Original Article | Published:

Whole-grain consumption, dietary fibre intake and body mass index in the Netherlands cohort study

European Journal of Clinical Nutrition volume 63, pages 3138 (2009) | Download Citation

Contributors: RAG and PAvdB were responsible for the study concept and design, and also for the acquisition of the data. LPLvdV, RAG and LMCvdB were responsible for the analysis and interpretation of data. LPLvdV and RAG drafted the manuscript. All authors took part in critical revision of the manuscript. Statistical expertise was provided by RAG. Funding was obtained by RAG and PAvdB.

Subjects

Abstract

Objectives:

To assess the association of whole-grain and (cereal) fibre intake with body mass index (BMI) and with the risk of being overweight (BMI25) or obese (BMI30 kg m−2).

Subjects:

A total of 2078 men and 2159 women, aged 55–69 years, were included in the analysis, after exclusion of subjects with diagnosed cancer or deceased within 1 year after baseline or with missing dietary information.

Results:

We found an inverse association between whole-grain consumption and BMI and risk of overweight and obesity in men as well as women. The association in men was stronger than in women; the risk of being obese as compared to normal weight was 10% (95% CI: 2–16%) and 4% (95% CI: 1–7%) lower for each additional gram of (dry) grain consumption in men and women, respectively. Fibre and cereal fibre intake were inversely associated with BMI in men only. Associations were similar after exclusion of likely under- and overreporters of energy. A retrospective analysis of baseline fibre intake and weight gain after the age of 20 years also showed a slight inverse association.

Conclusions:

Whole-grain consumption may protect against becoming overweight or obese; however, the cross-sectional design of the study does not allow conclusions about the causality of the association.

Introduction

Overweight and obesity are the result of long-term imbalance between energy intake and energy expenditure. Besides lack of physical activity (energy expenditure), dietary habits (energy intake) are assumed to be the cause of overweight and obesity. Mean body mass index (BMI) as well as the prevalence of overweight (BMI 25–29 kg m−2) and obesity (BMI30 kg m−2) are increasing over time, not only in the United States, but also in Europe, although to a somewhat lesser extent. In the Netherlands, about 9% of the adult population is obese (RIVM, 2004). Persons who are overweight or obese have an increased risk of hypertension, type II diabetes, cancer and other chronic diseases (Anonymous, 1998, 2000). Strategies to improve nutrition and to increase physical activity are important to prevent and force back overweight, thus restricting present and future health-care costs.

One of the strategies to improve nutrition is to increase consumption of whole grains. Whole grains might help in preventing overweight and obesity due to their effect on satiety, although for fibre the evidence for this effect was considered insufficient (Institute of Medicine, 2002). Another potential mechanism is to slow down starch digestion and absorption, which may lead to lower glucose and insulin responses, favouring oxidation and lipolysis of fat rather than its storage (Slavin et al., 1999).

The association between the consumption of whole grains and several diseases has been studied (Jacobs et al., 1998; Jacobs and Gallaher, 2004; McKeown, 2004; Merchant et al., 2006), but thus far only few studies have reported on the association between whole-grain intake and BMI (McKeown et al., 2002; Koh-Banerjee and Rimm, 2003; Steffen et al., 2003) and (changes in) whole-grain intake and weight gain (Liu et al., 2003; Steffen et al., 2003; Koh-Banerjee et al., 2004; Bazzano et al., 2005) and dietary fibre intake and body weight (Slavin, 2005).

We have assessed the association between the consumption of whole-grain foods and dietary fibre and BMI as well as overweight and obesity in the population of the Netherlands Cohort Study (NLCS), both in a cross-sectional setting and in a prospective setting in a smaller sample.

