.67 (2.13–3.34)<0.001
      Age-/sex-adjusted1.62 (1.25–2.09)<0.001
      Multivariable adjusted
      Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (0 versus 1), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dl), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).
      1.74 (1.32–2.31)<0.001
      Hazard ratios (HRs) and 95% confidence intervals (95%CI) were derived from Cox regression models and calculated for a 1-SD higher loge transformed dp-ucMGP level. Vascular death included CVD mortality, death from haemorrhagic stroke, heart failure, pulmonary embolism and other CVD deaths. Non-vascular/non-cancer death included death from trauma, violence, infections, mental, neurological, liver, digestive, renal, respiratory and endocrine diseases, as well as deaths from unknown causes.
      MI, myocardial infarction; SCD, sudden cardiac death.
      a Adjusted for age (years), sex (male versus female), smoking (loge transformed pack years), hypertension (0 versus 1), body mass index (kg/m2), low-density lipoprotein cholesterol (mmol/L), high-density lipoprotein cholesterol (mmol/L), loge transformed triglycerides level (mg/dl), physical activity (score), diabetes (no versus yes), social status (low versus high), and estimated glomerular filtration rate (ml/min).

      4. Discussion

      In this study on data from 684 individuals aged 50–89 years of the prospective population-based Bruneck Study, we assessed the association of plasma levels of dp-ucMGP with incident CVD and all-cause mortality over a median follow-up of 15.5 years. Dp-ucMGP was significantly associated with both incident CVD and all-cause mortality, with multivariable adjusted HRs of 1.23 (95%CI: 1.02–1.47) and 1.40 (95%CI: 1.20–1.64) per 1-SD higher level of loge transformed dp-ucMGP, respectively. The magnitude of association was stronger for all-cause mortality, a finding that also sustained in sensitivity analyses. We did not find an association between dp-ucMGP and myocardial infarction and sudden cardiac deaths, but a strong association with other vascular deaths and non-vascular/non-cancer deaths.
      To date, five studies [
      • Dalmeijer G.W.
      • van der Schouw Y.T.
      • Magdeleyns E.J.
      • Vermeer C.
      • Verschuren W.M.M.
      • Boer J.M.A.
      • Beulens J.W.J.
      Circulating desphospho-uncarboxylated matrix γ-carboxyglutamate protein and the risk of coronary heart disease and stroke.
      ,
      • Liu Y.-P.
      • Gu Y.-M.
      • Thijs L.
      • Knapen M.H.J.
      • Salvi E.
      • Citterio L.
      • Petit T.
      • Carpini S.D.
      • Zhang Z.
      • Jacobs L.
      • Jin Y.
      • Barlassina C.
      • Manunta P.
      • Kuznetsova T.
      • Verhamme P.
      • Struijker-Boudier H.A.
      • Cusi D.
      • Vermeer C.
      • Staessen J.A.
      Inactive matrix gla protein is causally related to adverse health outcomes.
      ,
      • van den Heuvel E.G.H.M.
      • van Schoor N.M.
      • Lips P.
      • Magdeleyns E.J.P.
      • Deeg D.J.H.
      • Vermeer C.
      • den Heijer M.
      Circulating uncarboxylated matrix Gla protein, a marker of vitamin K status, as a risk factor of cardiovascular disease.
      ,
      • Riphagen I.J.
      • Keyzer C.A.
      • Drummen N.E.A.
      • De Borst M.H.
      • Beulens J.W.J.
      • Gansevoort R.T.
      • Geleijnse J.M.
      • Muskiet F.A.J.
      • Navis G.
      • Visser S.T.
      • Vermeer C.
      • Kema I.P.
      • Bakker S.J.L.
      Prevalence and effects of functional vitamin K insufficiency: the PREVEND study.
      ,
      • Shea M.K.
      • Booth S.L.
      • Weiner D.E.
      • Brinkley T.E.
      • Kanaya A.M.
      • Murphy R.A.
      • Simonsick E.M.
      • Wassel C.L.
      • Vermeer C.
      • Kritchevsky S.B.
      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).
      ] have examined the association between plasma dp-ucMGP and incident CVD and all-cause mortality in the general population. The Longitudinal Aging Study Amsterdam [
      • van den Heuvel E.G.H.M.
      • van Schoor N.M.
      • Lips P.
      • Magdeleyns E.J.P.
      • Deeg D.J.H.
      • Vermeer C.
      • den Heijer M.
      Circulating uncarboxylated matrix Gla protein, a marker of vitamin K status, as a risk factor of cardiovascular disease.
