Abstract
Introduction Matrix Gla protein (MGP) is an inhibitor of lung tissue calcification. The plasma level of dephosphorylated-uncarboxylated MGP (dp-ucMGP) is a biomarker of vitamin K status. The present study assessed whether lower vitamin K status (reflected by higher dp-ucMGP) was associated with lung function and lung disease/symptoms.
Methods A general population sample of 4092 individuals, aged 24 to 77 years, underwent a health examination including questionnaires, spirometry and measurements of plasma dp-ucMGP. Associations of dp-ucMGP with lung function and self-reported disease/symptoms were estimated using regression models adjusted for age, sex and height. Associations were expressed as β-estimates or odds ratios (ORs) per doubling in dp-ucMGP.
Results Lower vitamin K status (higher dp-ucMGP) was associated with lower forced expiratory volume in 1 s (FEV1) (98 mL; 95% CI: 54–141 mL) and lower forced vital capacity (FVC) (136 mL; 95% CI: 85–187 mL). Dp-ucMGP was not associated with the FEV1/FVC ratio (0.0 percentage points higher than the expected value; 95% CI: −1.0–1.0). Furthermore, lower vitamin K status was associated with COPD (OR 2.24, 95% CI: 1.53–3.27), wheezing (OR 1.81, 95% CI: 1.44–2.28) and asthma (OR 1.44, 95% CI: 1.12–1.83).
Conclusion Lower vitamin K status was associated with lower ventilatory capacity (lower FEV1 and FVC), and with higher risk of self-reported asthma, COPD and wheezing. Vitamin K status was not associated with airflow obstruction (FEV1/FVC ratio).
Tweetable abstract
Lower vitamin K status was associated with lower lung function in a cross-sectional general population study. Individuals with lower vitamin K status were more likely to report COPD, wheezing and asthma. https://bit.ly/44b4xEB
Introduction
Matrix Gla protein (MGP) is a mineral-binding protein secreted by chondrocytes and vascular smooth muscle cells in the extracellular matrix [1]. MGP is present in various tissues, including lung tissue [2], and is a powerful inhibitor of soft-tissue calcification [3]. MGP may have a special role in preventing calcification of the lung tissue, since the size of the MGP is smaller in contrast to other calcification inhibitors, and it functions inside small fibres such as elastin and collagen [4].
One of the current theories on how COPD develops is based on the interaction between elastases and anti-elastases [5]. Elastases are inflammatory proteases that degrade elastin fibres. In the absence of anti-elastases such as α-1 antitrypsin (which is strongly associated with COPD), degradation continues untamed, and due to elastin's affinity for calcium, elastin degradation further contributes to elastin calcification [5]. Elastin is eventually reproduced by fibroblasts in the lungs, but the intricate structure of the fibres cannot be reestablished. The gradual unravelling of the fibre structure eventually leads to emphysema, a condition which often coexists with COPD [5, 6].
The function of elastin in the lung parenchyma is to contribute to the elasticity of the lungs. This affects not only the expansion of the lungs during inspiration, but also the elastic recoil during expiration [5]. Elastin degradation may therefore contribute to both obstructive and restrictive lung pathologies [5].
Gla proteins, such as MGP, depend on vitamin K for activation and are therefore often referred to as vitamin K-dependent proteins [7]. The unmodified isoform of MGP, dephosphorylated-uncarboxylated MGP (dp-ucMGP), is physiologically inactive [8] and is used as an inverse biomarker of vitamin K status in peripheral tissues. High dp-ucMGP plasma levels reflect the amount of inactive/uncarboxylated MGP and thus low vitamin K status [6]. Measurement of dp-ucMGP is preferred to direct measurements of vitamin K vitamers for assessment of functional vitamin K status since plasma vitamers are more transient and reflect recent food intake [9].
