点击显示 收起
1 From the Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, MA (LMT, PFJ, AHL, and SLB); the Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA (LMT, PFJ, AHL, ETK, and SLB); the Institute for Aging Research, Hebrew SeniorLife, Boston, MA (MTH and DPK); and BIDMC Division of Gerontology, Harvard Medical School, Boston, MA (MTH and DPK)
2 Any opinions, findings, conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the US Department of Agriculture. 3 Supported by the National Institute of Aging (AG14759 SLB); the National Institute of Arthritis and Musculoskeletal and Skin Diseases (AR/AG 41398 DPK); the National Heart, Lung, and Blood Institute's Framingham Heart Study (NO1-HC-25195); and the US Department of Agriculture under agreement 58-1950-001. 4 Reprints not available. Address correspondence to SL Booth, Jean Mayer USDA Human Nutrition Research Center, Tufts University, 711 Washington Street, Boston, MA, 02111. E-mail: sarah.booth{at}tufts.edu.
ABSTRACT
Background: Poor diet may affect bone status by displacing nutrients involved in bone health. Dihydrophylloquinone, a form of vitamin K present in foods made with partially hydrogenated fat, is a potential marker of a low-quality dietary pattern.
Objective: Our objective was to examine the cross-sectional associations between dihydrophylloquinone intake and bone mineral density (BMD) of the hip and spine in men and women.
Design: Dihydrophylloquinone intake was estimated with a food-frequency questionnaire, and BMD (in g/cm2) was measured by dual-energy X-ray absorptiometry in 2544 men and women (mean age: 58.5 y) who had participated in the Framingham Offspring Study. General linear models were used to examine the associations between dihydrophylloquinone intake (in tertiles: <15.5, 15.5–29.5, and >29.5 µg/d) and hip and spine BMD after adjustment for age, body mass index, energy intake, calcium intake, vitamin D intake, smoking status, physical activity score, and, for women, menopause status and estrogen use.
Results: Higher dihydrophylloquinone intakes were associated with lower mean BMD at the femoral neck [lowest-to-highest tertiles (95% CI): 0.934 (0.925, 0.942), 0.927 (0.919, 0.935), and 0.917 (0.908, 0.926), P for trend = 0.02], the trochanter [lowest-to-highest tertiles (95% CI): 0.811 (0.802, 0.820), 0.805 (0.797, 0.813), and 0.795 (0.786, 0.804), P for trend = 0.02], and the spine [lowest-to-highest tertiles (95% CI): 1.250 (1.236, 1.264), 1.243 (1.242, 1.229), and 1.227 (1.213, 1.242), P for trend = 0.03] in men and women after adjustment for the covariates. Further adjustment for markers of healthy and low-quality dietary patterns did not affect the observed associations.
Conclusions: Higher dihydrophylloquinone intakes are associated with lower BMD in men and women. This association remains significant after adjustment for other markers of diet quality.
Key Words: Dihydrophylloquinone • vitamin K • partially hydrogenated fat • trans fatty acids • diet patterns • bone mineral density • osteoporosis
INTRODUCTION
Evidence indicates that several foods, such as fruit and vegetables, and nutrients beyond calcium and vitamin D, such as phylloquinone (vitamin K1), positively influence bone status (1-4). Conversely, the components of a low-quality dietary pattern, such as sweet baked products and candy, have been shown to be associated with lower bone mineral density (BMD; 5), possibly by displacing foods that contain important nutrients involved in bone health.
Dihydrophylloquinone is formed when phylloquinone-rich plant oils are synthetically hydrogenated. Because much of the vegetable oil consumed in the United States is in the partially hydrogenated form, dihydrophylloquinone is a common source of vitamin K and contributes as much as 30% of total vitamin K intake (6). However, unlike phylloquinone, which is found in foods associated with a healthy dietary pattern, the major dietary sources of dihydrophylloquinone are commercially baked snack and fried foods (6, 7). Consequently, dihydrophylloquinone is a marker of snack food and fast food intakes, which are elements of a low-quality dietary pattern.
We hypothesized that higher intakes of dihydrophylloquinone, as a marker of a low-quality dietary pattern, are associated with lower BMD. To test this hypothesis, we examined cross-sectional associations between dihydrophylloquinone intake and BMD of the hip and spine in men and women.
