AJCN Tufts Nutrition Symposium, Boston Sept 24-26
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Koh-Banerjee, P.
Right arrow Articles by Rimm, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Koh-Banerjee, P.
Right arrow Articles by Rimm, E.
Agricola
Right arrow Articles by Koh-Banerjee, P.
Right arrow Articles by Rimm, E.
American Journal of Clinical Nutrition, Vol. 80, No. 5, 1237-1245, November 2004
© 2004 American Society for Clinical Nutrition

ORIGINAL RESEARCH COMMUNICATION

Changes in whole-grain, bran, and cereal fiber consumption in relation to 8-y weight gain among men1,2,3

Pauline Koh-Banerjee, Mary Franz, Laura Sampson, Simin Liu, David R Jacobs, Jr, Donna Spiegelman, Walter Willett and Eric Rimm

1 From the Departments of Nutrition (PK-B, MF, LS, WW, and ER), Epidemiology (SL, DS, WW, and ER), and Biostatistics (DS), Harvard School of Public Health, Boston; the Department of Preventive Medicine, The University of Tennessee Health Science Center, Memphis (PK-B); the Division of Epidemiology, University of Minnesota, Minneapolis (DRJ); Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital, and Harvard Medical School, Boston (SL, WW, and ER); and the Institute for Nutrition Research, University of Oslo, Norway (DRJ)

2 Supported by research grants CA 55075 and HL 35464 and an unrestricted grant from the Kellogg Company.

3 Reprints not available. Address correspondence to P Koh-Banerjee, Department of Preventive Medicine, The University of Tennessee Health Science Center, 66 North Pauline Street, Suite 608, Memphis, TN 38163. E-mail: pkohbane{at}utmem.edu.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 
Background: Epidemiologic studies that directly examine changes in whole-grain consumption in relation to weight gain are sparse, and characterization of this association has been obscured by methodologic inconsistencies in the assessment of whole grains.

Objective: We aimed to ascertain the associations between changes in new quantitative estimates of whole-grain intake and 8-y weight gain among US men.

Design: The study was conducted in a prospective cohort of 27 082 men aged 40–75 y at baseline in 1986. Data on lifestyle factors were obtained periodically by using self-reported questionnaires, and participants measured and reported their body weight in 1986 and 1994.

Results: In multivariate analyses, an increase in whole-grain intake was inversely associated with long-term weight gain (P for trend < 0.0001). A dose-response relation was observed, and for every 40-g/d increment in whole-grain intake from all foods, weight gain was reduced by 0.49 kg. Bran that was added to the diet or obtained from fortified-grain foods further reduced the risk of weight gain (P for trend = 0.01), and, for every 20 g/d increase in intake, weight gain was reduced by 0.36 kg. Changes in cereal and fruit fiber were inversely related to weight gain. No associations were observed between changes in refined-grain or added germ consumption and body weight.

Conclusions: The increased consumption of whole grains was inversely related to weight gain, and the associations persisted after changes in added bran or fiber intakes were accounted for. This suggests that additional components in whole grains may contribute to favorable metabolic alterations that may reduce long-term weight gain.

Key Words: Whole grains • cereal fiber • bran • obesity • weight gain


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 
Epidemiologic evidence suggests an inverse relation between the consumption of whole-grain foods and body weight (1-4) and serum biomarkers of obesity, including insulin, C-peptide, and leptin concentrations (1-5). However, investigation of whole grains is relatively nascent, and characterization of the relation between whole-grain consumption and weight has been clouded by several methodologic issues. Currently, there is no uniform definition of a whole-grain food (6, 7), and the term "whole grain" has been equally applied to foods containing the intact grain and to foods that contain the appropriate proportions of all the milled grain constituents (ie, bran, endosperm, and germ) (8). Jacobs et al (3) proposed a definition of whole-grain foods as those with ≥25% whole-grain or bran content by weight. Several studies have adopted this definition and semiquantitatively assessed the number of servings of whole grain–rich foods. However, as a limitation to this method, the actual whole-grain content may vary substantially per serving of food deemed high in whole grains, which results in potential misclassification and residual bias in the estimation of whole-grain effects. This issue is further compounded by the absence of standard serving sizes in various dietary questionnaires across studies (9).

To provide more standardized nomenclature and to increase consumer awareness of high-whole-grain foods, the FDA approved a whole-grain health claim (10) for foods with ≥51% whole-grain content (by weight) per reference amount customarily consumed (RACC; 11). RACC is a measurement established by the FDA for policy purposes: the amount of foods customarily consumed per eating occasion for each category on the list is based primarily on national food-consumption surveys conducted by the US Department of Agriculture. With regard to whole-grain content, foods must contain all portions of the kernel and provide a minimum of 16 g whole grain/RACC. Compliance with the definition is determined by reference to the fiber content of whole wheat (10). However, data to support these regulatory requirements are lacking, in part because of limitations in existing whole-grain databases. To date, no studies have apportioned foods into their respective whole-grain and non-whole-grain ingredients and calculated the percentage of whole-grain concentration across foods. Further, estimates of whole-grain concentrations in cereals obtained by using the method of Jacobs et al did not separate out the bran content added during processing to increase the dietary fiber or micronutrient composition (or both). In this regard, it will be of interest to ascertain whether the associations between whole-grain foods and body weight are primarily driven by bran, some other dietary component of whole-grain foods, or another, related aspect of diet. Whether added germ is associated with body weight independent of whole grains is unknown, and no prior studies showed whether the types or food sources of fiber are differentially related to changes in body weight.