      ] including 577 individuals with a mean baseline age of 60 years found that higher dp-ucMGP concentration is significantly associated with an increased risk of CVD over a mean follow-up of 5.6 years. In contrast, the Health ABC [
      • Shea M.K.
      • Booth S.L.
      • Weiner D.E.
      • Brinkley T.E.
      • Kanaya A.M.
      • Murphy R.A.
      • Simonsick E.M.
      • Wassel C.L.
      • Vermeer C.
      • Kritchevsky S.B.
      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).
      ] and FLEMENGHO [
      • Liu Y.-P.
      • Gu Y.-M.
      • Thijs L.
      • Knapen M.H.J.
      • Salvi E.
      • Citterio L.
      • Petit T.
      • Carpini S.D.
      • Zhang Z.
      • Jacobs L.
      • Jin Y.
      • Barlassina C.
      • Manunta P.
      • Kuznetsova T.
      • Verhamme P.
      • Struijker-Boudier H.A.
      • Cusi D.
      • Vermeer C.
      • Staessen J.A.
      Inactive matrix gla protein is causally related to adverse health outcomes.
      ] studies did not find a significant association with CVD. The race composition with 44% of blacks in the former and the younger age of participants in the latter study might in part explain the discrepancy in results with our study. No significant association was found between dp-ucMGP and coronary heart disease in a case-cohort study nested within EPIC-NL [
      • Dalmeijer G.W.
      • van der Schouw Y.T.
      • Magdeleyns E.J.
      • Vermeer C.
      • Verschuren W.M.M.
      • Boer J.M.A.
      • Beulens J.W.J.
      Circulating desphospho-uncarboxylated matrix γ-carboxyglutamate protein and the risk of coronary heart disease and stroke.
      ], and no and an even inverse association with cardiac events and coronary events in the FLEMENGHO [
      • Liu Y.-P.
      • Gu Y.-M.
      • Thijs L.
      • Knapen M.H.J.
      • Salvi E.
      • Citterio L.
      • Petit T.
      • Carpini S.D.
      • Zhang Z.
      • Jacobs L.
      • Jin Y.
      • Barlassina C.
      • Manunta P.
      • Kuznetsova T.
      • Verhamme P.
      • Struijker-Boudier H.A.
      • Cusi D.
      • Vermeer C.
      • Staessen J.A.
      Inactive matrix gla protein is causally related to adverse health outcomes.
      ] study. In our study, we found no significant association between dp-ucMGP and myocardial infarction or sudden cardiac death, which is the main cardiovascular disease in younger individuals, however, higher dp-ucMGP levels were associated with an increased risk of other cardiovascular diseases. The discrepant findings for myocardial infarction and other cardiovascular diseases are a potential explanation for the divergent findings in previous studies. A possible explanation could be that myocardial infarction primarily arises from non-calcified plaque rupture [
      • Virmani R.
      • Burke A.P.
      • Farb A.
      • Kolodgie F.D.
      Pathology of the unstable plaque.
      ] and active MGP inhibits plaque calcification [
      • Roumeliotis S.
      • Dounousi E.
      • Eleftheriadis T.
      • Liakopoulos V.
      Association of the inactive circulating matrix gla protein with vitamin K intake, calcification, mortality, and cardiovascular disease: a review.
      ]. Furthermore, elevated dp-ucMGP levels may be associated with less plaque hemorrhage, suggestive of more stable plaques [
      • Zwakenberg S.R.
      • van der Schouw Y.T.
      • Vermeer C.
      • Pasterkamp G.
      • den Ruijter H.M.
      • Beulens J.W.J.
      Matrix gla protein, plaque stability, and cardiovascular events in patients with severe atherosclerotic disease.
      ].
      We found a strong association between higher plasma dp-ucMGP levels and increased risk of all-cause mortality. This is in accordance with results from high-risk population studies [
      • Chen H.-G.
      • Sheng L.-T.
      • Zhang Y.-B.
      • Cao A.-L.
      • Lai Y.-W.
      • Kunutsor S.K.
      • Jiang L.
      • Pan A.
      Association of vitamin K with cardiovascular events and all-cause mortality: a systematic review and meta-analysis.
      ,
      • Mayer O.
      • Bruthans J.
      • Seidlerová J.
      • Karnosová P.
      • Mateřánková M.
      • Gelžinský J.
      • Rychecká M.
      • Opatrný J.
      • Wohlfahrt P.
      • Kučera R.
      • Trefil L.
      • Cífková R.
      • Filipovský J.
      • Vermeer C.
