Whole grain, fiber and health

Whole grain, fiber, and health David Jacobs, PhD Professor of Epidemiology, School of Public Health, University of Minnesota Guest Professor, Department of Nutrition, University of Oslo University of Class lecture September 23, 2005 University of Oslo, Norway What is a whole cereal grain? Go to Fulcher slide show Fiber amounts in different whole cereal grains Fiber per 100 g of grain wheat 12 g oats 10.6 g rye 14.6 g corn 7.3 g brown rice 3.5 g refined wheat <2 g white rice <2 g How can we tell whether whole grain consumption is good for health? 1. Epidemiology 1. 2. 3. 4. Cross-section Case-control Prospective Long term clinical trial 2. Feeding study Overview of Study Design, 1 • Cross-section – – – – Everyone sampled at one time Can’t define whether exposure or disease comes first (temporality) Any disease has already happened, so disease can cause changes in exposure Effect measures: risk difference, relative prevalence Overview of Study Design, 2 • Case-control – – – – Sample separately from cases and from controls Can’t define whether exposure or disease comes first (temporality) Any disease has already happened, so disease can cause changes in exposure Effect measures: prevalence odds ratio (approximately relative risk if disease is rare) Overview of Study Design, 3 • Prospective – – – – – Cross-sectional sample excluding prevalent disease Exposure stated before disease is known (temporality well-defined) Disease has not happened, so it can not cause changes in exposure As in cross-section and case-control, level of nutritional exposure is naturally occurring and correlated with demographics and other behaviors Effect measures: risk difference or relative risk or relative hazard (if time to event is considered) Overview of Study Design, 4 • Long term clinical trial – Cross-sectional sample excluding prevalent disease, randomly assigned to different nutritional exposures Temporality well-defined Assigned nutrition exposure uncorrelated with other factors, but other nutrition exposures are naturally occurring and correlated with demographics and other behaviors Effect measures: risk difference or relative risk or relative hazard (if time to event is considered) – – – Overview of Study Design, 5 • Long term clinical trial (continued) – Very difficult to carry out over years – Successful examples include Lyon Diet Heart Study and PREDIMED (in recruitment phase) Overview of Study Design, 6 • Feeding study – Short term (hours, days or weeks) – Exposure assigned randomly – All food provided or supplemental food only provided – Excellent for study short term effects such as changes in body weight or plasma insulin (disease intermediaries) How do we know what people are eating? • Epidemiology – Ask questions about diet • Short term recall (detailed, not representative of long term) • Long term frequency (intuitive, less detailed, representative, guess work) • Study what people actually eat • Feeding – Provide all or some food – Study what the researcher is interested in Common epidemiologic representations of whole grain • • • • • • • Dark bread Brown bread Whole wheat bread Brown rice Usual breakfast cereal Hot cereal Must ask about food in terms the participant understands, so if whole grain is not understood, the participant can’t tell you about it Norwegian example using bread (1) how many slices of bread do you usually eat per day? (<2 slices per day; 2 ± 4; 5 ± 6; 7 ± 8; 9 ± 12; 13+) (2) what kind of bread do you eat most often? (store bought; home-baked) (3) If you buy, what type most often? (white bread, fine or light bread, whole grain or dark bread) (4) If you bake at home, what proportion of the flour is whole grain (dark)? (don't use whole grain flour; <1/4 whole grain flour; 1/4- 1/2 whole grain flour; more than 1/2 whole grain flour). Based on a 24 hour recall, store bought whole grain bread was overestimated and used a recipe with less whole grain Whole grain bread score: the number of slices of bread eaten per day times the proportion of whole grain flour. American example using breakfast cereal Ask usual breakfast cereal Refer to coding list for whole grain content of several hundred brand name products Does not account for people eating multiple products American example using bread The term whole grain bread is not well known and is confused with organic (which is irrelevant) and multigrain (which be all refined) Dark bread in US and brown bread in UK identify most whole grain breads, but misclassifies breads containing dark color, such as molasses Whole grain and cancer: casecontrol experience Chatenoud L, Tavani A, La Vecchia C, Jacobs DR Jr, Negri E, Levi F, Franceschi S. Whole grain food intake and cancer risk. Int J Cancer. 1998 Jul 3;77(1):24-8. Jacobs DR Jr, Marquart L, Slavin J, Kushi LH. Whole-grain intake and cancer: an expanded review and meta-analysis. Nutr Cancer. 