Population and methods

Population

The NLCS was initiated in 1986 among more than 120 000 Dutch men and women aged 55–69 years, derived from a random sample from municipal population registries across the country (van den Brandt et al., 1990). The study was originally designed to assess the relation between dietary habits and cancer. The NLCS is designed as a case–cohort study; questionnaires were processed for all cases plus a random sample of 5000 subjects drawn from the total cohort at baseline (the subcohort). The cross-sectional study presented in this paper is based on the baseline data of the subcohort. For data analysis, we excluded cases with prevalent cancer at baseline, cancer diagnosed or deceased within 1 year after baseline, subjects with missing information on weight or height or with incomplete or inconsistent dietary questionnaires (Goldbohm et al., 1994). The remaining subcohort comprised 4237 subjects (2078 men, 2159 women).

The reproducibility of the self-administered questionnaire was determined from five annually repeated questionnaire administrations in independent random samples from the subcohort of about 300 subjects each. The data from the reproducibility study (Goldbohm et al., 1995) were used to calculate changes in BMI and changes in diet between the two measurements. The entire reproducibility data set contains 1546 men and women. For this prospective analysis, we additionally excluded subjects deceased between the first year after baseline and the first year after completing the repeated questionnaire. The remaining sample from the reproducibility study contained 1257 subjects (50% women, 50% men).

Food-frequency questionnaire

At baseline, the participants completed an extensive self-administered questionnaire on dietary habits and other risk factors for cancer, such as height, weight, smoking habits, physical activity, medical history and so on. The 150-item semi-quantitative food-frequency section concentrated on habitual consumption during the preceding year and has been validated against a 9-day diet record (Goldbohm et al., 1994). In this validation study, the Pearson correlation coefficient for dietary fibre (both unadjusted and adjusted for energy and sex) was 0.74. The individual frequencies and serving sizes of the food items were recalculated into mean daily intake (g day−1) for each subject. For the data analyses, several exposure variables were created. The variable ‘All grain’ was calculated as the sum score of the food items: bran, wheat germs, muesli, porridge (oats or whole wheat), brown rice and cooked grains (for example, millet, buckwheat and so on). Porridge, brown rice and cooked grains were recalculated as the dry product to prevent unbalanced weighting to the sum score. The variable ‘Whole grain’ differs from the variable ‘All grain’ in not including bran and wheat germs. The variable ‘Total brown bread’ is the sum score of brown, wholemeal and rye bread. Brown bread is made of a mixture of wholemeal flour and white flour. Although not strictly considered as wholemeal bread, brown bread was added to this variable, since respondents do not always know the difference between the two types of bread. Besides these composite variables, four fibre variables are included in the data analysis: ‘total fibre’ and ‘fibre from grain’, both expressed as grams per day and as grams per MJ (that is, fibre density).

Outcome variables

At baseline, the following self-reported variables were available: height, weight and weight at age 20 years. As dependent variables for the statistical analysis, BMI (continuous), overweight and obesity as categorical variables (BMI25 and BMI30 kg m−2) and the change in BMI between the age of 20 years and baseline (continuous) were used. In the reproducibility data set, change in BMI (continuous) between the two measurements was calculated. To account for different time intervals between baseline and repeated measurement, the change in BMI was divided by the length of the interval in years, resulting in change in BMI per year.

Statistical analysis

Descriptive data for all variables were calculated. The analyses were performed for men and women separately. For the analysis with continuous outcome variables, multivariate regression (least squares) was employed, adjusting for potential confounders in several steps. For the analyses with categorical outcome variables, logistic regression was used to estimate odds ratios (ORs). A 10% change in the regression coefficient was taken as a cutoff point to determine whether in this population a potential confounder indeed had a confounding effect. If so, these variables were included in the full models. To enhance comparability, a fixed set of confounding variables was used for all data analyses, both for men and women. In the regression analysis of change in BMI on change in intake, the association was additionally adjusted for the baseline levels of these variables.

We considered the following potential confounders: age; smoking (yes/no) and number of cigarettes/day for current smokers; physical activity (based on leisure time and for men additionally based on occupation at baseline); education (four classes); other dietary variables, that is, intake of energy, nutrients (animal and vegetable protein, fats and carbohydrates), alcohol, and consumption of vegetables and fruit; dieting and eating behaviour; disease history. The final set of confounders in the multivariate analysis included age, energy intake, intake of animal protein, education, smoking status, number of cigarettes and consumption of fruit and vegetables.