Therefore, we prospectively examined the associations between changes in whole-grain intakes with long-term weight gain among a cohort of 27 082 men by using new quantitative estimates of grams of consumption from foods that meet various classification criteria. We further examined the associations of added bran, germ, and the various types of fiber with weight gain.


   SUBJECTS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 
Study population
The Health Professionals Follow-up Study (HPFS) is a prospective investigation of 51 529 male health professionals who were 40–75 y old at baseline in 1986. This cohort includes 29 683 dentists, 10 098 veterinarians, 4185 pharmacists, 3745 optometrists, 2218 osteopathic physicians, and 1600 podiatrists. In 1986, participants completed a detailed questionnaire regarding medical history, dietary intake, and physical activity. The participants reported their age, current height (in inches), weight (in pounds), current and past smoking history, marital status, and medical history. On a biennial basis thereafter, participants were followed with mailed questionnaires to update information on exposures and to ascertain newly diagnosed disease.

We excluded from the analysis men who died or developed heart disease, stroke, cancer, or diabetes before 1994 (n = 13 169) because the development of these diseases may alter weight measures, dietary intake, and physical activity. Furthermore, we excluded men who failed to report their body weight or dietary data (n = 11 278) for the study period. Our analysis is therefore based on 27 082 healthy men for whom we have a complete set of predictor and outcome information for the study period of 1986 to 1994. The Institutional Review Board of the Harvard School of Public Health approved the protocol for this study.

Outcome assessment
The outcome was defined as the difference between the body weights reported by the subjects on the 1986 and 1994 questionnaires. Previously, we evaluated the reproducibility and validity of the self-reported measures of weight by using a subset of the cohort on whom technician-assessed measurements were taken 6 mo apart (12). The validity of self-reported height was not evaluated because that measurement was previously reported as highly valid (13). Self-reported weight and the average of 2 technician measures of weight were highly correlated at 0.97, and there were no significant linear trends in accuracy of reported weight across quartiles of age or body mass index (12).

Dietary exposures
Detailed dietary information was obtained in 1986, 1990, and 1994 through the use of a semi-quantitative food-frequency questionnaire (FFQ) described previously (14, 15). The FFQ was used to assess typical food intake over the previous year and included questions regarding the consumption of grain foods such as cooked and cold breakfast cereals, white and dark bread, white and brown rice, and pasta. The men were asked to specify the brand and type of cold breakfast cereal usually eaten. For each food, a commonly used unit or portion size was specified, and these portions were converted to gram weights per serving. The participant was asked how often, on average, he had consumed the specified amount during the previous year; 9 responses were possible, ranging from "never" to "≥6 times/d." The FFQ included open-ended questions regarding the usual serving size and frequency of consumption of foods not listed on the FFQ.

Using time-specific product labels and US Department of Agriculture national nutrient database information, we partitioned all foods on the FFQ into their whole-grain and non-whole-grain ingredients, consistent with the classification developed by the US Department of Agriculture (16), and we determined the percentage of whole-grain concentration per food serving on a dry weight basis. Complete nutrient and ingredient data were available from breakfast cereal manufacturers for all 3 time periods, which allowed the successful characterization of >250 brand-name breakfast cereals. Manufacturers’ product information was also used to develop whole-grain profiles for commercially prepared foods such as bran muffins, and cookbooks (eg, The Joy of Cooking) were used to create recipes for home-prepared bakery items.

For the survey question on "dark bread," we used a composite of recipes to represent commercial wheat (not whole-wheat) and wheatberry breads, which were selected on the basis of their proportion of marketplace shelf space during those periods. The nutrient profile was derived from US Department of Agriculture nutrient data and labels from 4 commercial wheat breads, including Home Pride Honey Wheat Berry Bread (Interstate Bakeries Corp, Kansas City, MO), Albertson’s Wheat Bread (Albertson’s Inc, Boise, ID), Pepperidge Farm Light Style Wheat Bread (Pepperidge Farm, Inc, Norwalk, CT), and Country Farms Wheat Berry Bread (Earthgrains Baking Companies, Inc, St Louis). The recipe for "other cooked breakfast cereal" on the FFQ was based on cream of wheat, which was selected on the basis of the proportion of marketplace shelf space for hot cereals from 1986 to 1994.

Whole grains were considered in their intact and pulverized forms, and each whole-grain ingredient was required to satisfy the content of an individual type of grain; ie, whole-grain wheat must have the proper content of bran, endosperm, and germ (C Pratt, personal communication, 2002). The whole-grain designation was assigned to the following items in the database: whole wheat and whole-wheat flour, whole oats and whole-oat flour, whole cornmeal and corn flour, brown rice and brown rice flour, whole barley, whole rye and rye flour, bulgur, buckwheat, popcorn, amaranth, and psyllium. Wheat bran, corn bran, oat bran, rice bran, and wheat germ that were added to foods to increase the fiber content were further considered as added bran and germ, respectively.

The whole-grain percentage for each food was then multiplied by the gram weight per serving to obtain the grams of whole-grain content per RACC. Total whole-grain consumption (in g/d) was then calculated by summing the whole-grain intakes from 1) high-whole-grain foods with ≥51% whole-grain content by weight in accordance with the FDA whole-grain claim; 2) moderate whole-grain foods with ≥25% whole-grain or bran content by weight according to the classification of Jacobs et al; and 3) all foods consumed. Total bran and germ consumption were calculated in a similar manner and further subdivided into their natural and added components. The natural portion included the amount of bran and germ that would correspond to the amount of whole grain in the food, and any amount exceeding the naturally occurring portion was considered to have been added.

