Abstract
Background
Circulating biomarkers are often used to infer nutrient intakes and overall nutrient status, but may not accurately reflect nutrient concentrations in specific tissues. Since direct measurement of human tissue nutrient concentrations is rarely feasible, researchers rely on biomarkers as proxies. Studies linking diet, circulating biomarkers, and tissue nutrient concentrations are needed to enhance biomarker interpretability.
Objectives
We sought to evaluate the interrelationships among dietary intakes, circulating biomarkers, and human brain tissue stores of vitamin D and vitamin K, two nutrients implicated in cognitive health.
Methods
Brain 25-hydroxyvitamin D3 [25(OH)D3] and menaquinone-4, metabolites of vitamin D and vitamin K, were measured in 288 and 322 deceased participants, respectively, of the Rush Memory and Aging Project. Plasma 25(OH)D3, phylloquinone (vitamin K1), and uncarboxylated matrix Gla protein (a functional vitamin K status biomarker) were measured from samples collected 3.4 ± 1.9 y before death. Dietary intakes were derived from food frequency questionnaires (FFQs) obtained at the same visit, and in secondary analysis, the mean across ≤11 FFQs spanning ≤14 y before death was used.
Results
When based on the most recent FFQ, vitamin D or vitamin K intakes were not associated with plasma or brain concentrations of their respective metabolites. However, when averaged across all available FFQs, total vitamin D intake was positively associated with plasma and brain 25(OH)D3 concentrations (P ≤ 0.033). Phylloquinone intake was positively associated with plasma phylloquinone and inversely associated with plasma uncarboxylated matrix Gla protein (P ≤ 0.006). Phylloquinone intake was not associated with brain menaquinone-4 concentrations (P ≥ 0.567).
Conclusions
Repeated dietary assessments strengthened the associations for vitamin D, supporting the use of circulating 25(OH)D3 as an indicator of vitamin D intake and brain 25(OH)D3 concentrations. For vitamin K, the inter-relationship among dietary intakes, circulating biomarkers, and brain tissue stores is more complex, highlighting a need for further research to clarify its role and biomarkers in cognitive health.
Keywords: vitamin K, vitamin D, biomarkers, dietary intakes, brain, older adults
Introduction
The role of diet in mitigating cognitive decline and dementia has garnered a significant amount of research attention. However, translating this research into dietary guidance for dementia risk reduction is not straightforward due to various methodological and biological factors [1]. Dietary intake assessment carries inherent limitations [2], and dietary intake data alone capture only one aspect of the complex diet–disease relationship. Blood-based biomarkers, which are objectively measured, are commonly used to infer nutrient intakes and total-body nutrient stores. However, these biomarkers may not always reflect tissue stores or account for variations in tissue-specific nutrient utilization. Moreover, logistical and ethical constraints make the direct measurement of tissue nutrient stores in humans difficult. As a result, researchers often rely on circulating biomarkers as proxies. To enhance the utility and interpretability of these biomarkers, studies that can address the interrelationships among diet, circulating biomarkers, and tissue concentrations are essential.
We sought to address this gap, focusing on vitamin D and vitamin K, two nutrients that have been increasingly recognized for their potential roles in brain health. In rodent models, altering vitamin D intakes influenced cognition, behavior, as well as the development of Alzheimer’s disease and related neuropathology [[3], [4], [5]]. Recently, low dietary vitamin K intake was found to impair learning and memory-related cognitive function in aging mice [6]. In the Rush Memory and Aging Project (MAP), a population-based study of older adults, higher postmortem brain concentrations of 25-hydroxyvitamin D3 [25(OH)D3] and menaquinone-4 (MK4), metabolites of vitamin D and vitamin K, respectively, were associated with better cognitive function and less cognitive decline prior to death [7,8]. However, it is not known how diet is associated with these brain nutrient concentrations in humans.
Circulating 25(OH)D is the most widely used biomarker of vitamin D status [9]. For vitamin K status, multiple biomarkers are used, including circulating phylloquinone (vitamin K1) and carboxylation of a vitamin K-dependent protein, such as matrix Gla protein (MGP) [10]. Leveraging data available in the Rush MAP, we integrated dietary intakes, circulating biomarkers of vitamin D and vitamin K status, and their respective brain nutrient metabolites, to address the overarching question of how diet and circulating biomarkers of nutrient status are associated with brain nutrient stores. Because nutrient intakes often reflect overall dietary patterns, in addition to vitamin D and vitamin K intakes, we also considered Mediterranean, dietary approaches to stop hypertension (DASH), and Mediterranean-DASH intervention for neurodegenerative delay (MIND) diet adherence.
Methods
Rush MAP
Antemortem and postmortem measures were conducted in Rush MAP participants. MAP is an ongoing community-based longitudinal study with the overall objective of identifying risk factors for Alzheimer’s disease and related disorders and cognitive decline. MAP began in 1997 and continues to enroll males and females 65 y and older from northeastern Illinois [11,12]. At enrollment, MAP participants are free of dementia and agree to participate in detailed clinical evaluations annually and organ donation upon death. All participants signed an informed consent and the Anatomic Gift Act. The institutional review boards of Rush University Medical Center and Tufts University approved this study.
Dietary intakes
Beginning in 2004, MAP participants were invited to complete a food frequency questionnaire (FFQ) at the time of their annual examination [13]. The FFQ queried participants’ usual intake frequencies for >144 food items in the past year and has been validated in older community-dwelling residents [14].
To estimate vitamin D and vitamin K intakes from foods, the consumption frequency and portion size of each food reported on the FFQ were multiplied by the vitamin D and vitamin K (phylloquinone) content of that food based on the USDA Food Data Central database. Portion sizes were based on natural sizes (e.g., 1 egg) or according to age- and sex-specific serving sizes reported in national diet surveys [15]. Nutrient intakes were then averaged across the available FFQs for each participant. Energy intakes (kilocalories per day) were similarly estimated by multiplying the kilocalorie content of each food by the consumption frequency and portion size, then averaging across all FFQs available for each participant. Supplemental vitamin D intake was based on the use of vitamin D supplements and vitamin D–containing multivitamin supplements. Supplemental vitamin K intake was based on phylloquinone-containing multivitamins, since no participants reported taking supplements that contained menaquinones or that exclusively contained vitamin K.