To assess impact of under- and overreporting of energy intake on the investigated associations, we conducted sensitivity analyses excluding the most likely under- and overreporters in the data set. We used the regression equations published by the Institute of Medicine (IoM) (Institute of Medicine, 2002; Huang et al., 2005) to predict energy requirement, based on sex, age, height, weight and physical activity level. Different equations are used for subjects under and over BMI 25 (kg m−2). On the basis of the available (nonoccupational) physical activity variable, each subject was assigned to one of three activity levels used in the IoM equations as follows: <30, 30–60 and >60 minday−1 in levels 1, 2 and 3, respectively. Reported energy intake was adjusted with a factor of 100/86, to account for the incomplete listing of all available foods on the food-frequency questionnaire. This factor was based on the mean difference in energy intake between the 9-day dietary record and the food-frequency questionnaire in the validation study (Goldbohm et al., 1994). The ratio of reported intake over predicted requirement (rEI/pER) was used to test several cutoff levels. We used the cutoff level, where R2 between rEI and pER was 20%, corresponding to an accepted range of the ratio of 78–128% in both men and women. Approximately 75% of the population had energy intakes thus considered acceptable. This percentage was slightly lower in obese (but not in overweight) subjects, that is 66% of obese men and women had acceptable energy intakes.

Two-sided 5% significance levels were used throughout the paper. All analyses were done in SAS (version 8).

Results

In Table 1, the characteristics of the study population are presented per sex and category of BMI. About 47% of the men and 44% of the women were overweight or obese, but the percentage of obese women (9%) was much higher than that of men (4%). The mean age in the different BMI categories was comparable. A tendency of higher protein intake, lower carbohydrate intake (in men), lower alcohol intake (in women) and a decrease in percentage smokers (in men) over the successive BMI categories was seen. More subjects suffered from cardiovascular disease, hypertension, diabetes and gallstones (women only) in the higher BMI categories, whereas the percentage of subjects suffering from intestinal disorders was lower. As expected, more subjects in the high BMI classes reported to have followed a (energy-restricted) diet in the past 5 years.

Table 1: Baseline demographic, lifestyle and dietary characteristics of the study population according to sex and BMI strata

Both in men and women, whole-grain consumption consistently decreased over increasing BMI categories. The proportion consumers of whole-grain foods decreased from 40 to 24% in men and from 47 to 32% in women, while the amount of whole-grain foods consumed by them decreased likewise. No consistent trend was observed for the amount of brown bread, which was consumed by more than 90% of the population. Dietary fibre intake was only slightly lower in overweight and obese men.

In Table 2, the results of the linear regression analyses performed in the baseline data (cross-sectional) are presented, with BMI as a continuous variable. For men, statistically significant inverse associations were observed between the whole-grain variables as well as fibre intake and BMI in the age-adjusted model, the age- and energy-adjusted model (results not shown) and in the multivariate model. The fibre variables were more strongly inversely associated with BMI in the multivariate model. The consumption of brown bread was not significantly associated with BMI. In women, inverse associations for the whole-grain variables were observed both in the age- and multivariate-adjusted models. The consumption of brown bread was significantly positively associated with a higher BMI only after multivariate adjustment and no associations were found for fibre or fibre from grain or their densities. Both in men and women it was estimated that a 1 g day−1 higher intake of dry whole grains was associated with a 0.03 kg m−2 lower BMI; a decrease of 1 unit BMI thus corresponds to a 33 g day−1 increase in dry whole grain. When the analyses for the grain variables were repeated for the consumers of those foods only (575 men and 769 women), the results were comparable with the results presented in Table 2. The exclusion of subjects who reported digestive diseases (283 men and 174 women), gallstones (107 men and 303 women) and subjects who reported to have followed a dietary regimen in the 5 years prior to the baseline measurement (354 men and 449 women), had only minor impact on the results (not shown).

