PHYSIOLOGICAL EFFECTS AND HEALTH CONSEQUENCES OF DIETARY FIBER

June 1987

Prepared for

CENTER FOR FOOD SAFETY AND APPLIED NUTRITION FOOD AND DRUG ADMINISTRATION

DEPARTMENT OF HEALTH AND HUMAN SERVICES WASHINGTON, D.C. 20204

under

Contract No. FDA 223-84-2059

LIFE SCIENCES RESEARCH OFFICE FEDERATION OF AMERICAN SOCIBTIBS

FOR EXPERIMENTAL BIOLOGY 9650 Rockville Pike

Bethesda, Maryland 20814

0

PHYSIOLOGICAL EFFECTS AND HEALTH CONSEQUENCES OF DIETARY FIBER

June 1987

Prepared for

CENTER FOR FOOD SAFETY AND APPLIED NUTRITION FOOD AND DRUG ADMINISTRATION

DEPARTMENT OF HEALTH AND HUMAN SERVICES WASHINGTON, D.C. 20204

under

Contract Number FDA 223-84-2059

edited by

Susan M. Pilch, Ph.D.

LIFE SCIENCES RESEARCH OFFICE FEDERATION OF AMERICAN SOCIETIES

FOR EXPERIMENTAL BIOLOGY 9650 Rockville Pike

Bethesda, Maryland 20814

11

\.

0

FOREWORD

The Life Sciences Research Office (LSRO), Federation of American Societies for Experimental Biology (FASEB), provides scientific assessments of topics in the biomedical sciences. Reports are based upon literature reviews and the scientific opinions of knowledgeable investigators engaged in work in specific areas of biology and medicine.

This report was developed for the Center for Food Safety and Applied Nutrition, Food and Drug Administration (FDA) in accordance with the provisions of Contract No. FDA 223-84-2059. It was prepared and edited by Susan M. Pilch, Ph.D., Senior Staff Scientist, LSRO, FASEB, under the direction of the ad hoc Expert Panel on Dietary Fiber. Scientists selected as members of the Panel were chosen for their qualifications, experience, and judg­ment, with due consideration for balance and breadth in appro­priate professional disciplines. Members of the Panel and others who assisted in the preparation of this report are listed in Section XI. The Panel and LSRO staff acknowledge the cooperation of scientific staff of the Center for Food Safety and Applied Nutrition, FDA, who identified the scope of the Agency's questions in regard to the physiological and health effects of dietary fiber.

The Expert Panel met three times between January 1986 and December 1986 to obtain background information, identify pertinent issues related to the interpretation of data on dietary fiber, and develop drafts of the report. Members of the Expert Panel reviewed each draft of the report and provided additional documentation and viewpoints for incorporation into the final report. The Expert Panel and LSRO accept responsibility for the study conclusions and accuracy of the report; however, the list­ing of these individuals in Section XI does not imply that they specifically endorse all statements in the report.

The final report was reviewed and approved by the LSRO Advisory Committee (which consists of representatives of each constituent Society of FASEB) under authority delegated by the Executive Committee of the Federation Board. Upon completion of these review procedures, the report was approved by the Executive Director, FASEB and transmitted to FDA.

While this is a report of the Federation of American Societies for Experimental Biology, it does not necessarily reflect the opinion of each individual member of the FASEB constituent Societies.

Date

iii

Kenneth D. Fisher, Ph.D. Director Life Sciences Research Office

EXECUTIVE SUMMARY

This report reviews the physiological effects and health consequences of consumption of dietary fiber components. It was prepared by an Expert Panel of the Life Sciences Research Office for the Food and Drug Administration and evaluates the scientific evidence regarding possible health benefits and risks of dietary fiber intake.

The Expert Panel used the generally accepted definition of dietary fiber as the endogenous components of plant materials in the diet which are resistant to digestion by enzymes produced by man. Analytically, dietary fiber includes predominantly non­starch polysaccharides and lignin. In addition, the analytical definition encompasses a variety of isolated polysaccharide-rich, plant-derived products, which may or may not be chemically modi­fied, including brans, pectins, and gums. Because dietary fiber has diverse chemical constituents, no one analytical method is entirely satisfactory for quantifying dietary fiber from all sources. A satisfactory analytical method for dietary fiber should measure all the nonstarch polysaccharides and lignin; characterization of individual polysaccharides is not currently realistic for a routine method. Methods that separate the cellulosic and soluble and insoluble noncellulosic fractions with characterization of the monosaccharide, mixed-linkage S-glucan, and uronic acid components are available and do provide useful compositional informations For applications in which a single value for total dietary fiber is sought, several methods show promise but need further validation with a wide variety of foods. The heterogeneity of the complex mixture of substances termed dietary fiber and their varied physiological effects can be addressed adequately only by careful comparison of fiber values obtained by different methods and determination of recoveries.

Physical and Ch~mical Properties of Dietary Fiber

The in vitro physical and chemical properties of dietary fibers, such as water-holding capacity, viscosity, ion-exchange capacity, binding of organic compounds, and fermentability by gastrointestinal bacteria, have been examined. Standardized methods for measuring these properties are not available but would enhance their usefulness for characterizing dietary fiber sources. However, the data suggest that these physical and chemical characteristics are generally poor predictors of the in vivo physiological effects of various dietary fiber sources. The physical and chemical properties, as well as the physio­logical effects of dietary fiber; can be modified by particle size and various processing techniques, but these effects have not been studied extensively.

V

Physiological Effects of Dietary Fiber

The physiological effects of fiber-containing foods, fiber supplements, and isolated fiber components occur over the length of the entire alimentary canal and include subsequent systemic effects. Diets containing large amounts of dietary fiber tend to be bulky, require longer times for ingestion, and add effort to eating. Foods rich in dietary fiber also seem to confer greater satiety than foods depleted of dietary fiber, although objective methods for the assessment of satiety have not yet been developed. Results of studies of the effects of dietary fiber on gastric emptying rate have been variable and dependent on the method of measurement used. The viscous fibers (such as guar gum and pectin) have been shown frequently, but not con­sistently, to have marked effects on delaying gastric emptying. Evidence also suggests that certain dietary fibers may affect the digestion and absorption of other nutrients.

In general, there is a reasonably good relationship between the ability of various fiber components to sequester bile acids in vitro and their overall effects on fecal bile acid excretion. Dietary fiber sources such as pectin, oat bran, guar, and psyllium mucilloid have been shown to produce increases in daily fecal bile acid excretion, but wheat bran has no effecto Dietary fiber sources can also influence the concentration of bile acids in the feces and alter the composition of biliary bile acidso The effects of various sources of dietary fiber on serum lipids have been studied extensively. Insoluble fiber sources have generally been found to be ineffective in lowering serum cholesterol levels in normal subjects; preparations with soluble, viscous fiber components have exhibited hypolipidemic effectse Soluble fiber sources also have been found to enhance glucose tolerance and increase insulin sensitivity in normal subjects in short-term studies, but results after longer feeding trials have been equivocal.

Data on small bowei mucosal morphology in humans show geographic variations and developmental changes that may be related to environmental factors, including diet. There is no direct evidence to indicate that consumption of fiber produces these changes or that these changes are pathological. In animal studies, fermentable fibers generally appear to increase intesti­nal weight and length, and most such fiber sources stimulate cell proliferation. Currently, there is no evidence in humans that fiber stimulates small or large bowel epithelial cell growth.

Dietary fiber has substantial effects on colonic func­tion and colonic bacterial activities. Although the composition of the bacterial flora is minimally changed by fiber intake, the microbial mass appears to be increased. Fermentable fiber sources are degraded by the colonic bacteria to gases and short­chain fatty acids. The roles played by short-chain fatty acids in the human colon and the influence of the colonic bacteria on bile acid turnover remain to be elaborat~d. Wheat bran and other

vi

sources of insoluble fiber have been found to decrease gastro­intestinal transit time and increase stool weight over a wide range of doses and by different measurement methods.

Potential Beneficial and Adverse Health Effects of Dietary Fiber

The effects of various dietary fiber sources in the prevention, amelioration, or treatment of specific diseases and disorders have been examined and the Expert Panel 1 s conclusions are summarized below:

Obesity. The limited data from clinical trials sug­gesting that dietary fiber supplements or high-fiber diets are useful for weight reduction are contradictory. When a positive effect has been found, the total weight loss is modest. Long­term follow-up studies of the ability of subjects to maintain the weight loss have not been conducted. Dietary fiber may have a limited role as an adjunct in the treatment of obesity, but con­trolled, long-term clinical trials are needed before this role can be established.

Diabetes. Clinical trials have shown that supplements of soluble fiber sources, such as guar gum, pectin, and oat bran, can be useful in decreasing insulin requirements, improving glycemic control, and lowering serum cholesterol levels in sub­jects with diabetes. High-fiber, high-carbohydrate diets which contain large amounts of complex carbohydrates also have been shown to have similar effects. These effects cannot necessarily be attributed solely to the fiber content of such diets.