The Mediterranean diet score includes 11 dietary components (fruits, vegetables, legumes, fish, red meat, poultry, non-refined grains, potatoes, full-fat dairy, olive oil, and alcoholic beverages), each scored 0–5, which are summed for a total score ranging from 0 to 55 [13,16]. The DASH diet adherence was based on the exercise and nutrition interventions for cardiovascular health trial in which 10 dietary components (whole grains, fruits, vegetables, nuts/seeds/legumes (combined), dairy, meat, total fat, saturated fat, sweets, and sodium) were each scored 0, 0.5, or 1 and summed for a total score ranging from 0 to 10 [13,17]. The MIND diet score ranges from 0 to 15, by summarizing 15 dietary components including 10 healthy food groups (green leafy vegetables, other vegetables, nuts, berries, beans, whole grains, fish, poultry, olive oil, and wine) and 5 unhealthy food groups (red meats, butter and stick margarine, cheese, pastries/sweets, and fried/fast food) [13]. For all 3 diet patterns, higher scores represent better adherence to the dietary pattern.
Plasma
Antemortem plasma 25(OH)D3 was measured using liquid chromatography-tandem mass spectrometry [Waters Acquity ultraperformance liquid chromatography with triple quadrupole mass spectrometer; coefficient of variation (CV): 6%] and National Institute of Standards and Technology-traceable standards for assay calibration at Tufts Medical Center [7]. Plasma phylloquinone was measured using HPLC at the USDA Human Nutrition Research Center on Aging at Tufts University. The laboratory participates in the vitamin K external quality assurance program (KEQAS). In >15 y of KEQAS participation, the laboratory consistently generates serum/plasma phylloquinone data within the acceptable range of expected values (analyses occur every 4 mo, for >30 cycles of verification). Low and high control specimens had average values of 1.1 nmol/L and 4.5 nmol/L, with interassay CVs of 8.1% and 7.7%, respectively [8]. Plasma dephospho-uncarboxylated MGP (ucMGP) was measured using a commercially available automated sandwich ELISA, which uses 2 anti-MGP monoclonal antibodies directed against dephosphorylated ucMGP (Immunodiagnostics Systems); CV: 4.0%‒4.5%. The lower limit of detection (LLOD) of the assay is 300 pmol/L. (Higher plasma ucMGP reflects lower vitamin K status.) Two fasted plasma samples (both stored at −80°C until analysis) were available for each participant and analyzed for 25(OH)D3, phylloquinone, and ucMGP. The last available sample, obtained on average 3.3 (SD = 2.1) y before death, was used for the primary analyses. The first available sample, obtained on average 7.7 (SD = 2.8) y before death, was included in sensitivity analyses.
Brain
The 25(OH)D3 and MK4, the primary forms of vitamin D and vitamin K detected in postmortem brain tissue in Rush MAP, were measured in 4 brain regions [mid-temporal cortex (MT), mid-frontal cortex (MF), cerebellum, and anterior watershed (AWS) white matter] using liquid chromatography-tandem mass spectrometry and HPLC, respectively, as described previously [18,19]. The LLOD for both analytes was 0.1 pmol/g. All samples were stored at ‒80○C protected from light exposure until the time of analysis. We previously found 25(OH)D3 in human brain tissue stored under these conditions is stable in samples stored for ≤6 y, and MK4 is stable in brain tissue samples stored for ≤8 y [19]. Therefore, participants with brain tissue stored >6 y were excluded from the brain 25(OH)D3 analyses, and participants with brain tissue stored >8 y were excluded from the brain MK4 analyses. This left 292 participants available for inclusion in the brain 25(OH)D3 analyses and 376 available for inclusion in the brain MK4 analyses.
Covariates
Sex was self-reported at the baseline visit. Age at death was documented at autopsy using dates of birth and death. Smoking history was categorized as never or former/current smoker, based on self-report at the baseline visit. Self-reported alcohol intake (grams per day) was derived from the FFQ administered at the baseline visit. BMI was computed from measured weight and height (weight in kilograms per height in meters squared) (averaged across visits). Apolipoprotein E (APOE) genotype was determined as described previously [20]. Triglycerides were measured as part of the standard lipid panel by Quest Diagnostics. Season of death was categorized based on the month of death, also documented at autopsy (December-February, March-May, June-August, September-November). For plasma 25(OH)D3 analyses, the season of blood draw was categorized similarly, based on the clinic visit date.
Statistical approach
Statistical analyses of the brain regions focused on 25(OH)D3 and MK4 since these were the main forms of vitamin D and vitamin K, respectively, in all human brain regions evaluated.
The associations of vitamin D and vitamin K intakes from diet and supplements, as well as the Mediterranean, DASH, and MIND diet scores, with brain 25(OH)D3 and MK4 concentrations, respectively, were determined using multiple linear regression. To satisfy linearity assumptions, brain MK4, plasma phylloquinone, and plasma ucMGP concentrations were natural log-transformed, and brain and plasma 25(OH)D3 concentrations were square-root transformed. Vitamin K intake from foods and total intake (from foods and supplements) were also natural log-transformed for the analysis, and vitamin D intake from supplements, foods, and total were square-root transformed. Covariates included age, sex, BMI, smoking history, alcohol use, energy intake, APOE4 genotype (present/absent), and use of lipid-lowering medication. Season of death was also included as a covariate in the analyses of brain 25(OH)D3. Associations of dietary and supplement use exposures with plasma phylloquinone, ucMGP, and 25(OH)D3 concentrations followed a similar approach, with season of blood draw included in the circulating 25(OH)D3 models and serum triglycerides included in the circulating phylloquinone and ucMGP models (because phylloquinone is transported on triglyceride-rich lipoproteins and was also correlated with plasma ucMGP). For the primary analyses, dietary exposures were based on the last available FFQ, and plasma 25(OH)D3, phylloquinone, and ucMGP were based on the last available plasma sample, which was obtained a mean (SD) of 3.3 (2.1) y before death.
To evaluate if the associations based on FFQs administered and/or plasma biomarkers measured at a single time point were consistent with those assessed repeatedly, we conducted the following sensitivity analyses: 1) the plasma 25(OH)D3, phylloquinone, and ucMGP concentrations from the first and last available plasma samples were averaged, 2) dietary intakes were averaged across all available FFQs for each participant, the mean (range) of which was 4 (1‒11) y, and 3) Dietary data from each participant’s first available FFQ were assessed separately. (None of the MAP participants with brain MK4 or plasma phylloquinone measures reported taking vitamin K-containing supplements in the first available FFQ, so vitamin K intake was based on diet alone for that analysis.) Plasma ucMGP values below the LLOD (300 pmol/L, n = 4) were replaced with 150 pmol/L (LLOD/2). All analyses were conducted using SAS (version 9.4), and statistical significance was set at P < 0.05.