Table 2: Regression coefficients (β) and 95% confidence intervals (95% CIs) in the baseline data for BMI (continuous) and exposure variables

After exclusion of likely under- and overreporters of energy (that is, subjects with a ratio rEI/pER <72 or >128%), the associations of BMI with the grain variables were unchanged. The associations with fibre and fibre from grain became slightly more inverse in men (β: −0.05, 95% confidence interval (CI): −0.07, −0.03), but were unchanged in women. Associations with fibre densities became more inverse (men) or less positive (women), due to a change in the association between energy intake and BMI in a positive direction.

The results for overweight and obesity according to categories of BMI are presented in Table 3. For men, a high intake of whole grain was associated with a lower risk of being obese or overweight. The associations between all fibre intake variables and overweight were statistically significant inverse. Even stronger, but, due to the small number of obese men, not statistically significant inverse associations were observed between fibre intake and the risk of being obese. No association was observed between the consumption of brown bread and the risk of being obese or overweight. In women, a high consumption of whole-grain products was associated with a lower risk of being overweight or obese. A high intake of brown bread and fibre was associated with a higher risk of being obese.

Table 3: Multivariate odds ratios (ORs) and 95% confidence intervals (95% CIs) in baseline data for the outcome variables overweight and obesity compared to normal weight (BMI<25)

Results after exclusion of likely under- and overreporters of energy were entirely in agreement with those described for the linear regression analyses.

We also conducted regression analysis on the change in BMI between age 20 years and baseline as the dependent and the dietary variables at baseline as independent variables, considering that the change in BMI during adult life is the result of eating habits during adult life, which are assessed by the baseline dietary questionnaire. In men, inverse associations between the exposure variables and BMI were found. One additional gram of fibre per day was associated with a 0.05 kg m−2 smaller BMI increase in the period between the age of 20 years and the baseline measurement at age 55–69 years. The direction and magnitude of the regression coefficients were comparable to the results shown in Table 2. Only in women, the estimates for the fibre density variables deviated from those in the cross-sectional analyses and were −0.01 (−0.23, 0.22) and 0.01 (−0.25, 0.26) for fibre and fibre from grain (g MJ−1), respectively.

In the subjects with repeated measurements, the change (s.d.) in body weight between the repeated measurement and the baseline was 0.1 (2.3) kg and 0 (1.7) kg per year for men and women, respectively. This corresponded to a change in BMI per year of 0.1 (0.8) kg m−2 in men and 0.0 (0.8) kg m−2 in women. Between the two measurements, energy intake decreased by 312 (1575) kJ in men and 213 (1325) kJ in women. Fibre intake did not change on average, whereas vegetable consumption decreased and fruit consumption increased. No association was observed between the baseline values of the exposure variables and the change in BMI between repeated measurements and baseline (data not shown). Changes in fibre intake were not associated with a change in BMI either.

Discussion

We investigated the association between the consumption of whole-grain foods and dietary fibre and BMI in the population of the Netherlands Cohort Study. The cross-sectional analyses in the data collected at baseline indicated that men with a relatively high intake of whole-grain products or fibre had a lower BMI and less overweight and obesity. In women, an inverse association was observed for whole grain only and a positive association for brown bread.

Overall, the results for men are more consistent than those for women. In men, the grain and, less pronounced, fibre variables except total brown bread, showed an inverse association with BMI. Surprisingly, in women, the fibre density variables tended to be positively associated with BMI. It is known that obese persons tend to underreport energy intake and overestimate healthy behaviour (Lichtman et al., 1992) and nutrient density (Livingstone and Black, 2003). It is therefore possible that women more often than men underreport the energy-containing foods and possibly overreport fibre intake. From Table 1, it can be seen that overweight and obese women indeed reported lower energy intake. However, the results of the analyses excluding the most likely under- and overreporters of energy were very similar, in particular for the whole-grain variables. The exclusion did have an effect in the anticipated direction on the association between energy intake and BMI in the multivariate models: in both sexes, a positive association was observed, whereas the direction of the association was inverse for women in the complete group, indicating differential underreporting of energy intake.