SUBJECTS AND METHODS
Subjects
The Framingham Offspring Study is a longitudinal community-based study that began in 1971, with the primary goal of evaluating heart disease risk factors in 5124 adult offspring (and the offspring's spouses) of the original Framingham Heart Study cohort participants. Follow-up study visits occur every 3–4 y and include physical examinations, anthropometric measurements, biochemical assessments, questionnaires, and the evaluation of cardiovascular disease and other risk factors. For dietary information, participants were mailed a food-frequency questionnaire (FFQ) and were asked to bring the completed FFQ with them to the sixth (1995–1998) and seventh (1998–2001) follow-up examinations. The Framingham Osteoporosis Study, an ancillary study to the Framingham Offspring Study, conducted baseline BMD assessments during 1996–2001 to collect osteoporosis-related information in the Framingham Offspring Study cohort.
BMD measurements, valid FFQ data, and covariates were available for 2636 Framingham Osteoporosis Study participants. For purposes of the present analysis, we excluded 92 participants who reported using warfarin or other oral anticoagulants, which act as vitamin K antagonists, osteoporosis medications, or both at the same time as the BMD measurements were obtained. A total of 2544 subjects (1130 men and 1414 women) remained and were included in the analyses. The present study was approved by the institutional review boards of Hebrew Senior Life, Tufts–New England Medical Center, and Boston University School of Medicine.
Dietary assessment
Usual dietary intake over the previous year was reported with the use of the Willett FFQ. The FFQ consisted of a list of 126 food items with specified standard serving sizes and a selection of 9 frequency of consumption categories, which ranged from never or <1 serving/mo to 6 servings/d, as described elsewhere (8, 9). We excluded from the analyses those subjects who reported an energy intake <600 kcal/d (2.51 MJ/d) or >4000 kcal/d (16.74 MJ/d) for women or 4200 kcal/d (17.57 MJ/d) for men. Subjects who left 12 items blank on the FFQ were also excluded.
Daily dietary dihydrophylloquinone intake was calculated by multiplying the dihydrophylloquinone content per serving of each food (6, 10, 11) by the reported frequency of consumption from the FFQ and summing over all foods. The dihydrophylloquinone content of foods was based on laboratory analyses of geographically represented foods collected as part of the US Department of Agriculture National Food and Nutrient Analyses Plan, as described elsewhere (10). We previously reported that a higher dihydrophylloquinone intake was associated with higher plasma dihydrophylloquinone concentrations and with higher trans fatty acid intakes in the Framingham Offspring Study (12).
Various measures of diet quality were calculated. Baked, fried, and snack foods, which reflect a low-quality dietary pattern, were split into 2 categories: commercially baked, fried, and snack foods that contained dihydrophylloquinone (brownies, cakes, chips, cookies, crackers, doughnuts, French fries, muffins or biscuits, pies, popcorn, and sweet rolls) and snack foods that were unlikely to contain dihydrophylloquinone (candy and nuts). The number of servings of baked, fried, and snack foods per day was calculated by summing the number of servings per day of each of these food items from the FFQ. Fruit and vegetable intake, and magnesium and potassium intakes, which are found primarily in whole and unrefined foods, reflect a healthy dietary pattern. The number of servings of fruit and vegetables per day was calculated by summing the number of servings per day of canned, fresh, and frozen fruit and vegetables from the FFQ. Magnesium and potassium intakes were calculated by summing the contribution of magnesium or of potassium from each food item from the FFQ. In addition, we calculated a measure of overall diet quality, the 2005 Dietary Guidelines Adherence Index. The 2005 Dietary Guidelines Adherence Index, as described elsewhere (13), was developed to assess the adherence to key dietary intake recommendations of the US government's 2005 Dietary Guidelines for Americans, which describes a dietary pattern intended to promote health and to prevent chronic disease (14). The possible range of the 2005 Dietary Guidelines Adherence Index scores is a minimum of 0 to a maximum of 20.
Bone mineral density
BMD of the proximal right femur, which includes the femoral neck and the trochanter, and the lumbar spine (L2–L4) was measured by dual-energy X-ray absorptiometry and with standard positioning as recommended by the manufacturer (DPX-L; Lunar, Madison, WI). The right hip region was scanned unless there was a history of previous fracture or hip joint replacement, in which case the left hip was scanned. The CVs for the DPX-L measurements were 1.7% for the femoral neck, 2.5% for the trochanter, and 0.9% for the lumbar spine (15). Monthly measurements of a bone phantom throughout the period of BMD data collection showed no machine drift.