1998;30(2):85-96. Review. Meta-analysis of case-control studies of whole grain food intake and cancer, 1998 Cancer Site Studies Reduced Risk? Pooled Odds Ratio Gonadotrophic hormone-related Prostate 1 1 0.9 Breast, endometrium, ovary 6 6 0.85 Aerodigestive Upper aerodigestive 15 14 0.6 Colon, rectum, and polyps 13 10 0.8 Pancreas 4 4 0.7 Other Brain 3 2 0.7 Lymphoma 3 3 0.5 Soft tissue sarcoma 2 2 0.3 Bladder 1 1 0.5 Myeloma 1 1 0.5 Liver 1 1 0.6 Thyroid 1 1 0.6 Whole grain and heart disease/type 2 diabetes: prospective study experience Jacobs DR Jr, Gallaher DD. Whole grain intake and cardiovascular disease: a review. Curr Atheroscler Rep. 2004 Nov;6(6):415-23. Review. Whole grain confounding with other behaviors Iowa Women’s Health Study 34,492 postmenopausal Iowa women, 1986; levels adjusted for age and energy intake Whole grain intake Refined grain intake Quintile Quintile 1 5 1 5 Quintile Quintile Health Behaviors Current smoker, % Vitamin supplement use, % Physical activity % engaging in regular activity Hormone replacement therapy % ever 24.7 53.7 30.3 12.1 68.7 45.5 16.0 70.9 49.5 17.1 54.4 30.0 34.0 42.6 40.4 34.7 34,492 postmenopausal Iowa women, 1986; levels adjusted for age and energy intake Whole Grain Quintile Quintile Quintile Refined Grain Quintile 1 5 1 5 Other Aspects of Diet Fruits and vegetables, no juice mean, servings/week Red meat, mean, servings/week 34.1 39.9 6.6 45.5 5.1 28.6 6.0 42.8 5.4 Keys score, mean, mg/dl 41.0 Sucrose, mean, g/day Constituent Nutrients Dietary fiber, mean, g/day Vitamin E, mean, IU/day 9.0 Folate, mean, µg/day 45.7 38.9 42.2 16.3 8.9 270.7 38.8 22.3 10.1 332.3 37.6 21.7 10.2 342.1 47.3 17.1 264.7 Whole grain and cancer – Norwegian and Iowa prospective studies Unpublished talk give in Montreal, American Association of Cereal Chemists, 2002 Findings much less clear than in case-control studies Same questions and design as for heart disease and diabetes Iowa Women’s Health Study: Fiber in 2 groups with similar total grain fiber 71% grain fiber from whole 77% grain fiber from refined grain 3.6 - 6 g / 2000 kcal whole fiber grain <3.6 g / 2000 kcal whole fiber <3.6 g / 2000 kcal refined fiber 3.6 g / 2000 kcal refined fiber 7481 3559 4.7 ± 0.7 1.9 ± 0.8 1.3 ± 1.0 4.5 ± 0.9 Sample size Whole grain fiber Refined grain fiber Total grain fiber Total dietary fiber 6.6 ± 1.0 23.3 ± 5.3 5.8 ± 1.2 19.2 ± 4.5 Jacobs DR, Pereira MA, Meyer KA, Kushi LH. J Amer Col Nutr 2000 Iowa Women’s Health Study: 11-year relative risk of mortality in 11040 women: cereal fiber in itself does not predict death Hazard Rate Ratio (95% CI) in 71% whole grain fiber eaters relative to 77% refined grain fiber eaters Total mortality Coronary heart disease All cancer All other mortality 0.83 (0.73, 0.94) 0.89 (0.66,1.20) 0.92 (0.75,1.12) 0.83 (0.71, 0.97) Which direction from here? Problems with whole grain theory, 1 • Supported by selected feeding • • • • • and supplement studies Reliability Validity Residual confounding A synergy model is consistent with the complexity of biology In vitro studies also suggest synergy Which direction from here? Problems with whole grain theory, 2 • • • • Fiber not the whole answer Myriad phytochemicals coexist with fiber in all plant foods These include enzymes, pesticides, antioxidant defense, signal transducers Though the nutrient model has some value, it seems more helpful to think of foods and food combinations than of nutrients. Other plant foods are broadly similar – fruit, vegetables, nut, legumes, spices Plant-based food patterns are more reliably ascertained epidemiologically and less prone to dietary confounding, inverse to many chronic diseases • • Nutrients vs Foods Fiber is probably not as active as its accompanying phytochemicals The value of carbohydrate depends on whether it is fiber rich (with accompanying phytochemicals) or fiber poor The value of fat depends on saturation and location of the double bond if unsaturated Some single nutrient conditions •A single nutrient deficiency disease would be alleviated by introduction of an isolated nutrient in the form of a supplement – Scurvy and vitamin C – Pellagra, beri-beri and B-vitamins – Rickets and vitamin D – Neural tube defects and folate – Trans fatty acids, increased cholesterol and coronary heart disease – Hypertension and salt Nutrition in the etiology of disease •Focus on food and the synergy of its components may be a method for better understanding nutrition and disease etiology – Reductionism looks for simple, biochemically-based associations – Some diseases are the result of deficiency or excess of single nutrients – Food synergy is defined as additive or more than additive influences of foods and food constituents on health – Chronic diseases such as atherosclerosis, ischemic heart disease, and cancers, are complex with multiple etiologies and not simple deficiencies The food synergy approach A hierarchical structure of dietary patterns, foods, and nutrients for study of food synergy. “Top down” research begins with the higher levels, searching for combinations of foods and their constituents that influence health “Bottom up” research begins at the lowest level, searching for individual constituents that influence health. Examples of dietary component at each synergy level Food synergy level Level 5: Dietary pattern Level 4: Food groups Level 3: Whole grain Level 2: Whole wheat ‘Prudent diet’, ‘Western diet’, other combinations of food groups Whole grain, dairy, fruit, vegetables, meat Whole wheat, brown rice, rolled oats Bran, germ, endosperm; extract of fat soluble portion Level 1: Bran or a Specific nutrients or phytochemicals single phytochemical Maximize nutritional value per bite • We are increasingly inactive – We are getting fat from positive energy balance – Even so, energy intake is less than it used to be – We cannot afford to waste bites eaten food with low nutritional value, such as nutrient-poor carbohydrate – Eat a varied diet rich in plant foods