      The coincidence of low vitamin K status and high expression of growth differentiation factor 15 may indicate increased mortality risk in stable coronary heart disease patients.
      ] and the general population-based FLEMENGHO [
      • Liu Y.-P.
      • Gu Y.-M.
      • Thijs L.
      • Knapen M.H.J.
      • Salvi E.
      • Citterio L.
      • Petit T.
      • Carpini S.D.
      • Zhang Z.
      • Jacobs L.
      • Jin Y.
      • Barlassina C.
      • Manunta P.
      • Kuznetsova T.
      • Verhamme P.
      • Struijker-Boudier H.A.
      • Cusi D.
      • Vermeer C.
      • Staessen J.A.
      Inactive matrix gla protein is causally related to adverse health outcomes.
      ] and PREVEND [
      • Riphagen I.J.
      • Keyzer C.A.
      • Drummen N.E.A.
      • De Borst M.H.
      • Beulens J.W.J.
      • Gansevoort R.T.
      • Geleijnse J.M.
      • Muskiet F.A.J.
      • Navis G.
      • Visser S.T.
      • Vermeer C.
      • Kema I.P.
      • Bakker S.J.L.
      Prevalence and effects of functional vitamin K insufficiency: the PREVEND study.
      ] studies showing positive linear or J-shaped associations. When analyzing cause-specific deaths we did not find a significant association of plasma dp-ucMGP with vascular deaths. It is plausible that lack of this association is explained by the fact that 32.6% of vascular deaths were due to myocardial infarction and sudden cardiac death, since the association became significant after excluding these two components. In our study, higher dp-ucMGP levels were significantly associated with increased risk of non-vascular/non-cancer death, which comprised deaths from dementia, infection and renal failure among others. The FLEMENGHO study [
      • Liu Y.-P.
      • Gu Y.-M.
      • Thijs L.
      • Knapen M.H.J.
      • Salvi E.
      • Citterio L.
      • Petit T.
      • Carpini S.D.
      • Zhang Z.
      • Jacobs L.
      • Jin Y.
      • Barlassina C.
      • Manunta P.
      • Kuznetsova T.
      • Verhamme P.
      • Struijker-Boudier H.A.
      • Cusi D.
      • Vermeer C.
      • Staessen J.A.
      Inactive matrix gla protein is causally related to adverse health outcomes.
      ] did find a significant association with plasma dp-ucMGP and CVD mortality and non-cancer mortality risk. Likewise, the PREVEND study [
      • Riphagen I.J.
      • Keyzer C.A.
      • Drummen N.E.A.
      • De Borst M.H.
      • Beulens J.W.J.
      • Gansevoort R.T.
      • Geleijnse J.M.
      • Muskiet F.A.J.
      • Navis G.
      • Visser S.T.
      • Vermeer C.
      • Kema I.P.
      • Bakker S.J.L.
      Prevalence and effects of functional vitamin K insufficiency: the PREVEND study.
      ] showed a significantly increased CVD mortality risk with higher dp-ucMGP levels.
      The mechanisms linking dp-ucMGP with CVD and all-cause mortality remains to be fully determined. With regard to CVD, higher levels of dp-ucMGP reflect less efficient inhibition of vascular calcification, consequently leading to arterial stiffness, increased central blood pressure and finally organ dysfunction [
      • Roumeliotis S.
      • Dounousi E.
      • Eleftheriadis T.
      • Liakopoulos V.
      Association of the inactive circulating matrix gla protein with vitamin K intake, calcification, mortality, and cardiovascular disease: a review.
      ]. A vascular independent role of MGP has been suggested as well. Experimental studies have shown an involvement of MGP in apoptosis, tissue growth and development [
      • Cario-Toumaniantz C.
      • Boularan C.
      • Schurgers L.J.
      • Heymann M.-F.
      • Le Cunff M.
      • Léger J.
      • Loirand G.
      • Pacaud P.
      Identification of differentially expressed genes in human varicose veins: involvement of matrix gla protein in extracellular matrix remodeling.
      ,
      • Nishimoto S.K.
      • Nishimoto M.
      Matrix Gla protein C-terminal region binds to vitronectin. Co-localization suggests binding occurs during tissue development.
      ]. MGP expression is increased in response to cardiac pressure overload, potentially induced by angiotensin II [
      • Mustonen E.
      • Pohjolainen V.
      • Aro J.
      • Pikkarainen S.
      • Leskinen H.
      • Ruskoaho H.
      • Rysä J.
      ], suggesting an involvement in cardiac remodeling. MGP is differentially expressed in various human cancer cells [
      • Chen L.
      • O'Bryan J.P.
