Impact of Vitamin K supplementation on cardiovascular health outcomes associated with kidney disease – A systematic review and meta-analysis

Eligibility criteria

A PECOS framework was used to establish inclusion and exclusion criteria, in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses reporting guidelines.^[13] Studies were included if they were patient studies of CKD or ESKD (Population), supplemented with Vitamin K (menaquinone forms: MK-4, MK-7 or MK-9) (Exposure), compared with placebo or no treatment (Comparator) and measured cardiovascular health outcomes such as arterial stiffness (pulse wave velocity), CAC (Agatston Score), all-cause and cardiovascular mortality and vascular function markers (Outcome). RCTs, prospective cohorts and case–control studies in isolation were included for analysis, but cross-sectional studies, reviews, animal studies and studies with poor control groups were excluded from the study.

Search strategy

A systematic review of the existing literature was carried out using various databases such as PubMed, Embase, Cochrane Library, Web of Science and Scopus using a strategic combination of MeSH terms and Boolean operators [Table 1]. The specific search terms utilised were ‘Vitamin K2’ OR ‘menaquinone’ AND ‘chronic kidney disease’ OR ‘end-stage renal disease’ AND ‘cardiovascular disease’ OR ‘arterial stiffness’ OR ‘vascular calcification’ OR ‘mortality.’ The search was restricted to peer-reviewed articles published in English from the inception of the databases until the completion date of the final search (February 2025).

Study selection and data extraction

Two independent reviewers screened titles and abstracts, and subsequently full texts for assessment of study eligibility. Disagreements were settled by discussion with a third reviewer. Data extraction was performed using a standard data collection form, extracting information such as study design, sample size, description of participants, Vitamin K2 dose and duration, control interventions and cardiovascular health outcomes.

Risk of bias evaluation

ROBINS-I tool^[14] for the non-randomised trials and Cochrane Risk of Bias 2.0 assessment tool^[15] for the RCTs were applied, respectively to assess the risk of bias across the included trials. We based the studies on methodological quality as low risk of bias, moderate risk of bias or high risk of bias.

Statistical analysis

Meta-analysis was conducted with Review Manager (RevMan 5.4) and Stata 17. Continuous variables such as pulse wave velocity and CAC were pooled through mean differences (MDs) with 95% confidence intervals (CIs), while binary outcomes such as all-cause mortality and cardiovascular mortality were calculated as odds ratios (ORs) or risk ratios (RR). Random-effects models were used to account for heterogeneity, and heterogeneity size was calculated through the I² statistic. The study selection process is shown in Figure 1 (PRISMA flowchart).

Close

Figure 1:

PRISMA flowchart representing the study selection process for the review. PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Export to PPT

Bias assessment observations

Among the RCTs included [Figure 2], De Vriese et al.^[16] Lees et al.^[19] and Eelderink et al.^[17] had some issues in some domains, notably D1 (process of randomisation) and D4 (measurement of the outcome), but were at low risk in other domains. Kurnatowska et al.^[18] and Levy-Schousboe et al.^[20] were at low risk of bias in most domains but had some issues in D3 (missing outcome data) and D5 (selection of reported results). Witham et al.^[23] had issues in D2 (deviations from intended interventions) but otherwise had a low overall risk of bias.

Close

Figure 2:

Bias assessment using the RoB 2.0 tool.

Export to PPT

For non-randomised trials, as evaluated by ROBINS-I [Figure 3], Palmer et al.^[21] had a general moderate risk of bias, mainly due to moderate risk in D1 (confounding), D4 (outcome measurement) and D6 (intervention classification), whereas in all the other domains, low risk was present. Shea et al.^[22] had low risk in most of the domains, whereas only in D6 (intervention classification), it was moderate, and in general, it was low.

Close

Figure 3:

Bias assessment using the ROBINS-I tool.

Export to PPT

Demographic attributes

The trials in this review were a broad range of RCTs and observational cohort studies, varying by sample size, follow-up duration and demographic profile of the participants. The RCTs by De Vriese et al.,^[16] Eelderink et al.,^[17] Kurnatowska et al.,^[18] Lees et al.,^[19] Levy-Schousboe et al.^[20] and Witham et al.^[23] mainly examined the effect of interventional Vitamin K2 supplementation. The prospective cohort studies by Palmer et al.^[21] and Shea et al.^[22] aimed to examine dietary Vitamin K intake as well as biomarker-related correlations with cardiovascular events.