Covariate information
Potential factors that affect BMD status were identified, and information was collected either at the time of the BMD scans or at a date closest in time to both the Framingham and BMD examinations. The factors included age; sex; weight; height; total daily intakes of energy, calcium, and vitamin D (diet plus supplements); alcohol intake (number of drinks per day); current cigarette smoking status (smoked regularly in the past year: yes or no); and physical activity as estimated by the Physical Activity Scale for the Elderly (PASE). PASE is a 7-d record of household activities (eg, housework) and the number of hours spent on usual daily activities (eg, standing) and on light, moderate, and strenuous sports and recreation. PASE is a commonly used brief survey to assess physical activity in the elderly (16). For women, current hormone replacement therapy (HRT; reported use of oral conjugated estrogen, patch, or cream: yes or no) and menopause status (defined as no menstrual bleeding for 1 y, reported current HRT, or age >55 y) were also included as covariates. Body mass index (BMI; in kg/m2) was calculated from weight and height recorded at the closest follow-up examination to the BMD measurements.
Statistical methods
All the statistical analyses were performed with the Statistical Analysis System (version 9.1; SAS Institute Inc, Cary, NC). The analyses were conducted for men and women together because there was no significant interaction between sex and dihydrophylloquinone intake. Statistical significance refers to P < 0.05 unless otherwise noted.
All analyses used tertiles of dihydrophylloquinone intake (<15.5, 15.5–29.5, >29.5 µg/d), which were based on the intakes reported by the entire cohort. Least-squares means from the general linear models procedure (PROC GLM, version 9.1; SAS Institute Inc) were used to examine the descriptive comparisons by tertile of dihydrophylloquinone intake. The resulting least-squares means for each of the subject characteristics and dietary measures were compared across all pairwise combinations with the use of the Tukey-Kramer option to adjust for multiple comparisons in the GLM procedure.
Least-squares mean BMD at the femoral neck, trochanter, and lumbar spine (L2–L4) were calculated overall and by tertile of dihydrophylloquinone intake, with adjustment for the covariates age; BMI; smoking status; physical activity score; intakes of total energy, calcium and vitamin D (diet plus supplements); alcohol intake; and, for women, menopause status and current HRT. In addition, a test for linear trend by tertile of dihydrophylloquinone intake was performed by assigning the median value of dihydrophylloquinone intake for each tertile and by treating these median values as continuous variables, after adjustment for the covariates listed above. To assess the influence of calcium and phylloquinone intakes on the associations between BMD and dihydrophylloquinone intake, we ran separate GLM models with interaction terms of dihydrophylloquinone intake as a continuous measure and median calcium intake (800 or >800 mg/d), calcium supplement use (yes or no), and phylloquinone intake (diet plus supplements) as a continuous measure.
To corroborate our hypothesis of dihydrophylloquinone intake as a marker of a low-quality dietary pattern, we conducted additional analyses that controlled for markers of a healthy dietary pattern (number of fruit and vegetable servings, potassium intake, and magnesium intake separately) and of an overall healthy dietary pattern (as estimated by the 2005 Dietary Guidelines Adherence Index; 13) and for an independent marker of a low-quality dietary pattern (snack foods unlikely to contain dihydrophylloquinone) in GLM models.
RESULTS
Subject characteristics are summarized in Table 1 as overall means ± SEMs (unless otherwise noted) by tertile of dihydrophylloquinone intake. Men and women had a mean age of 58.5 y and a mean BMI in the overweight and obese classification (BMI > 25). Forty-one percent of the entire population was overweight (BMI: 25-30), and 28% was obese (BMI > 30). Eighty-one percent of the women were postmenopausal (n = 1139), 37% of whom were receiving HRT (n = 422) at the time these data were collected.
View this table:
TABLE 1. Subject characteristics and dietary intakes, overall and by tertile of dihydrophylloquinone intake1
Unadjusted values for age, vitamin D intake, and number of fruit and vegetable servings were not associated with dihydrophylloquinone intake across tertile categories. Total energy and unadjusted calcium intake and number of servings of baked, fried, and snack foods were positively associated with dihydrophylloquinone (Table 1). After adjustment for age and sex, the values for calcium intake and number of servings of baked, fried, and snack foods remained significantly different across tertile categories (data not shown).
A significant inverse association was found between dihydrophylloquinone intake and BMD at all bone sites for men and women (Table 2). No significant interactions were found between dihydrophylloquinone intake and calcium intake above or below 800 mg, calcium supplement use, or phylloquinone intake (data not shown).