Results
Vitamin D
Of the 292 participants available for analyses of brain 25(OH)D3, 4 were excluded due to missing covariate data, leaving 288 included in the analyses of diet and brain 25(OH)D3. Of these, 270 had plasma 25(OH)D3 measurements available and were included in the analyses of diet and plasma 25(OH)D3 (Supplemental Figure 1). Seventy-seven percent of the participants included in the brain 25(OH)D3 analyses were female. Their mean (SD) age at death was 92.4 (5.9) y. Their median (IQR) plasma 25(OH)D3 was 85.0 (52.5, 112.5) nmol/L, and vitamin D intake from dietary sources was 171 (126, 218) IU/d. Most participants (n = 230) reported taking vitamin D‒containing dietary supplements, which contributed 483 (333, 876) IU/d [median (IQR)] to their total vitamin D intake (Table 1).
TABLE 1.
Participant characteristics [baseline mean (SD) or n (percentage), unless indicated otherwise]
| Included in brain 25(OH)D3 analyses (n = 288) | Included in brain MK4 analyses (n = 322) | |
|---|---|---|
| Age at death, years | 92.4 (5.9) | 92.2 (6.0) |
| Female, n (%) | 221 (77) | 242 (75) |
| BMI, kg/m21 | 26.3 (4.7) | 26.4 (4.7) |
| Triglycerides, mg/dL2 | 117 (89, 159) | 119 (90, 157) |
| APOE4 allele, % | ||
| Present | 66 (23) | 71 (22) |
| Absent | 222 (77) | 251 (78) |
| Smoking history, % | ||
| Never | 179 (62) | 204 (63) |
| Current or former | 109 (38) | 118 (37) |
| Statin use (at ≥50% visits) n (%) | ||
| Yes | 141 (49) | 150 (47) |
| No | 147 (51) | 172 (53) |
| Alcohol intake, g/d | 5.1 (11.5) | 4.8 (11.7) |
| Energy intake, kcal/d3 | 1874 (627) | 1868 (645) |
| Vitamin D intake, IU/d | ||
| Total3,4 | 502 (174, 889) | 443 (174, 810) |
| From diet3,4 | 163 (103, 224) | 166 (106, 233) |
| From supplements3,4 (users only, n = 230) | 586 (400, 1000) | 500 (400, 1000) |
| Brain 25(OH)D3, pmol/g | ||
| Mean of MF and MT4 | 1.3 (0.8, 1.8) | 1.3 (0.9, 1.9) |
| AWS4 | 1.1 (0.7, 1.6) | 1.0 (0.6, 1.5) |
| CR4 | 1.3 (0.8, 1.9) | 1.3 (0.8, 1.9) |
| Season of death, % | ||
| Fall | 76 (26) | 91 (28) |
| Spring | 72 (25) | 81 (25) |
| Summer | 66 (23) | 64 (20) |
| Winter | 74 (26) | 86 (27) |
| Plasma 25(OH)D3, nmol/L2,4 | 85.0 (52.5, 112.5) | 85.0 (55.0, 113.8) |
| Season of blood draw, n (%) | ||
| Fall | 75 (35) | 82 (34) |
| Spring | 50 (24) | 56 (23) |
| Summer | 64 (30) | 75 (31) |
| Winter | 24 (11) | 27 (11) |
| Missing data (n) | 75 | 82 |
| Phylloquinone intake, mcg/d | 143 (86, 233) | 143 (94, 233) |
| Total3,4 | ||
| From diet3,4 | 139 (84, 217) | 139 (92, 224) |
| From supplements3 (users only, n = 41) | 30 (30, 43) | 30 (30, 43) |
| Brain MK4, pmol/g | ||
| Mean of MF and MT4 | 1.7 (1.0, 2.8) | 1.7 (1.0, 2.9) |
| AWS4 | 0.9 (0.4, 1.4) | 0.9 (0.4, 1.4) |
| CR4 | 1.6 (0.8, 2.8) | 1.6 (0.8, 3.0) |
| Plasma phylloquinone, nmol/L2,4 | 1.0 (0.7, 1.5) | 1.0 (0.6, 1.5) |
| Plasma ucMGP, pmol/L2,4 | 776 (597, 1323) | 733 (578, 1009) |
| Mediterranean diet score (0‒55)3 | 29.3 (5.1) | 28.9 (5.1) |
| DASH diet score (0‒9)3 | 3.6 (1.2) | 3.6 (1.1) |
| MIND diet score (0‒15)3 | 6.8 (1.5) | 6.7 (1.4) |
Abbreviations: APOE4, apoliprotein E4 genotype; AWS, anterior watershed; BMI, body mass index; CR, cerebellum; DASH, dietary approaches to stop hypertension; FFQ, food frequency questionnaire; IQR, interquartile range; MF, mid-frontal cortex; MIND, Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay; MK4, menaquinone-4; MT, mid-temporal cortex; SD, standard deviation; ucMGP, uncarboxylated matrix Gla protein; 25(OH)D3, 25-hydroxyvitamin D3.
Based on average across all visits.
Measured from the last available plasma sample.
Based on the last available FFQ.
Reported as median (IQR).
Vitamin D intake (based on the last available FFQ) was not associated with brain or plasma 25(OH)D3 (Table 2). Similarly, none of the dietary patterns analyzed were associated with brain or plasma 25(OH)D3. Plasma 25(OH)D3 and brain 25(OH)D3 were significantly positively correlated, as reported previously [8]. When dietary intakes were based on the average of all available FFQs, total vitamin D intake was significantly positively associated with plasma 25(OH)D3 concentrations and with the mean 25(OH)D3 concentrations of the MF and MT regions and the AWS 25(OH)D3 (Table 3). Vitamin D intake from diet alone was associated with plasma 25(OH)D3, whereas vitamin D intake from supplements alone was associated with brain 25(OH)D3. Of the 3 dietary pattern scores analyzed, only the MIND diet score was significantly associated (positively) with plasma 25(OH)D3 (Table 3). The results were generally similar whether the analysis was based on the last available sample or the average of the first and last available samples. The Mediterranean, DASH, and MIND diet scores were not significantly associated with the 25(OH)D3 concentrations in any brain region measured (Table 3). When dietary intakes were based on the first available FFQ, vitamin D intake (from diet, supplements, or total) was not significantly associated with brain 25(OH)D3 in any region, but total vitamin D intake was positively associated with plasma 25(OH)D3 (Supplemental Table 1). DASH diet adherence was positively associated with brain 25(OH)D3 in all regions measured based on the first available FFQ (Supplemental Table 1).
TABLE 2.
