Dietary Calcium, Vitamin D, and the Prevalence of Metabolic Syndrome in Middle-Aged and Older U.S. Women

Of 39,876 female health professionals aged ≥45 years in the Women’s Health Study (18), ∼98% of participants provided detailed information on their diet by completing a 131-item semiquantitative food frequency questionnaire (SFFQ), and 71% provided baseline blood samples collected in EDTA, which were stored in liquid nitrogen (19). To minimize the effects of postmenopausal hormone therapy on metabolic parameters, we excluded women who ever used postmenopausal hormones, leaving 10,066 women who were also free of diabetes at study entry and contributed complete data for all five components of the metabolic syndrome for this analysis.

Participants were asked how often on average they had consumed individual foods of commonly used portions during the previous year. Nutrient intakes were computed by multiplying the frequency of consumption of each unit of food from the SFFQ by the nutrient content of the specified portion size (20). The average daily intakes of individual dairy items were combined to compute dairy intake: low-fat dairy products, including skim or low-fat milk, sherbet, yogurt, and cottage/ricotta cheese; high-fat dairy foods, including whole milk, cream, sour cream, ice cream, cream cheese, and other cheese; and total dairy products, including all of the above. Total milk included skim or low-fat milk and whole milk.

Data on the use of individual and multivitamin supplements were used to assess intake of supplemental calcium and vitamin D. Total calcium and vitamin D intakes were calculated from both dietary and supplemental sources. In populations of nurses and health professionals, this SFFQ has demonstrated reasonably good validity as a measure of long-term average dietary intakes (20). Pearson’s correlation coefficients between responses from the SFFQ and those from four 1-week dietary records spaced over a year were 0.56 for total calcium and 0.51 for dietary calcium (20). With respect to milk intake, the correlation coefficients between the SFFQ and dietary records were 0.69 for skim or low-fat milk and 0.56 for whole milk (21).

The metabolic syndrome was defined using a modification of the criteria proposed by the Adult Treatment Program III of the National Cholesterol Education Program, as previously described (22). Women with three or more of the following conditions were defined as having the metabolic syndrome: 1) triglycerides ≥150 mg/dl, 2) HDL cholesterol <50 mg/dl, 3) blood pressure ≥135/85 mmHg, 4) obesity as defined by a BMI ≥30 kg/m², and 5) abnormal glucose metabolism as defined by incident type 2 diabetes. In the Women’s Health Study, however, waist circumference was not reported until year 6 of follow-up. To assess the robustness of our definition for obesity, we performed sensitivity analysis using a definition of the metabolic syndrome based on a waist circumference of 88 cm (>35 inch). Because fasting glucose levels were not available, we instead used the diagnosis of incident type 2 diabetes during an average of 8.8 years of follow-up as an alternative measure of baseline abnormal glucose metabolism. The validity of self-reported type 2 diabetes has been confirmed by a validation study, as previously reported (23).

Total, HDL, and LDL cholesterol and triglyceride levels were measured with direct measurement assays (Roche Diagnostics, Indianapolis, IN). All samples were identically handled and analyzed in random order to reduce systematic bias and interassay variation. Blinded quality control specimens were simultaneously analyzed with the study sample.

The age-adjusted (in 5-year groups) baseline characteristics according to quintiles of total calcium and vitamin D intake were compared. Proportions were tested with the stratified Mantel-Haenszel test, and mean values were compared using multiple linear regression. We applied a logistic regression model to examine the association between calcium and vitamin D intake and the risk of metabolic syndrome. The odds ratios (OR) and 95% CIs for the prevalence of the metabolic syndrome were calculated. The initial model was adjusted for age (continuous), total calorie intake (continuous), and randomized treatment. In multivariate models, we further adjusted for smoking status (current, past, and never), exercise (rarely/never, less than once per week, one to three times per week, and four or more times per week), alcohol intake (rarely/never, one to three drinks per month, one to six drinks per week, and one or more drinks per day), multivitamin use (never, past, and current), and parental history of myocardial infarction before age 60 years (yes/no). The second multivariable model added dietary factors, including intakes of total fat, cholesterol, protein, and glycemic load (all categorized as quintiles). In the final multivariable-adjusted model, we additionally adjusted for total calcium intake or total vitamin D intake. Tests of linear trend were conducted by assigning the medians of intakes in quintiles treated as a continuous variable. Furthermore, potential joint effects or effect modifications were evaluated by subgroup analyses stratified by the prespecified factors including age (>65 vs. ≤65 years) and smoking status (never, past, or current smokers). Likelihood ratio test was used to assess the significance of interaction. All statistical analyses were conducted using SAS (version 8.0; SAS Institute, Cary, NC). Statistical significance was set at P < 0.05, using two-sided tests.