Although differential underreporting can bias the results of our analysis, in particular for dietary fibre and fibre density, we think that the results for whole-grain foods were much less susceptible to such biases. The assessment of consumption of whole-grain foods was more qualitative than quantitative and considering the special nature of these foods, it seems very unlikely that overweight persons would underreport such foods. This theory is supported by the results of the sensitivity analysis excluding under- and overreporters. We think therefore that the inverse association between whole-grain consumption and BMI is real.

The prospective analysis that we performed was hampered by the relatively small time interval (varying between 1 and 5 years) and therefore small changes in BMI between the two measurements. As weight measurements are subject to short-term variations and self-report increases inaccuracy even more, the measurement error of the change in BMI is large compared to the true changes in BMI, resulting in a small signal-to-noise ratio. This may well explain the absence of a longitudinal association between grain variables and change in BMI.

A limitation of the study is the self-reported height and weight, resulting in non-differential and probably also differential misclassification according to BMI. Both would result in underestimation of the true association. The same is true for the majority of studies published to date.

Although we adjusted in the analysis for many potential confounders, we cannot exclude the possibility of residual confounding, in particular by determinants of a healthy lifestyle or attitudes that are not captured by the measured covariates. This is a drawback of all observational studies, which can only be overcome by randomized controlled intervention studies.

In the tables, we have expressed the change in BMI and the risk of being overweight or obese per increment of 1 g of grains or fibre. Men with a 10-g higher intake of dry grains have a 0.34 kg m−2 lower BMI. For fibre and fibre from grains, men with a 10-g higher intake have a 0.38 kg m−2 lower BMI. This seem to be relatively weak associations. However, the logistic regression demonstrated that the risk of being obese as compared to normal weight was 10% lower for each additional gram of (dry) whole-grain consumption. In women, this risk was 4% lower. One serving of cooked grain (1/2 cup) per day, corresponding to 1 ounce (=28.3 g) dry grain according to the USDA Food Guide Pyramid (www.mypyramid.com), corresponds to ORs (95% CI) for overweight of 0.51 (0.36–0.73) in men and 0.55 (0.36–0.83) in women; ORs (95% CI) for obesity are 0.06 (0.01–0.53) in men and 0.32 (0.12–0.85) in women for the same amount of whole grain.

A limited number of studies, mostly from North America, have investigated the association between intake of whole-grain products in relation to BMI or body weight. In two cross-sectional studies an inverse association was observed between whole-grain intake and measured BMI (McKeown et al., 2002; Steffen et al., 2003). In large prospective studies, a consistent inverse association between whole-grain consumption and self-reported weight gain was reported in women (Liu et al., 2003) and men (Koh-Banerjee et al., 2004). In a prospective study among male physicians, consumption of breakfast cereals was inversely associated with self-reported weight gain (Bazzano et al., 2005). These results correspond to our cross-sectional findings in both men and women.

With respect to dietary fibre, we found an inverse association between fibre intake and BMI in men only. Slavin (2005) summarized the evidence on the association between dietary fibre intake and body weight. Most cross-sectional studies found an inverse association between fibre intake and body weight, which was confirmed by the few available longitudinal studies (Liu et al., 2003; Ludwig et al., 1999). More recent cross-sectional studies not included in the review by Slavin and conducted in France (Lairon et al., 2005) and Spain (Bes-Rastrollo et al., 2006) also observed inverse associations between fibre intake and BMI.