Coronary heart disease/hyperlipidemias/hypertension. Epidemiological data on the relationship of dietary fiber to coronary heart disease are inconclusive. Clinical studies on the effects of dietary fiber in hyperlipidemias show that soluble fibers such as pectin, guar gum, locu~t bean gum, oat gum, or psyllium mucilloid reduce significantly serum total cholesterol and low-density-lipoprotein-cholesterol levels with little effect on high-density-lipoprotein-cholesterol levels. Insoluble fibers such as bran or cellulose have essentially no effect. In gen­eral, _clinical studies show a lowering of blood pressure in response to diets with increased amounts of fiber from various sources These trials suffer the same shortcoming as the epidemiological studies in that other components of the diet in addition to fiber differed.

Gallstones. The influence of dietary fiber on gallstone formation and regression in humans is still unresolved. Data suggest that other components of the diet (i.e., sugar, energy, and alcohol) play important roles in lithogenesis and that the association with fiber may be a reflection of changes in the overall diet. The possible effects of wheat bran on the cholesterol saturation of bile deserve further study.

vii

Peptic and duodenal ulcers. Epidemiological studies on dietary fiber intake and ulcer are inconclusive. Clinical evi­dence currently suggests little benefit of high-fiper diets in the treatment of peptic or duodenal ulcer.

Constipation. Many clinical studies have shown that dietary fiber, particularly wheat fiber and other insoluble fiber sources, is useful in the prevention and treatment of constipa­tion.

Irritable bowel syndrome. Evidence at this time sug­gests that patients with irritable bowel syndrome most likely to benefit from an increased intake of dietary fiber (as wheat bran) are those whose chief complaint is constipation. Neither bran nor a high-fiber diet is a panacea; treatment of accompanying anxiety or depression should also be undertaken. Further controlled dietary studies are desirable, particularly with selection of more homogeneous groups of patients.

Diverticular disease. An etiological relationship of low-fiber intake to development of diverticular disease has been suggested by epidemiological data. Clinical evidence suggests that a high-fiber diet, especially one containing wheat bran, may relieve the symptoms of uncomplicated diverticular disease.

Inflammatory bowel disease. Results of available studies are equivocal. Additional studies are required to determine the appropriate diet for and the role of dietary fiber in the treatment of Crohn disease.

Colon neoplasms. There are many plausible mechanisms to explain a relationship between dietary fiber intake and colon cancer risk in humans. Nutritional epidemiological data from ecological (correlational) studies show a fairly consistent inverse relationship between intake of fiber or fiber-containing foods and colon cancer. However, results of case-control studies on the relationship of colon cancer to consumption of dietary fiber or fiber-containing foods are contradictory. Reanalyzing the existing data based on the evolving knowledge about types of dietary fibers (soluble, insoluble, etc.) and gathering more high quality and better validated data about individual diets in observational studies would contribute to understanding of the possible relationships. The metabolic epidemiological studies conducted thus far have revealed that prevalence of fecal mutagen excretion and the concentrations of fecal secondary bile acids are higher in populations at high risk for the development of colon cancer than in populations at low risk. Dietary fiber, particularly whole-grain cereals and bread, may reduce the production and/or excretion of fecal mutagens and decrease the concentrations of secondary bile acids that seemingly play a role in colon carcinogenesis. Further research is needed to elucidate the role of fecal mutagens in colon cancer and the influence of various dietary· fibers.

viii

Studies of dietary fiber and colon carcinogenesis in animal models suggest that the inhibitory or enhancing effects depend in part on the type of fiber. Wheat bran appears to inhibit colon tumor development in animal models more consis­tently than other fiber sources. The varied conclusions on the other types of fiber appear to be due to differences in the experimental protocols. The experimental animal models are highly sensitive to factors such as dietary change and the timing, dose and nature of the carcinogen. The relevance of these animal models to cancer in humans requires further critical study.

Potential deleterious effects. Possible deleterious effects of dietary fiber were considered as well. Evidence that fiber has an adverse effect on mineral bioavailability is con­flicting. Although some test meals result in decreased absorp­tion of iron and zinc due to fiber, the actual amount of decrease appears small in relation to the total daily intake of these minerals. Some reports indicate that fecal mineral content was increased by fiber in the diet, but others showed no effect. Results of balance studies showed that with high intakes of wheat fiber, mineral balances were decreased or negative. The addition of cellulose resulted in negative balances or had no effect. Diets containing large amounts of fruits and vegetables resulted in no effect on mineral balances unless oxalic acid in spinach was included. Locust bean gum, karaya, carboxymethylcellulose, and pectin did not affect mineral balances. Given the possibil­ity that there may be adaptation to any alteration in mineral availability caused by an increased fiber intake, a moderate level of fiber intake does not appear to pose a problem with respect to mineral balance. Isolated cases of other deleterious effects have been reported with the consumption of large amounts of purified fiber sources.

Intake and Recommendations

The Expert Panel also reviewed estimates of dietary fiber intake in the United States and offered recommendations for consumption of foods containing dietary fiber by the healthy, adult population of the United States. These recommendations are based on the collective experience and judgment of the Expert Panel, as well as their review of the physiological effects and potential beneficial and adverse consequences of dietary fiber consumption, and should be considered in the context of the caveats outlined in Section VIII of this report. The Expert Panel recommended the consumption of a wide variety of whole­grain products, fruits, and vegetables, leading to a dietary fiber intake range of 20=35 g/d (10=13 g/1000 kcal) for the healthy, adult population. The Panel emphasized that this range of intakes may not be appropriate for children, the elderly, or persons consuming special diets.

ix

TABLE OF CONTENTS

Foreword

Executive Summary

I Introduction

Background e, 0 0 0 0 GOG O 0 SO G GIG GOO 8 G IJ 9 0 e O CG 8 G O GOG

B Scope of study ·••o••o•oo••··••o••

II. Dietary Fiber Terminology ... o e o. o ••••••••••••••••

Page

iii

V

1

1

2

5

III. Analytical Methodology for Measuring Dietary Fiber ·••o••oo••o••····················· 11

Ae Evaluation and comparison of methods ........ 11

1. Gravimetric methods .................... 11

2. Chemical methods ·······••o••······ .... 13

3 Comparison of methods 14

B. Conclusions 15

IV. In vitro Physical and Chemical Properties of Dietary Fiber .................................. 19

A. Water-holding capacity ·••o••····••o••····~·. 19

B. Viscosity eeecooeooeeooeeeGt!tOOGOOG000G900GGGG 20

C. Ion-exchange capacity ....................... 21

D. Binding of organic compounds ................ 22

E. Bacterial fermentation ...................... 23

F. Modulation of effects by particle size ···••o 24

G. Modulation of effects by processing ··••ee••o 24

xi

Page

V. Effects of Dietary Fiber on Normal

A.

B.

C.

Physiological Functions . . . . . . . . . . . . . . . .. . . . . . . . . 27

Chewing

Satiety

Gastric

and salivation ...................... .

emptying ············•·e••e••••eeeo••

27

27

28

D. Intestinal digestion and absorption of nutrients .............................. 29

1. Changes in.nutrient balance/fecal excretion ........... e •••••••••• •• 0... 29

2. Proposed mechanisms .................... 33

E. Bile acid excretion ......................... 36

F. Serum lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

G. Postprandial serum glucose and hormone levels . . . . . . . . . . . . . . . . . . . . . . . . 51

H. Intestinal morphology and cell

I.

J.

proliferation • . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

1. Morphology ············••e••············ 55

2. Cell proliferation ..................... 57

3. Relationship of changes in intestinal cell proliferation to the

Colon

Colon

1.

2.

3.

4.

development of disease ............... 58

microflora ............................ 59

function ···········••o••······••o••··· 61

Transit time

Stool weight

62

64

Stool frequency . . . . . . . . . . . . . . . . . . . . . . . . 66

Fecal constituents ..................... 67

VI. Epidemiological Studies and Human Trials of Health Effects of Dietary Fiber .......... 69

xii

Page

A. Strengths and weaknesses of epidemiological studies ................... 69

1. Ecological studies ..................... 70

2. Individual-based studies ............... 71

B. Evaluation of methodologies for human studies .............. e •••••• o •• Q.......... 73

C. Studies of potential beneficial effects of dietary fiber ················••e••••oo• 74

L Weight reduction/control .. e o • • • • • • 74

2e Diabetes OOGGCGOl\i,OGOOOOO&ceeeeeeaeeemeee 79

a. b.

