点击显示 收起
1 From the Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston (NMM and PFJ); the General Medicine Division and Department of Medicine, Massachusetts General Hospital and Harvard Medical School, Boston (JBM); the Division of Preventive Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston (SL); and The Boston University School of Public Health, Department of Epidemiology and Biostatistics, and the National Heart, Lung, and Blood Institute’s Framingham Heart Study (PWFW).
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 a clinical research grant from the American Diabetes Association and a Junior Faculty Development Award from SmithKline Beecham and based on work supported in part by the US Department of Agriculture (agreement no. 58-1950-9-001); the National Institutes of Health, National Heart, Lung, and Blood Institute’s Framingham Heart Study (contract N01-HC-38038); and an NIH postdoctoral training grant in Human Nutrition and Metabolism (T32 DK 07651). 4 Reprints not available. Address correspondence to PF Jacques, Epidemiology Program, Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging, 711 Washington Street, Tufts University, Boston, MA 02111. E-mail: paul{at}hnrc.tufts.edu.
ABSTRACT
Background: The influence of whole grains on cardiovascular disease risk may be mediated through multiple pathways, eg, a reduction in blood lipids and blood pressure, an enhancement of insulin sensitivity, and an improvement in blood glucose control.
Objective: The objective was to examine the association between diets rich in whole- or refined-grain foods and several metabolic markers of disease risk in the Framingham Offspring Study cohort.
Design: Whole-grain intake and metabolic risk markers were assessed in a cross-sectional study of 2941 subjects.
Results: After adjustment for potential confounding factors, whole-grain intake was inversely associated with body mass index (
Key Words: Whole grains • refined grains • risk factors • survey • Framingham Offspring Study • food-frequency questionnaire • type 2 diabetes • cardiovascular disease
INTRODUCTION
There is substantial interest in the role of nutrients in chronic disease (1, 2) but comparatively little emphasis on the specific contributions of food groups, particularly whole-grain foods. Although current dietary guidelines (3) recommend the consumption of several daily servings of whole-grain foods, in the United States most grain products consumed are refined, and the average consumption of whole grains is only one serving per day (4). Refined-grain foods contain lower amounts of vitamin E, fiber, and magnesium than do whole-grain foods, but the enrichment of refined flour with thiamine, riboflavin, niacin, and iron and the recent fortification with folic acid have improved the nutritional value of refined grains (5, 6). However, in contrast with whole grains, there appears to be no evidence of a protective association between refined-grain intake and risk of heart disease (7) or type 2 diabetes (8, 9).
Several epidemiologic studies found that diets rich in whole grains may protect against cardiovascular disease (7, 10, 11), stroke (12), type 2 diabetes (8, 9), and certain cancers (13, 14). The protective effects of whole grains may depend on the presence or interaction of several biologically active constituents, including dietary fiber, vitamin E, magnesium, folate, and other nutrients and nonnutrients (15). Dietary fiber has been shown to decrease glucose, insulin, and serum lipid concentrations in both diabetic and nondiabetic persons (16, 17). Magnesium, a rich constituent of the grain germ, is associated with low insulin concentrations (18, 19) and a low incidence of type 2 diabetes (8, 20, 21). In epidemiologic studies, vitamin E, folate, and fiber have independently been associated with a reduced risk of coronary heart disease (1, 2, 22). However, diets rich in whole-grain foods have been associated with a reduction in the risk of coronary heart disease and type 2 diabetes, independent of the effects of select nutrients found in whole grains (7, 9, 10).
The influence of whole grains on cardiovascular disease risk may be mediated through multiple pathways, eg, a reduction in blood lipids (23, 24) and blood pressure (25, 26), an enhancement of insulin sensitivity, and an improvement in blood glucose control (27). To explore mechanisms whereby diets enriched in whole-grain foods may protect against cardiovascular disease and type 2 diabetes, we examined the association between whole-grain and refined-grain intakes on several metabolic markers of disease risk in the Framingham Offspring Study.