It has been hypothesised that the allocation of vitamin K in the body follows a triage-based distribution, i.e. functions related to short-term survival will be prioritised at the expense of other functions [10]. When vitamin K levels are scarce, hepatic functions (such as activation of prothrombotic proteins) will be favoured at the expense of extra-hepatic functions such as lung tissue [11]. While inactive hepatic Gla proteins are often used as biomarkers of coagulation and vitamin K status, they are insensitive markers for extra-hepatic vitamin K status, making dp-ucMGP the preferable extra-hepatic biomarker for vitamin K status [12].
Previous studies in smaller patient cohorts have tried to establish a relationship between low vitamin K status and COPD [13, 14], COVID-19 [15, 16] and cystic fibrosis [17]. All studies found lower vitamin K status (higher dp-ucMGP levels) in patient populations compared with controls. Less is known about the associations between vitamin K status and lung function in individuals from the general population. This study examines the association between low vitamin K status (as reflected by higher dp-ucMGP levels) and lung parameters, including lung function measures, in a large sample from the general population in Denmark.
Methods
Setting
The DanFunD cohort
The Danish study of Functional Disorders (DanFunD) cohort was a random sample of the general adult population comprising men and women born in Denmark and living in the western part of Greater Copenhagen [18]. The DanFunD baseline cohort consisted of a part 1 (n=2308) and a part 2 (n=7493), and the present study consisted of participants from the 5-year follow-up of DanFunD part 2 (n=4291) [19]. All participants, who underwent health examinations between 2017 and 2020 at The Center for Clinical Research and Prevention, Denmark, were included in the present study (n=4092) (figure 1 and table 1). Participants underwent health examinations including a lung function examination (spirometry) and blood and urine samples. Prior to the health examination, participants filled in questionnaires on general health, chronic diseases and lifestyle factors such as smoking, alcohol consumption, physical activity and diet.
Flowchart of study participation and analyses. dp-ucMGP: dephosphorylated-uncarboxylated matrix Gla protein.
Characteristics of the study population according to quartile levels of dephosphorylated-uncarboxylated matrix Gla protein (dp-ucMGP)
Vitamin K status
Biochemical determination of vitamin K status was performed according to the manufacturer's instructions using the IDS-iSYS InaKtif MGP assay, which is an in vitro diagnostic test intended for the quantitative determination of the inactive isoform of MGP, i.e. dp-ucMGP, in human plasma [20]. 26 participants had missing dp-ucMGP measurements, 10 of which had missing blood samples. Dp-ucMGP values below the lower limit of quantitation (300 pmol·L−1) were set to 299 pmol·L−1.
Lung function (spirometry)
Lung function was measured through spirometries performed with Spiro USB Spirometers (MicroMedical Limited, Rochester, Kent, UK). A daily calibration check was made with a 3-L calibrated syringe, and every 6 months a calibration check was made with a decompression flow simulator [21]. The spirometries and measurements of stature were performed according to the American Thoracic Society and the European Respiratory Society (ATS/ERS) standard [22]. Owing to the COVID-19 pandemic, spirometries were halted on 11 March 2020. As a result, 904 participants had missing information on lung function parameters (figure 1).
Lung function measurements included forced expiratory volume in 1 s (FEV1, litres) and forced vital capacity (FVC, litres). FEV1/FVC ratio was calculated as the ratio between these measurements. Predicted values for FEV1, FVC and FEV1/FVC ratio were calculated according to the Global Lung Initiative (GLI) recommendations [23]. FEV1, FVC and FEV1/FVC were then calculated as percentages of the predicted values. The GLI equations account for participants’ sex, age, height and ethnicity. Ethnicity was set to caucasians for all, as data on ethnicity was not obtained and the vast majority of individuals were caucasians. Lung function outcomes were also defined as dichotomous measures (normal or decreased) according to the “lower limit of normal” (LLN). As recommended by an ERS task force, LLN was not set as a fixed threshold (e.g. 80% of predicted, or 0.7 for FEV1/FVC ratio). Instead, the 2.5th percentile was used (z-score −1.970) to define LLN [23].