Associations of vitamin D intake and dietary pattern adherence (based on intakes reported using the last available food frequency questionnaires) with brain and plasma 25-hydroxyvitamin D3 in the Rush Memory and Aging Project
| n | Brain 25(OH)D31 |
Plasma 25(OH)D32 |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| MF and MT |
AWS |
CR |
||||||||
| β (SE)3 | P | β (SE)3 | P | β (SE)3 | P | n | β (SE) 4 | P | ||
| Total vitamin D intake from diet + supplements | 288 | 0.002 (0.002) | 0.265 | 0.002 (0.002) | 0.108 | 0.002 (0.002) | 0.136 | 270 | 0.007 (0.007) | 0.066 |
| Vitamin D intake from supplements only | 160 | ‒0.0005 (0.002) | 0.800 | 0.001 (0.002) | 0.612 | 0.0003 (0.002) | 0.900 | 149 | 0.009 (0.009) | 0.315 |
| Vitamin D intake from diet only | 288 | 0.001 (0.005) | 0.784 | ‒0.003 (0.005) | 0.618 | 0.002 (0.005) | 0.754 | 270 | 0.037 (0.023) | 0.112 |
| Mediterranean diet score | 288 | ‒0.001 (0.004) | 0.844 | 0.001 (0.004) | 0.816 | 0.0007 (0.005) | 0.881 | 270 | 0.003 (0.020) | 0.863 |
| DASH diet score | 288 | 0.007 (0.017) | 0.693 | 0.004 (0.017) | 0.801 | 0.009 (0.018) | 0.630 | 270 | 0.025 (0.079) | 0.754 |
| MIND diet score | 288 | ‒0.015 (0.013) | 0.258 | ‒0.006 (0.014) | 0.642 | ‒0.010 (0.014) | 0.479 | 270 | 0.062 (0.060) | 0.310 |
Abbreviations: AWS, anterior watershed; BMI, body mass index; CR, cerebellum; DASH, dietary approaches to stop hypertension; MF, mid-frontal cortex (average of); MIND, Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay; MT, mid-temporal cortex (average of); SE, standard error; 25(OH)D3, 25-hydroxyvitamin D3.
Covariates: age at death, season of death, sex, BMI, APOE4 status, smoking history, energy intake, alcohol intake, lipid-lowering medication use.
Based on the last available sample, covariates were the same as1, except the season of blood draw was included instead of the season of death.
Unstandardized β coefficients (SE).
TABLE 3.
Associations of vitamin D intake and dietary pattern adherence (based on the average intakes reported using all available food frequency questionnaires) with brain and plasma 25-hydroxyvitamin D3 in the Rush Memory and Aging Project
| n | Brain 25(OH)D31 |
Plasma 25(OH)D32 |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MF and MT |
AWS |
CR |
Last available sample |
Average of first and last available samples |
||||||||
| β (SE)3 | P | β (SE)3 | P | β (SE)3 | P | n | β (SE)3 | P | β (SE)3 | P | ||
| Total vitamin D intake from diet + supplements | 288 | 0.004 (0.002) | 0.026 | 0.004 (0.002) | 0.033 | 0.004 (0.002) | 0.058 | 270 | 0.047 (0.014) | 0.001 | 0.042 (0.011) | <0.001 |
| Vitamin D intake from supplements only | 230 | 0.005 (0.008) | 0.015 | 0.005 (0.002) | 0.019 | 0.005 (0.002) | 0.027 | 218 | 0.021 (0.015) | 0.178 | 0.015 (0.012) | 0.232 |
| Vitamin D intake from diet only | 288 | 0.004 (0.008) | 0.600 | ‒0.008 (0.008) | 0.326 | 0.0001 (0.008) | 0.993 | 270 | 0.128 (0.055) | 0.022 | 0.086 (0.044) | 0.055 |
| Mediterranean diet score | 288 | ‒0.002 (0.005) | 0.758 | 0.002 (0.005) | 0.653 | 0.001 (0.005) | 0.851 | 270 | 0.038 (0.038) | 0.313 | 0.026 (0.030) | 0.394 |
| DASH diet score | 288 | 0.028 (0.019) | 0.141 | 0.034 (0.020) | 0.079 | 0.039 (0.020) | 0.058 | 270 | 0.262 (0.026) | 0.065 | 0.158 (0.113) | 0.164 |
| MIND diet score | 288 | ‒0.007 (0.015) | 0.637 | ‒0.001 (0.016) | 0.945 | 0.001 (0.017) | 0.968 | 270 | 0.223 (0.112) | 0.048 | 0.179 (0.090) | 0.047 |
Abbreviations: AWS, anterior watershed; BMI, body mass index; CR, cerebellum; DASH, dietary approaches to stop hypertension; MF, mid-frontal cortex (average of); MIND, Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay; MT, mid-temporal cortex (average of); SE, standard error; 25(OH)D3, 25-hydroxyvitamin D3.
Covariates: age at death, season of death, sex, BMI, APOE4 status, smoking history, energy intake, alcohol intake, lipid-lowering medication use.
Covariates are the same as1, except the season of blood draw was included instead of the season of death.
Unstandardized β coefficients (SE).
Vitamin K
Of the 376 participants available for analysis of brain MK4, 5 were excluded due to missing pertinent covariate data. Since warfarin is a vitamin K antagonist and appears to influence brain MK4 concentrations [21], we further excluded 49 participants who reported taking warfarin at their last 2 clinic visits before death. This left 322 participants included in the analysis of diet with brain MK4. Of these, 295 had reliable measures of plasma phylloquinone, and 297 had reliable measures of plasma ucMGP and were included in the analyses of diet with those outcomes (Supplemental Figure 1). Seventy-five percent of the participants included in the brain MK4 analyses were female, and their mean (SD) age at death was 92.2 (6.0) y. Their median (IQR) plasma phylloquinone was 1.0 (0.6, 1.5) nmol/L, plasma ucMGP was 733 (578, 1009) pmol/L, and total phylloquinone intake was 155 (108, 248) mcg/d (Table 1). Only 41 participants reported taking dietary supplements that contained vitamin K. These supplements provided, on average (SD), 31 (15) mcg/d phylloquinone.
Vitamin K intake from the diet, based on the last available FFQ, was not associated with brain MK4 concentrations in any region measured or with plasma phylloquinone (Table 4). Vitamin K intake from supplements, but not diet, was positively associated with plasma ucMGP. None of the diet patterns analyzed were associated with brain MK4, plasma phylloquinone, or plasma ucMGP (Table 4). Plasma ucMGP was significantly inversely associated with brain MK4 concentrations in all regions measured: adjusted unstandard β MT and MF = ‒0.554 (P < 0.001), AWS = ‒0.576 (P < 0.001), cerebellum = ‒0.504 (P < 0.001). Plasma ucMGP was also inversely correlated with plasma phylloquinone (partial r = ‒0.18, adjusted for triglycerides, P = 0.002) (both measured from the last available plasma sample). As reported previously, plasma phylloquinone (measured from the last available sample) was not correlated with MK4 in any brain region [7]. When dietary intakes were based on the average of all available FFQs, dietary vitamin K intake and total vitamin K intake, as well as Mediterranean and MIND diet adherence, were positively associated with plasma phylloquinone (Table 5). Total and dietary phylloquinone intake were inversely associated with plasma ucMGP. However, none of the diet pattern scores analyzed were associated with plasma ucMGP. These results were similar whether the analysis was based on the last available sample or the average of the first and last available samples (Table 5). Phylloquinone intake from diet, supplements, or combined (based on the average of all available FFQs) was not associated with the MK4 concentrations in any brain region measured (Supplemental Table 2). None of the dietary patterns analyzed was associated with brain MK4 concentrations either (Supplemental Table 2).