Whole grains might help in preventing overweight and obesity due to their effect on satiety, although evidence for this effect was considered insufficient for fibre (Institute of Medicine, 2002). Another potential mechanism is to slow down starch digestion and absorption, which leads to lower glucose and insulin responses that favour oxidation and lipolysis of fat rather than its storage (Slavin et al., 1999). Diets rich in whole-grain products have a low glycaemic index (GI) (Jenkins et al., 1986, 1988). Several studies have investigated the effect of low compared to high glycaemic diets on weight gain or reduction (reviewed in Ludwig, 2002). Short-term studies in humans showed lower satiety, increased hunger or higher voluntary food intake in subjects consuming high compared to low GI meals (Ludwig, 2002; Roberts et al., 2002). Weight loss in obese hyperinsulinaemic women was higher in those consuming an energy restricted low compared to high glycaemic diet for 12 weeks (Slabber et al., 1994), and in men lower adiposity was found after consumption of an energy-restricted low glycaemic diet for 5 weeks (Bouche et al., 2002). Although these were all short-term findings in mostly obese subjects, this mechanism may play a role in normal subjects to maintain normal weight over a longer period.

The lack of inverse associations between brown bread consumption and BMI in our study agrees with the low-glycaemic load explanation. Wholemeal bread does not appear to have a lower GI than white bread (GI of 70 relative to glucose), despite its much higher fibre content (Foster-Powell et al., 2002). This is due to the absence of whole-grain kernels in bread (Jenkins et al., 1988); actually, in the majority of Dutch wholemeal or brown breads, the fibre-rich fraction is added back to the white flour after removal of the germs to prolong shelf life. The whole-grain products eaten by our population, in contrast, have a GI varying between 40 and 60. As brown bread contributes substantially to the fibre intake in this population, this may also explain the deviating associations observed for fibre intake in this population.

In conclusion, the results of this study in a healthy middle-aged population in the Netherlands indicate that men and women with a high intake of whole grains have a lower BMI and a lower risk of overweight and obesity than men or women with a low intake of whole grains. The cross-sectional design of the study does not allow conclusions about the causality of the associations. Nevertheless, the consistency of the association between whole-grain consumption and BMI and its biological plausibility are in line with a causal association. Intervention studies are needed to find out whether consumption of whole grain as such decreases the risk of becoming overweight. The results for dietary fibre and wholemeal bread were less clear. This may be due to methodological problems in dietary and BMI assessments, residual confounding, physiological reasons or a combination of these factors.

References

  1. Anonymous (1998). Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults.The Evidence Report. National Institutes of Health, National Heart, Lung and Blood Institute: Bethesda, MD.

  2. Anonymous (2000). Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser 894.

  3. , , , , , (2005). Dietary intake of whole and refined grain breakfast cereals and weight gain in men. Obes Res 13, 1952–1960.

  4. , , , , (2006). Association of fiber intake and fruit/vegetable consumption with weight gain in a Mediterranean population. Nutrition 22, 504–511.

  5. , , , , , et al. (2002). Five-week, low-glycemic index diet decreases total fat mass and improves plasma lipid profile in moderately overweight nondiabetic men. Diabetes Care 25, 822–828.

  6. , , (2002). International table of glycemic index and glycemic load values: 2002. Am J Clin Nutr 76, 5–56.

  7. , , , , , et al. (1994). Validation of a dietary questionnaire used in a large-scale prospective cohort study on diet and cancer. Eur J Clin Nutr 48, 253–265.

  8. , , , , , et al. (1995). Reproducibility of a food frequency questionnaire and stability of dietary habits determined from five annually repeated measurements. Eur J Clin Nutr 49, 420–429.

  9. , , , (2005). Effect of screening out implausible energy intake reports on relationships between diet and BMI. Obes Res 13, 1205–1217.

  10. Institute of Medicine (2002). Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein and Amino Acids. National Academy Press: Washington.

  11. , (2004). Whole grain intake and cardiovascular disease: a review. Curr Atheroscler Rep 6, 415–423.

  12. , , , (1998). Whole-grain intake and cancer: an expanded review and meta- analysis. Nutr Cancer 30, 85–96.

  13. , , , , , et al. (1988). Wholemeal versus wholegrain breads: proportion of whole or cracked grain and the glycaemic response. BMJ 297, 958–960.