Self-reported estimates
Self-reported asthma and COPD were based on a yes/no response to the questions “Has a doctor ever told you that you have: asthma?” and “Has a doctor ever told you that you have: COPD/smoker's lungs?”, respectively. Asthma was also separately defined from questionnaire answers according to ECRHS criteria [24]. “Wheezing” was included as a confirmatory answer to both questions “Have you had wheezing or whistling in your chest at any time during the last 12 months?” and “If yes, have you had this wheezing or whistling when you did not have a cold?”.
Ethics
All participants provided oral and written informed consent for participation. The study was approved by the Ethical Committee of Copenhagen County (Ethics Committee: H-17021141; H-3-2012-0015) and the Danish Data Protection Agency.
Statistical analysis
Background characteristics were reported according to dp-ucMGP quartiles. For continuous variables, characteristics were reported as mean and as median, according to distribution. Categorical variables were reported as frequencies. Differences in characteristics between dp-ucMGP quartiles were assessed by one-way analysis of variance for normally distributed variables, Kruskal–Wallis test for skewed variables and Chi square test for categorical variables.
Dp-ucMGP levels were log2 transformed to achieve an approximate normal distribution. Multivariable linear regressions were used to assess associations between dp-ucMGP levels and lung function outcomes: FEV1 (mL), FVC (mL) and FEV1/FVC ratio. The associations between dp-ucMGP and FEV1, FVC and FEV1/FVC ratio were also calculated as differences in percentage points of the expected value. Multivariable logistic regressions were used to assess the associations between dp-ucMGP and dichotomous variables: COPD, asthma and wheezing, and reported as odds ratios (ORs). Estimates were reported with 95% confidence intervals (CIs).
Analyses were adjusted for potential confounders in three models adjusting for: 1) age, sex and height; 2) further adjusted for weight and estimated glomerular filtration rate (eGFR, used as a marker of kidney function) calculated with the CKD-EPI creatinine equation [25]; 3) further adjusted for smoking status (current, occasional, former, never), leisure-time physical activity [26], self-rated fitness [27], increased alcohol consumption (>14/>7 units/week for men/women) and healthy eating index [28]. Similar adjustment models were used for outcomes with percentage of predicted values for FEV1, FVC and FEV1/FVC ratio, except that these were not adjusted for age, sex and height, since those are included in the derivation from the GLI equations [23].
Sensitivity analyses were performed to assess: 1) the potential impact of vitamin K antagonists (VKAs) by excluding all participants with self-reported anticoagulant use; 2) whether censoring dp-ucMGP outliers (>3 sd from the mean) affected the results; 3) whether associations were affected by applying dp-ucMGP as a dichotomous outcome (>500 pmol·L−1 versus ≤500 pmol·L−1); 4) whether associations existed for participants with normal kidney function by excluding participants with eGFR values <60 mL/min/1.73 m2; 5) whether associations were different for never-smokers only; 6) whether associations were different for participants without airflow obstruction, FEV1/FVC≥LLN; 7) whether associations were different for younger participants (<65 years); and 8) whether associations were different when adjusting for smoking status in all adjustment models. The data were analysed using SAS Enterprise Guide 7.15 (SAS Institute Inc., Cary, NC, USA).
Results
A total of 4092 individuals, aged 24–77 years, participated in the DanFunD 5-year follow-up health examinations. Dp-ucMGP measurements were available for 4066 participants. Of these, 897 participants had no lung function measurements, resulting in 3169 participants in analyses on lung function measurements. Among the 4066 participants, 18 participants had missing information on self-reported asthma, COPD and wheezing, resulting in 4048 participants in analyses on self-reported outcomes (figure 1).
Lower vitamin K status (higher dp-ucMGP levels) was associated with higher body mass index, higher age, higher systolic blood pressure, lower kidney function, higher albumin-to-creatinine ratio (ACR, >30 mg·g−1), high alcohol consumption, lower physical activity, lower self-rated fitness, less healthy eating, former smoking and anticoagulant use (table 1).