Epidemiological studies . . .. Clinical trials and other

79

human studies .................... 81

3o Hyperlipidemia and cardiovascular disease O CG. GI GI e e. e e e e e El e e e e e & e O Ge Ge O GO 92

ao Epidemiological studies of coronary heart disease . . . . . . . . . 92

b. Trials of lipid-lowering effects eoeoeeooeeoeeeeeooeeeeeoo 94

c Hypertension ..... oo•o••••·•••···· 99

4e Gallstones ............................. 99

5. Gastrointestinal disorders ........ e. a• 104

ae Peptic and duodenal ulcers ........ 104 b. Constipation ···••o•••o•o••········ 104 c. Irritable bowel syndrome .......... 109 d. Diverticular disease ......... a •••• 110 e. Inflammatory bowel disease ........ 116

6a Colon neoplasms ···············••e••···· 118

a. Nutritional epidemiology ······••o• 118 b. Metabolic (biochemical)

epidemiology .................... 126 c. Colon carcinogenesis in

laboratory animals ······••e••··· 129

xiii

7.

8.

Other neoplasms

Other disorders

Page

135

136

De Studies of potential undesirable

1.

2.

effects of dietary fiber ·········••oeeem•• 136

Effects on mineral bioavailability 136

a. Mineral status of vegetarians ..... 136 b. Blood mineral levels .............. 137 c. Mineral availability from

test meals ...................... 138 d. Fecal excretion of minerals ....... 139 e. Results of mineral balance

studies - supplements to self-selected diets ............. 140

f. Results of mineral balance studies - controlled diets ...... 140

g. Summary ........................... 146

Other potential deleterious effects 146

VII. Estimates of Fiber Intake in the United States e•111•e•o••••e••············•·o•e•••e 149

A. Dietary methodologies used to estimate fiber intake ............................. e 149

B. Food composition databases for dietary fiber ............................. 149

C. Estimates of intake ................. e••····· 156

D. Groups at risk of inadequate or excessive intake of dietary fiber ......... 156

E. Additional research needed .................. 156

VIII. Recommendations for Fiber Intake in the United States .................................. 159

A. Introduction 159

B. Food sources of dietary fiber ............... 160

C. Recommended level of intake ................. 162

xiv

IX. Literature Cited

x. Glossary

XL Study Participants

xv

Page

165

231

235

I. INTRODUCTION

A. BACKGROUND

The importance of dietary fiber components in the diet has been known for many years (Trowell, 1978) but has received increased attention in the past decade. Study of the role of dietary fiber in health and disease has been stimulated by the work of Burkitt (1971, 1973a,b) and Painter et al. (1972) as well as the rediscovery of the earlier work of Cleave (1956). These and other investigators hypothesized that the relatively low level of plant fiber in diets consumed by Western populations may predispose these populations to a variety of diseases and dis­orders which differ from the diseases occurring in less developed regions. Interest in this hypothesis has led to a number of laboratory, clinical, and epidemiological studies; suggestions for health benefits of dietary fiber; development of new food products and dietary regimens; calls for guidelines on the fiber content and labeling of food products; and revision of nutri­tional recommendations (Health and Welfare Canada, 1985; National Advisory Committee on Nutrition Education, 1983; National Cancer Institute, 1984; Talbot, 1980; U.S. Department of Agriculture/ U.S. Department of Health and Human Services, 1985). However, dietary fiber intake is only one aspect that must be considered in making dietary recommendations and is difficult to address in isolation from the total diet.

Despite increasing public and scientific interest, several problems have impeded research on the health effects of dietary fiber. One major issue has been the absence of a universally accepted definition of dietary fiber. Most defini­tions encompass a wide variety of compounds with different chemical characteristics and physiological functionse As might be expected from the diverse chemical constituents, no one analytical methodology has been entirely satisfactory for identification and characterization of the many components of dietary fiber from all sources.

Various beneficial health effects have been suggested for dietary fiber, individual components of dietary fiber, and fiber-containing foods. Epidemiological studies and/or clinical trials have been conducted to examine the effects of fiber on glycemic response, lipid metabolism, laxation, diverticular disease, colon cancer, weight loss, and many other conditions. Interpretation of these studies is complicated by differences in the methods used for assessing dietary fiber intake in epidemio­logical studies and differences in the type and level of fiber components and fiber-containing foods used in clinical trials. The same factors also complicate studies of the potential adverse effects of high amounts of dietary fiber, e.g., altered avail­ability of minerals and trace elements, altered absorption of drugs, changes in bowel function, and others.

l

8. SCOPE OF STUDY

For the current study, the Food and Drug Administration (FDA) requested that the Life Sciences Research Office assemble an ad hoc Expert Panel on Dietary Fiber to review knowledge about the benefits and risks associated with dietary fiber intake in the United States. Specific issues the Panel was asked to address in this review include the following:

1. Guidance in the selection of appropriate terminology which is consistent with current scientific research. Terminology in future regulatory activities must be supported by available analytical methods to identify the components or properties of fiber which can be routinely measured and evaluated with regard to their physiological effects.

2. Review of clinical methodologies used to assess the physiological effects of dietary fiber components on human health.

3. Review of data collected during epidemiological surveys and clinical trials. The following items were of par­ticular interest to FDA:

a. The appropriateness of the study design and research tools selected to assess health outcome.

b. What beneficial as well as harmful effects have been linked to dietary fiber intake in these studies? Critical reviews of animal research, particularly that related to toxicology issues, should be included.

c. Which components of dietary fiber are believed to be responsible for the effects noted in 3(b)? How does food processing alter the properties of dietary fiber components? What other components or ingredients in food are believed to alter the effects of dietary fiber?

d. Which population groups are likely to be at risk of inadequate or excessive dietary fiber intake? What are the dietary fiber intake levels?

4. With respect to dietary estimates of dietary fiber intake by U.S. population groups, FDA requested the following:

a. What dietary methodologies have been used to estimate dietary fiber intake?

b. How useful are existing food composition tables· for developing estimates of dietary fiber intake?

2

c. How well can dietary fiber ·intake be estimated using existing dietary intake and food composition data?

d. Recommendations for future research approaches in this area.

5. General recommendations for dietary fiber intake by U.S. population groups:

a. What recommendation does the Expert Panel have for the general, healthy U.S. population in terms of dietary fiber intake?

b. What types of dietary fiber are most beneficial to good health?

c. What food sources of dietary fiber are recommended for inclusion in the diet?

d. What conclusions were reached by the Panel based on their evaluation of available data pertaining to health claims for fiber and fiber components?

In addressing the questions posed by FDA, the Expert Panel has included in its report a summary of current dietary fiber terminology; reviews of the in vitro physical and chemical properties and normal physiological functions of dietary fibers; evaluation of the potential beneficial and deleterious effects of dietary fiber, as well as the methods used to assess these effects; estimates of dietary fiber intake in the United States; and recommendations for dietary fiber intake by the general, healthy U.S. population.

3

II. DIETARY FIBER TERMINOLOGY

In their deliberations, the Expert Panel proposed a definition of dietary fiber similar to the one given by the Canadian Expert Advisory Committee on Dietary Fibre (Health and Welfare Canada, 1985). For the purposes of this study, dietary fiber is defined as the endogenous components of plant materials in the diet which are resistant to digestion by enzymes produced by humans. Analytically, dietary fiber includes predominantly nonstarch polysaccharides and lignin. In addition, the analyt­ical definition encompasses a variety of isolated polysaccharide­rich, plant-derived products, which may or may not be chemically modified, including brans, pectins, and gums. The physical and chemical properties of these isolated materials, as well as physiological responses to them, differ from their properties when present in the original complex matrix of the plant. Table 1 lists the major dietary fiber components from plant cell walls together with other substances which form an integral part of the plant cell wall structure or are associated with it. The definition for dietary fiber given by the Expert Panel excludes fiber-associated substances in the plant cell wall (phytates, cutins, saponins, lectins, proteins, waxes, silicon, and other inorganic constituents), indigestible compounds formed during cooking or processing (nonenzymatic browning products, resistant starch), oligosaccharides and carbohydrate polymers of less than 50 or 60 degrees of polymerization that have not been recovered in dietary fiber analysis (Theander and Westerlund, 1986), and nonplant-derived substances or synthetic carbohydrate polymers (chitin, chitosans, pharmaceutical resins) that may or may not possess some ·of the properties of dietary fiber. Some investi­gators consider that at least some of these substances should be included in the definition of dietary bere Regardless of their inclusion in the definition, these substances are capable of influencing the physiological effects of dietary fiber.

Current descriptions of the physiological functions of dietary fiber frequently classify fibers as soluble or insoluble, depending on ·the plant food or manufactured material and the method of analysis. In general, soluble fibers include gums, pectins, mucilages, and some hemicelluloses. Insoluble fibers generally include cellulose, lignin, and some hemicelluloses. [Pectin may also be detected in the insoluble fraction of some foods (James and Theander, 1981; Neilson and Marlett, 1983).J Both the classification of soluble and insoluble fibers and the classification of constituents into pectin, hemicellulose, and cellulose are operational depending on the extraction methods used. Classification according to the sugar composition can partly overlap the fractions defined by extraction. For example, nonstarch glucose can usually be used as an estimate of cellu­lose, but B-glucans also contain significant amounts of glucose of nonstarch origin. The B-glucans, however, are found in significant amounts only in oat and barley products. Uronic

5

Table 1. Components of dietary fiber obtained from plants.*

Structural components of plant cell wall

Nonstructural components either found naturally or used as food additives

Protein Lipids, (e.g., waxes) Inorganic constituents Lignin Cellulose Noncellulosic**

polysaccharides

l Pectin Gums Mucilages Algal polysaccharides Modified cellulose

* Adapted from Southgate (1982) and Ellis (1985).