  14. , , , , , et al. (1986). Low glycemic response to traditionally processed wheat and rye products: bulgur and pumpernickel bread. Am J Clin Nutr 43, 516–520.

  15. , (2003). Whole grain consumption and weight gain: a review of the epidemiological evidence, potential mechanisms and opportunities for future research. Proc Nutr Soc 62, 25–29.

  16. , , , , , et al. (2004). Changes in whole-grain, bran, and cereal fiber consumption in relation to 8-y weight gain among men. Am J Clin Nutr 80, 1237–1245.

  17. , , , , , et al. (2005). Dietary fiber intake and risk factors for cardiovascular disease in French adults. Am J Clin Nutr 82, 1185–1194.

  18. , , , , , et al. (1992). Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med 327, 1893–1898.

  19. , , , , , (2003). 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 78, 920–927.

  20. , (2003). Markers of the validity of reported energy intake. J Nutr 133 (Suppl 3), 895S–920S.

  21. , , , , , et al. (1999). Dietary fiber, weight gain, and cardiovascular disease risk factors in young adults. JAMA 282, 1539–1546.

  22. (2002). The glycemic index: physiological mechanisms relating to obesity, diabetes, and cardiovascular disease. JAMA 287, 2414–2423.

  23. (2004). Whole grain intake and insulin sensitivity: evidence from observational studies. Nutr Rev 62, 286–291.

  24. , , , , (2002). 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 76, 390–398.

  25. , , , (2006). Whole-grain and fiber intakes and periodontitis risk in men. Am J Clin Nutr 83, 1395–1400.

  26. RIVM 2004. Volksgezondheid Toekomst Verkenning, Nationaal Kompas Volksgezondheid (versie 2.7). 19 2004.

  27. , , (2002). The influence of dietary composition on energy intake and body weight. J Am Coll Nutr 21, 140–145.

  28. , , , , (1994). Effects of a low-insulin-response, energy-restricted diet on weight loss and plasma insulin concentrations in hyperinsulinemic obese females. Am J Clin Nutr 60, 48–53.

  29. (2005). Dietary fiber and body weight. Nutrition 21, 411–418.

  30. , , , (1999). Plausible mechanisms for the protectiveness of whole grains. Am J Clin Nutr 70, 459S–463S.

  31. , , , , , et al. (2003). Whole grain intake is associated with lower body mass and greater insulin sensitivity among adolescents. Am J Epidemiol 158, 243–250.

  32. , , , , , (1990). A large-scale prospective cohort study on diet and cancer in the Netherlands. J Clin Epidemiol 43, 285–295.

Download references

Acknowledgements

We are indebted to the participants of this study and further wish to thank C Rubingh for additional data analysis, Dr L Schouten, S van de Crommert, H Brants, J Nelissen, C de Zwart, M Moll, W van Dijk, M Jansen and A Pisters for assistance and D van der Doest, H van Montfort and T van Moergastel for programming assistance. This study was funded by the General Mills. The sponsor did not at any time contribute to the analysis, interpretation and reporting of the data and was not involved in formulation of the conclusions. Publication of the results was contractually agreed a priori.

Author information

Author notes

    • L P L van de Vijver

    Current address: Louis Bolk Institute, Driebergen, the Netherlands.

    • R A Goldbohm

    Current address: Department of Prevention and Health, TNO Quality of Life, Leiden, the Netherlands.

Affiliations

  1. Department of Food and Chemical Risk Analysis, TNO Quality of Life, Zeist, the Netherlands

    • L P L van de Vijver
    • , L M C van den Bosch
    •  & R A Goldbohm
  2. Department of Epidemiology, Maastricht University, Maastricht, the Netherlands

    • P A van den Brandt

Authors

  1. Search for L P L van de Vijver in:

  2. Search for L M C van den Bosch in:

  3. Search for P A van den Brandt in:

  4. Search for R A Goldbohm in:

Corresponding author

Correspondence to R A Goldbohm.

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/sj.ejcn.1602895

Further reading