Lung function
Lower vitamin K status (as reflected by higher dp-ucMGP levels) was associated with lower FEV1 and lower FVC in all models in the multivariable linear regressions (tables 2 and 3). In model 1 (adjusting for age, sex and height), FEV1 was 97.8 mL (95% CI: 54.4–141.3 mL) lower per doubling in dp-ucMGP, and FVC was 136.0 mL (95% CI: 84.9–187.2 mL) lower per doubling in dp-ucMGP (table 2).
Changes in lung function per doubling in dephosphorylated-uncarboxylated matrix Gla protein (dp-ucMGP)
Odds ratios (OR) and 95% confidence intervals (CI) of lung function abnormalities, self-reported disease and symptoms per doubling in dephosphorylated-uncarboxylated matrix Gla protein (dp-ucMGP)
Using percentages of predicted values of FEV1 and FVC presented a similar pattern. FEV1 was 3.3 percentage points (95% CI: 2.0–4.7) lower than the predicted value per doubling of dp-ucMGP, and FVC was 3.4 percentage points (95% CI: 2.2–4.7) lower than the predicted value per doubling of dp-ucMGP (table 2).
A doubling in dp-ucMGP was associated with an OR of 2.41 (95% CI: 1.65–3.52) for having FEV1 below LLN (figure 2). The association was attenuated when further adjusting for weight, kidney function and lifestyle factors (figure 2): OR: 2.09 (95% CI: 1.35–3.23). Dp-ucMGP was associated with FVC<LLN (OR 2.48, 95% CI: 1.46–4.22); the association was attenuated when adjusting for weight, kidney function and lifestyle factors (figure 2 and table 3).
Forest plot of lung function measurements below lower limit of normal (LLN) when doubling dephosphorylated-uncarboxylated matrix Gla protein (dp-ucMGP). FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity; eGFR: estimated glomerular filtration rate.
Dp-ucMGP was not associated with the FEV1/FVC ratio in linear regression models. The FEV1/FVC ratio was 0.0 (95% CI: −1.0–1.0) percentage points higher than the predicted value without adjustments (table 2). However, a doubling of dp-ucMGP was associated with an OR of 1.48 (95% CI: 1.02–2.15) of having a FEV1/FVC ratio <LLN when adjusted for weight and kidney function (table 3).
Asthma, COPD and wheezing
Lower vitamin K status was associated with self-reported COPD (OR 2.09, 95% CI: 1.35–3.23) per doubling in dp-ucMGP and wheezing (OR 1.40, 95% CI: 1.08–1.81) per doubling in dp-ucMGP when adjusted for age, sex, height, weight, kidney function and lifestyle factors (figure 3 and table 3).
Forest plot of self-reported outcomes when doubling dephosphorylated-uncarboxylated matrix Gla protein (dp-ucMGP). The European Community Respiratory Health Survey (ECRHS) defines asthma as a confirmatory answer to at least one of the following: being awoken by shortness of breath in the last 12 months, having had an asthma attack in the last 12 months and current use of any type of asthma medication. eGFR: estimated glomerular filtration rate.
Lower vitamin K status was associated with asthma, both self-reported (OR 1.44, 95% CI: 1.12–1.83) and defined according to ECRHS criteria (OR 1.62, 95% CI: 1.27–2.06), adjusted for age, sex and height. The associations were attenuated to an OR of 1.29 (95% CI: 0.99–1.67) for self-reported asthma and an OR of 1.32 (95% CI: 1.01–1.71) for asthma defined by the ECRHS criteria, respectively, when further adjusted for weight, kidney function and lifestyle factors (figure 3 and table 3).
Sensitivity analyses
The sensitivity analyses did not alter conclusions. However, associations were attenuated when limiting the analysis to participants without a smoking history suggesting that vitamin K status may be particularly important for individuals with a smoking history (supplementary tables 1–16).
Discussion
The present study found that while low vitamin K status (reflected by higher dp-ucMGP) was associated with lower ventilatory capacity (reflected by lower FEV1 and FVC), it generally did not affect the FEV1/FVC ratio. Lower vitamin K status was also associated with higher risk of self-reported asthma, COPD and wheezing.