Dietary fiber

** The noncellulosic polysaccharides of the plant cell wall include hemicellulose and pectic substances.

6

acids, which are the main components of pectins, also occur in hemicelluloses and gums. These points are illustrated in Tables 2 and 3, which show the fractions of dietary fiber that can be obtained by various extractions and the major chemical constituents of dietary fiber components, respectively.

The composition of dietary fiber varies with the type of plant tissue that is the source of the cell wall material (see Table 4.) The proportions of different dietary fiber constitu­ents found in plant foods depend on the maturity of the plants. In typical plant cell walls, the percentages of cellulose, lignin, and ash are higher, and the percentages of noncellulosic polysaccharides, waxes, and protein are lower in mature than immature plants (Siegal, 1968). The portion of the plant con­sumed and its relative maturity, as well as storage and ripening may influence the dietary fiber composition of plant foods, and food processing techniques may also alter the native fiber composition.

In describing the physiological effects and health consequences of dietary fiber in this report, distinctions are made among fiber-containing foods (e.g., whole grains and grain products, legumes, beans, fruits, and vegetables); fiber-rich supplements (e.g., wheat, oat, corn, and rice brans; bagasse; ispaghula; psyllium; and soy isolates); and isolated or purified fibers (e.g., cellulose, methylcellulose, carboxymethylcellulose, pectins, guar gum, gum arabic, locust bean gum, karaya gum, carrageenan, and lignin). In some cases, the classification of fibers as soluble and insoluble is useful in describing their characteristics and effects.- Terminology based on special properties (such as viscous fibers and fermentable fibers) or on analytical techniques (such as neutral-detergent fiber) is useful for some applications. See the Glossary (Section X) for detailed definitions of terms used in this report. Although the use of such diverse terminology may be undesirable, it is necessary because varied terms are used to describe fiber sources in the scientific literature and the relationships among the various terms and descriptions and function are not fully understood.

7

Table 2. Schematic representation of relationships among fractionations of dietary fiber.*

Other Soluble polysaccharides fiber

Total Nonstarch Noncellulosic Pectin ----------------dietary polysaccharides polysaccharides Some soluble fiber Some insoluble Cl) ----------------Hemicellulose

Insoluble fiber

Cellulose Cellulose -

Lignin Lignin Lignin

* Adapted from Asp and Johansson (1984).

\0

Table 3. Chemical composition of dietary fiber components.

Component

Cellulose

Hemicelluloses

Pectins

Mucilages

Gums

Algal poly­saccharides

Lignin

Major Chemical Constituents

Primary Chains

Glucose

Mannose, galactose, glucose, xylose, arabinose

Galacturonic acid

Galactose-mannose, glucose-mannose, arabinose-xylose, galacturonic acid

Galactose, glucuronic­mannose, galacturonic acid, glucose

Mannose, xylose, glucuronic acid

Sinapryl alcohol, conferyl alcohol, p-coumaryl alcohol

Secondary Chains

A:rabinose, galactose, glucuronic acid

Rh~nnose, arabinose, xylose, fucose

Galactose

Xylose, fucose, galactose

Galactose

Description

Linear polymer with S-(1,4)-linkages; forms crystalline matrix; selective depolymerization yields microcrystalline cellulose (used as emulsifier); derivatization to carboxy­methyl- and methylcelluloses results in polymers with gelling properties; some digestion by colonic bacteria

Primarily B-(1,4)-linked pyranosides; acid and neutral hemicelluloses distinguished by contents of uronic acids; digestion by colonic bacteria

Primarily galacturonic acid b,-(1,4)-0-galacturons] with rhamnose i~sertions; carbonyl groups of primary chain vary in extent of methoxylation; highly methoxylated pectins (naturally occurring) more viscous; extensive digestion by colon bacteria

Complex polysaccharides; emulsification and stabilization properties

Complex polysaccharides; emollient, thickening, and emulsification properties

Vary in uronic acid and sulfate content

Complex, cross-linked phenylpropane polymer; hydrophobic; resistant to degradation by colon bacteria

Table 4. Influence of source on the composition of dietary fiber.*

Major food group

Cereals

Vegetables and fruits

Seeds other than cereals, e.g., seed legumes

Polysaccharide food additives

Type of Tissue Present

Parenchymatous endosperm; seed coats, partially ligni fied

Parenchymatous flesh; partially lignified vascular. tissues; cutinized epidermal tissues

Parenchymatous cotyledons; thickened endospermal walls

Amorphous, soluble, or dispersible

* Adapted from Selvendran (1984)0

10

Major Polymers Present

Noncellulosic polysaccharides, arabinoxylans, S-D-glucans, glucuronoarabinoxylans; cellulose, phenolic esters, lignin

Noncellulosic polysaccharides, pectic substances, xyloglucans, glucuronoxylans; cellulose, lignin, cutin, and waxes

Noncellulosic polysaccharides, pectic substances, xyloglucans, galactomannans; cellulose

Gums; algal polysaccharides, alginates, sulfated galactans; cellulose esters and ethers; modified starches

III. ANALYTICAL METHODOLOGY FOR MEASURING DIETARY FIBER

The diversity of chemical constituents of dietary fiber precludes a single analytical method that is entirely satisfac­tory for measuring dietary fiber from all sources. In general, methods are designed to yield either a single value for "total" dietary fiber or values for the individual constituents of dietary fiber. Methods for dietary fiber determination can be divided into two broad categories:

Gravimetric methods, in which the residue remaining after chemical or enzymatic solubilization of nonfiber components is weighed. Usually the residue is corrected for nitrogen con­tent which is removed incompletely. Although most gravimetric methods yield a single value for total dietary fiber, some recent methods are capable of measuring soluble and insoluble dietary fiber components, the sum of which is considered total dietary fiber.

Chemical methods, in which the neutral sugar constit­uents of fiber polysaccharides are released by acid hydrolysis and measured colorimetrically, by high-pressure liquid chroma­tography (HPLC), or by gas chromatography (GC) following deriva­tization. Uronic acids are determined colorimetrically or as the carbon dioxide released during decarboxylation of the uronic acid moiety. Lignin is usually determined gravimetrically as the residue insoluble in 72% sulfuric acid, that is, Klason lignin

A. EVALUATION AND COMPARISON OF METHODS

1. Gravimetric methods

The earliest method for the determination of fiber was the crude fiber method, developed in 1806 by Einhof (van Thaer, 1809). This method gives an approximate estimate of cellulose plus lignin and was originally designed for the determination of indigestible material in feeds and forages. Because simple alternative methods were lacking, it came into use for human foods as well and has remained an approved method (Association of Official Analytical Chemists, 1984). Values for crude fiber still appear in food composition tables, but the relationship between the crude fiber and dietary fiber content of a single food is variable.

The acid-detergent fiber (ADF) method (Van Soest, 1963) measures essentially the crude lignin and cellulose fraction of plants, but may also include indigestible Maillard products, tannins, animal skin and hair, plastic, and silica. Hemicellu­loses are usually excluded. This method was originally developed for use on animal forages which contain negligible pectic

11

substances. When it was applied to human foods, it became apparent that variable amounts of pectin and starch were retained in the ADF residue (Robertson and Van Soest, 1981).

The neutral-detergent fiber (NDF) method (Van Soest and Wine, 1967) measures gravimetrically a residue containing lignin, ' cellulose, and most, but not all, of the hemicelluloses in a food. Protein and pectin are efficiently removed by this method, but the original procedure requires modification to eliminate starch contamination (AACC Committee on Dietary Fiber, 1981; Robertson and Van Soest, 1977, 1981). Starch is most effectively removed by incubating the NDF residue with an amylase (Marlett and Lee, 1980; Robertson and Van Soest, 1981). The use of hog pancreatic amylase first proposed by Schaller (1976) avoids the possibility that the enzyme may contain fiber-hydrolyzing enzymes, as is the case with bacterial sources.

The difference between the· ADF and NDF methods can be an estimate of the hemicellulose in the insoluble fraction of fiber if the analyses are performed properly. Because the ADF residue contains pectin that is removed by neutral-detergent solution, the difference between the NDF and ADF can be an estimate of the insoluble hemicellulosic fraction only if the ADF analysis is conducted on a NDF residue (Robertson and Van Soest, 1981). This method of estimating the insoluble hemicellulosic fraction has not been validated. A separate determination of lignin may be made on the ADF residue (Robertson and Van Soest, 1981). The main advantages of the detergent methods are their simplicity and rapidity; the main disadvantage is that soluble fiber components are not measured and cannot be easily recovered from the fil­trate. Although the number of comparisons is limited, available data suggest the NDF recovers a fraction of fiber similar in amount and composition to the insoluble fiber fraction recovered by the Theander (1983) or Southgate (1969) method of chemical analysis (Marlett and Chesters, 1985; Neilson and Marlett, 1983).