SUBJECTS AND METHODS
Subjects
The Framingham Offspring Study is a longitudinal, community-based study of cardiovascular disease among the offspring of the original participants of the Framingham Heart Study cohort and their spouses (28). In 1971, 5135 participants were enrolled in the study (29); since then, the cohort has been examined every 3–4 y. Between 1991 and 1995, during the fifth examination cycle of the Framingham Offspring Study, 3799 participants underwent a standardized medical history and physical examination. Valid data from food-frequency questionnaires (FFQs) were available for 3418 participants. Participants were excluded from these analyses if they had a diagnosis of diabetes based on the use of insulin or oral hypoglycemic medication (n = 126) or if they were taking cholesterol-lowering medication (n = 214). Participants with missing covariate information were also excluded, which reduced the final sample to 2941 (1338 men and 1603 women). The final sample was younger (
Usual dietary intake during the previous year was assessed during the fifth examination cycle with the use of a semiquantitative, 126-item FFQ (30). The questionnaires were mailed to the participants before the examination, and the participants were asked to bring the completed FFQ with them to their appointment. The FFQ consisted of a list of foods with a standard serving size and a selection of 9 frequency categories ranging from never or < 1 serving/mo to > 6 servings/d. Participants were asked to report their frequency of consumption of each food item during the previous year. Separate questions about the use of vitamin and mineral supplements and the type of breakfast cereal most commonly consumed were also included in the FFQ. Nutrient intakes were calculated by multiplying the frequency of consumption of each unit of food from the FFQ by the nutrient content of the specified portion. Dietary information was judged as unreliable and excluded from further analysis if reported energy intakes were < 2.51 MJ/d (600 kcal/d) or > 16.74 MJ/d (4000 kcal/d) for women and > 17.57 MJ/d (4200 kcal/d) for men or if 12 food items were left blank. The relative validity of the FFQ for both nutrients and foods was examined previously in several populations (30–32).
Breakfast cereal intake was subdivided into whole and refined grain based on the content of whole grain or bran of the cereal, as reported by Jacobs et al (7) and others (10). A breakfast cereal was considered whole grain if it contained 25% whole grain or bran by weight. Other whole-grain foods included dark bread, popcorn, cooked oatmeal, wheat germ, brown rice, and other grains (eg, bulgur, kasha, and couscous). Refined-grain foods included cold breakfast cereals (< 25% whole grain or bran), muffins, cakes, cookies, white bread and rolls, white rice, pancakes, waffles, pasta, and pizza. In the absence of information on brand names, breakfast cereals were classified as refined grain. Crackers could not be distinguished as whole- or refined-grain products and thus were classified as unspecified grains and were not included in the analyses.
Outcome measurements
Height, weight, waist, and hip circumferences were measured while the subjects were standing. BMI was calculated, and obesity was defined as a BMI 30 (33, 34). Fasting blood samples were drawn after the subjects had fasted overnight for the measurement of glucose, insulin, and lipid concentrations. Fasting insulin concentrations were measured in plasma as total immunoreactive insulin. Among patients without diagnosed diabetes, a 75-g oral-glucose-tolerance test was administered according to World Health Organization standards (35), and 2-h postchallenge glucose and insulin concentrations were measured. Previously undiagnosed diabetes was defined as a fasting plasma glucose concentration of 7.0 mmol/L or a 2-h postchallenge glucose concentration of 11.1 mmol/L (36). Glycated hemoglobin (Hb A1c) was measured as a marker of long-term glucose homeostasis (37). Serum lipid profiles included enzymatic measurement of total cholesterol and triacylglycerol concentrations (38) and the measurement of the HDL-cholesterol fraction after precipitation of LDL and VLDL cholesterol with dextran sulfate–magnesium (39). LDL-cholesterol concentrations were calculated with the use of the Friedewald equation (40) for individuals with triacylglycerol concentrations < 4.5 mmol/L (400 mg/dL). Blood pressure was measured twice after the participants sat for 5 min. Subjects were classified as hypertensive if both of the 2 measurements for diastolic or systolic blood pressure were > 90 mm Hg or > 140 mm Hg, respectively, or if use of antihypertensive medication was reported (41). Additional covariate information included age, smoking dose (0, 1–15, 16–25, or > 25 cigarettes/d), alcohol intake (g/d), current multivitamin use (yes or no), physical activity score (42), and current use of estrogen replacement therapy in postmenopausal women.