There was an approximately three- to fivefold difference in median intake of calcium or vitamin D between the highest and lowest quintiles of the study population (514 vs. 1,583 mg/day for total calcium and 120 vs. 667 IU/day for total vitamin D). The median intake was 857 mg/day (87.9% from diet and 12.1% from supplements) for total calcium and 266 IU/day (83.6% from diet and 16.4% from supplements) for vitamin D in these middle-aged women.

As shown in Table 1, women with high calcium intake were slightly older and leaner and less likely to smoke, drink alcohol, or have a history of hypertension and were more likely to exercise and use multivitamins than those with lower calcium intake. Calcium intake was also positively associated with dietary protein, fiber, and glycemic load and inversely associated with dietary fat and cholesterol. Similar associations were found for vitamin D intake, except that vitamin D intake was positively associated with cholesterol intake.

Table 2 displays age-adjusted metabolic syndrome components according to quintiles of calcium and vitamin D intake at baseline. Overall, the prevalence of each of the five components was lower in women in the highest quintile of calcium and vitamin D intake than in those in the lowest quintile. All these associations reached statistical significance with several exceptions; hypertriglyceridemia was not associated with total intakes of calcium and vitamin and incident type 2 diabetes was not associated with dietary vitamin D intake. Similarly, there were inverse associations between calcium and vitamin D intakes and the prevalence of metabolic syndrome defined as the total number of components. Because the proportion of women with all five metabolic syndrome components was small, we may not have adequate statistical power to detect significant trends for both calcium and vitamin D intakes.

In the age- and total calorie–adjusted model, total calcium intake was significantly associated with the prevalence of metabolic syndrome (Table 3). Similar associations were observed for calcium intake from either diet or supplements; the multivariable ORs (model 1) for the highest relative to the lowest category were 0.69 (95% CI 0.58–0.82) (P for trend<0.0001) for dietary calcium and 0.84 (0.70–0.99) (P for trend = 0.005) for supplemental calcium. Further adjustment for dietary total fat, cholesterol, protein, and glycemic load (model 2) did not materially change these inverse associations. Neither did additional adjustment for total vitamin D (model 3). In contrast, a significant inverse association was consistently evident only for dietary vitamin D and the metabolic syndrome. However, this association appeared to be entirely explained by adding total calcium intake. In sensitivity analysis, we repeated statistical models using a waist circumference (>88 cm) and found similar results (data not shown).

Because intake of calcium was highly correlated to intake of vitamin D (correlation coefficient = 0.73), we further examined the joint association of dietary vitamin D and calcium together as a nine-category variable with risk of metabolic syndrome. An inverse association between dietary calcium intake and the metabolic syndrome was observed in women in each tertile of dietary vitamin D intake without significant interaction (P = 0.60) (data not shown). We also found no apparent modification of the relation between dietary vitamin D intake and the prevalence of metabolic syndrome by smoking status and age (data not shown).

To assess the internal consistency of our observations, we also examined the direct relations of major dairy products with the metabolic syndrome (Table 4). There were trends for lower prevalence of the metabolic syndrome associated with total diary products, high-fat dairy products, low-fat dairy products, and total milk intake.

In this large cohort of middle-aged and older U.S. women, we observed a significantly lower prevalence of the metabolic syndrome among those with higher calcium intake (both food and supplements). Similar inverse associations were also evident for intake of dairy products and the metabolic syndrome. There was a suggestion of an inverse association between vitamin D intake and metabolic syndrome, but this association was no longer apparent after adjustment for other major lifestyle risk factors including calcium intake.

Our findings are consistent with previous observations (1,2,5,24) of an inverse association between calcium intake and blood pressure and insulin sensitivity. Numerous epidemiologic studies have suggested that low calcium intake may be a risk factor for primary hypertension (25). Most trial data have shown that calcium supplementation led to a reduction in blood pressure among hypertensive patients (26). Recently, animal and human studies indicated that high calcium intake might decrease levels of parathyroid hormone and 1,25(OH) vitamin D and thus influence adipocyte metabolism by inhibiting lipogenesis and stimulating lipolysis (7). However, few studies have specifically examined the association between calcium intake and the metabolic syndrome, although it is plausible that a potential beneficial effect of calcium intake is partially mediated by its effects on blood pressure, insulin sensitivity, and body weight.