The gravimetric assessment of total dietary fiber has been the objective of the recent procedure of the Association of Official Analytical Chemists (AOAC) (Prosky et al., 1984, 1985), which is based on the common experience of three groups (Asp et al., 1983; Furda, 1981; Schweizer and Wursch, 1981). The method consists of enzymatic treatment of the sample with a protease and two amylases and gravimetric determination of the residue after correction for protein and ash content. One dis­advantage of the method is that it does not measure the soluble and insoluble fractions of dietary fiber, although a modification (Arrigoni et al., 1984) proposes a dialysis step for the separa­tion of soluble components. The interlaboratory collaborative studies have used primarily grain products and have found repro­ducibility to be poor when the amount of fiber in a sample is low (Prosky et al., 1985). This method shows promise for determina­tion of total dietary fiber, but it needs to be applied to a wide range of foods and validated by comparison with other methods

12

before it can be accepted without reservation. However, for many purposes, a single value for total dietary fiber is less useful than information about the individual components.

Mongeau and Brassard (1986) have recovered the soluble fiber fraction as that material soluble at 100°c and precipit­able with 100% ethanol after enzymatic starch removal and com­bined .this gravimetric determination with the weight of the NDF residue to obtain a value for total dietary fiber~ Several aspects of the method proposed by Mongeau and Brassard (1986) need further study. These include completeness of starch hydrolysis, the modifications of the procedure to extract the soluble fraction which may alter the composition and yield, the use of filtering crucibles of differing porosities depending on the sample being analyzed, and the use of 100% instead of 80% ethanol more commonly used to precipitate polysaccharides. The validity of the assumption that performing the determinations of the soluble and insoluble fiber fractions on separate aliquots of sample yields the same result as performing both analyses on a single aliquot needs to be tested. Comparisons of the fiber values obtained by this method with those generated by the AOAC first-action gravimetric method (Prosky et al., 1985) and the chemical procedure of Englyst et al. (1982a) reveal significant differences in results that will need to be explained by further methodological work (Mongeau and Brassard, 1986).

2 Chemical methods

For the determination of individual components of dietary fiber, the Southgate fractionation procedure (Southgate, 1969) has been used to measure "unavailable carbohydrate." This method involves the enzymatic hydrolysis of starch and the sequential acid hydrolysis of the noncellulosic polysaccharide and cellulosic fractions. Klason lignin is determined from the cellulose fraction. Colorimetric reactions are applied to the soluble and insoluble noncellulosic fractions and to the cellu­lose fraction; the results are expressed as hexoses and pentoses. This method gives satisfactory values for many foods, but the colorimetric determination of sugars is relatively nonspecific and subject to interferences (Southgate and Englyst, 1985). In addition, starch is inadequately removed from many foods with this procedure, leading to erroneously high values of the cellulose fraction. These two sources of error have made it difficult to determine the accuracy of the dietary fiber values in the British Food Composition Tables (Paul and Southgate, 1978), most of which were determined by this method. More recently, Southgate has recommended sugar quantitation by GC and the use of more effective enzymes to remove starch more adequately@

The fractionation procedure of Englyst (1981) is a modification of the Southgate (1969) method in which the poly­saccharides are characterized as cellulose and noncellulosic

13

materials and the constituent sugars are determined by GC. A further modification (Englyst et al., 1982a) also includes a determination of resistant starch. Lignin and other noncarbo­hydrates are not measured in this procedure.

Another fractionation procedure is that of Theander (1983); it includes the use of three methods applied to various residues obtained during fractionation. Neutral sugars are analyzed by GC and uronic acids by a decarboxylation method. Results are usually expressed as soluble and insoluble polysac­charides and lignin (Klason lignin). One notable advantage of this procedure is its excellent starch hydrolysis.

3. Comparison of methods

Most of the chemical approaches to f~ber analysis have several steps in common. All define the soluble fiber fraction operationally, that is, as the carbohydrate material that remains in solution or is filterable after a sample has been treated to remove simple sugars, lipids, and starch. Most recover lignin as the residue remaining after an acid hydrolysis procedure involving 72% sulfuric acid. Because the determination of lignin is not specific and other acid-insoluble material such as tannins, cutins, and Maillard-reacted proteins may be included, Theander (1983) has proposed use of the term "noncarbohydrate residue'' for this fraction. Gravimetric measurement of the acid­detergent fiber residue after oxidation with potassium permanga­nate (Robertson and Van Soest, 1981) is probably a more specific assay for lignin, although comparisons between this and Klason lignin have not been made. Most foods for humans contain very little lignin; thus, some investigators consider that its neg­ligible contribution may be omitted. Comparisons of the Klason lignin recovered from an NDF residue from which most of the pro­tein has been removed with those remaining after acid hydrolysis of insoluble fiber obtained with the Theander or Southgate method, however, supports Theander's conclusi~n that Maillard products are formed when protein is present and are recovered as Klason lignin (Marlett and Chesters, 1985; Neilson and Marlett, 1983). These Maillard products contain some sugars derived from the fiber poly­saccharides. In addition, Klason lignin must be included in order to obtain satisfactory recoveries of starting material (Marlett and Chesters, 1985; Neilson and Marlett, 1983). Even though Klason lignin is not an accurate measure of true lignin, it does contain fiber-derived materials and should be included in the measurement of dietary fiber.

Determination of recoveries, which is not routinely done by most laboratories, also includes measurement of the residual starch in the fiber fraction not removed by enzymatic hydrolysis (Marlett and Chesters, 1985; Neilson and Marlett, 1983). Most procedures do not adequately remove starch from legumes. This starch has been termed "resistant starch" and some have argued

.14

that it should be considered part of the dietary fiber. That conclusion may be premature since the fate of such starch in the mammalian gastrointestinal tract is unknown.

Separation of cellulosic and noncellulosic polysaccha­rides in the insoluble fraction of fiber is also operationally defined; those materials hydrolyzed by 1-2 N sulfuric acid are referred to as noncellulosic fiber components, although this approach has not been validated. Resistant starch, as well as mixed-linkage S-glucans, would likely appear as glucose in the noncellulosic polysaccharide fraction. There may be merit to Theander 1 s suggestion that the glucose in the entire insoluble fraction of fiber should be considered an estimate of cellulose and that further separation of this fraction provides little additional information. Mixed-linkage s-glucans can be measured separately by an enzymatic procedure for the foods in which they are found, i.e., oats and barley (Theander, 1983).

Relatively few comparisons have been made of the compo­sition and yield of fiber obtained with the various gravimetric and chemical methods. Several methods were used to analyze a common group of foods under the sponsorship of the European Economic Communities Committee on Medical Research (James and Theander, 1981). The results of the AOAC gravimetric procedure for total dietary fiber and those of several chemical methods have been compared, although the presentation of data in a graph limits its usefulness (Prosky et al., 1984). Mongeau and Brassard (1986) have compared results of their gravimetric NDF plus soluble method with analyses using the AOAC first-action and Englyst methods. Both these investigators and Marlett and coworkers have applied the chemical ·approach of fiber analysis to residues obtained by gravimetric procedures as a means of validating gravimetric assessment of fiber content (Marlett and Chesters, 1985; Mongeau and Brassard, 1986; Neilson and Marlett, 1983). In general, these comparative studies suggest that many of the methods of fiber analysis obtain similar results for several foods. However, results differ significantly for some foods, and the number and kinds of foods analyzed thus far are very limited. The available data suggest that there are sig­nificant differences in the yield of fiber from some foods by chemical vs. gravimetric methods that need further study before conclusions can be drawn about the validity of many of the methods for dietary fiber analysis currently used.

8. CONCLUSIONS

Analytical methods for dietary fiber should measure all the nonstarch polysaccharides and lignin. Because both physio­logical and physical properties of dietary fiber are related to specific components, a method that provides information about the composition of dietary fiber is desirable for research studies aimed at understanding the mode of action of dietary fiber. (In many studies in the past, physiological effects were related to

15

total dietary fiber or to the amount of fiber-containing foods or supplements.) Characterization of individual polysaccharides is not currently practical for a routine method, but separation of the cellulosic and soluble and insoluble noncellulosic fractions with characterization of the monosaccharide, mixed-linkage s-glucan, and uronic acid components is possible and does provide useful compositional information (Southgate and Englyst, l985)e Thus, a general analysis scheme that includes the steps outlined in Figure 1 is recommended for most purposes. Insufficient evaluation and application to a diverse group of foods limits specification of details of the general approach. Efficient removal of starch, including resistant starch and amylase-lipid complexes (found naturally or formed during food processing), as well as inclusion of Klasen lignin, is essential for accurate results. The Panel recommends that consideration be given to combining such an analysis with measurement of the in vitro physical properties (water-holding and cation exchange capac­ities, in vitro fermentability, and particle size) for the characterization of a fiber source.