      • Smith H.S.
      • Liu E.
      Overexpression of matrix Gla protein mRNA in malignant human breast cells: isolation by differential cDNA hybridization.
      ,
      • Mertsch S.
      • Schurgers L.J.
      • Weber K.
      • Paulus W.
      • Senner V.
      Matrix gla protein (MGP): an overexpressed and migration-promoting mesenchymal component in glioblastoma.
      ,
      • Fan C.
      • Sheu D.
      • Fan H.
      • Hsu K.
      • Allen Chang C.
      • Chan E.
      Down-regulation of matrix Gla protein messenger RNA in human colorectal adenocarcinomas.
      ], whereby its role is still unknown. MGP has been shown to protect against pulmonary injury by inhibiting calcification and degradation of elastic fibers, and to predict poor outcome and mortality in patients with chronic obstructive pulmonary disease [
      • Piscaer I.
      • van den Ouweland J.M.W.
      • Vermeersch K.
      • Reynaert N.L.
      • Franssen F.M.E.
      • Keene S.
      • Wouters E.F.M.
      • Janssens W.
      • Vermeer C.
      • Janssen R.
      Low vitamin K status is associated with increased elastin degradation in chronic obstructive pulmonary disease.
      ] and, most recently, in patients with SARS-CoV-2 pneumonia [
      • Dofferhoff A.S.M.
      • Piscaer I.
      • Schurgers L.J.
      • Visser M.P.J.
      • van den Ouweland J.M.W.
      • de Jong P.A.
      • Gosens R.
      • Hackeng T.M.
      • van Daal H.
      • Lux P.
      • Maassen C.
      • Karssemeijer E.G.A.
      • Vermeer C.
      • Wouters E.F.M.
      • Kistemaker L.E.M.
      • Walk J.
      • Janssen R.
      Reduced vitamin K status as a potentially modifiable risk factor of severe coronavirus disease 2019.
      ]. Finally, dp-ucMGP, a marker of low vitamin K status, might be a feature of the function of other vitamin K dependent proteins, such as osteocalcin and Protein S, which are known to act against age-related syndromes including osteoporosis and bone fractures, neurodegenerative diseases, infections, cancers as well as metabolic, inflammatory and coagulative disorders [
      • Popa D.-S.
      • Bigman G.
      • Rusu M.E.
      The role of vitamin K in humans: implication in aging and age-associated diseases.
      ].
      Median dp-ucMGP concentration was 478 pmol/L in our study, which is higher than reported from other general population studies (for example 114 pmol/L in the EPIC-NL study, 335 pmol/L in the Longitudinal Aging Study Amsterdam study, and 372 pmol/L in the PREVEND study) [
      • Dalmeijer G.W.
      • van der Schouw Y.T.
      • Magdeleyns E.J.
      • Vermeer C.
      • Verschuren W.M.M.
      • Boer J.M.A.
      • Beulens J.W.J.
      Circulating desphospho-uncarboxylated matrix γ-carboxyglutamate protein and the risk of coronary heart disease and stroke.
      ,
      • van den Heuvel E.G.H.M.
      • van Schoor N.M.
      • Lips P.
      • Magdeleyns E.J.P.
      • Deeg D.J.H.
      • Vermeer C.
      • den Heijer M.
      Circulating uncarboxylated matrix Gla protein, a marker of vitamin K status, as a risk factor of cardiovascular disease.
      ,
      • Riphagen I.J.
      • Keyzer C.A.
      • Drummen N.E.A.
      • De Borst M.H.
      • Beulens J.W.J.
      • Gansevoort R.T.
      • Geleijnse J.M.
      • Muskiet F.A.J.
      • Navis G.
      • Visser S.T.
      • Vermeer C.
      • Kema I.P.
      • Bakker S.J.L.
      Prevalence and effects of functional vitamin K insufficiency: the PREVEND study.
      ] and lower as compared to high risk populations [
      • Chen H.-G.
      • Sheng L.-T.
      • Zhang Y.-B.
      • Cao A.-L.
      • Lai Y.-W.
      • Kunutsor S.K.
      • Jiang L.
      • Pan A.
      Association of vitamin K with cardiovascular events and all-cause mortality: a systematic review and meta-analysis.
      ,
      • Schurgers L.J.
      • Barreto D.V.
      • Barreto F.C.
      • Liabeuf S.
      • Renard C.
      • Magdeleyns E.J.
      • Vermeer C.
      • Choukroun G.
      • Massy Z.A.
      The circulating inactive form of matrix gla protein is a surrogate marker for vascular calcification in chronic kidney disease: a preliminary report.