Strengths and weaknesses
The present study possessed several strengths, such as a large sample size compared with previous studies. Further, we had detailed information on biochemical data, lung function and respiratory symptoms. The study also had limitations. Information on dietary or supplementary intake of vitamin K was not available, as estimation of this is challenging due to incomplete food composition databases. However, as phylloquinone from green leafy vegetables is the main source of vitamin K in western populations [29], dp-ucMGP could be a marker of healthy diet and lifestyle. The baseline cohort in DanFunD part 2 had a relatively low participation rate (29.5%), which potentially could bias associations or decrease the representativity of the results. Non-responders had lower socioeconomic status, meaning the study population may under-represent this population segment [30]. Dp-ucMGP is often elevated in patients using VKAs [31]. However, it was not possible to distinguish between VKAs and other anticoagulant treatments in the present data. In a previous study involving the general population from the same area, <1% of the participants used VKAs [20], and so participants on anticoagulants were therefore included in the main analyses. Sensitivity analysis excluding all participants on anticoagulant treatment did not alter the conclusions.
Although the cross-sectional design of the study does not allow for causal or temporal inferences, our results suggest that vitamin K may play a role in respiratory diseases and lung health, not confined to COPD. Future studies could benefit from including total lung capacity (TLC), diffusing capacity of the lung for carbon monoxide and intrathoracic gas volume, which are better suited for examining restrictive pathologies and pulmonary vascular disorders [32]. There is a need for longitudinal studies to clarify the relationship between loss of lung tissue and vitamin K status, preferably with CT-scored evaluations of emphysema and pulmonary fibrosis. Future studies should also consider measuring and adjusting for total MGP (tMGP), since dp-ucMGP is positively associated with tMGP [33] and higher dp-ucMGP levels may not only reflect low vitamin K status but also an induction in MGP synthesis [29].
Comparison with other studies
Only a few studies have investigated the association between vitamin K status (as reflected by dp-ucMGP) and lung function or respiratory disease, and these have almost exclusively focused on COVID-19 [9, 15, 16] and COPD [13, 14]. Previous studies indicate that COPD patients have higher dp-ucMGP levels compared with controls, which corresponds with our finding that low vitamin K status was associated with self-reported COPD. One of the studies also assessed lung function parameters in a cohort with 192 COPD patients and 186 age-matched controls and found 10% higher dp-ucMGP levels to be associated with 3.0% lower FEV1 (95% CI: −1.0–6.8%) [14]. This corresponds with our finding that higher dp-ucMGP levels were associated with lower FEV1. The same study, however, did not find associations between dp-ucMGP and FEV1, FVC or FEV1/FVC ratio in a separate cohort of 290 COPD patients [14].
Other studies have focused on desmosin, a unique waste product of elastin degradation, which is used as a biomarker of the rate of elastic fibre degradation and is part of a proposed central mechanism in the development of COPD [34]. These studies found that lower vitamin K status (reflected by higher dp-ucMGP) was associated with higher desmosin levels [9, 13–15]. In the present study, we did not measure desmosin, however the results are in line with our finding that higher dp-ucMGP was associated with self-reported COPD, and potentially also that dp-ucMGP was associated with lower FEV1, as FEV1 is used in the grading of COPD severity [35].
Others have also suggested that low vitamin K status may be relevant in restrictive lung pathologies [36]. Studies have found an association between use of VKAs and increased mortality in patients with idiopathic pulmonary fibrosis [37], although this could be due to an increased risk of thrombosis rather than parenchymal damage. Case studies have shown that patients with idiopathic pulmonary fibrosis have experienced fatal excacerbations of fibrosis shortly after receiving VKAs [38]. Protein S and GAS6 (both Gla proteins) are present in lung tissue and may have protective effects against fibrotic diseases [36], and this could be an explanation for the association between vitamin K and idiopathic pulmonary fibrosis. The suggested association between vitamin K status and lung fibrosis is in line with our findings that lower vitamin K status is associated with both lower FEV1 and FVC, as seen in restrictive lung diseases. This is supported by a study where low vitamin K status was associated with lower TLC [14], a characteristic feature of restrictive ventilatory defects [32].