Statistical methods
SAS statistical software (release 8.0; SAS Institute, Cary, NC) was used for all statistical analyses. Dependent variables were waist-to-hip ratio (WHR), systolic and diastolic blood pressure, LDL cholesterol, and the natural logarithms of BMI, total cholesterol, HDL cholesterol, fasting glucose, 2-h glucose, fasting insulin, 2-h insulin, and triacylglycerols. To express transformed variables in their natural scale, geometric means were computed by exponentiation of adjusted least-squares means. In separate models, first-order interactions between sex and whole- and refined-grain intakes were entered to determine whether associations were similar between men and women. There were no significant (P < 0.05) interactions by sex on the association of whole grain and metabolic risk factors.
We determined age- and sex-adjusted and energy-adjusted geometric means for lifestyle and dietary characteristics across increasing quintiles of whole- and refined-grain intakes by using SAS PROC GLM. We assessed the statistical significance (defined as a two-tailed P value < 0.05) of trends across categories of grain consumption with linear (for continuous outcome variables) or logistic (for dichotomous outcome variables) regression models by using grain consumption (in servings/wk) as an ordinal variable with the median grain intake value in each category assigned as a score.
Multivariable models included sex, age (y), energy intake (kcal/d), multivitamin supplementation use (yes or no), alcohol intake (g/d), use of blood pressure medication (yes or no), current cigarette smoking (categorical), physical activity score (continuous), and current estrogen replacement therapy (yes or no) among women. Multivariate models were also used to evaluate the potential role of BMI as a mediator or confounder in the association between whole-grain intake and these metabolic risk factors. For dependent variables in which a positive association with whole-grain intake was observed, we further examined the possibility that dietary patterns associated with diets high in whole grains would explain these associations by adjusting for the percentage of polyunsaturated fatty acids and intakes of meat, fish, fruit, and vegetables. We also considered constituents found in whole grains that may be potential mediating nutrients. For these analyses, we further adjusted for intakes of dietary fiber, vitamin E, folate, magnesium, and vitamin B-6. We examined whether there was a significant interaction between BMI and whole grains on insulin concentrations. Insulin curves over BMI were smoothed by using the locally weighted regression scatter plot smoothing procedure (43). All correlation coefficients reported were calculated as Spearman rank-order correlation coefficients.
RESULTS
The reported mean weekly intakes of whole-grain foods were 8.3 and 8.8 servings/wk in men and women, respectively (Table 1). The reported median intake was lower, 6.0 and 6.5 servings/wk in men and women, respectively. The food items on the FFQ that contributed most to whole-grain intakes were dark bread, cold cereals, and popcorn. Both men and women reported higher intakes of refined grain than of whole grain: 22.0 and 18.5 servings/wk in men and women, respectively. Of the 2086 subjects who reported eating breakfast cereals, 63% (n = 1313) reported eating breakfast cereals that contained 50% whole grain or bran by weight, whereas 4% (n = 74) reported eating cereals that contained between 25% and 50% whole grain. A total of 33% (n = 699) reported eating refined-grain cereals, which included those cereals that contained a small amount of whole grain (1–24%).
