⚠️ Closed-access (Thrombosis and Haemostasis 2004; no PMC; not obtainable via local archive). Verified against the published abstract (Crossref JATS abstract, PMID 14961167): direction of effect, all four percentage contrasts (DC 8.8%, CC 8.6%, PP 6.3%, E 13.2%), p-values, n’s (181 enrolled / 150 completed / 108 analysed), three-arm design, 3-year duration, K1 (not K2) form, and IMT null result are all abstract-confirmed. Absolute within-arm values for DC, CC, PP, and E at baseline and year 3, dose of vitamin K1 and vitamin D, participant age range, and attrition reasons remain unconfirmed without full-text access. gap/no-fulltext-access
Beneficial effects of vitamins D and K on the elastic properties of the vessel wall in postmenopausal women
Braam LA, Hoeks AP, Brouns F, Hamulyák K, Gerichhausen MJ, Vermeer C · Thrombosis and Haemostasis 91(2):373–380 · 2004 · DOI: 10.1160/TH03-07-0423
Closed access — not in PMC. 135 citations as of 2026-06-14 (Europe PMC).
TL;DR
A 3-year, three-arm RCT in 181 postmenopausal women (108 analysed) found that a combined minerals + vitamin D + vitamin K1 (MDK) supplement preserved carotid artery wall elasticity over 3 years, while the minerals + vitamin D alone (MD) and placebo arms both showed deteriorating vessel-wall elastic properties. Between-group contrasts for MDK vs placebo: distensibility coefficient (DC) −8.8% (p<0.05), compliance coefficient (CC) −8.6% (p<0.05), pulse pressure (PP) +6.3% (p<0.05), Young’s modulus (E) +13.2% (p<0.01); IMT was not significantly different between groups. The MD (D alone) arm did not differ significantly from placebo on any measure. Three key caveats for wiki use: (1) the vitamin K component was K1 (phylloquinone), not K2/MK-7 — important when citing this as evidence for D3+K2 synergy; (2) vessel-wall elasticity is a functional surrogate endpoint, not a hard cardiovascular outcome; (3) only 108 of 181 enrolled women were analysed (~40% loss) — differential attrition could bias results. These percentage contrasts are abstract-confirmed (PMID 14961167); absolute within-arm values and doses require full-text access. gap/no-fulltext-access
Background and rationale
Vascular stiffness — declining vessel-wall elasticity — is a hallmark of arterial aging and an independent predictor of cardiovascular events. Matrix Gla protein (MGP) is a vitamin K-dependent inhibitor of vascular calcification whose expression is also regulated by vitamin D; the co-dependence of these two nutrient-dependent pathways motivates combined D + K supplementation as an intervention. This trial, conducted by the Vermeer group at Maastricht University (a leading centre for vitamin K vascular biology), was among the first RCTs to test combined D + K on a vascular functional endpoint.
Study design
| Feature | Detail |
|---|---|
| Design | Randomised, controlled, three-arm |
| Population | Postmenopausal women (age range not confirmed from abstract; supplement trial context implies typical postmenopausal range ~50–65 years) gap/needs-fulltext |
| n enrolled | 181 |
| n completed | 150 |
| n analysed | 108 (see Attrition note) |
| Arms | Arm 1: Placebo; Arm 2: Minerals + vitamin D (MD); Arm 3: Minerals + vitamin D + vitamin K1 (MDK) |
| Vitamin K form | Phylloquinone (K1) — dose not confirmed from available abstract gap/needs-fulltext |
| Vitamin D form/dose | Not confirmed from available abstract gap/needs-fulltext |
| Duration | 3 years |
| Primary endpoint | Carotid artery wall elastic properties: distensibility coefficient (DC), compliance coefficient (CC), intima-media thickness (IMT), Young’s modulus (E), pulse pressure (PP) |
| Measurement method | Non-invasive ultrasound (Hoeks AP, co-author, is a specialist in vascular ultrasound — vessel-wall tracking method) |