As indicated in the above discussion, total dietary fiber may not be adequate for some experimental purposes. For applications in which a single value for total dietary fiber is needed, e.g., for certain labeling purposes, the AOAC method of Prosky et al. (1985) shows promise, as does the procedure of Mongeau and Brassard (1986) which combines NDF with a measure of the soluble fiber fraction. However, both methods need further validation with a wide variety of foods. The heterogeneity of the extraordinarily complex mixture of substances termed dietary fiber can be addressed adequately only by careful comparison of fiber values obtained by different methods and determination of recoveries.

16

FOOD SAMPLE

bwATER DRY SAMPLE

bsuGARS RESIDUE .

b LIPIDS

RESIDUE

~ PROTEINS

RESIDUE

~STARCH

RESIDUE

Potential problems

Avoid heat

Representative sampling

No acid

Gelatinization Pure enzymes Completeness

SOLUBLE FIBER

RESIDUE (INSOLUBLE FIBER)

l

Procedural details

80% Ethanol or methanol

Hexanes, ether, or acetone

Protease - not common

GC or HPLC for neutral sugars Colorimetry for uronic acid

GC or HPLC for neutral sugars Colorimetry for uronic acid

LIGNIN Gravimetric, as material insoluble in 72% H2so4

Figure 1. General scheme for the chemical analysis of dietary fiber

17

IV. IN VITRO PHYSICAL AND CHEMICAL PROPERTIES OF DIETARY FIBER

The physical and chemical properties of dietary fiber necessarily have been examined by in vitro procedures; investi­gators have studied intact food fibers, individual fiber supple­ments, and isolated fiber components. The properties expressed in vitro may not be identical to those found in vivo because of factors such as osmolality of other food components, adsorption, bacterial action, and integrity of the plant cell wall in the gastrointestinal tract. In vitro studies can only serve as an indication of physical properties under specific experimental conditions. Nonetheless, they do define important characteris­tics of fiber not available from analytical methods. The physi­cal and chemical properties of dietary fibers, plus certain factors that can influence them, will be discussed in this sec­tion.

Ae WATER-HOLDING CAPACITY

This property reflects the extent of hydratability of the fiber, either by adsorption of water to the fiber surface or by entrapment of water within the fiber matrix. Studies of the water-holding capacity of various fiber sources have empha­sized the variability of different fibers. In addition, the techniques used for the determination of water-holding capacity have been varied. Methods h~ve included removal of free water from hydrated fiber sources by centrifugation or filtration, equilibration of a fiber suspension with a "probe" substance (for example, dextran blue), dialysis with a solution formulated to simulate the terminal ileum and right colon contents, and measurement of water-holding capacity against an osmotic pres­sure. A reliable method is not yet available for the deter­mination of in vitro water-holding capacity. Development of a standard, validated method would provide a useful tool for further characterization of a fiber's physical properties.

The chemical composition of a dietary fiber is one determinant of its water-holding capacity (McConnell et al., 1974; Stephen and Cummings, 1979). In general, water holding is higher in fiber sources with high levels of free polar sugar residues exposed and is lower in fiber sources with increased intramolecular linking (such as cellulose). Lignin, which is relatively apolar, is substantially less hygroscopic than the polysaccharide fiber components. The pH and electrolyte content of the test solution can also influence water-holding capacity (McConnell et al., 1974).

The physical structure of the fiber source may exert a greater •impact on its water-holding capacity than its chemical composition. Decreasing the particle size of wheat brans and other particulate fiber sources changes their water-holding

19

capacities (Mongeau and Brassard, 1982; Stephen and Cummings, 1979). The method of fiber preparation (i.e., never dried, air dried, or freeze dried) can alter water-holding capacity even though the chemical composition is unchanged (Robertson and Eastwood, 1981). Other influences include the variety and age of the plant source of the fiber material (Robertson et al., 1980). Food processing also might be expected to have a signif­icant impact on water-holding capacity of fibers, although it has been studied inadequately.

The physiological consequences of water-holding capac­ity of fibers measured in vitro are not entirely predictablee Water-holding capacity was originally hypothesized to be directly related to effects on fecal volume, but the work of Stephen and Cummings (1979) and others suggested an inverse relationship. Eastwood et al. (1983) have proposed several reasons for the lack of a direct relationship between water-holding capacity in vitro and fecal bulking. Fermentation by bacteria in the colon may alter the physical and chemical properties responsible for the water-holding capacity of the fibers, or the partitioning of the bound water in different phases of the fiber may be more directly related to physiological functions than is the total amount of water bound. The former suggestion was tested by McBurney et ale (1985) who found alterations in the water-holding capacity of several dietary fiber sources after in vitro fermentation by a human fecal inoculum.

B. VISCOSITY

Viscosity is a property related to water-holding capac­ity. When water is added to certain polysaccharides they swell to form a gel, a semi-rigid jelly-like mass, which holds all of the water present (Eastwood, 1973). The gel-forming or viscous fibers include pectins, gums, and mucilages. The chemical compo­sition is an important determinant of the viscosity of a fiber source. Highly methoxylated pectins are viscous, whereas poorly methoxylated compounds have little gelling capacity. Derivati­zation of cellulose to carboxymethyl- and methylcelluloses pro­duces polymers with gel-forming ability.

Jenkins et al. (1978a) have reported that viscosity may play a role in mediating several of the physiological effects of dietary fiber. Its potential impact in gastric emptying, intes­tinal absorption, blood lipid concentrations, and postprandial glucose levels is discussed in Sections V-C, V-D, V-F, and V-Ge In vitro studies with everted intestinal sacs from rats have shown that various viscous fibers can impair the transport of monosaccharides and neutral amino acids (Elsenhans et al., 1980; Johnson and Gee, 1981).

20

C. ION-EXCHANGE CAPACITY

Although anion-exchange capability has not been detected in dietary fibers, in vitro cation exchange has been documented (McConnell et al., 1974) and is apparently related to the car­boxyl groups of the acidic sugars of polysaccharides (James et al., 1978). This exchange capacity has been demonstrated for monovalent cati6ns, and for calcium, copper, zinc, and iron (Ismail-Beigi et al., 1977a; James et al., 1978; McBurney et al., 1983; McConnell et ale, 1974). In a review of the recent litera­ture on the in vitro binding of minerals by fiber, Kelsay (1986) noted that studies with fiber from wheat, corn, pinto beans, and cereals with different phytate contents, as well as isolated hemicellulose, cellulose, lignin, and.pectin, have indicated that the presence of other ligands can have a significant effect on the binding of minerals by fiber.

Reinhold et al. (1981) found that iron was bound to the NDF (cellulose, hemicelluiose, lignin) of wheat and corn, and the amount bound depended upon the iron concentration, pH, quantity of fiber, and the presence of inhibitors such as citrate, ethyl= enediamine tetraacetic acid (EDTA), ascorbate, phytate, cysteine, and hydrazine sulfate. Binding of iron by ADF (cellulose, lig­nin) was determined largely by binding to cellulose. Kojima et al. (1981) examined the solubilization of iron from cooked pinto beans and also found a pH dependence and inhibition of iron binding by ascorbate and citrate. Tea was found to decrease iron solubility. Lignin and hemicellulose (from psyllium seed) were found to bind iron; cellulose and pectin were less effective (Fernandez and Phillips, 1982). In this experiment, the binding of iron to cellulose and lignin was inhibited by citrate and EDTA, but not by fructose, cysteine, and ascorbate. In studies of isolated wheat bran, iron binding was found to be inhibited by ascorbate, citrate, nitrotriacetic acid, and EDTA (Leigh and Miller, 1983).

Mod et al. (1982) found that calcium, magnesium, and manganese were bound to water- and alkali-soluble hemicelluloses from rice bran. Large amounts of the bound minerals were released by changes in pH and treatment with hemicellulase, pep­sin, and trypsin. Lyon (1984) studied the solubility of calcium, magnesium, zinc, and copper extracted from a variety of cereal products with hydrochloric acid (pH= 1). Neutralizing the acid extract caused precipitation of large amounts of zinc and iron, and moderate amounts of calcium, especially from the cereals with a high phytate content. Addition of the organic ligands EDTA and citric acid prevented precipitation.