      ]. Older age, higher prevalence of chronic disease and higher intake of vitamin K antagonists are associated with higher dp-ucMGP levels, which may explain the differences of dp-ucMGP concentrations across studies.
      Strengths of the present study include its population-based prospective design, >90% participation rate, long-term follow-up, rigorous assessment of a large number of clinical and laboratory characteristics and complete follow-up of CVD and all-cause mortality endpoints. Moreover, the majority of the previous studies, measured dp-ucMGP at a single time point. One study examined the stability of the dp-ucMGP level over time and found a very high correlation between dp-ucMGP at baseline and after one year [
      • Vermeer C.
      • Vik H.
      Effect of Menaquinone-7 (vitamin K2) on vascular elasticity in healthy subjects: results from a one-year study.
      ]. Our study is unique in extending the finding of a high correlation to a 5- and 10-year period. There are limitations as well. Our cohort was entirely of white origin, thus, the generalizability to other ethnic populations and geographical regions remains uncertain. Even though we adjusted for a large number of clinical and biochemical variables, residual confounding cannot be ruled out. Since the amount of dp-ucMGP in circulation also depends on the total amount of MGP available, the observed association of dp-ucMGP with CVD and all-cause mortality may not be reflective solely of vitamin K status. Assays that measure total MGP and allow for this correction, however, have not yet been developed.
      In conclusion, in this prospective population-based study, we found a significant association of plasma dp-ucMGP with CVD risk and a significant and even stronger association with risk for all-cause mortality. Clinical trials are needed to investigate whether vitamin K substitution results in improved health outcomes.

      Financial support

      This study was supported by the ‘Pustertaler Verein zur Prävention von Herz-und Hirngefaesserkrankungen’, the ‘Gesundheitsbezirk Bruneck’ and the ‘Südtiroler Sanitätsbetrieb’, Province of Bolzano , Italy and by an excellence initiative (Competence Centers for Excellent Technologies - COMET) of the Austrian Research Promotion Agency FFG: “Research Center of Excellence in Vascular Ageing, VASCage” (K-Project Nr. 843536 and 868624 ) funded by the BMVIT , BMWFW , the Wirtschaftsagentur Wien and the Standortagentur Tirol .

      Author contributions

      Conceptualization, K.W., J.W., S.K.; methodology, K.W., J.W., S.K.; formal analysis, K.W. and L.T; data curation, K.W., P.S., R.P., C.V., J.W., S.K.; writing—original draft preparation, K.W.; writing—review and editing, P.S, L.T., R.P., C.V., J.W., S.K.; funding acquisition, J.W. and S.K. All authors have read and agreed to the published version of the manuscript.

      Declaration of interests

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgments

      We are very grateful to all participants in the Bruneck Study.

      References

        • Rennenberg R.J.M.W.
        • Kessels A.G.H.
        • Schurgers L.J.
        • van Engelshoven J.M.A.
        • de Leeuw P.W.
        • Kroon A.A.
        Vascular calcifications as a marker of increased cardiovascular risk: a meta-analysis.
        Vasc. Health Risk Manag. 2009; 5: 185-197https://doi.org/10.2147/vhrm.s4822
        • Sallam T.
        • Cheng H.
        • Demer L.L.
        • Tintut Y.
        Regulatory circuits controlling vascular cell calcification.
        Cell. Mol. Life Sci. 2013; 70: 3187-3197https://doi.org/10.1007/s00018-012-1231-y
        • Roumeliotis S.
        • Dounousi E.
        • Eleftheriadis T.
        • Liakopoulos V.
        Association of the inactive circulating matrix gla protein with vitamin K intake, calcification, mortality, and cardiovascular disease: a review.
        Int. J. Mol. Sci. 2019; 20: E628https://doi.org/10.3390/ijms20030628
        • Chen H.-G.
        • Sheng L.-T.
        • Zhang Y.-B.
        • Cao A.-L.
        • Lai Y.-W.
        • Kunutsor S.K.
        • Jiang L.
        • Pan A.
        Association of vitamin K with cardiovascular events and all-cause mortality: a systematic review and meta-analysis.
        Eur. J. Nutr. 2019; 58: 2191-2205https://doi.org/10.1007/s00394-019-01998-3
        • Dalmeijer G.W.
        • van der Schouw Y.T.
        • Magdeleyns E.J.
        • Vermeer C.
        • Verschuren W.M.M.
        • Boer J.M.A.
        • Beulens J.W.J.
        Circulating desphospho-uncarboxylated matrix γ-carboxyglutamate protein and the risk of coronary heart disease and stroke.