View this table:
TABLE 1 . Distribution of grain foods for men and women in the Framingham Offspring Study
The age-, sex-, and energy-adjusted descriptive characteristics of the study population across quintile categories of whole-grain intake are shown in Table 2. The median intake of whole-grain foods ranged from 0.9 to 20.5 serving/wk, from the lowest to the highest quintile category of whole-grain intake. Compared with participants in the lower category, those in the upper category of whole-grain intake tended to be older, tended to have a lower BMI and WHR (the latter even after adjustment for BMI), were less likely to smoke (if they did smoke, they smoked fewer cigarettes on average), were more likely to take multivitamin supplements, and were more likely to take estrogen replacement therapy (women only). A higher intake of whole grains was associated with a healthier diet, with those in the upper category also consuming less saturated fatty acids, alcohol, and meat and more dietary fiber, polyunsaturated fatty acids, fruit, and vegetables. Favorable trends in lifestyle and dietary factors across quintile categories of whole-grain intake were significant for most factors. The intake of whole grains was positively associated with total intakes of dietary fiber (r = 0.54), magnesium (r = 0.51), folate (r = 0.45), vitamin E (r = 0.40), and vitamin B-6 (r = 0.45), which are important constituents of whole grains.
View this table:
TABLE 2 . Characteristics of the Framingham Offspring Study participants by quintiles of whole-grain intake1
The median intake of refined-grain foods ranged from 6.9 to 38.9 servings/wk across quintile categories (Table 3). Associations between lifestyle and dietary factors were opposite to associations observed for whole grains for multivitamin and current estrogen use; percentage intakes of monounsaturated and saturated fatty acids; and intakes of protein, folate, magnesium, fruit, vegetables, and fish.
View this table:
TABLE 3 . Characteristics of the Framingham Offspring Study participants by quintiles of refined-grain intake1
Multivariate-adjusted mean concentrations of metabolic risk factors across quintile categories of whole grain are shown in Table 4. After adjustment for potential confounding variables and dietary factors associated with diets high in whole grains, whole-grain intake was significantly inversely associated with WHR (mean difference: 0.1; P for trend < 0.005), LDL cholesterol (mean difference: 0.12 mmol/L; P for trend = 0.02), and fasting insulin (mean difference: 6 pmol/L; P for trend = 0.03). Marginally significant inverse associations remained between wholegrain intake and BMI (mean difference: 0.5; P for trend = 0.06) and HDL-cholesterol (mean difference: 0.3 mmol/L; P for trend = 0.16), total cholesterol (mean difference: 0.11 mmol/L; P for trend = 0.06), and triacylglycerol concentrations (mean difference: 0.11 mmol/L; P for trend = 0.07) after adjustment for covariates. The inverse associations between whole-grain intake, systolic blood pressure, diastolic blood pressure, fasting glucose, and 2-h insulin were attenuated after BMI was controlled for (model 2). There was no association between whole-grain intake and Hb A1c or 2-h postchallenge glucose concentrations.
View this table:
TABLE 4 . Metabolic risk factors by quintile of whole-grain intake1
To better understand the association between whole-grain intakes and risk factors related to nutrients found in whole grains, we individually adjusted metabolic risk factors that were significant in model 3 for potential mediating nutrients, including fiber, folate, magnesium, vitamin E, and vitamin B-6 (data not shown). The association between whole-grain intake and both WHR and LDL cholesterol was not attenuated after adjustment for these nutrients and the covariates in model 2. The association between whole-grain intake and fasting insulin (207 and 198 pmol/L for the highest compared with the lowest quintile category, respectively; P = 0.002) was attenuated after adjustment for intake of magnesium (206 and 202 pmol/L, respectively; P = 0.30) or dietary fiber (205 and 201 pmol/L, respectively; P = 0.16). When fiber was separated into soluble and insoluble fiber, the inverse association between fasting insulin and whole-grain intake was attenuated after insoluble, rather than soluble fiber, was included in the model.