The extent of mineral binding by individual fibers becomes increasingly unpredictable as additional variables are introduced in experimental conditions (Kelsay, 1986)9 In addi­tion, other factors should be considered in assessing the possi-

. ble relationships between in vitro vs. in vivo fiber-mineral interactions. The gut digestive processes are unlikely to yield

21

isolated fibers and NDF residues from ingested material and, although the effects of the gastric and small bowel secretions on the cell wall matrix are unknown, the effects are undoubtedly significant. As with water-holding capacity, standardization of the method to estimate cation exchange capacity of dietary fibers , is needed in order to use this property as a means of character­izing the physical properties of fiber. The possible ramifica­tions of in vitro mineral-binding capacity on mineral balance are discussed in Section VI-D.l of this report.

D. BINDING OF ORGANIC COMPOUNDS

In addition to interacting with metal ions, dietary fibers can adsorb a great variety of organic substances, most notably bile acids and neutral sterols. In vitro binding of bile acids to various grains, food fiber sources, and isolated fiber components has been demonstrated repeatedly (e.g., Birkner and Kern, 1974; Eastwood and Hamilton, 1968; Kritchevsky and Story, 1974; Story and Kritchevsky, 1976; Vahouny et al., 1980, 1981) and reviewed by Kay (1982). This binding is believed to be primarily an adsorption phenomenon and is influenced by bile acid structure, pH and osmolality of the test solution, and the chemical and physical forms of the fiber sources. In general, binding is greater at lower pH, is probably hydrophobic (Eastwood and Hamilton, 1968), and appears to be reversible (Eastwood and Mowbray, 1976). Some, although not all, fiber sources display a preference for binding the unconjugated bile acids in vitro, rather than the taurine or glycine conjugates, and may show some specificity for the di- and trihydroxylated bile acid analogues (Story and Kritchevsky, 1976; Vahouny, 1985).

Among the individual fiber components, lignin has most often been suggested to be responsible for observed bile acid adsorption (Eastwood and Hamilton, 1968; Story et al., 1982). Isolated lignin has been shown to adsorb bile salts (Kay et al., 1979; Story and Kritchevsky, 1976). However, as Southgate (1986) notes, the bile acid-binding effects of complex fiber prepara­tions should not be ascribed to defined fractions without con­sidering the possible involvement of associated materials such as saponins. The tertiary structure of the plant cell matrix needs to be considered as well.

The sequestering ability of various types of fiber is not limited to bile acids. Vahouny et al. (1980, 1981) reported adsorption of all components of mixed micelles (cholesterol, lecithin, monoolein, and oleic acid, as well as taurocholate) by a variety of dietary fiber sources. The greatest binding was observed with lignin and guar gum, and the least with cellulose and wheat bran.

22

Other organic compounds, such as drugs, hormones, and amino acids, also have been reported to bind to dietary fibers (Floyd, 1978; Howard et al., 1986; Rotstein et al., 1980; Rubio et al., 1979; Shultz and Howie, 1986). Physiological consequenc­es of this property are discussed in Sections V-E, V-F, and VI-C.

E. BACTERIAL FERMENTATION

Depending upon the chemical and physical nature of the fiber and the bacterial flora, dietary fiber can be extensively degraded in the human colone The largely anaerobic bacterial flora contains enzymes required to digest many plant polysaccha­rides. Typically the resulting di- and monosaccharides (xylose, arabinose, glucose, and galactose) are used as substrates for growth and produce two major types of end products, volatile short-chain fatty acids (acetate, propionate, and butyrate) and gases (carbon dioxide, hydrogen, and methane). Further discus­sion of the physiological significance of fermentation of die­tary fiber by the colon microflora is contained in Section V-Ie

A bacterial flora that can digest cellulose and hemi­cellulose is present in humans (Bryant, 1978; Slavin and Marlett, 1980). These polysaccharides and other fiber components are fermented to varying degrees (Cummings, 1982). Lignin is almost completely resistant to degradation; however, its presence can influence the fermentation of the polysaccharide components. Cellulose can be fermented extensively, but different sources are degraded to varying extents. Cellulose in fruits and vegetables is degraded more than cellulose in wheat bran or purified cellu­losee Highly crystalline celluloses isolated from wood (iee., most of the purified celluloses) are less susceptible to bacte­rial enzymatic action than the celluloses in fruit and vegetable plant cell walls and in many instances remain essentially unfer­mented. Pectin is nearly completely degraded~ The ferment­ability of hemicelluloses is highly dependent on their source.

Clearly, fermentation can alter the characteristics of dietary fibers. McBurney et al. (1985) have studied the water­holding capacity of several fiber sources before and after in vitro fermentation with a human fecal inoculum. Water-hold­ing capacity of the NDF residues after fermentation was lower for fiber from lucerne (alfalfa) and cabbage, but unchanged for cellulose; no fibrous residue remained after the fermentation of pectin. McBurney et al= (1985) also measured the water-holding capacity of a gravimetrically determined estimate of microbial mass residue remaining after fermentation, and suggested that including this value in the calculation of potential water-hold­ing capacity could improve prediction of in vivo effects of dif­ferent fiber sources on fecal mass. Determining the in vitro fermentability- of a fiber would also add to the characterization of a fiber, although the methodology needs further development.

F. MODULATION OF EFFECTS BY PARTICLE SIZE

Particle size of dietary fibers is a factor which can markedly affect the properties described above, and thus alter the physiological effects of fibers as well. Kirwan et al. (1974) found a higher water-holding capacity in a coarse wheat bran preparation than in a fine bran preparation from another source (with a different composition). Brodribb and Groves (1978) showed that coarse wheat bran (80% of particles >1.5 mm) had a higher water-holding capacity than fine bran (80% of par­ticles <0.71 mm). Mongeau and Brassard (1982) demonstrated a decrease in both water-holding capacity and bile salt-binding with decreasing particle size in a variety of breakfast cereals and brans. These investigators employed a centrifugation tech­nique to measure water-holding capacitye On the other hand, Stephen and Cummings (1979) found greater water-holding capacity with a dialysis bag technique in smaller particles (0.075 to 0.1 mm) of wheat bran, carrot, and cabbage than in larger parti­cles (0.71 to 1.0 mm). These conflicting data suggest that the effects of particle size on water-holding capacity depend on the methods used to measure it. Centrifugation would likely remove more water surrounding the particles than would the dialysis bag technique. Which of these two methods more closely approximates conditions in vivo is not clear.

A standardized method for measuring particle size would facilitate determination of the significance of particle size as a physical characteristic of fiber. Additional discussion of the effects of fiber particle size on physiological functions is found in Section V-J of this report.

G. MODULATION OF EFFECTS BY PROCESSING

Processing can affect the physical and chemical prop­erties, as well as the physiological functions, of various fiber sources. Grinding or blending can disrupt the physical structure of intact cell walls and individual fiber components. Cooking can alter the chemical composition of foods and change their hydration. A systematic study of such processing effects has not been undertaken, but some examples may be given.

Jenkins et al. (1982) tested the effects of processing of lentils on digestion of starch by saliva in vitro. They found no difference in starch digestion of lentils boiled for 20 or 60 min, or blended after boiling; starch digestion was significantly increased, however, when boiled lentils were blended, dried, and ground. Wong and □ 'Dea (1983) found that adding cooked lentils to a glucose solution did not impair the transport of glucose across a dialysis membrane, and that grinding lentils or brown rice before cooking produced an increase in starch hydrolysis in vitro. Whole brown rice

24

was shown to elicit lower postprandial glucose, insulin, and gastric inhibitory polypeptide (GIP) responses in diabetic subjects than did ground brown rice or glucose (Collier and □ 'Dea, 1982).

Matthee and Appledorf (1978) showed an increase in the water content and in the NDF and ADF content (on a ·dry weight basis) after boiling various vegetables. Mild extrusion cooking of a bran-starch-gluten mixture did not produce a change in the digestion of starch and fiber, but decreased digestion of phytate in ileostomy patients (Sandberg et al., 1986). Consumption of raw bran, however, had a greater effect on fecal dry weight than consumption of cooked bran (a commercial cereal product) with an equal amount of dietary fiber (Wyman et al., 1976). Lykken et al. (1986) found that less zinc was absorbed from cornflakes (browned) made from corn intrinsically labeled with 65zn than from corn grits (unbrowned) made from the same labeled product.

25

V. EFFECTS OF DIETARY FIBER ON NORMAL PHYSIOLOGICAL FUNCTIONS

The physical and chemical properties of dietary fiber form the basis for in vivo responses to fiber-containing foods, fiber supplements, and isolated fiber components. These responses occur over the length of the entire alimentary canal and include systemic effects as well. In the discussion of physiological effects below, emphasis is placed on results of studies with human subjects.

A. CHEWING AND SALIVATION

Diets containing large amounts of plant fiber tend to be bulky, require longer times for ingestion, and add effort to the act of eating (Heaton, 1980). When subjects were asked to consume equicaloric amounts of whole apple, apple puree, and apple juice as quickly as they comfortably could, they required 17.2 min for fruit, 5.9 min for puree, and 1.5 min for juice (Haber et al., 1977). Mccance et al. (1953) reported that sub­jects required an average of 45 min to consume a large meal of brown bread and 34 min to consume an equicaloric amount of white bread. Duncan et al. (1983) found that both total eating time and chewing time were significantly longer when obese and non­obese subjects consumed a low-energy-density diet (high in fresh fruits and vegetables, whole grains, and dried beans) than when they consumed a high-energy-density diet (low in dietary fiber). The stimulation of chewing can lead to an increase in the flow of saliva and dilution of the oral contents. Little research _has been done to evaluate the effects of increased chewing on oral health or determine the impact of increased salivation.