        J. Thromb. Haemostasis. 2014; 12: 1028-1034https://doi.org/10.1111/jth.12609
        • Liu Y.-P.
        • Gu Y.-M.
        • Thijs L.
        • Knapen M.H.J.
        • Salvi E.
        • Citterio L.
        • Petit T.
        • Carpini S.D.
        • Zhang Z.
        • Jacobs L.
        • Jin Y.
        • Barlassina C.
        • Manunta P.
        • Kuznetsova T.
        • Verhamme P.
        • Struijker-Boudier H.A.
        • Cusi D.
        • Vermeer C.
        • Staessen J.A.
        Inactive matrix gla protein is causally related to adverse health outcomes.
        Hypertension. 2015; 65: 463-470https://doi.org/10.1161/HYPERTENSIONAHA.114.04494
        • van den Heuvel E.G.H.M.
        • van Schoor N.M.
        • Lips P.
        • Magdeleyns E.J.P.
        • Deeg D.J.H.
        • Vermeer C.
        • den Heijer M.
        Circulating uncarboxylated matrix Gla protein, a marker of vitamin K status, as a risk factor of cardiovascular disease.
        Maturitas. 2014; 77: 137-141https://doi.org/10.1016/j.maturitas.2013.10.008
        • Riphagen I.J.
        • Keyzer C.A.
        • Drummen N.E.A.
        • De Borst M.H.
        • Beulens J.W.J.
        • Gansevoort R.T.
        • Geleijnse J.M.
        • Muskiet F.A.J.
        • Navis G.
        • Visser S.T.
        • Vermeer C.
        • Kema I.P.
        • Bakker S.J.L.
        Prevalence and effects of functional vitamin K insufficiency: the PREVEND study.
        Nutrients. 2017; 9: 1334https://doi.org/10.3390/nu9121334
        • Shea M.K.
        • Booth S.L.
        • Weiner D.E.
        • Brinkley T.E.
        • Kanaya A.M.
        • Murphy R.A.
        • Simonsick E.M.
        • Wassel C.L.
        • Vermeer C.
        • Kritchevsky S.B.
        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. Nutr. 2017; 147: 888-895https://doi.org/10.3945/jn.117.249375
        • Kiechl S.
        • Schett G.
        • Schwaiger J.
        • Seppi K.
        • Eder P.
        • Egger G.
        • Santer P.
        • Mayr A.
        • Xu Q.
        • Willeit J.
        Soluble receptor activator of nuclear factor-kappa B ligand and risk for cardiovascular disease.
        Circulation. 2007; 116: 385-391https://doi.org/10.1161/CIRCULATIONAHA.106.686774
        • Stegemann C.
        • Pechlaner R.
        • Willeit P.
        • Langley S.R.
        • Mangino M.
        • Mayr U.
        • Menni C.
        • Moayyeri A.
        • Santer P.
        • Rungger G.
        • Spector T.D.
        • Willeit J.
        • Kiechl S.
        • Mayr M.
        Lipidomics profiling and risk of cardiovascular disease in the prospective population-based Bruneck study.
        Circulation. 2014; 129: 1821-1831https://doi.org/10.1161/CIRCULATIONAHA.113.002500
        • Willeit P.
        • Kiechl S.
        • Kronenberg F.
        • Witztum J.L.
        • Santer P.
        • Mayr M.
        • Xu Q.
        • Mayr A.
        • Willeit J.
        • Tsimikas S.
        Discrimination and net reclassification of cardiovascular risk with lipoprotein(a): prospective 15-year outcomes in the Bruneck Study.
        J. Am. Coll. Cardiol. 2014; 64: 851-860https://doi.org/10.1016/j.jacc.2014.03.061
        • Kiechl S.
        • Wittmann J.
        • Giaccari A.
        • Knoflach M.
        • Willeit P.
        • Bozec A.
        • Moschen A.R.
        • Muscogiuri G.
        • Sorice G.P.
        • Kireva T.
        • Summerer M.
        • Wirtz S.
        • Luther J.
        • Mielenz D.
        • Billmeier U.
        • Egger G.
        • Mayr A.
        • Oberhollenzer F.
        • Kronenberg F.
        • Orthofer M.
        • Penninger J.M.
        • Meigs J.B.
        • Bonora E.
        • Tilg H.
        • Willeit J.
        • Schett G.
        Blockade of receptor activator of nuclear factor-κB (RANKL) signaling improves hepatic insulin resistance and prevents development of diabetes mellitus.