The refined-grain group was calculated in servings per day and included grain foods with <25% whole-grain content by weight, such as cooked and other cold breakfast cereals, dark bread made from wheat flour rather than whole-wheat flour, white bread, English muffins, bagels or rolls, pancakes or waffles, white rice, pasta, cookies, doughnuts, brownies, sweet rolls, coffee cake, and pizza.

The FFQ was validated among a subset of the study participants as reported previously (14, 17). Two FFQs were administered 1 y apart, with two 1-wk diet records administered 6 mo apart during this 1-y period. The performance of the FFQ in assessing the intakes of individual grain products was further evaluated, and the Pearson correlation coefficients between the FFQs and the dietary records ranged from 0.30 for servings of refined-grain foods to 0.86 for those of cold breakfast cereals (17, 18).

Statistical analysis
The differences in mean grain intakes from 1986 to 1994 were compared by using Wilcoxon’s signed-rank tests. We determined the age- and energy-adjusted means for lifestyle and dietary characteristics across increasing quintiles of whole-grain intakes. Tests for linear trend were calculated by assigning the median value for each quintile of intake treated as a continuous variable by using linear (for continuous outcome variables) or logistic (for dichotomous outcome variables) regression models.

Using multivariate linear regression, we examined how changes in intakes of whole grain, refined grain, and added bran and germ (1986–1994) were associated with the dependent variable, namely the change in weight (in kg) in the same period. We used the robust variance estimate (19) to obtain valid inferences and avoid the necessary assumptions of normality for linear regression. In all analyses, there was one observation per participant. Whole grains were modeled as continuous variables when tests for nonlinearity using spline regression were not statistically significant; otherwise, the exposures were categorized. We calculated the multivariate adjusted least-squares means for changes in body weight across categories of change in whole grains and fixed the continuous covariates at their respective median values and the categorical values at their modal values.

Multivariable models included age (in 5-y categories), the respective baseline exposure measure (quintiles or continuous), baseline weight (quintiles), and baseline and change in total calories (continuous variables), alcohol (quintiles), and total physical activity (both in quintiles). Change in smoking status between 1986 and 1994 was included as a categorical variable. Men who were nonsmokers at both time periods were classified as "nonsmokers," and men who were smokers at both periods were categorized as "habitual smokers." Men who reported smoking in 1986 and nonsmoking in 1994 were classified as "quitters," and men who reported never smoking in 1986 and smoking in 1994 were classified as "new smokers."

We adjusted for the percentages of energy derived from protein and monounsaturated, polyunsaturated, saturated, and trans unsaturated fatty acids to examine the possibility that dietary patterns associated with diets high in whole grains would explain any apparent inverse associations. We also considered constituents found in whole grains that may potentially mediate the association with weight gain. For these analyses, we further adjusted for intakes of dietary fiber, folate, magnesium, and vitamins E and B-6.

To evaluate the influence of measurement error on the association between changes in whole grains and weight gain, we used a subset of participants from a separate but similar study (15, 20) for whom repeated dietary records and FFQs were available in both 1980 and 1986. Using the regression calibration approach (Appendix A) together with data from our main study, we adjusted the regression coefficients and least-squares means for measurement error. All statistical analyses were conducted by using SAS software (version 8.2; SAS Institute Inc, Cary, NC), and statistical significance was defined as a two-tailed P value < 0.05.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
DISCUSSION
 APPENDIX A
 REFERENCES
 
The mean (±SD) whole-grain intakes (in g/d) from foods with various whole-grain concentrations are listed in Table 1Go. The high-whole-grain foods (≥51% whole-grain content by weight) that contributed most to intake (in g/d) included select cold breakfast cereals (37.6%), brown rice (25.3%), cooked oatmeal (21.8%), and popcorn (11.6%). Of the participants who reported eating breakfast cereals (76.3% of the cohort), 44.2% reported eating breakfast cereals with ≥51% whole-grain content by weight, 26% reported eating breakfast cereals with 25–50% whole-grain content by weight, and 29.7% reported eating refined-grain cereals (<25% whole-grain content by weight). The foods that fulfilled the ≥25% whole-grain content criterion included specific cold breakfast cereals, crackers, and other grains. On the basis of the composite of commercial wheat bread formulations used, "dark bread" had <25% whole-grain content by weight, and it was therefore classified as a refined grain. Dark bread contributed to the participants’ total whole-grain consumption; each slice provides 2 g whole grain.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Quantitative whole-grain intakes assessed in 1986 and 1994 food-frequency questionnaires in the Health Professionals Follow-up Study1

 
Between 1986 and 1994, added bran from all foods increased significantly (P < 0.001), from 2.2 ± 6.0 to 5.3 ± 7.9 g/d, and this increase derived primarily from wheat and oats. The intakes of added germ significantly decreased from baseline to 0.3 ± 1.3 g/d in 1994 (P < 0.001) and were obtained largely from wheat germ (62.7%).