Attrition note. 181 enrolled → 150 completed → 108 analysed. The gap between completers (150) and analysed (108) is notable (28% further exclusion after completion). The reasons are not confirmed from the available abstract and should be verified against Table 1 or Methods in the primary source. Differential attrition or exclusion criteria could bias the per-protocol analysis. gap/no-fulltext-access
Results
Primary vascular elasticity outcomes
The abstract and structured metadata confirm the following directional findings 1:
| Parameter | MDK group (n≈36 analysed) | MD group | Placebo group | MDK vs placebo p-value | Confirmation level |
|---|---|---|---|---|---|
| Distensibility coefficient (DC) | Unchanged over 3 years | Decreased | Decreased | <0.05 (8.8% difference) | abstract-confirmed |
| Compliance coefficient (CC) | Unchanged over 3 years | Decreased | Decreased | <0.05 (8.6% difference) | abstract-confirmed |
| Pulse pressure (PP) | Unchanged over 3 years | Increased | Increased | <0.05 (6.3% difference) | abstract-confirmed |
| Young’s modulus (E) | Unchanged over 3 years | Increased* | Increased* | <0.01 (13.2% difference) | abstract-confirmed |
| Intima-media thickness (IMT) | No change | No change | No change | NS (no between-group differences) | abstract-confirmed |
*Higher Young’s modulus = stiffer vessel = worse outcome. Higher pulse pressure = worse arterial compliance.
The key finding is preservation, not absolute improvement: the MDK group maintained baseline elasticity over 3 years; control groups deteriorated. The “beneficial effect” in the title describes attenuated age-related decline, not a reversal of stiffness.
The percentage contrasts above (DC 8.8%, CC 8.6%, PP 6.3%, E 13.2%) are stated in the published abstract (Crossref JATS text, PMID 14961167) and are abstract-confirmed. Absolute within-arm values (DC, CC, PP, and E at baseline and year 3 per arm) are not confirmed from full-text tables — these require primary PDF access. gap/no-fulltext-access
MD-alone arm
Minerals + vitamin D without K1 did not significantly differ from placebo on elasticity outcomes — a finding consistent with the hypothesis that vitamin K is a necessary co-factor for MGP carboxylation to protect vessel walls, and that vitamin D alone is insufficient. This D-alone null result is the mechanistic anchor for the D+K synergy hypothesis. gap/needs-replication
Critical caveats for wiki cross-referencing
1. Vitamin K1 (phylloquinone), NOT K2/MK-7
The intervention used phylloquinone (K1), not the menaquinone-7 (MK-7) form that has become standard in vascular supplementation trials and is the subject of the wiki’s vitamin-d3-k2 and related pages. K1 and MK-7 differ in half-life (1–2 hours vs ~3 days), tissue distribution (K1 is preferentially retained by liver; K2/MK-7 reaches vascular and bone tissue more effectively), and potentially in carboxylation efficiency at extrahepatic sites including the vessel wall. This trial cannot be directly cited as evidence for K2/MK-7’s vascular benefits without acknowledging the form difference.
2. Surrogate, functional endpoint — not hard events
Carotid artery compliance and distensibility are validated surrogates for arterial aging and are predictive of cardiovascular events in population studies, but the trial was not powered to demonstrate reduction in myocardial infarction, stroke, or cardiovascular mortality. The clinical relevance of preserving these elasticity measures over 3 years in a supplement trial context requires extrapolation.
3. High attrition in the analysed sample
Only 108 of 181 enrolled participants were analysed — a loss of ~40% from enrolment to analysis. Whether this attrition is differential across arms, and the criteria for exclusion from analysis after completion, are load-bearing details that require primary-source verification.