        Nat. Med. 2013; 19: 358-363https://doi.org/10.1038/nm.3084
        • Walker A.E.
        • Robins M.
        • Weinfeld F.D.
        The national Survey of stroke. Clinical findings.
        Stroke. 1981; 12: I13-I44
        • World Health Organization
        Ischaemic Heart Disease Registers: Report of the Fifth Working Group (Including a Second Revision of the Operating Protocol).
        Regional Office for Europe, World Health Organization, Copenhagen -AC, Denmark1971 (n.d.)
      1. Report of the expert committee on the diagnosis and classification of diabetes mellitus.
        Diabetes Care. 1997; 20: 1183-1197https://doi.org/10.2337/diacare.20.7.1183
        • Baecke J.A.
        • Burema J.
        • Frijters J.E.
        A short questionnaire for the measurement of habitual physical activity in epidemiological studies.
        Am. J. Clin. Nutr. 1982; 36: 936-942https://doi.org/10.1093/ajcn/36.5.936
        • Remme W.J.
        • Swedberg K.
        Task force for the diagnosis and treatment of chronic heart failure, European society of cardiology, guidelines for the diagnosis and treatment of chronic heart failure.
        Eur. Heart J. 2001; 22: 1527-1560https://doi.org/10.1053/euhj.2001.2783
        • Levey A.S.
        • Stevens L.A.
        • Schmid C.H.
        • Zhang Y.L.
        • Castro A.F.
        • Feldman H.I.
        • Kusek J.W.
        • Eggers P.
        • Van Lente F.
        • Greene T.
        • Coresh J.
        CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration), A new equation to estimate glomerular filtration rate.
        Ann. Intern. Med. 2009; 150: 604-612https://doi.org/10.7326/0003-4819-150-9-200905050-00006
        • McCullough M.L.
        • Feskanich D.
        • Stampfer M.J.
        • Giovannucci E.L.
        • Rimm E.B.
        • Hu F.B.
        • Spiegelman D.
        • Hunter D.J.
        • Colditz G.A.
        • Willett W.C.
        Diet quality and major chronic disease risk in men and women: moving toward improved dietary guidance.
        Am. J. Clin. Nutr. 2002; 76: 1261-1271https://doi.org/10.1093/ajcn/76.6.1261
        • Cranenburg E.C.M.
        • Koos R.
        • Schurgers L.J.
        • Magdeleyns E.J.
        • Schoonbrood T.H.M.
        • Landewé R.B.
        • Brandenburg V.M.
        • Bekers O.
        • Vermeer C.
        Characterisation and potential diagnostic value of circulating matrix Gla protein (MGP) species.
        Thromb. Haemostasis. 2010; 104: 811-822https://doi.org/10.1160/TH09-11-0786
        • Pivin E.
        • Ponte B.
        • Pruijm M.
        • Ackermann D.
        • Guessous I.
        • Ehret G.
        • Liu Y.-P.
        • Drummen N.E.A.
        • Knapen M.H.J.
        • Pechere-Bertschi A.
        • Paccaud F.
        • Mohaupt M.
        • Vermeer C.
        • Staessen J.A.
        • Vogt B.
        • Martin P.-Y.
        • Burnier M.
        • Bochud M.
        Inactive matrix gla-protein is associated with arterial stiffness in an adult population-based study.
        Hypertension. 2015; 66: 85-92https://doi.org/10.1161/HYPERTENSIONAHA.115.05177
        • Studies Collaboration Fibrinogen
        Correcting for multivariate measurement error by regression calibration in meta-analyses of epidemiological studies.
        Stat. Med. 2009; 28: 1067-1092https://doi.org/10.1002/sim.3530
        • Virmani R.
        • Burke A.P.
        • Farb A.
        • Kolodgie F.D.
        Pathology of the unstable plaque.
        Prog. Cardiovasc. Dis. 2002; 44: 349-356https://doi.org/10.1053/pcad.2002.122475
        • Zwakenberg S.R.
        • van der Schouw Y.T.
        • Vermeer C.
        • Pasterkamp G.
        • den Ruijter H.M.
        • Beulens J.W.J.
        Matrix gla protein, plaque stability, and cardiovascular events in patients with severe atherosclerotic disease.
        Cardiology. 2018; 141: 32-36https://doi.org/10.1159/000493006
        • Mayer O.
        • Bruthans J.
        • Seidlerová J.
        • Karnosová P.
        • Mateřánková M.
        • Gelžinský J.
        • Rychecká M.
        • Opatrný J.
        • Wohlfahrt P.
        • Kučera R.
        • Trefil L.
        • Cífková R.