Sample sizes were extremely heterogeneous, ranging from a large study by Palmer et al.^[21] (n = 56,048, follow-up time: 23 years) to a very small RCT by Eelderink et al.^[17] (n = 40, follow-up time: 12 weeks). Mean age of the participants varied from 57 years (Eelderink et al.^[17]) to 66 years (Witham et al.^[23]), which implies that the population was largely middle-aged to elderly.

Sex distribution among participants was variably presented, with Kurnatowska et al.^[18] presenting a precise male-to-female ratio (22 M:20 F), while other studies presented approximate percentage females (e.g., Witham et al.^[23]: 39% female; Eelderink et al.^[17]: 35% female). Sex distribution was not presented by some studies (De Vriese et al.,^[16] Lees et al.^[19] and Levy-Schousboe et al.^[20]), which may limit the possibility of conducting gender-specific outcome analyses [Table 2].

Vitamin K2 supplementation and study interventions

The research included mainly Vitamin K2 supplementation, mainly in the form of menaquinone-7 (MK-7), varying in dose and frequency of administration. De Vriese et al.^[16] used 2000 mcg of MK-7 3 times a week together with rivaroxaban, while Eelderink et al.^[17] gave 360 mcg MK-7 daily. Kurnatowska et al.^[18] used 90 mcg MK-7/day co-administered with Vitamin D, while Witham et al.^[23] gave 400 mcg MK-7/day. A different approach was taken in Lees et al.^[19] which used Menadiol diphosphate (Vitamin K3) at 5 mg 3 times a week, differing from the MK-7 form mainly used. Control groups used were placebo or standard treatment, except in Kurnatowska et al.^[18] where Vitamin D was used as the control. The population targeted was varied, with post-transplant patients being included in Eelderink et al.^[17] and Lees et al.^[19] but dialysis patients were studied solely in De Vriese et al.^[16] and Levy-Schousboe et al.^[20] The estimated glomerular filtration rate (eGFR) was varied in range, with Witham et al.^[23] using patients with eGFR 15–45 mL/min/1.73 m², while other trials had more severe CKD or dialysis-dependent patients [Table 3].

Impact on vascular health and calcification markers

Several studies have also looked at vascular stiffness and arterial calcification by measuring parameters such as pulse wave velocity (PWV), carotid intima-media thickness (CCA-IMT), CAC and abdominal aortic calcification (AAC). Eelderink et al.^[17] reported a significant reduction in PWV (−0.06 ± 0.26 m/s) in comparison to the placebo group (+0.27 ± 0.43 m/s, P = 0.010), demonstrating increased arterial elasticity following Vitamin K2 supplementation. On the contrary, Witham et al.^[23] did not see statistically significant changes in PWV (−0.1 m/s, 95% CI: −0.9–0.7, P = 0.77), indicating the vascular response may be dose-dependent or based on the severity of CKD. Kurnatowska et al.^[18] documented a significant reduction in CCA-IMT (0.06 ± 0.08 mm) in comparison to the control group (0.136 ± 0.05 mm, P = 0.005), indicating the beneficial effect of Vitamin K2 on arterial remodelling. On the contrary, Lees et al.^[19] and Levy-Schousboe et al.^[20] noted no significant changes in aortic distensibility, CAC or AAC despite documenting a measurable decrease in dp-ucMGP (−1380 pmol/L, P < 0.05 as per Levy-Schousboe et al.^[20]), the marker of Vitamin K status. These findings indicate that even though Vitamin K2 has been linked with benefits in biomarkers of VC, the clinical significance of arterial stiffness and plaque formation remains uncertain.

Effects on cardiovascular and overall mortality rates

Longitudinal cohort studies have clarified the associations of Vitamin K status and risk of mortality. Palmer et al.^[21] found a robust inverse association between dietary Vitamin K and all-cause mortality (HR: 0.76, 95% CI: 0.72–0.79, P < 0.001) with a protective effect. Similarly, Shea et al.^[22] noted reduced risk of all-cause mortality with lower levels of dp-ucMGP (HR: 0.71, 95% CI: 0.61–0.83, P < 0.001). In RCTs, De Vriese et al.^[16] noted a significant reduction in fatal and nonfatal cardiovascular events (HR: 0.34, 95% CI: 0.19–0.61, P = 0.0003) in dialysis patients treated with Vitamin K2 supplementation. Conversely, Witham et al.^[23] and Lees et al.^[19] did not find statistically significant differences in all-cause mortality or cardiovascular events and noted heterogeneity of Vitamin K2 effects in different subpopulations with CKD.