4. Relative timing and independence from Shea 2009
This trial was published 5 years before shea-2009-vitamin-k-cac. Both test vitamin K1 on vascular endpoints but use different measures (arterial elasticity here vs CAC by CT in Shea). They are independent cohorts with different vascular outcomes — neither replicates the other directly.
| Dimension | Status | Notes |
|---|---|---|
| Pathway conserved in humans? | yes | Human RCT; vessel-wall elasticity endpoint directly measured |
| Phenotype conserved? | yes | Arterial stiffening is a human aging phenotype |
| Replicated in humans? | partial | MK-7 trials (Knapen 2015, VitaKats) show bone effects; dedicated arterial-elasticity MK-7 RCT replication needed |
Context and limitations
Strengths:
- Three-arm design isolates D alone vs D+K1 effect; the MD-null result is informative mechanistically
- 3-year duration; non-invasive validated endpoint methodology (Hoeks vascular ultrasound group)
- Published by the Vermeer group — authoritative in vitamin K vascular biology
- 135 citations; widely referenced in supplement literature
Limitations:
- High attrition (181 enrolled → 108 analysed; ~40%); reasons for exclusion of completers (150 → 108) not confirmed from abstract gap/no-fulltext-access
- Closed-access journal (Thrombosis and Haemostasis 2004); no PMC deposit; not open-access
- K1, not K2/MK-7 — limits extrapolation to the primary supplement form of current interest
- Doses of K1 and vitamin D not confirmed from available abstract gap/no-fulltext-access
- Surrogate endpoint; no hard cardiovascular outcome data
- All postmenopausal women; no male cohort, no pre-menopausal data
- Participant age range not stated in abstract gap/no-fulltext-access
- Absolute within-arm values for DC, CC, PP, E not confirmed without full text — only percentage contrasts and directionality are abstract-confirmed
Significance and wiki role
This study is the earliest RCT evidence for a vitamin D + vitamin K1 combination preserving vascular functional properties. Its finding that D alone was insufficient — while D+K1 preserved elasticity — is the founding experimental anchor for the D+K synergy hypothesis in vascular aging. The K1-not-K2 caveat must be preserved in all downstream citations. Combined with shea-2009-vitamin-k-cac, it forms the RCT foundation for the vitamin K vascular calcification evidence base, but both trials used K1 and neither directly tests the K2/MK-7 supplementation strategy most commonly discussed in current aging interventions.
Cross-references
| Page | Relationship |
|---|---|
| vascular-calcification | Primary process page for the vascular calcification biology this study tests |
| vitamin-k1 | Compound page for the K form tested here |
| vitamin-k2-mk7 | MK-7 compound page — distinct from K1; this trial’s results require form-caveat when cited in K2 context |
| vitamin-d | The D co-supplement; D-alone null result is the mechanistic separator |
| matrix-gla-protein | Proposed primary mechanistic mediator (vitamin K-dependent carboxylation) |
| shea-2009-vitamin-k-cac | Companion trial; K1 on CAC endpoint; read in tandem |
| chronic-inflammation | Arterial stiffening intersects with inflammaging mechanisms |
Gaps and open questions
gap/no-fulltext-access — Thrombosis and Haemostasis 2004 is closed-access; no PMC entry. The published abstract (Crossref JATS, PMID 14961167) confirms: n’s (181/150/108), three-arm design, 3-year duration, K1 form, all four percentage contrasts (DC 8.8%, CC 8.6%, PP 6.3%, E 13.2%) and their p-values, and the IMT null result. Remaining unconfirmed items requiring full-text access: absolute DC/CC/PP/E values per arm at baseline and year 3, participant age range, vitamin K1 and D doses, and reasons for the 150→108 exclusion after completion.
gap/needs-replication — This is the only three-arm D vs D+K1 RCT for arterial elasticity. A direct replication using MK-7 (K2) with the same endpoint and lower attrition is needed to support K2 vascular claims.
gap/dose-response-unclear — The K1 dose used in this trial is not confirmed from available metadata. Whether the effect is dose-dependent, and how it compares to the 500 µg/d dose in Shea 2009, is unknown without full-text access.
Footnotes
Footnotes
-
doi:10.1160/TH03-07-0423 · n=108 analysed (of 181 enrolled, 150 completed) · rct (three-arm: placebo / MD / MDK) · p<0.05 (DC 8.8%, CC 8.6%, PP 6.3%); p<0.01 (E 13.2%) for MDK vs placebo; MD vs placebo NS on all measures · model: postmenopausal women; 3-year follow-up; common carotid artery ultrasound · closed-access; percentage contrasts abstract-confirmed (PMID 14961167); absolute within-arm values and doses not confirmed from full text gap/no-fulltext-access ↩