The age and energy-adjusted descriptive characteristics of the study population across quintile categories of whole-grain intake from all foods are shown in Table 2Go. The median intakes of whole grains ranged from 3.0 g/d in the lowest category of intake to 42.7 g/d in the highest category. Favorable trends in lifestyle and dietary factors across quintile categories of whole-grain intake were significant for all factors (P for trend < 0.001 for all). Compared with participants in the lowest category, those in the highest category of consumption tended to have a lower BMI, were less likely to smoke, and were more physically active. Higher intakes of whole grains were associated with lower intakes of total fat, saturated fat, dietary cholesterol, and alcohol and higher intakes of dietary fiber, magnesium, vitamin E, vitamin B-6, and folate. Glycemic load was higher in those with greater whole-grain intakes.


View this table:
[in this window]
[in a new window]
  		TABLE 2 Demographic and behavioral characteristics by quintiles of whole-grain intake for US men in the Health Professionals Follow-up Study at baseline in 19861

 
Multivariate models
Multivariate-adjusted mean weight gains across quintile categories of whole grains are shown in Table 3Go. All participants tended to gain weight over the 8-y follow-up period, and the mean weight gain was 1.9 ± 5.2 kg. After adjustment for potential confounding variables and dietary factors associated with diets high in whole grains, increases in consumption of whole grains (in g/d) over the 8-y period were inversely related to weight gain (P for trend < 0.01, regardless of the whole-grain concentration of the food sources). For every 40-g/d increment in whole grains from all foods (the approximate difference in median intakes between the highest and lowest quintiles), long-term weight gain was reduced by 0.49 kg.


View this table:
[in this window]
[in a new window]
  		TABLE 3 Multivariate-adjusted mean 8-y weight change (kg) according to quintiles of changes in whole-grain intakes from foods that meet various whole-grain content criteria among US men in the Health Professionals Follow-up Study1

 
To better understand the relation between whole-grain consumption and weight gain related to nutrients found in whole grains, we adjusted for potential mediating nutrients, including food sources of fiber, magnesium, folate, and vitamins E and B-6. The association between changes in whole-grain intake and weight gain (0.75 and 1.24 kg weight gain for the highest and the lowest quintile category, respectively; P for trend < 0.0001) was mildly attenuated after adjustment for these nutrients (0.93 and 1.33 kg weight gain, respectively; P for trend < 0.01).

Using the results of our validation study, we further corrected for covariate measurement error in the average changes in body weight associated with quintile changes in grain intakes. After correction for measurement error, the 8-y weight gain associated with a 40-g/d difference in whole-grain intake was 1.1 kg (compared with 0.49 kg, uncorrected).

Dietary fiber was inversely related to weight gain independent of whole grains (P for trend < 0.0001). The men in the lowest quintile of change gained 1.40 kg, whereas the men in the highest quintile of change gained 0.39 kg (Table 4Go). For every 20-g/d increase in dietary fiber, weight gain was reduced by 1.18 kg. After correction for measurement error, long-term weight gain was reduced by 5.5 kg for each 20-g/d increment in dietary fiber.


View this table:
[in this window]
[in a new window]
  		TABLE 4 Multivariate-adjusted mean 8-y weight change (kg) according to quintiles of changes in food sources of fiber among US men in the Health Professionals Follow-up Study1

 
Because the physiologic consequences of a high-fiber diet may depend on the type or food source of fiber, we further examined the relations between cereal, fruit, and vegetable fiber and weight change. After simultaneous adjustment for each of the 3 main sources of fiber, significant inverse associations were observed for cereal and fruit fiber, but not vegetable fiber. The dose-response relation was strongest for fruit fiber. For every 20 g/d increment in fruit fiber, weight gain was reduced by 2.51 kg (P for trend < 0.001). Because apples predominantly accounted for the fruit fiber intake (26.7%), we also calculated the mean weight gain for changes in apple consumption. Men who added a single daily apple to their diet gained 0.67 kg less weight over the 8-y period, if all other factors were held constant (P for trend < 0.0001).

Cereal fiber was inversely related to weight gain (P for trend < 0.001). For every 20 g/d increment in cereal fiber, weight gain was reduced by 0.81 kg. Similarly, changes in cold breakfast cereal consumption (servings/d) significantly predicted weight gain. Increases in intake of cereals with ≥51% whole-grain content by weight protected against weight gain (P for trend < 0.001), whereas the association with changes in intake of cereals with 25–50% whole-grain content was only marginally significant (P for trend = 0.05 for trend). In contrast, refined-grain cereals (<25% whole-grain content) were positively related to long-term weight gain (P for trend < 0.001).

Added bran consumption was also inversely associated with weight gain in a dose-response manner, and men in the highest quintile of change gained 15% less weight than did men in the lowest quintile over 8 y (P for trend = 0.01; Table 5Go). For every 20-g/d increment in added bran intake, weight gain was reduced by 0.36 kg. In contrast, there was no significant trend across categories of refined-grain or added germ intakes.


View this table:
[in this window]
[in a new window]
  		TABLE 5 Multivariate-adjusted mean 8-y weight change (kg) according to tertiles of changes in added bran and germ intakes among US men in the Health Professionals Follow-up Study1

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
APPENDIX A
 REFERENCES
 
In this population of men, an increase in consumption of whole grains protected against long-term weight gain, and the dose-response relation appeared strongest for foods with ≥25% whole-grain or bran content by weight. Changes in added bran consumption also negatively predicted weight gain, whereas no association was observed for changes in servings of refined grains or in grams of added germ intake. To our knowledge, this is the first prospective cohort study to report the independent associations between changes in quantitative estimates of whole grains (in g/d) and weight gain among men.