Biochemical markers and follow-up results

Several studies have compared biomarkers of Vitamin K to assess the efficacy of supplementation. Eelderink et al.^[17] and Levy-Schousboe et al.^[20] both indicated significant decreases in dp-ucMGP levels, with Levy-Schousboe et al.^[20] finding a mean decrease of 1380 pmol/L (P < 0.05). Kurnatowska et al.^[18] also assessed OC and osteoprotegerin, indicating significant reductions after Vitamin K2 supplementation. Witham et al.^[23] assessed markers of arterial stiffness, such as augmentation index (AIx) and blood pressure (BP), but indicated no significant changes in these measurements. Overall, these results suggest that although Vitamin K2 universally increases biochemical markers that are predictive of vascular health, the direct clinical significance for cardiovascular outcomes is uncertain.

Arterial stiffening and VC

The carotid-femoral pulse wave velocity analysis [Figure 4] showed a pooled MD of 0.05 [−0.34, 0.45], no statistically significant difference being found by Vitamin K supplementation versus controls (Z = 0.27, P = 0.79). Low heterogeneity (I² = 0%) existed, with homogenous results being reported in studies (Eelderink et al.,^[17] Lees et al.,^[19] Levy-Schousboe et al.^[20] and Witham et al.^[23]).

Close

Figure 4:

Carotid-femoral pulse wave velocity (Vitamin K vs. controls). The bold value represents the mean difference (MD) in pulse wave velocity (PWV) between the Vitamin K2 supplementation and control groups. SD: Standard deviation, IV: Inverse variance, CI: Confidence interval.

Export to PPT

Likewise, coronary arterial calcification (Agatston Score) [Figure 5] yielded a pooled MD of 25.05 [−57.05, 107.14] with no statistically significant difference between the Vitamin K and control groups (Z = 0.60, P = 0.55). The heterogeneity was very low in the analysis (I² = 0%), in support of the consistency of the results. The same trend was noted in abdominal aortic calcification (Agatston Score) [Figure 6] with the pooled effect size of 10.35 [−30.76, 51.45], with no significant reduction in VC (Z = 0.49, P = 0.62).

Close

Figure 5:

Coronary arterial calcification - Agatston score (Vitamin K vs. controls). The bold value represents the mean difference (MD) in pulse wave velocity (PWV) between the Vitamin K2 supplementation and control groups. SD: Standard deviation, IV: Inverse variance, CI: Confidence interval.

Export to PPT

Close

Figure 6:

Abdominal aortic calcification – Agatston score (Vitamin K vs. controls). The bold value represents the mean difference (MD) in pulse wave velocity (PWV) between the Vitamin K2 supplementation and control groups. SD: Standard deviation, IV: Inverse variance, CI: Confidence interval.

Export to PPT

Mortality outcomes

The combined OR for all-cause mortality [Figure 7] was calculated as 0.96 [0.93, 0.98], showing a small but statistically significant reduction in mortality in those cohorts treated with Vitamin K (Z = 3.31, P = 0.0009). The biggest share of this effect was provided by cohort studies (Palmer et al.,^[21] Shea et al.,^[22]), whereas RCTs (Eelderink et al.,^[17] Witham et al.,^[23]) provided non-significant results (OR = 0.81 [0.45, 1.47], Z = 0.69, P = 0.49).

Close

Figure 7:

All-cause mortality across randomised controlled trials and cohort studies (Vitamin K vs. controls). The bold value represents the mean difference (MD) in pulse wave velocity (PWV) between the Vitamin K2 supplementation and control groups. SD: Standard deviation, IV: Inverse variance, CI: Confidence interval.

Export to PPT

For cardiovascular mortality [Figure 8], the global OR was 0.98 [0.94, 1.02], which did not indicate any protective effect of Vitamin K (Z = 1.09, P = 0.28). RCT and cohort studies revealed comparable findings, with low heterogeneity (I² = 0%).

Close

Figure 8:

Cardiovascular mortality across randomised controlled trials (RCT) and cohort studies (Vitamin K vs. controls). The bold value represents the mean difference (MD) in pulse wave velocity (PWV) between the Vitamin K2 supplementation and control groups.