TABLE 4.
Associations of vitamin K intake and dietary pattern adherence (based on intakes reported using the last available food frequency questionnaires) with brain menaquinone-4 and plasma vitamin K biomarkers in the Rush Memory and Aging Project
| n | Brain MK41 |
Plasma biomarkers2 |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MF and MT |
AWS |
CR |
Phylloquinone |
ucMGP |
|||||||||
| β (SE)3 | P | β (SE)3 | P | β (SE)3 | P | n | β (SE)3 | P | n | β (SE)3 | P | ||
| Phylloquinone intake from diet + supplements | 322 | ‒0.114 (0.010) | 0.240 | ‒0.100 (0.113) | 0.395 | ‒0.083 (0.095) | 0.384 | 295 | 0.104 (0.055) | 0.061 | 297 | ‒0.062 (0.042) | 0.151 |
| Phylloquinone intake from supplements only | 31 | ‒0.004 (0.010) | 0.732 | ‒0.009 (0.013) | 0.484 | ‒0.002 (0.011) | 0.823 | 30 | ‒0.009 (0.006) | 0.186 | 30 | 0.020 (0.007) | 0.011 |
| Phylloquinone intake from diet only | 322 | ‒0.121 (0.100) | 0.209 | ‒0.104 (0.113) | 0.355 | ‒0.079 (0.095) | 0.407 | 295 | 0.094 (0.055) | 0.089 | 297 | ‒0.059 (0.043) | 0.169 |
| Mediterranean diet score | 322 | 0.0003 (0.013) | 0.983 | ‒0.013 (0.015) | 0.379 | 0.008 (0.013) | 0.524 | 295 | 0.007 (0.007) | 0.348 | 297 | ‒0.011 (0.006) | 0.070 |
| DASH diet score | 322 | ‒0.024 (0.055) | 0.670 | ‒0.051 (0.065) | 0.429 | 0.024 (0.054) | 0.654 | 295 | ‒0.002 (0.033) | 0.951 | 297 | ‒0.033 (0.025) | 0.193 |
| MIND diet score | 322 | ‒0.029 (0.042) | 0.497 | ‒0.021 (0.050) | 0.674 | 0.028 (0.042) | 0.509 | 295 | 0.004 (0.024) | 0.864 | 297 | ‒0.142 (0.018) | 0.455 |
Abbreviations: AWS, anterior watershed; BMI, body mass index; CR, cerebellum; DASH, dietary approaches to stop hypertension; MF, mid-frontal cortex (average of); MIND, Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay; MK4, menaquinone-4; MT, mid-temporal cortex (average of); SE, standard error; ucMGP, uncarboxylated matrix Gla protein.
Covariates: age at death, sex, BMI, APOE4 status, smoking history, energy intake, alcohol intake, lipid-lowering medication use.
Based on the last available sample, covariates are the same as1 plus triglycerides.
Unstandardized β coefficients (SE).
TABLE 5.
Associations of vitamin K intake and dietary pattern adherence (based on the average intakes reported using all available food frequency questionnaires) with plasma phylloquinone and plasma uncarboxylated matrix Gla protein in the Rush Memory and Aging Project
| n | Plasma phylloquinone1 |
Plasma ucMGP1 |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Last available sample |
Average of first and last available samples |
Last available sample |
Average of first and last available samples |
|||||||
| β (SE)2 | P | β (SE)2 | P | n | β (SE)2 | P- | β (SE)2 | P | ||
| Phylloquinone intake from diet + supplements | 295 | 0.279 (0.062) | <0.001 | 0.238 (0.052) | <0.001 | 297 | ‒0.137 (0.049) | 0.006 | ‒0.174 (0.046) | <0.001 |
| Phylloquinone intake from supplements only | 40 | ‒0.009 (0.006) | 0.123 | ‒0.009 (0.005) | 0.093 | 40 | 0.012 (0.007) | 0.083 | 0.012 (0.006) | 0.051 |
| Phylloquinone intake from diet only | 295 | 0.273 (0.062) | <0.001 | 0.285 (0.052) | <0.001 | 297 | ‒0.136 (0.049) | 0.006 | ‒0.180 (0.046) | <0.001 |
| Mediterranean diet score | 295 | 0.023 (0.009) | 0.013 | 0.027 (0.008) | <0.001 | 297 | ‒0.009 (0.007) | 0.238 | ‒0.013 (0.007) | 0.068 |
| DASH diet score | 295 | 0.028 (0.037) | 0.460 | 0.067 (0.032) | 0.035 | 297 | ‒0.032 (0.029) | 0.269 | ‒0.037 (0.028) | 0.179 |
| MIND diet score | 295 | 0.063 (0.029) | 0.029 | 0.079 (0.024) | 0.001 | 297 | ‒0.021 (0.022) | 0.347 | ‒0.031 (0.021) | 0.132 |
Abbreviations: BMI, body mass index; DASH, dietary approaches to stop hypertension; MIND, Mediterranean-dietary approaches to stop hypertension intervention for neurodegenerative delay; SE, standard error; ucMGP, uncarboxylated matrix Gla protein.
Covariates: age at death, sex, BMI, APOE4 status, smoking history, energy intake, alcohol intake, lipid-lowering medication use, and triglycerides.
Unstandardized β coefficients (SE).
When dietary intakes were based on the first available FFQ, neither phylloquinone intake nor any diet score was significantly associated with brain MK4 in any region (Supplemental Table 3). Dietary phylloquinone intake as well as Mediterranean and MIND diet adherence were positively associated with plasma phylloquinone (Supplemental Table 3). Plasma ucMGP was not associated with phylloquinone intake or any diet pattern scores based on the first available FFQ (Supplemental Table 3).
Discussion
The interrelationships among diet, circulating nutritional biomarkers, and tissue nutrient stores are predicated on the assumption that nutrient intakes predict blood concentrations and tissue concentrations. This assumption is foundational to understanding how dietary intakes affect nutrient status and related physiological functions, but it is difficult to directly assess in humans. We leveraged data from the Rush MAP to evaluate these assumptions with respect to vitamin D and vitamin K, and brain tissue concentrations of these two nutrients implicated in cognitive health.