Perspectives
It may be speculated that the association between lower vitamin K status and lower lung function is due to a fibrotic response from recurrent occult lung injury. The body's supply of vitamin K is limited, and it may be quickly depleted [39] at the expense of extra-hepatic functions [40], meaning an increased demand for vitamin K from degradation of lung fibres is causing further vitamin K depletion [9]. Paradoxically, this would also increase thrombogenicity, since a large fraction of the anticoagulative Gla protein, protein S, is produced extra-hepatically, and could therefore be affected before the pro-coagulant proteins [39]. Elastin degradation and increased thrombogenicity from vitamin K depletion may help explain why during the COVID-19 pandemic, patients with COVID-19 had higher concentrations of dp-ucMGP and frequently developed both respiratory failure and thromboembolisms [9]. An alternative hypothesis could be that low vitamin K status is associated with reduced lung volumes due to poor lung development earlier in life as a result of vitamin K deficiency or low vitamin K status being a marker of an unhealthy lifestyle. However, the mechanisms of the present findings need further investigation.
The literature on elastinolysis focuses on COPD, an obstructive lung disease, but elastin degradation is also increased in the lungs of patients with restrictive diseases, such as idiopathic pulmonary fibrosis [41]. The associations between vitamin K status and lower ventilatory capacity (reflected by lower FEV1 and FVC) and absence of association with FEV1/FVC ratio could suggest a restrictive pattern. However, vitamin K status has been shown to be inversely associated with elastin degradation and could reflect COPD activity, which would not necessarily be captured by FEV1/FVC ratio or FEV1 [42]. Thus, the existing literature and our results suggest that the relationship between vitamin K status and lung health should be investigated, beyond the development of COPD.
Conclusion
This cross-sectional population-based study found lower vitamin K status to be associated with lower lung function (lower FEV1 and FVC), but not with airflow obstruction (FEV1/FVC ratio). Low vitamin K status was also associated with higher risk of self-reported asthma, COPD and wheezing.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material 00208-2023.SUPPLEMENT
Footnotes
Provenance: Submitted article, peer reviewed.
Data availability: The data are not publicly available due to Danish regulations on personal data protection. Data are available on a collaborative basis; please contact allan.linneberg{at}regionh.dk
Conflict of interest: F.B. Kampmann, A. Linneberg and S.M. Thysen have received an unrestricted grant and investigational products from Kappa Bioscience A/S for an intervention trial (The InterVitaminK Trial; clinicatrials.gov identifier NCT05259046) using vitamin K supplements as the active intervention.
Conflict of interest: The other authors declare no other conflicts of interest.
Support statement: The present study was funded by a Steno Collaborative Grant 2019 from the Novo Nordisk Foundation (0058130), the Center for Clinical Research and Prevention, and by a grant from the Danish Cardiovascular Academy, which is funded by the Novo Nordisk Foundation (grant number NNF20SA0067242) and The Danish Heart Foundation. Funding information for this article has been deposited with the Crossref Funder Registry.
Author contributions: F.B. Kampmann, A. Linneberg and S.M. Thysen conceived the idea for the study. T. Jespersen, F.B. Kampmann, A. Linneberg and S.M. Thysen designed the study. T.M. Dantoft, L.L. Kårhus and A. Linneberg were responsible for data collection and curation. N.R. Jørgensen was responsible for biochemical analyses of blood samples. F. Madsen set up the lung function measurements. T. Jespersen analysed the data with supervision from S.M. Thysen. T. Jespersen wrote the first draft of the paper with input from F.B. Kampmann, A. Linneberg and S.M. Thysen. All authors read, reviewed and approved the final manuscript.
- Received April 2, 2023.
- Accepted June 28, 2023.
- Copyright ©The authors 2023
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org
References