Whereas whole grains have been inversely associated with body weight and fat distribution (1-4), few experimental studies have been conducted to directly examine this relation. In one randomized crossover trial, body weight was reduced after 6 wk on a whole-grain diet, compared with a refined-grain diet (21). The difference was not significant, but the follow-up period may have been too brief for significant weight loss to occur. Prospective data directly linking changes in consumption of whole grains to weight gain are further lacking, particularly among men (22). In a recent study among female nurses, Liu et al (4) reported that women in the highest quintile of change in whole-grain intake gained 1.07 kg over two 4-y intervals, whereas women in the lowest quintile gained 1.58 kg. The magnitude of the reported association is consistent with our findings, because the median difference in intake between the extreme quintiles of whole-grain intake ({approx}40 g/d increment) was associated with a 0.49-kg weight gain.

In the current study, the effect of whole grains was mildly attenuated after adjustment for food sources of fiber and micronutrients including magnesium, which suggests that the apparent effects of whole grains may be partially mediated by these nutrients. Fiber may decrease energy intake and induce weight loss by inducing satiety (23, 24) and reducing postprandial glucose concentrations (25-27). Moreover, because of their high fiber and water content, whole-grain foods contain fewer calories gram-for-gram than does an isovolemic amount of corresponding refined-grain food (21). Dietary fiber generally has been inversely associated with body weight (28) and body fat (29, 30). In the Coronary Artery Risk Development in Young Adults Study, fiber was inversely associated with BMI at all levels of fat intake, and it predicted insulin concentrations and 10-y weight gain (31). Whereas, in our study, increases in fiber consumption had a stronger association with long-term weight change than did increases in whole-grain consumption, measurement error may partially account for the difference in the magnitude of the estimated effects. Consumers often have difficulty in discerning whole-grain products (32), and the limited number of whole-grain foods listed on the FFQ may contribute to measurement error and attenuate the weight change estimates for whole grains (33).

Whereas fiber has been inversely linked to body weight, the food sources of fiber have not previously been related to body weight. In our study, cereal fiber significantly protected against weight gain independent of fruit and vegetable fibers. In the Framingham Offspring Study, cereal fiber was favorably related to fasting insulin concentrations (1). Furthermore, in several prospective cohort studies, low cereal fiber content and a high glycemic load were positively associated with risk of type 2 diabetes (33-37). The mechanisms by which cereal fiber may affect insulin sensitivity are unknown because cereals are predominantly sources of insoluble fibers that do not result in highly viscous intestinal contents (25-27). Instead, the reductions in plasma glucose and insulin that are related to reductions in body weight have been attributed to the viscosity of soluble fiber sources that delays gastric emptying, intestinal absorption, or both (25-27).

In support of the potential effect of soluble fiber sources, fruit fiber and apples were inversely associated with long-term weight gain. In a recent feeding study, 12 wk of supplementation with apples and pears consumed 3 times/d contributed to weight loss among overweight women aged 30–50 y (38). Surprisingly, vegetable fiber did not appear to reduce long-term weight gain in this population. It is plausible that increases in vegetable fiber consumption did not confer added benefit because of high baseline intakes.

Because the inverse relation between whole grains and weight gain was not entirely explained by types of fiber, other benefits likely pertain to properties of whole-grain as compared with refined-grain products, including a lower glycemic index and greater particle size (1, 39). Moreover, compounds that are botanically linked to fiber may confer benefits beyond the effect of fiber itself (40). In particular, vitamins, minerals, phenolic compounds, and phytoestrogens are abundant in the bran (41), in contrast to the fiber derived from the starchy endosperm. This may explain why fiber from whole-grain foods has been inversely related to adverse health outcomes, unlike the fiber derived from refined grains (41). Furthermore, whereas fortification of flours may replace B vitamins, other nutrients lost during processing are not replaced (42). These observations underscore both the importance of the grain source of fiber and the potentially synergistic effect of constituents naturally found in whole grains (41).

Added bran, which was inversely related to weight gain in our study, has been purported to improve carbohydrate tolerance through antioxidant mechanisms (43) by reducing serum iron concentrations and increasing magnesium concentrations (44). Although we did not separately examine the sources of bran on weight gain, most of the bran that was consumed in this population was derived from wheat and oats. Epidemiologic studies have consistently shown a beneficial effect of wheat fiber in reducing the risk of diabetes (33-37), but the findings for wheat bran supplementation on glycemic control have been mixed. In a recent crossover study, wheat bran did not influence glycemic control in diabetics (43), although, in other feeding studies, wheat bran significantly improved glucose tolerance (45, 46). Oat bran has been found to enhance glycemic control, attributed in part to the unbranched ß-glucan component of soluble oat fiber (47). Whereas no effects were observed for the intake of supplemental germ, there may have been insufficient variation in intakes to detect any association.

Because many dietary factors may induce metabolic alterations within a relatively short latency period, we assessed changes in whole grains that were concurrent with weight gain (15). This study design may limit the drawing of causal inferences. However, the consistency of our findings with those of prior studies and the likely effects induced by whole grains on weight gain lend credence to important causal relations. Confounding also poses a challenge in the study of whole grains, as an increased consumption may be a marker for changes in other health behaviors. However, we adjusted the regression models for changes in several other lifestyle and dietary behaviors, including smoking, physical activity, and fat consumption with only a modest attenuation of the relation between whole grains and weight gain.