Because insulin concentrations may mediate the association between whole-grain intake and total or regional adiposity, we further adjusted estimates of these characteristics for concentrations of fasting insulin. Adjustment for fasting insulin concentrations attenuated the association between BMI and whole-grain intake (BMI: 26.9 in lowest quintile category and 26.6 in the highest quintile category; P for trend = 0.38), but did not change the association between whole-grain intake and WHR. Fasting insulin did not attenuate the association between whole-grain intake and total or LDL cholesterol (data not shown). To further explore the association between BMI and insulin, an interaction term for BMI by whole-grain intake was included in the model 2. Because this interaction was significant (P = 0.02) for insulin concentrations, the relation between whole-grain intake and insulin was stratified by BMI, as shown in Figure 1. This inverse association between whole-grain intake and fasting insulin was much stronger for those with a higher BMI ( 30; P for trend = 0.003) than for those with a lower BMI (< 30; P for trend = 0.10).
FIGURE 1. . Association between whole-grain intake and fasting plasma insulin concentrations by BMI: BMI < 30 (- - - -; n = 2123) and BMI 30 (___; n = 624). Values were adjusted for age, sex, energy intake, treatment of hypertension, smoking, alcohol intake, multivitamin use, current estrogen replacement therapy, physical activity, and BMI. P for trend = 0.10 for BMI < 30 and 0.003 for BMI 30; P = 0.02 for the interaction between BMI and whole-grain intake.
We also examined the relation between mean concentrations of these metabolic risk factors and intakes of refined grains (data not shown). In contrast with whole-grain intakes, there were no significant trends across quintile categories of refined-grain intake for these risk factors, except for Hb A1c. After adjustment for potential confounders, mean Hb A1c (%) values in the lowest to highest intake categories of refined grains were 5.4, 5.4, 5.3, 5.4, and 5.3 (P for trend = 0.05).
DISCUSSION
In the current study, we examined associations between diets rich in whole-grain or refined-grain foods in relation to several metabolic risk factors for cardiovascular disease or type 2 diabetes. We found favorable associations between whole-grain intake and WHR and total cholesterol, LDL-cholesterol, and fasting insulin concentrations. In contrast, refined-grain intake was not related to concentrations of these metabolic risk factors.
Although the risk of both cardiovascular disease and type 2 diabetes is mediated in part by tissue sensitivity to the effects of insulin (44), relatively few epidemiologic studies have examined the association of whole-grain intake and insulin sensitivity (45). In the current study, fasting insulin concentrations were lower in those with a higher intake of whole-grain foods, after control for BMI and other potential confounders. The association between whole-grain intake and insulin was most striking among obese subjects, such that the highest fasting insulin concentrations were among those subjects with higher BMIs and lower intakes of whole grains. Other studies have shown that insulin resistance is higher among obese persons (46). A recent crossover study involving 11 overweight hyperinsulinemic subjects found that insulin sensitivity, as measured by the euglycemic hyperinsulinemic clamp, improved more after 6 wk of a whole-grain diet than after a refined-grain diet, independent of body weight (47). Other studies found that diets rich in whole-grains foods are associated with lower insulin concentrations (45, 48).
However, consistent with other findings (45), we found that the association between whole-grain intake and fasting insulin concentrations was attenuated after adjustment for dietary fiber and magnesium. This suggests that the apparent insulin-sensitizing effect of whole grains might be partially mediated by the effect of these nutrients.
Improved insulin sensitivity with elevated fiber intake may be one mechanism whereby diets rich in whole-grain foods may protect against the development of type 2 diabetes. Fiber has been shown to improve the glycemic response and circulating insulin concentrations both in healthy subjects and in those with type 2 diabetes (17, 49–51). Fiber intake is not only inversely associated with fasting insulin (52), but insoluble and cereal fiber intakes significantly reduce the risk of type 2 diabetes (8, 20, 21). The improved insulin sensitivity with high-fiber diets may occur because the gel-forming properties of soluble fibers delays the rate of carbohydrate absorption (53). However, in the current study, it appears that insoluble cereal fiber rather than soluble fiber was the predominant fiber having a favorable effect on fasting insulin. Magnesium is another component in whole grains that may improve insulin sensitivity. Intracellular magnesium has also been linked to insulin sensitivity in metabolic studies (54, 55), and clinical studies have shown that supplementation with magnesium improves insulin sensitivity (56, 57). Furthermore, an inverse association between dietary and serum magnesium and incidence of type 2 diabetes is supported by some epidemiologic data (8, 20, 58).