        • Filipovský J.
        • Vermeer C.
        The coincidence of low vitamin K status and high expression of growth differentiation factor 15 may indicate increased mortality risk in stable coronary heart disease patients.
        Nutr. Metabol. Cardiovasc. Dis. 2021; 31: 540-551https://doi.org/10.1016/j.numecd.2020.09.015
        • Cario-Toumaniantz C.
        • Boularan C.
        • Schurgers L.J.
        • Heymann M.-F.
        • Le Cunff M.
        • Léger J.
        • Loirand G.
        • Pacaud P.
        Identification of differentially expressed genes in human varicose veins: involvement of matrix gla protein in extracellular matrix remodeling.
        J. Vasc. Res. 2007; 44: 444-459https://doi.org/10.1159/000106189
        • Nishimoto S.K.
        • Nishimoto M.
        Matrix Gla protein C-terminal region binds to vitronectin. Co-localization suggests binding occurs during tissue development.
        Matrix Biol. 2005; 24: 353-361https://doi.org/10.1016/j.matbio.2005.05.004
        • Mustonen E.
        • Pohjolainen V.
        • Aro J.
        • Pikkarainen S.
        • Leskinen H.
        • Ruskoaho H.
        • Rysä J.
        Upregulation of Cardiac Matrix Gla Protein Expression in Response to Hypertrophic Stimuli. vol. 18. Blood Press, 2009: 286-293https://doi.org/10.3109/08037050903244643
        • Chen L.
        • O'Bryan J.P.
        • Smith H.S.
        • Liu E.
        Overexpression of matrix Gla protein mRNA in malignant human breast cells: isolation by differential cDNA hybridization.
        Oncogene. 1990; 5: 1391-1395
        • Mertsch S.
        • Schurgers L.J.
        • Weber K.
        • Paulus W.
        • Senner V.
        Matrix gla protein (MGP): an overexpressed and migration-promoting mesenchymal component in glioblastoma.
        BMC Cancer. 2009; 9: 302https://doi.org/10.1186/1471-2407-9-302
        • Fan C.
        • Sheu D.
        • Fan H.
        • Hsu K.
        • Allen Chang C.
        • Chan E.
        Down-regulation of matrix Gla protein messenger RNA in human colorectal adenocarcinomas.
        Cancer Lett. 2001; 165: 63-69https://doi.org/10.1016/s0304-3835(01)00416-5
        • Piscaer I.
        • van den Ouweland J.M.W.
        • Vermeersch K.
        • Reynaert N.L.
        • Franssen F.M.E.
        • Keene S.
        • Wouters E.F.M.
        • Janssens W.
        • Vermeer C.
        • Janssen R.
        Low vitamin K status is associated with increased elastin degradation in chronic obstructive pulmonary disease.
        J. Clin. Med. 2019; 8: E1116https://doi.org/10.3390/jcm8081116
        • Dofferhoff A.S.M.
        • Piscaer I.
        • Schurgers L.J.
        • Visser M.P.J.
        • van den Ouweland J.M.W.
        • de Jong P.A.
        • Gosens R.
        • Hackeng T.M.
        • van Daal H.
        • Lux P.
        • Maassen C.
        • Karssemeijer E.G.A.
        • Vermeer C.
        • Wouters E.F.M.
        • Kistemaker L.E.M.
        • Walk J.
        • Janssen R.
        Reduced vitamin K status as a potentially modifiable risk factor of severe coronavirus disease 2019.
        Clin. Infect. Dis. 2021; 73: e4039-e4046https://doi.org/10.1093/cid/ciaa1258
        • Popa D.-S.
        • Bigman G.
        • Rusu M.E.
        The role of vitamin K in humans: implication in aging and age-associated diseases.
        Antioxidants. 2021; 10: 566https://doi.org/10.3390/antiox10040566
        • Schurgers L.J.
        • Barreto D.V.
        • Barreto F.C.
        • Liabeuf S.
        • Renard C.
        • Magdeleyns E.J.
        • Vermeer C.
        • Choukroun G.
        • Massy Z.A.
        The circulating inactive form of matrix gla protein is a surrogate marker for vascular calcification in chronic kidney disease: a preliminary report.
        Clin. J. Am. Soc. Nephrol. 2010; 5: 568-575https://doi.org/10.2215/CJN.07081009
        • Vermeer C.
        • Vik H.
        Effect of Menaquinone-7 (vitamin K2) on vascular elasticity in healthy subjects: results from a one-year study.
        Vasc. Dis. Ther. 2020; 5https://doi.org/10.15761/VDT.1000179