View this table:
TABLE 2. Least-squares mean BMD measurements for general linear models (GLM), by tertile of dihydrophylloquinone intake for men and women
Further adjustment for other markers of diet quality, which include the 2005 Dietary Guidelines Adherence Index and the intake of snack foods unlikely to contain dihydrophylloquinone (candy and nuts), did not affect these associations, after adjustment for covariates (Table 2). Similar trends were observed after adjustment for the number of fruit and vegetable servings, potassium intake, and magnesium intake (data not shown).
DISCUSSION
The major finding of the present study was that higher dihydrophylloquinone intake was significantly associated with lower femoral neck, trochanter, and lumbar spine (L2–L4) BMD in men and women. This observation is consistent with a previous study that examined food groups, in which the participants who reported consuming sweet baked products, a primary source of dihydrophylloquinone, had lower BMD measures than did the individuals who had a healthy dietary pattern (5). However, in the current study, the number of fruit and vegetable servings was not lower with higher dihydrophylloquinone intakes, which suggests that baked, fried, and snack food intakes did not displace fruit and vegetable intakes.
Our findings do not support the hypothesis that the association between dihydrophylloquinone intake and BMD is due to dietary pattern. The significant associations between higher dihydrophylloquinone intake and lower BMD remained after additional adjustment for multiple markers of healthy and low-quality dietary patterns. Instead, the observed inverse associations between dihydrophylloquinone intake and BMD may indicate a biological phenomenon.
Vitamin K is a cofactor for vitamin K–dependent carboxylase, which converts certain glutamic acid residues to -carboxylated glutamic acid residues in specific proteins, such as osteocalcin, which is involved in bone metabolism (17). The active site for this -carboxylation reaction is conserved during the synthetic hydrogenation process that forms dihydrophylloquinone from phylloquinone-rich oils (18). However, the absorption and metabolism of dihydrophylloquinone are not well understood. Although dihydrophylloquinone is absorbed and supports -carboxylation of vitamin K–dependent coagulation proteins, dihydrophylloquinone does not appear to have the capacity for -carboxylation of osteocalcin equivalent to that of the parent form, phylloquinone (19). Our findings support these previous observations insofar as a higher dihydrophylloquinone intake was shown to be associated with lower BMD. However, there are currently no data on the adequacy of dihydrophylloquinone in terms of its role as a cofactor when there is an adequate amount of phylloquinone in the diet, as was observed in the current study, nor has the possibility of a potential detrimental effect of dihydrophylloquinone on phylloquinone been explored. In the current study, higher dihydrophylloquinone intake was significantly associated with lower BMD at all bone sites, regardless of phylloquinone intake, which challenges the hypothesis that there is a direct competition between dihydrophylloquinone and phylloquinone as a cofactor for the -carboxylase reaction, particularly because the intakes of dihydrophylloquinone were low relative to phylloquinone intakes. Alternatively, we may not have adequately captured the intakes of all the unhealthy foods that contribute to poor bone health that are otherwise being tracked by the dihydrophylloquinone intakes.
Consistent positive associations between diets high in trans fatty acids and coronary heart disease (20, 21) have led to a shift away from the use of partially hydrogenated fat in commercial products (22-24), which may have positive effects on bone health, regardless of the actual mechanism by which dihydrophylloquinone has a role.
One limitation of the present study is that our cross-sectional data can identify associations but cannot establish a direct biological effect. Furthermore, dihydrophylloquinone intake may be tracking components in commercially baked, fried, and snack foods that may be detrimental to bone health.
In summary, many components of both healthy and low-quality dietary patterns are related to bone health, but the observed association between dihydrophylloquinone intake and BMD appears to be independent of diet quality. The data from the present study suggest that dihydrophylloquinone intake may have a detrimental effect on bone status or may act as a marker of other dietary or lifestyle factors not captured in our assessment of dietary patterns.
ACKNOWLEDGMENTS
We are grateful to the Framingham Study participants and staff, especially the densitometer technician Mary Hogan and the Framingham Study laboratory chief Patrice Sutherland and her staff. We are also grateful to Gail Rogers for programming assistance and James Peterson for technical assistance.
The authors' responsibilities were as follows—LMT: designed and performed the statistical analyses and drafted the manuscript; PFJ, MTH, and DPK: contributed to the design of the analyses, the interpretation of the data, and the writing of the manuscript; AHL and ETK: contributed to the interpretation of the data and the writing of the manuscript; SLB: designed the analyses, interpreted the data, and contributed to the writing of the manuscript; DPK: directs the Framingham Osteoporosis Study; DPK, MTH, and SLB: collected and managed the data used in these analyses; and all authors: reviewed the final manuscript. None of the authors reported any conflicts of interest.
REFERENCES