It is possible that the protective mechanism of whole grains can be attributed to yet additional components in whole grains. Liu et al (9) found that although fiber and magnesium intakes attenuate the association between whole-grain intake and the risk of type 2 diabetes, this observation was not entirely explained by either component. Hence, in addition to the previously discussed roles of nutrients, a low glycemic index and a greater particle size of most whole-grain products compared with refined-grain products (59) may have a beneficial effect on risk factors for cardiovascular disease (60).
Evidence from epidemiologic studies also suggests that refined-grain intake is not associated with fasting insulin concentrations (45) or an increased incidence of type 2 diabetes (8). On the other hand, a high intake of refined grain relative to whole grain has been related to an increased risk of type 2 diabetes (9). In the current study, refined-grain intake appeared to be unrelated to concentrations of fasting insulin; however, there was an unexplained weak inverse association with refined-grain intake on concentrations of Hb A1c.
Consistent with the results of earlier epidemiologic studies (7, 45), we found that whole-grain but not refined-grain intake was inversely association with body weight and fat distribution. The inherent high-fiber content of most whole-grain foods may prevent weight gain or promote weight loss (52, 61, 62). Metabolic studies suggest that high-fiber diets help control appetite by providing a longer feeling of satiety (63, 64). After adjustment for other factors associated with diets rich in whole grains, such as fruit and vegetable consumption, the inverse association between whole-grain intake and BMI was only marginally significant, perhaps suggesting that the association with BMI may have been attributable in part to an overall healthier lifestyle. However, the association between whole-grain intake and BMI was substantially attenuated after adjustment for insulin, which may suggest that insulin is a potential mediator in the causal pathway.
It is well documented that a reduction in total and LDL cholesterol and an increase in HDL cholesterol is associated with a decrease in the risk of future coronary events (65–68). Randomized clinical trials (24) and metabolic studies (23) have shown that oats and oat bran reduce total blood cholesterol. Moreover, rye bread, which is considered a whole-grain food, has been shown to reduce total cholesterol concentrations in moderately hypercholesterolemic men, whereas refined wheat bread had no favorable effect on cholesterol concentrations (69). In the current study, total cholesterol and LDL concentrations were lower in the highest quintile category of whole-grain intake. Although the association between total cholesterol and whole-grain intake was largely attributed to dietary fiber and other dietary factors associated with a healthier lifestyle, whole-grain intake remained inversely association with LDL cholesterol, independent of both dietary fiber and saturated fat intakes.
Methodologic issues in our study may have accounted for the presence or absence of associations between grain intake and intermediate metabolic markers. Because of the fixed food categories associated with the FFQ, it is difficult to separate whole- and refined-grain foods accurately from some foods. For example, dark breads such as pumpernickel or wheat bread may include breads made with refined-grain flour. Yet, despite this potential measurement error in exposure, which would tend to attenuate associations, we found significant associations between whole-grain intake and concentrations of several biomarkers of disease risk. Additionally, diets rich in whole-grain foods appear to reflect an overall healthier lifestyle that may not have been accurately captured and controlled in our analysis, resulting in residual confounding. Nevertheless, our data suggest that whole-grain diets are associated with favorable effects on several metabolic risk factors. Thus, an improvement in the global metabolic milieu may be one mechanism whereby whole-grain intake may reduce the risk of type 2 diabetes and cardiovascular disease.
REFERENCES