Because of measurement errors in assessing changes in diet, the observed associations are underestimations of the true effect estimates. After accounting for these errors, long-term weight gain was reduced by 1.1 kg for each 40-g/d increment in whole grains from all foods (0.49 kg, uncorrected). Moreover, the dose-response relation appeared strongest for whole grains with ≥25% whole grain or bran content by weight (1.6 kg for each 40 g/d increment, after correction for measurement error; 0.56 kg, uncorrected). Our findings suggest that the current FDA whole-grain health claim may not adequately capture all of the benefits of reduced body weight associated with whole-grain foods because added bran is excluded from the definition. Further scientific data are needed to establish optimal whole-grain intakes for various health outcomes. Studies may also examine how particular types of grains and bran fibers may affect body weight regulation among diverse populations.


   APPENDIX A
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
REFERENCES
 
Correction of least-squares means for covariate measurement error
This method is a straightforward extension of the methods detailed by Rosner et al (1) and Spiegelman et al (2). An SAS macro (available from DS) performs these calculations (stdls{at}channing.harvard.edu). Let Y be the dependent variable, Z be a p1 x 1 vector of covariates measured with error, X be a p1 x 1 vector of the corresponding perfectly measured covariates, and U be a p2 x 1 vector of covariates that are always measured without error. To apply this method, a main study must be available with data (Yi, Zi, Ui), i = 1, ..., n1, and a validation study must be available with data (Xi, Zi, Ui), i = 1, ..., n2. Typically, n1 >> n2.

We assume a linear model for Y:

(A1)
where ß* is a 1 x (1 + p) vector of regression coefficients, and p = p1 + p2.

In addition, we assume a linear measurement error model:

(A2)
where {Gamma} is a [p1 x (1 + p)] matrix of regression parameters describing the measurement error models for each of the p1 covariates, X, measured with error.

As stated by Rosner et al (1) and Spiegelman et al (2), a consistent estimator of ß*, *, is given by

(A3)
where is obtained from the regression of Y on (1, ZT, UT)T in the main study, and is obtained from the regression of X on (1, ZT, UT)T in the validation study.

The variance, {Sigma}ß*, for this estimator is given in references 1 and 2 as a function of the variances of and , which are the standard variances from the respective linear regression models.

Then the estimator of the least-squares mean at a specified value of X = x and U = u, Y | (X = x, U = u), is given by

(A4)
and

(A5)


   ACKNOWLEDGMENTS
 
We are indebted to Al Wing, Lydia Liu, and Ruifeng Li for their technical support.

The authors’ contributions to this publication are as follows: design of the study, analysis of the data, and writing of the manuscript (PK-B, DS, and SL); database development and writing of the manuscript (MF and LS); writing of the manuscript (DJ); and securing the funding, collection and analysis of the data, design of the study, and writing of the manuscript (WW and ER). None of the authors had any personal or financial conflicts of interest to report.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX A
 REFERENCES
 