Export to PPT

Comparative assessment of analysed evidence

The results across the included trials varied with regard to the impact of Vitamin K2 supplementation on cardiovascular events in CKD and ESKD patients. Eelderink et al.,^[17] Kurnatowska et al.^[18] and De Vriese et al.^[16] reported improved vascular function and a decrease in cardiovascular events, while Lees et al. ^[19] Levy-Schousboe et al.^[20] and Witham et al.^[23] had no significant effects on arterial stiffness or VC. Palmer et al.^[21] and Shea et al.^[22] the long-term cohort studies, described that increased Vitamin K intake was related to decreased mortality, partly in agreement with the interventional result of De Vriese et al.^[16] but in disagreement with the null results in the RCTs of Lees et al.^[19] and Witham et al.^[23]

Vitamin K status and cardiovascular risk in CKD

Insufficient consumption of both phylloquinone (Vitamin K1) and menaquinone (Vitamin K2) has been correlated with a heightened risk of cardiovascular mortality as well as overall mortality in patients diagnosed with CKD.^[24] The inadequacy of Vitamin K has been identified as an independent risk factor for CVD, with research indicating its involvement in the processes that lead to VC and arterial stiffness.^[25] Low levels of Vitamin K2, or the pharmacological blockade of Vitamin K function through warfarin administration, have been associated with an augmented accumulation of vascular calcium deposits, thereby exacerbating arterial injury in CKD populations.^[12] Some investigations have indicated that Vitamin K2 supplementation may lead to a modest increase in high-density lipoprotein cholesterol levels while concurrently reducing systemic inflammation.^[26,27] In light of its possible impacts on vascular health, it has been suggested that Vitamin K2 supplementation could decelerate the progression of VC and lower the risk of atherosclerosis, CVD and stroke.^[28-31] Observational evidence derived from dietary intake analyses has indicated an inverse association between menaquinone intake exceeding 21.6 mg/day and mortality associated with coronary heart disease and aortic calcification; however, such an association has not been observed with phylloquinone intake.^[32-34] Results from the PREVEND study indicated that 31% of participants demonstrated functional Vitamin K deficiency, with a markedly higher prevalence among older individuals, those afflicted by hypertension, type 2 diabetes, CKD and pre-existing cardiovascular conditions.^[35] Active RCTs are ongoing as well to evaluate the efficacy of Vitamin K1 and K2 supplementation for reducing VC in CKD, though optimal dose and clinical efficacy remain under investigation.^[36,37]

Guidelines pertaining to Vitamin K supplementation in CKD

International clinical practice guidelines suggest arterial BP below 130/80 mmHg in CKD patients with comorbid cardiovascular risk factors, in an attempt to avoid cardiorenal complications.^[38] Recent studies have envisioned that Vitamin K2 supplementation could have a role in the modulation of BP, particularly in primary hypertension.^[39,40] Mechanistic studies have employed 16S ribosomal RNA (rRNA) sequencing to examine the theoretical mechanisms by which Vitamin K2 modulates vascular function and BP homeostasis.^[41] Their results demonstrated interaction between Vitamin K2, the complement system, calcium signalling pathways and the renin-angiotensin-aldosterone system (RAAS) in an experimental model of salt-sensitive hypertension.^[41] The study proved RAAS participation in salt-induced hypertension, but also demonstrated Vitamin K2 administration blocked RAAS-mediated signals, and thus suggests a potential regulatory function in hypertension pathophysiology. In addition, microbial analysis demonstrated some gut bacteria, such as Dubosiella and Ileibacterium, were associated with beneficial modulation of RAAS, and thus suggest a potential connection between the composition of intestinal microbiota, Vitamin K2 metabolism and vascular homeostasis.^[42] These findings have prompted speculations that probiotic supplementation with bacterial strains enhancing Vitamin K2 production is implicated in vascular endothelium protection through immune and metabolic pathway regulation.^[25-42] Clinical trials are, however, needed to establish the long-term effect of Vitamin K2 supplementation on hypertension and vascular function in CKD populations.

Vitamin K deficiency and its significance in CKD

Our findings concurred with the postulations of Bellone et al.^[43] who emphasised that Vitamin K deficiency is an ubiquitous and potentially modifiable risk factor in CKD that is responsible for VC as well as bone fragility. In line with their study, our research confirmed that CKD patients with a low Vitamin K diet were at increased risk of all-cause mortality, as revealed by the pooled OR of 0.96 [0.93, 0.98] (Z = 3.31, P = 0.0009, I² = 0%). However, while Bellone et al.^[43] theorised that Vitamin K supplementation needs to be added to treatment strategies for CKD to prevent cardiovascular and skeletal complications, our meta-analysis was unable to identify any significant reduction in vascular stiffness or calcification biomarkers with such supplementation. This finding suggests that while Vitamin K deficiency is undoubtedly associated with unfavourable outcomes in the context of cardiovascular and bone morbidity, the immediate effect of supplementation is uncertain.