A key assumption in studies using nutrient biomarkers is that they reflect habitual dietary intakes. In many cohort studies, nutrient biomarkers are measured at a single point in time, under the assumption that a single measure captures long-term dietary exposure. Although we found overall, plasma 25(OH)D3, phylloquinone, and ucMGP reflect usual total and dietary vitamin D and vitamin K intakes, respectively, there were some inconsistencies when estimating supplement intake and/or relying on a single dietary assessment measure. When dietary intakes were based on the average obtained from ≤11 FFQs obtained over 14 y prior to death, total vitamin D and vitamin K intakes were significantly associated with their blood-based biomarkers. The results were generally similar when the plasma 25(OH)D3, phylloquinone, and ucMGP measurements were based on the average of two plasma samples (obtained a mean of 5.3 y apart), indicating that a single plasma sample sufficiently reflects longer-term total vitamin D and phylloquinone intakes in MAP. However, in our initial analyses, when dietary intakes were based on the last available FFQ obtained prior to death, neither vitamin D nor vitamin K intakes were associated with the plasma biomarkers of their respective nutrients. Unexpectedly, higher vitamin K intakes from supplements, but not diet, were associated with higher plasma ucMGP, an indicator of lower vitamin K status. However, given the small numbers of individuals reporting supplement use, we cannot discount the possibility that this finding is spurious. Total vitamin D intake from the first, but not the last, available FFQ, was positively associated with plasma 25(OH)D3. Similarly, the average dietary and total phylloquinone intakes were associated with both plasma phylloquinone and ucMGP, but phylloquinone intake from supplements was not. This was not surprising because supplements are not a key contributor to vitamin K intakes in the United States [22]. However, when dietary intakes were based on the last available FFQ only, neither plasma phylloquinone nor ucMGP was associated with dietary or total phylloquinone intake. These findings underscore the importance of basing dietary intakes on repeated dietary assessments as opposed to a single assessment, to better capture habitual intake and reduce measurement errors by averaging within-person variation.
Plasma 25(OH)D3 was significantly positively correlated with brain 25(OH)D3 concentrations in MAP [7], supporting the assumption that nutrient biomarkers measured in circulation reflect tissue nutrient stores. Higher average total vitamin D intake was also associated with higher brain 25(OH)D3 concentrations across all regions measured only when vitamin D intakes were based on the average of all available FFQs. This association appeared to be driven by supplement use because vitamin D intake from food sources alone was not associated with brain 25(OH)D3 in any region measured. These findings may indicate that higher doses of vitamin D than are typically obtained through the diet may be required to influence 25(OH)D3 concentrations in the brain. That vitamin D intakes derived from a single FFQ were not associated with brain 25(OH)D3 in any region may indicate brain 25(OH)D3 concentrations reflect longer-term intakes than are captured by a single FFQ, again highlighting the importance of obtaining repeated assessments of dietary intakes.
In contrast, plasma phylloquinone was not correlated with brain MK4 [8]. One contributing factor to the discrepant findings regarding vitamin K is that there are multiple forms of vitamin K. Phylloquinone is found in green leafy vegetables and vegetable oils, whereas most menaquinones are produced by bacteria, and some are found in fermented foods, including meats and cheeses [[23], [24], [25]]. In MAP, we found circulating phylloquinone reflected phylloquinone intake, as has been reported elsewhere [26,27]. Plasma ucMGP is a functional biomarker of vitamin K status that does not exclusively reflect phylloquinone or menaquinone intake [[28], [29], [30]]. Consistent with this, we found higher plasma ucMGP was significantly associated with lower brain MK4 concentrations and with lower circulating phylloquinone. In the brain (and other tissues), MK4 is converted from dietary phylloquinone and other vitamin K forms found in the diet. This was confirmed by a tracer experiment in which equimolar amounts of isotopically labeled phylloquinone, MK4, menaquinone-7, or menaquinone-9 were fed to mice for 1 wk. After the feeding period, MK4 was the only vitamin K form detected in brain tissue in all mice, despite the form fed in the diet. Moreover, over half of the brain MK4 was derived from the vitamin K form provided in the diet [31]. Despite preclinical evidence that dietary phylloquinone is a precursor to MK4 in brain tissue, we did not detect any significant associations between phylloquinone intake and brain MK4 in MAP participants. One possible explanation for this discrepancy is dose-related: rodents consumed a diet containing 2‒3 mg vitamin K per kilogram diet [31], whereas the median phylloquinone intake in MAP was 155 mcg/d (0.155 mg/d). A second explanation involves the regulation of phylloquinone conversion to MK4 in the brain, which has been demonstrated in rats [32]. For some nutrients, homeostatic regulation has been shown to influence the correlation between dietary intakes and brain nutrient concentrations [33], which may be the case for vitamin K.
The MIND diet was designed as a hybrid of the Mediterranean and DASH diets [13], and all 3 are generally similar healthy dietary patterns, including an emphasis on higher vegetable intakes. However, there are notable differences [1]. The MIND diet specifically encourages green leafy vegetables [1,13], which are rich in phylloquinone. The association of plasma phylloquinone with the Mediterranean and MIND diets indicates higher plasma phylloquinone is indicative of adhering to a diet high in green vegetables. However, plasma ucMGP does not appear to reflect adherence to a healthy diet. To the best of our knowledge, this is the first time the MIND diet is reported to be associated with circulating 25(OH)D3. Since the MIND diet also encourages fish and seafood, which are key dietary sources of vitamin D, this association is biologically plausible. However, caution is warranted in interpreting these findings since, although statistical significance was achieved with the MIND diet, the magnitude of association was similar for the DASH diet but did not achieve statistical significance. Moreover, none of the examined dietary patterns (including the MIND diet) were associated with brain 25(OH)D3 concentrations, suggesting overall diet quality may not significantly influence brain tissue 25(OH)D3 stores.
Important strengths of this study include the measurement of vitamin D and vitamin K metabolites in human brain tissue, biomarkers in plasma, as well as repeatedly administered FFQs to estimate dietary intakes ≤14 y prior to death. We acknowledge FFQs are based on self-report and carry inherent limitations [2]. We measured plasma 25(OH)D3, phylloquinone, and ucMGP from 2 separate samples obtained on average 5.3 y apart, and found the associations based on the average of the 2 measurements did not meaningfully differ from the associations based on the last sample alone, suggesting these nutrient biomarkers appear to be robust to short-term within-person variability in MAP. Sunlight exposure is known to influence circulating 25(OH)D3. Unfortunately, information on sunlight exposure or sunscreen use was not available in MAP. However, we adjusted our models for season to help mitigate potential confounding by seasonal sunlight exposure. We did not measure 25(OH)D2 (the hydroxylated derivative of vitamin D2) in brain or plasma. Vitamin D supplement use in MAP did not differentiate between vitamin D2 and D3 (although vitamin D3 elicits a greater increase in circulating 25(OH)D than vitamin D2 [34] and accounts for >90% of the vitamin D supplement market share in the United States [35]). Furthermore, on average, ∼10% of the total vitamin D intake in MAP was from vitamin D2, suggesting the observed associations were mostly driven by vitamin D3 exposure. We also acknowledge that, since MAP participants are nearly all White and 75% were females, the generalizability to other racial-ethnic groups and to males is uncertain. MAP participants’ mean age at death (92 y old) exceeds the average life expectancy in the United States by >10 y, so the generalizability of our findings to the broader aging population is also uncertain.