  1. McKeown NM, Meigs JB, Liu S, et al. Whole-grain intake is favorably associated with metabolic risk factors for type 2 diabetes and cardiovascular disease in the Framingham Offspring Study. Am J Clin Nutr 2002;76:390-8.[Abstract/Free Full Text]
  2. Pereira M, Jacobs D, Slatery M, et al. The association of whole grain intake and fasting insulin in a biracial cohort of young adults: The CARDIA study. CVD Prev 1998;1:231-42.
  3. Jacobs DR Jr, Meyer KA, Kushi LH, Folsom AR. Whole-grain intake may reduce the risk of ischemic heart disease death in postmenopausal women: the Iowa Women’s Health Study. Am J Clin Nutr 1998;68(2):248-57.[Abstract]
  4. Liu S, Willett WC, Manson JE, Hu FB, Rosner B, Colditz G. Relation between changes in intakes of dietary fiber and grain products and changes in weight and development of obesity among middle-aged women. Am J Clin Nutr 2003;78(5):920-7.[Abstract/Free Full Text]
  5. Fung TT, Rimm EB, Spiegelman D, et al. Association between dietary patterns and plasma biomarkers of obesity and cardiovascular disease risk. Am J Clin Nutr 2001;73:61-7.[Abstract/Free Full Text]
  6. Liu S. Whole-grain foods, dietary fiber, and type 2 diabetes: searching for a kernel of truth. Am J Clin Nutr 2003;77:527-9.[Free Full Text]
  7. Kantor LS, Variyam JN, Allshouse JE, Putnam JJ, Lin BH. Choose a variety of grains daily, especially whole grains: a challenge for consumers. J Nutr 2001;131:473S-86S.[Abstract/Free Full Text]
  8. Jenkins DJ, Wesson V, Wolever TM, et al. Wholemeal versus wholegrain breads: proportion of whole or cracked grain and the glycaemic response. BMJ 1988;297(6654):958-60.
  9. Lang R, Jebb SA. Who consumes whole grains and how much? Proc Nutr Soc 2003;62:123-7.[Medline]
  10. US Food and Drug Administration. Center for Food Safety and Applied Nutrition: A Food Labeling Guide. 1994, Docket 99P-2209. Editorial revisions: June 1999 and November 2000. Health Claim Notification for Wholegrain Foods. Internet: http://www.cfsan.fda.gov/label.html (accessed 10 January 2002).
  11. Code of Federal Regulations, Title 21, volume 2, revised as of 1 April 2002. 21 CFR [section]101.12. Internet: http://vm.cfsan.fa.gov/~lrd/CF101-12.HTML (accessed 19 August 2004).
  12. Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB, Willett WC. Validity of self-reported waist and hip circumferences in men and women. Epidemiology 1990;1:466-73.[Medline]
  13. Palta M, Prineas RJ, Berman R, et al. Comparison of self-reported and measured height and weight. Am J Epidemiol 1982;115:223-30.[Abstract/Free Full Text]
  14. Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol 1992;135:1114-26.[Abstract/Free Full Text]
  15. Willett WC. Nutritional epidemiology. 2nd ed. New York: Oxford University Press, 1998.
  16. Cleveland LE, Moshfegh AJ, Albertson AM, et al. Dietary intake of whole grains. J Am Coll Nutr 2000;19(suppl):331S-8S.[Abstract/Free Full Text]
  17. Hu FB, Rimm E, Smith-Warner SA, et al. Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. Am J Clin Nutr 1999;69(2):243-9.[Abstract/Free Full Text]
  18. Feskanich D, Rimm EB, Giovannucci EL, et al. Reproducibility and validity of food intake measurements from a semiquantitative food frequency questionnaire. J Am Diet Assoc 1993;93(7):790–6.[Medline]
  19. White H. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 1980;48:817-38.
  20. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol 1985;122(1):51-65.[Abstract/Free Full Text]
  21. Pereira MA, Jacobs DR Jr, Pins JJ, et al. Effect of whole grains on insulin sensitivity in overweight hyperinsulinemic adults. Am J Clin Nutr 2002;75(5):848-55.[Abstract/Free Full Text]
  22. Liu S. Intake of refined carbohydrates and whole grain foods in relation to risk of type 2 diabetes mellitus and coronary heart disease. J Am Coll Nutr 2002;21(4):298-306.[Abstract/Free Full Text]
  23. Howarth NC, Saltzman E, Roberts SB. Dietary fiber and weight regulation. Nutr Rev 2001;59(5):129-39.[Medline]
  24. Porikos K, Hagamen S. Is fiber satiating? Effects of a high fiber preload on subsequent food intake of normal-weight and obese young men. Appetite 1986;7(2):153-62.[Medline]
  25. Liese AD, Roach AK, Sparks KC, Marquart L, D’Agostino RB Jr, Mayer-Davis EJ. Whole-grain intake and insulin sensitivity: the Insulin Resistance Atherosclerosis Study. Am J Clin Nutr 2003;78(5):965-71.[Abstract/Free Full Text]
  26. Wood PJ, Braaten JT, Scott FW, Riedel KD, Wolynetz MS, Collins MW. Effect of dose and modification of viscous properties of oat gum on plasma glucose and insulin following an oral glucose load. Br J Nutr 1994;72(5):731-43.[Medline]
  27. Hallfrisch J, Behall KM. Mechanisms of the effects of grains on insulin and glucose responses. J Am Coll Nutr 2000;19:320S-5S.[Abstract/Free Full Text]
  28. Alfieri MA, Pomerleau J, Grace DM, Anderson L. Fiber intake of normal weight, moderately obese and severely obese subjects. Obes Res 1995;3(6):541-7.[Medline]
  29. Miller WC, Niederpruem MG, Wallace JP, Lindeman AK. Dietary fat, sugar, and fiber predict body fat content. J Am Diet Assoc 1994;94(6):612-5.[Medline]
  30. Nelson LH, Tucker LA. Diet composition related to body fat in a multivariate study of 203 men. J Am Diet Assoc 1996;96(8):771-7.[Medline]
  31. Ludwig DS, Pereira MA, Kroenke CH, et al. Dietary fiber, weight gain, and cardiovascular disease risk factors in young adults. JAMA 1999;282(16):1539-46.[Abstract/Free Full Text]
  32. Adams JF, Engstrom A. Helping consumers achieve recommended intakes of whole grain foods. J Am Coll Nutr 2000 Jun;19(suppl):339S-44S.[Abstract/Free Full Text]
  33. Fung TT, Hu FB, Pereira MA, et al. Whole-grain intake and the risk of type 2 diabetes: a prospective study in men. Am J Clin Nutr 2002;76:535-40.[Abstract/Free Full Text]
  34. Salmeron J, Manson JE, Stampfer MJ, et al. Dietary fiber, glycemic load, and risk of non-insulin dependent diabetes mellitus in women. JAMA. 1997;277:472-7.[Abstract]
  35. Salmeron J, Ascherio A, Rimm EB, et al. Dietary fiber, glycemic load, and risk of NIDDM in men. Diabetes Care 1997;20(4):545-50.[Abstract]
  36. Meyer KA, Kushi LH, Jacobs DR Jr, Slavin J, Sellers TA, Folsom AR. Carbohydrates, dietary fiber, and incident type 2 diabetes in older women. Am J Clin Nutr 2000;71(4):921-30.[Abstract/Free Full Text]
  37. Montonen J, Knekt P, Jarvinen R, et al. Whole-grain and fiber intake and the incidence of type 2 diabetes. Am J Clin Nutr 2003;77:622-9.[Abstract/Free Full Text]
  38. Conceicao de Oliveira M, Sichieri R, Sanchez Moura A. Weight loss associated with a daily intake of three apples or three pears among overweight women. Nutrition 2003;19(3):253-6.[Medline]
  39. Foster-Powell K, Miller JB. International tables of glycemic index. Am J Clin Nutr 1995;62(suppl):871S-90S.[Abstract]
  40. Jacobs DR, Pereira MA, Meyer KA, Kushi LH. Fiber from whole grains, but not refined grains, is inversely associated with all-cause mortality in older women: the Iowa Women’s Health Study. J Am Coll Nutr 2000;19(suppl):326S-30S.[Abstract/Free Full Text]
  41. Burkitt DP, Walker AR, Painter NS. Dietary fiber and disease. JAMA 1974;229:1068-74.[Medline]
  42. Murtaugh MA, Jacobs DR, Jacob B, et al. Epidemiological support for the protection of whole grains against diabetes. Proc Nutr Soc 2003;62:143-9.[Medline]
  43. Jenkins DJ, Kendall CQ, Augustin L, et al. Effect of wheat bran on glycemic control and risk factors for cardiovascular disease in type 2 diabetes. Diabetes Care 2002;25:1522-8.[Abstract/Free Full Text]
  44. Paolisso D, D’Amore A, Ceriello A, Varrichio M, D’Onofrio F. Pharmacologic doses of vitamin E improve insulin action in healthy subjects and non-insulin-dependent diabetic patients. Am J Clin Nutr 1993;57:650-6.[Abstract/Free Full Text]
  45. Brodribb AJ, Humphreys DM. Diverticular disease: three studies. Part III - Metabolic effect of bran in patients with diverticular disease. Br Med J 1976;1(6007):428-30.
  46. Bosello O, Ostuzzi R, Armellini F, Micciolo R, Scuro LA. Glucose tolerance and blood lipids in bran-fed patients with impaired glucose tolerance. Diabetes Care 1980;3(1):46-9.[Abstract]
  47. Braaten JT, Scott FW, Wood PJ, et al. High ß-glucan oat bran and oat gum reduce postprandial blood glucose and insulin in subjects with and without type 2 diabetes. Diabetic Med 1994;22:312-8.
  48. Rosner B, Spiegelman D, Willett WC. Correction of logistic regression relative risk estimates and confidence intervals for measurement error: the case of multiple covariates measured with error. Am J Epidemiol 1990;132:734-45.[Abstract/Free Full Text]
  49. Spiegelman D, McDermott A, Rosner B. The regression calibration method for correcting measurement error bias in nutritional epidemiology. Am J Clin Nutr 1997;65(suppl):1179S-86S.[Abstract/Free Full Text]
Received for publication March 18, 2004. Accepted for publication June 21, 2004.