In contrast, our findings did not show an inverse association between vitamin D intake and the metabolic syndrome, although an inverse association between serum concentrations of vitamin D and the presence of metabolic syndrome has been observed in several studies (4,27). Vitamin D is an essential nutrient for calcium homeostasis, which is largely consumed as vitamin D2 (ergocalciferol) or D3 (cholecalciferol). Accumulating data indicate that people with impaired glucose tolerance (16,28) and diabetes (12,16,28) have lower concentrations of vitamin D compared with those with normal glucose tolerance. Also, low concentrations of vitamin D are associated with β-cell dysfunction and impaired insulin secretion and action (4,16). In a clinical study of 126 participants, those with hypovitaminosis D were nearly three times as likely to have the metabolic syndrome as participants with normal concentrations of vitamin D (30 vs. 11% of participants, P < 0.001) (4). More recently, we have also reported an inverse relationship between serum concentrations of vitamin D and the prevalence of the metabolic syndrome in the Third National Health and Nutrition Examination Survey of the U.S. population (27). Our null results for vitamin D intake may be attributed to inadequate measures of overall vitamin D intake due to the lack of information on sun exposure, because the synthesis of vitamin D3 in the skin induced by the ultraviolet radiation from the sun is an important source of vitamin D. To date, serum concentrations of 25-(OH) vitamin D are considered as a reliable measure for vitamin D status in the human body, but the correlation between vitamin D intake assessed by the SFFQ and 25-(OH) vitamin D concentrations appeared to be relatively low (r² = 0.35 for total vitamin D and 0.25 for dietary vitamin D) (29).

It has been reported that supplementation with vitamin D improves insulin secretion (6,30). In our study, however, supplemental vitamin D contributed a relatively small proportion of total vitamin D intake (<15%) and did not significantly affect the metabolic syndrome. Moreover, the biological effects of vitamin D may well depend on the presence or absence of other highly correlated nutrients such as calcium. Both vitamin D and calcium are metabolically related and thereby contribute to the pathogenesis of metabolic abnormalities through common or different mechanisms. Given the strong evidence from both animal studies and human studies, a potential role of dietary vitamin D in the pathogenesis of metabolic disorders cannot be completely ruled out at present.

The 10th Recommended Dietary Allowances of calcium is 800 mg/day for adults aged ≥25 years; the recommended daily allowances of vitamin D is 200 IU/day up to the age of 50 years, 400 IU/day for people from age 51 to 70 years, and 600 IU/day for people aged >70 years (31). The median intake of total vitamin D is slightly higher than the 10th Recommended Daily Allowances. Recently, the Food and Nutrition Board raised the adequate daily intake of calcium intake to 1,200 mg/day for women aged >50 years (32).

Our findings that dairy products were significantly associated with risk of the metabolic syndrome were consistent with those reported previously from the Coronary Artery Risk Development in Young Adults study, where dairy consumption was inversely related to the development of the metabolic syndrome (3). Nevertheless, many major components in such foods are highly correlated so that it is difficult to completely separate the independent effects. Further research needs to determine what compounds are responsible for the potential beneficial effects of dairy products on metabolic syndrome components

Several limitations should be kept in mind when interpreting these findings. First, dietary assessment is not perfect and its associated measurement error may have biased findings toward the null. Second, our cross-sectional design cannot demonstrate the temporal relationship between calcium intake and risk of metabolic syndrome. Third, the Women’s Health Study included mainly Caucasian women; our findings may not be directly generalized to the general U.S. population.

In conclusion, we found that high calcium intake and dairy product consumption are associated with lower prevalence of the metabolic syndrome in middle-aged and older women. There was no significant inverse association between vitamin D intake and metabolic syndrome. These findings warrant further examination in prospective studies including clinical trials.

This study was supported by grants DK66401, CA-47988, and HL-43851 from the National Institutes of Health, Bethesda, MD.

We are indebted to the 39,876 dedicated and committed participants of the Women’s Health Study. We acknowledge the contributions of the entire staff of the Women’s Health study.