In addition, Bellone et al.^[43] suggested that Vitamin K status could serve as a biomarker of renal and cardiovascular well-being, a notion indirectly validated by our findings of reductions in dp-ucMGP within some of the trials (e.g., Levy-Schousboe et al.,^[20]) even without the attendant clinical benefits. This suggests that while Vitamin K-dependent proteins can be employed as a marker of vascular well-being, their alteration with supplementation does not necessarily equate to improved clinical results.

VC and activation of MGP

Our evidence supports Roumeliotis et al.^[44] who emphasised that VC is a progressive complication of CKD and ESRD and is mediated by a pro-calcific/anti-calcific imbalance, with MGP involved. Roumeliotis et al.^[44] further claimed that the inactive dp-ucMGP form of MGP, reflecting Vitamin K deficiency, is strongly associated with cardiovascular events and mortality in CKD patients. Our research also showed that cohort studies had a negative correlation between Vitamin K intake and all-cause mortality, which agrees with their evidence.

However, Roumeliotis et al.^[44] suggested that Vitamin K-dependent proteins can actively induce the reversal of VC by removing calcium deposits from arterial walls. The above suggestion conflicts with our findings, which revealed that MDs for carotid-femoral pulse wave velocity (0.05 [−0.34, 0.45], Z = 0.27, P = 0.79) and CAC (25.05 [−57.05, 107.14], Z = 0.60, P = 0.55) showed no significant improvement with Vitamin K2 supplementation. Such differences can be explained by differences in study populations, follow-up periods or suboptimal dosing of Vitamin K2 in the performed interventional trials. While mechanistic studies suggest that Vitamin K can have a protective advantage to vascular health, our evidence suggests that such an effect is not necessarily translatable to clinical significance in all CKD populations.

Vitamin K subclinical deficiency in CKD

The results shown by Cozzolino et al.^[45] concurred with our observation of prevalent subclinical Vitamin K deficiency among CKD patients. Their experiment demonstrated increased Vitamin K requirements upon activation of Vitamin K-dependent proteins involved in the regulation of calcification, which fact is consistent with increased dp-ucMGP levels reported among CKD patients in our considered studies (e.g., Levy-Schousboe et al.^[20]). Yet, while Cozzolino et al.^[45] indicated that the deficiency would increase the risk of cardiovascular complications in CKD patients, our research did not identify a lasting benefit of Vitamin K2 supplementation to reduce vascular stiffness or calcification of arteries. One of the key areas of contention is that Cozzolino et al.^[45] indicated that Vitamin K depletion is a direct stimulus to VC, whereas our study revealed that even with Vitamin K2 supplementation, no appreciable improvement in VC outcomes was evident (MD for AAC: 10.35 [−30.76, 51.45], Z = 0.49, P = 0.62). This indicates that although Vitamin K deficiency may be a causative factor for the phenomenon, it is doubtful that it is the only factor for VC in patients with CKD, and other pathophysiological mechanisms may be involved.

Limitations

This study was also faced with numerous limitations that influenced the interpretation of its results. Variability in the study design, Vitamin K2 dosing, follow-up periods and CKD stages led to heterogeneity of the reported effects. While improved biomarkers were documented, whether these changes have clinical significance which is uncertain, since there were no reported long-term decreases in arterial stiffness and VC. Most of the significant results relied on observational cohort studies, which are prone to residual confounding, and were thus precluded from making claims of causality. In addition, the relatively modest sizes of a number of the RCTs might have diluted the statistical power necessary to show significant differences between cardiovascular outcomes. Heterogeneity of the control interventions, which involved placebo and standard care, also made study comparisons difficult in a direct manner. Finally, the follow-up periods in a number of studies might not have been long enough to ascertain the long-term effects of Vitamin K2 on CVD.

Future implications

Larger controlled trials with controlled dosing of Vitamin K2, control interventions and follow-up times should be carried out in future research to determine its role in cardiovascular risk reduction in CKD and ESKD patients. Long-term follow-up studies will be required to determine the long-term effect of supplementation on VC and clinical outcomes. Investigation of potential interactions between Vitamin K2 and other cardiovascular risk modifiers, including anticoagulants and mineral metabolism regulators, may further determine its therapeutic role. Patient subgroups defined by CKD severity, dialysis status and baseline Vitamin K status should be studied to determine populations most likely to benefit from supplementation. Standardised reporting of arterial stiffness and VC outcomes will be required to maximise between-study comparability and to construct the evidence base to inform clinical recommendations.