In summary, total vitamin D intakes are positively associated with blood and brain 25(OH)D3 concentrations, and blood and brain 25(OH)D3 concentrations are also positively correlated, supporting core assumptions underlying the inter-relationship among vitamin D intakes, circulating 25(OH)D3, and human brain tissue 25(OH)D3 stores. However, for vitamin K, the inter-relationship among nutrient intakes, circulating biomarkers, and brain tissue stores is not straightforward, highlighting the complexity of vitamin K metabolism and the need for more refined biomarkers to clarify vitamin K’s role in cognitive health.
Author contributions
The authors’ responsibilities were as follows – MKS, SLB: study conceptualization; PA, XF: data collection; JLR: data analysis; MKS, SLB: writing, original draft; JLR, BD-H, PA, XF: critical review and editing; and all authors: read and approved the final manuscript.
Data availability
The data described in the manuscript, code book, and analytic code will be made available upon request, pending application and approval.
Funding
This project was supported by the National Institute of Aging (NIA) (R01AG051641, R01AG017917, R01AG085483) and the USDA Agricultural Research Service Cooperative Agreement 58-8050-3-003. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the NIA or USDA.
Conflict of interest
The authors report no conflicts of interest.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.tjnut.2026.101428.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
References
- 1.Booth S.L., English L.K., Reigh N.A., Jacques P.F., Forester B.P., Shea M.K. Dietary patterns and brain aging: enthusiasm before evidence? Annu. Rev. Nutr. 2025;45(1):251–268. doi: 10.1146/annurev-nutr-013125-030429. [DOI] [PubMed] [Google Scholar]
- 2.Natarajan L., Rock C.L., Major J.M., Thomson C.A., Caan B.J., Flatt S.W., et al. On the importance of using multiple methods of dietary assessment. Epidemiology. 2004;15(6):738–745. doi: 10.1097/01.ede.0000135178.36362.ef. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ali A., Shah S.A., Zaman N., Uddin M.N., Khan W., Ali A., et al. Vitamin D exerts neuroprotection via SIRT1/nrf-2/NF-kB signaling pathways against D-galactose-induced memory impairment in adult mice. Neurochem. Int. 2021;142 doi: 10.1016/j.neuint.2020.104893. [DOI] [PubMed] [Google Scholar]
- 4.Fan Y.G., Pang Z.Q., Wu T.Y., Zhang Y.H., Xuan W.Q., Wang Z., et al. Vitamin D deficiency exacerbates Alzheimer-like pathologies by reducing antioxidant capacity. Free Radic. Biol. Med. 2020;161:139–149. doi: 10.1016/j.freeradbiomed.2020.10.007. [DOI] [PubMed] [Google Scholar]
- 5.Bellettini-Santos T., Garcez M.L., Mina F., Magnus N.Q., Pereira N.S., Marques A.O., et al. Vitamin D3 improves spatial memory and modulates cytokine levels in aged rats. Metab. Brain Dis. 2023;38(4):1155–1166. doi: 10.1007/s11011-022-01152-9. [DOI] [PubMed] [Google Scholar]
- 6.Zheng T., Marschall S., Weinberg J., Fu X., Tarr A., Shukitt-Hale B., et al. Low vitamin K intake impairs cognition, neurogenesis, and elevates neuroinflammation in C57BL/6 mice. J. Nutr. 2025;155(4):1119–1126. doi: 10.1016/j.tjnut.2025.01.023. [DOI] [PubMed] [Google Scholar]
- 7.Shea M.K., Barger K., Dawson-Hughes B., Leurgans S.E., Fu X., James B.D., et al. Brain vitamin D forms, cognitive decline, and neuropathology in community-dwelling older adults. Alzheimers Dement. 2023;19(6):2389–2396. doi: 10.1002/alz.12836. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Booth S.L., Shea M.K., Barger K., Leurgans S.E., James B.D., Holland T.M., et al. Association of vitamin K with cognitive decline and neuropathology in community-dwelling older persons. Alzheimers Dement (N Y). 2022;8(1) doi: 10.1002/trc2.12255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Institute of Medicine (US) Committee to Review Dietary Reference Intakes for Vitamin D and Calcium. Ross A.C., Taylor C.L., Yaktine A.L., Valle H.B.D. National Academies Press (US); Washington (DC): 2011. Dietary reference intakes for calcium and vitamin D. [PubMed] [Google Scholar]
- 10.Shea M.K., Booth S.L. Concepts and controversies in evaluating vitamin K status in population-based studies. Nutrients. 2016;8(1):8. doi: 10.3390/nu8010008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bennett D.A., Schneider J.A., Buchman A.S., Barnes L.L., Boyle P.A., Wilson R.S. Overview and findings from the rush Memory and Aging Project. Curr. Alzheimer Res. 2012. 2012;9(6):646–663. doi: 10.2174/156720512801322663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bennett D.A., Schneider J.A., Buchman A.S., de Mendes L.C., Bienias J.L., Wilson R.S. The Rush Memory and Aging Project: study design and baseline characteristics of the study cohort. Neuroepidemiology. 2005;25(4):163–175. doi: 10.1159/000087446. [DOI] [PubMed] [Google Scholar]
- 13.Morris M.C., Tangney C.C., Wang Y., Sacks F.M., Barnes L.L, Bennett D.A., et al. MIND diet slows cognitive decline with aging. Alzheimers Dement. 2015;11(9):1015–1022. doi: 10.1016/j.jalz.2015.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Morris M.C., Tangney C.C., Bienias J.L., Evans D.A., Wilson R.S. Validity and reproducibility of a food frequency questionnaire by cognition in an older biracial sample. Am. J. Epidemiol. 2003;158(12):1213–1217. doi: 10.1093/aje/kwg290. [DOI] [PubMed] [Google Scholar]
- 15.Holland T.M., Agarwal P., Wang Y., Dhana K., Leurgans S.E., Shea K., et al. Association of dietary intake of flavonols with changes in global cognition and several cognitive abilities. Neurology. 2023;100(7):e694–e702. doi: 10.1212/WNL.0000000000201541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Panagiotakos D.B., Pitsavos C., Arvaniti F., Stefanadis C. Adherence to the Mediterranean food pattern predicts the prevalence of hypertension, hypercholesterolemia, diabetes and obesity, among healthy adults; the accuracy of the MedDietScore. Prev. Med. 2007;44(4):335–340. doi: 10.1016/j.ypmed.2006.12.009. [DOI] [PubMed] [Google Scholar]