This article has been cited by other articles:


Home page
 AMERICAN JOURNAL OF LIFESTYLE MEDICINEHome page
M. B. Andon and J. W. Anderson
State of the Art Reviews: The Oatmeal-Cholesterol Connection: 10 Years Later
American Journal of Lifestyle Medicine, February 1, 2008; 2(1): 51 - 57.
[Abstract] [PDF]

Home page
 J. Am. Coll. Nutr.Home page
C. K. Good, N. Holschuh, A. M. Albertson, and A. L. Eldridge
Whole Grain Consumption and Body Mass Index in Adult Women: An Analysis of NHANES 1999-2000 and the USDA Pyramid Servings Database
J. Am. Coll. Nutr., February 1, 2008; 27(1): 80 - 87.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
P. B Mellen, A. D Liese, J. A Tooze, M. Z Vitolins, L. E Wagenknecht, and D. M Herrington
Whole-grain intake and carotid artery atherosclerosis in a multiethnic cohort: the Insulin Resistance Atherosclerosis Study
Am. J. Clinical Nutrition, June 1, 2007; 85(6): 1495 - 1502.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
A. Regina, A. Bird, D. Topping, S. Bowden, J. Freeman, T. Barsby, B. Kosar-Hashemi, Z. Li, S. Rahman, and M. Morell
High-amylose wheat generated by RNA interference improves indices of large-bowel health in rats.
PNAS, March 7, 2006; 103(10): 3546 - 3551.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
S. N Rajpathak, E. B Rimm, B. Rosner, W. C Willett, and F. B Hu
Calcium and dairy intakes in relation to long-term weight gain in US men
Am. J. Clinical Nutrition, March 1, 2006; 83(3): 559 - 566.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
E. Lopez-Garcia, R. M van Dam, S. Rajpathak, W. C Willett, J. E Manson, and F. B Hu
Changes in caffeine intake and long-term weight change in men and women
Am. J. Clinical Nutrition, March 1, 2006; 83(3): 674 - 680.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
M. K Jensen, P. Koh-Banerjee, M. Franz, L. Sampson, M. Gronbaek, and E. B Rimm
Whole grains, bran, and germ in relation to homocysteine and markers of glycemic control, lipids, and inflammation 1
Am. J. Clinical Nutrition, February 1, 2006; 83(2): 275 - 283.
[Abstract] [Full Text] [PDF]

Home page
 Diabetes CareHome page
L. Qi, R. M. van Dam, S. Liu, M. Franz, C. Mantzoros, and F. B. Hu
Whole-Grain, Bran, and Cereal Fiber Intakes and Markers of Systemic Inflammation in Diabetic Women
Diabetes Care, February 1, 2006; 29(2): 207 - 211.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Clin. Nutr.Home page
N. R Sahyoun, P. F Jacques, X. L Zhang, W. Juan, and N. M McKeown
Whole-grain intake is inversely associated with the metabolic syndrome and mortality in older adults
Am. J. Clinical Nutrition, January 1, 2006; 83(1): 124 - 131.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Koh-Banerjee, P.
Right arrow Articles by Rimm, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Koh-Banerjee, P.
Right arrow Articles by Rimm, E.
Agricola
Right arrow Articles by Koh-Banerjee, P.
Right arrow Articles by Rimm, E.