- 17.Epstein D.E., Sherwood A., Smith P.J., Craighead L., Caccia C., Lin P.H., et al. Determinants and consequences of adherence to the dietary approaches to stop hypertension diet in African-American and White adults with high blood pressure: results from the ENCORE trial. J. Acad. Nutr. Diet. 2012;112(11):1763–1773. doi: 10.1016/j.jand.2012.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Fu X., Dolnikowski G.G., Patterson W.B., Dawson-Hughes B., Zheng T., Morris M.C., et al. Determination of vitamin D and its metabolites in human brain using an ultra-pressure LC-tandem mass spectra method. Curr. Dev. Nutr. 2019;3(7) doi: 10.1093/cdn/nzz074. nzz074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Fu X., Shea M.K., Dolnikowski G.G., Patterson W.B., Dawson-Hughes B., Holland T.M., et al. Vitamin D and vitamin K concentrations in human brain tissue are influenced by freezer storage time: the memory and aging project. J. Nutr. 2021;151(1):104–108. doi: 10.1093/jn/nxaa336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Oveisgharan S., Buchman A.S., Yu L., Farfel J., Hachinski V., Gaiteri C., et al. APOE ε2ε4 genotype, incident AD and MCI, cognitive decline, and AD pathology in older adults. Neurology. 2018;90(24):e2127–e2134. doi: 10.1212/WNL.0000000000005677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Shea M.K., Israel E.A., Fu X., James B.D., Agarwal P., Schneider J.A., et al. Warfarin treatment is associated with lower postmortem brain vitamin K concentrations in older adults. J. Nutr. 2025;155(8):2711–2716. doi: 10.1016/j.tjnut.2025.04.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kantor E.D., Rehm C.D., Du M., White E., Giovannucci E.L. Trends in dietary supplement use among US adults from 1999-2012. JAMA. 2016;316(14):1464–1474. doi: 10.1001/jama.2016.14403. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Beulens J.W., Booth S.L., van den Heuvel E.G., Stoecklin E., Baka A., Vermeer C. The role of menaquinones (vitamin K2) in human health. Br. J. Nutr. 2013;110(8):1357–1368. doi: 10.1017/S0007114513001013. [DOI] [PubMed] [Google Scholar]
- 24.Fu X., Harshman S.G., Shen X., Haytowitz D.B., Karl J.P., Wolfe B.E., et al. Multiple vitamin K forms exist in dairy foods. Curr. Dev. Nutr. 2017;1(6) doi: 10.3945/cdn.117.000638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Fu X., Shen X., Finnan E.G., Haytowitz D.B., Booth S.L. Measurement of multiple vitamin K forms in processed and fresh-cut pork products in the U.S. Food supply. J. Agric. Food Chem. 2016;64(22):4531–4535. doi: 10.1021/acs.jafc.6b00938. [DOI] [PubMed] [Google Scholar]
- 26.Shea M.K., Barger K., Booth S.L., Matuszek G., Cushman M., Benjamin E.J., et al. Vitamin K status, cardiovascular disease, and all-cause mortality: a participant-level meta-analysis of 3 US cohorts. Am. J. Clin. Nutr. 2020;111(6):1170–1177. doi: 10.1093/ajcn/nqaa082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.McKeown N.M., Jacques P.F., Gundberg C.M., Peterson J.W., Tucker K.L., Kiel D.P., et al. Dietary and nondietary determinants of vitamin K biochemical measures in men and women. J. Nutr. 2002;132(6):1329–1334. doi: 10.1093/jn/132.6.1329. [DOI] [PubMed] [Google Scholar]
- 28.Knapen M.H., Braam L.A., Drummen N.E., Bekers O., Hoeks A.P., Vermeer C. Menaquinone-7 supplementation improves arterial stiffness in healthy postmenopausal women. A double-blind randomised clinical trial. Thromb. Haemost. 2015;113(5):1135–1144. doi: 10.1160/TH14-08-0675. [DOI] [PubMed] [Google Scholar]
- 29.Shea M.K., O’Donnell C.J., Vermeer C., Magdeleyns E.J., Crosier M.D., Gundberg C.M., et al. Circulating uncarboxylated matrix gla protein is associated with vitamin K nutritional status, but not coronary artery calcium, in older adults. J. Nutr. 2011;141(8):1529–1534. doi: 10.3945/jn.111.139634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Shea M.K., Berkner K.L., Ferland G., Fu X., Holden R.M., Booth S.L. Perspective: evidence before enthusiasm-A critical review of the potential cardiovascular benefits of vitamin K. Adv. Nutr. 2021;12(3):632–646. doi: 10.1093/advances/nmab004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ellis J.L., Fu X., Karl J.P., Hernandez C.J., Mason J.B., DeBose-Boyd R.A., et al. Multiple dietary vitamin K forms are converted to tissue Menaquinone-4 in mice. J. Nutr. 2022;152(4):981–993. doi: 10.1093/jn/nxab332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Al Rajabi A., Booth S.L., Peterson J.W., Choi S.W., Suttie J.W., Shea M.K., et al. Deuterium-labeled phylloquinone has tissue-specific conversion to menaquinone-4 among Fischer 344 male rats. J. Nutr. 2012;142(5):841–845. doi: 10.3945/jn.111.155804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Spector R., Johanson C.E. Vitamin transport and homeostasis in mammalian brain: focus on vitamins B and E. J. Neurochem. 2007;103(2):425–438. doi: 10.1111/j.1471-4159.2007.04773.x. [DOI] [PubMed] [Google Scholar]
- 34.Tripkovic L., Lambert H., Hart K., Smith C.P., Bucca G., Penson S., et al. Comparison of vitamin D2 and vitamin D3 supplementation in raising serum 25-hydroxyvitamin D status: a systematic review and meta-analysis. Am. J. Clin. Nutr. 2012;95(6):1357–1364. doi: 10.3945/ajcn.111.031070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Mordor Intelligence . 2025. United States vitamin D supplements market size and share analysis – growth trends and forecast (2026-2031)https://www.mordorintelligence.com/industry-reports/united-states-vitamin-d-supplement-market-growth-trends-and-forecast [Cited December 1, 2025] Available from: [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data described in the manuscript, code book, and analytic code will be made available upon request, pending application and approval.