B. SATIETY

Objective methods for the assessment of satiety have not yet been developed; however, subjective evaluation of dietary fiber effects on satiety have been reported. Researchers have reported greater satiation after meals of various intact fruits than after energy-equivalent meals of the corresponding fruit juices (Bolton et al., 1981; Haber et al., 1977; Kay and Stitt, 1978). The time required for consumption of the various test foods was held constant in these studies, thereby eliminating one possible confounding variable. In one study (Haber et al., 1977), intact apples conferred greater satiation than a puree of apples, and the puree was more satisfying than juice, leading the investigators to conclude that both removal and disruption of dietary fiber in a food could influence its satiating effect. There are few studies of other sources of fiber; one single-meal comparison study showed that whole-meal bread was more satiating than white bread (Grimes and Gordon, 1978), but another study failed to confirm this effect (Bryson et al., 1980). In a longer

study with diets formulated to have either low-energy density/ high fiber or high-energy density/low fiber, Duncan et al. (1983) found that subjects achieved satiety on the low-energy-density/ high-fiber diet at a mean daily caloric intake of one-half of that required to achieve satiety on the high-energy-density/low­fiber diete

Whether fiber supplements and isolated fiber components also exhibit the same satiating effects as fiber in foods is uncertain. Krotkiewski (1984) has reported that hunger ratings of obese subjects consuming their usual diets were reduced more by guar supplements than by wheat bran. Careful studies with rigorous designs are required to determine the significance of the satiety effects of dietary fibers observed in short-term studies.

Krotkiewski (1985) has listed some factors which may contribute to the satiating effects of fiber-containing foods and high-fiber diets. These include caloric dilution of foods; triggering of afferent signals, including satiety, by chewing (Anand, 1974); operation of time-dependent satiety signals caused by the greater time and effort of ingestion; induction of satiety signals caused by distension of the stomach; influence on the feedback and interplay of the gastroentero-pancreatic axis related to delayed gastric emptying (see Section V-C); and stim­ulation of the release of gastrointestinal hormones (see Section V-G) which influence both insulin production and hunger feelingse

C. GASTRIC EMPTYING

Ingested solids are subject to two processes in the stomach: grinding of the swallowed particles by antral con­traction and propulsion of the finely suspended particles into the duodenum. Fiber particles are probably resistant to mechan­ical degradation in the stomach (Eastwood and Brydon, 1985). Little is known about the potential chemical degradation of fiber in the stomach. The shape, size, mechanical resilience, and tex­ture of solid ~articles from different sourc~s vary, and these factors may influence gastric emptying.

Results of studies of the effects of dietary fiber on gastric emptying rate have been variable depending on the method of measurement used. One of the older methods was aspiration of the gastric contents and measurement of the amount of meal remaining and·the gastric secretions. Most methods currently used involve the feeding of a radioisotope with a test meal anduse of various detecting devices to follow the loss of radio­activity in the stomach or the appearance of the label in the blood as the meal enters the duodenum. The pattern of gastric emptying detected may depend on whether the label partitions with the solid or liquid phase of the meal (Heading et al., 1976). Liquids are passed more rapidly from the stomach than are solids (MacGregor et al., 1977) and the rate of gastric emptying depends

28

in part on the solid-liquid separation in the postprandial state. Using cellulose particles 1 mm long, labeled with I-131, Carryer et al. (1982) showed that the solid particles left the stomach more slowly than liquid, but that there was some emptying of the fiber at all time periods, suggesting that the solid-liquid phase separation is a continuous process. Another source of variation is the effects of the other components of the test meal on gas­tric emptying; it may be expected that the rate of emptying of a fiber source given alone will be modified when it is fed with other dietary constituents, such as lipids, which are known to produce a significant delay in gastric emptying. Finally, gas­tric emptying has been expressed in terms of the half-time of emptying, the percentage emptying per minute, the time of com­plete emptying, the shape of the disappearance curve, and the percentage of the meal left in the stomach at the end of the study, making it difficult to compare the results of different studies.

Table 5 shows the results of studies of the effects of various dietary fibers on gastric emptying in humans. Among the individual fiber types, the viscous fibers (such as guar gum and pectin) have been shown frequently, but not consistently, to have marked effects on delaying gastric emptying. Many of these studies employed large, unphysiological doses of the isolated fibers, and most employed liquid test meals.

When a fiber-associated delay in gastric emptying is observed, it may have some influence on satiety and consequences for subsequent nutrient absorption from the small intestine. A slowing in gastric emptying is not always beneficial; for example, it may. promote gastroesophageal reflux in susceptible individuals (Harju and Larmi, 1985).

D. INTESTINAL DIGESTION AND ABSORPTION OF NUTRIENTS

As noted by Vahouny and Cassidy (1985), consumption of dietary fiber generally results in increases in fecal volume, weight, and energy content, although the level and kind of fiber are important determinants of the magnitude of the changes. These changes are in part attributable to increased excretion of indigestible plant cell wall material and increased bacterial mass, but evidence also suggests that certain dietary fibers may affect the digestion and absorption of other nutrients.

1. Changes in nutrient balance/fecal excretion

Many st~dies have shown that increased intake of dietary fiber and type of fiber consumption are associated with an increase in fecal nitrogen loss. Southgate and Durnin (1970) showed that increasing (approximately doubling) the amount of fiber in the diet by substitution of fiber-rich foods increased fecal nitrogen loss and decreased the apparent digestibility of

29

\J.J b

Table 5. Studies on the effect of dietary fiber on gastric emptying in humans.

Reference

Mccance et al. (1953)

Fiber Source (and Test Dose)

Whole wheat

Test Meal

300 g brown bread or 285 g white bread with 200 ml water

Method

Paste of barium sulfate fed with meal; radiographs taken

Subjects

6 healthy volunteers

Results

Gastric emptying time (time from midpoint of meal to emptying of stomach) longer with brown bread

Hunt (1954) Pectin or sucrose solution with Withdrawal and measurement 4 volunteers Emptying patterns the same for control, pectin, and gum

Holt et al. (1979)

Gum tragacanth

Guar (16 g) Pectin (10 g)

Wilmshurst & Guar (2 g) Crawley (1980)

Leeds et al. (1981)

Leatherdale et al. ( 1982)

Schwartz et al. (1982)

Pectin (10 g) (high methoxy)

Guar (10 g)

Pectin or a.­Cellulose fed 4 wk (20 g/d)

phenol red of stomach contents

400 ml orange juice+ Sequential scinti-paracetamol and scanning In-113m DTPA*

200 ml of low-energy liquid test m~!l labeled with Na

75 g glucose+ 150 ml water with In-113 label

porridge labeled with In-113 DTPA*

Sodium iodide crystal in ~~ad collimeter detected

Na appearance in blood

Gamma camera to detect label in stomach

Measured peak volume in stomach (early phase) and time for half decrease from peak volume

white bread+ tuna Gamma scintilation camera (Tc-99m surface label) to detect label & 360 ml grape juice

* DTPA = diethylenetriamine pentaacetic acid.

14 volunteers Rate of gastric emptying slower with guar and pectin

12 obese subjects

12 patients with dumping syndrome

10 diabetic patients and 10 matched controls

Gastric transit time increased by guar

Gastric emptying time slowed with pectin; mean plasma volume fall was lower

Peak volume lower in diabetic patients but not affected by guar; half-time decreased by guar in controls but not in diabetic patients

13 volunteers Pectin prolonged gastric emptying time; no effect of cellulose

Table 5. Studies on the effect of dietary fiber on gastric emptying in humans (continued).

Reference

Tadesse (1982)

Harju et al. (1984)

Kasper et al. (1985)

Rydning et al. (1985)

Fiber Source (and Test Dose)

Pectin or Carboxymethyl­cellulose (4 g)

Guar (5 g) or placebo

Guar (Li g)

Pectin (5-10 g) Guar (2 g) Wheat bran (15 g)

Test Meal

400 ml glucose and phenol red solution

yogurt (labeled with Tc-99m), orange juice, wheat bread, & butter

200 ml apple juice with Tc-99m label

200 ml liquid meal with Tc-99m label

Pectin (5 g) 200 ml liquid meal Wheat bran (15 g) with Tc-99m label

Guar (5 g) or Fiberform® (lo, 5 g)

Guar (5 g) or Fiber form® (10, 5 g)

200 g porridge (with Tc-99m label) and 200 ml juice

200 g porridge (with Tc-99m label) and 200 ml juice

Method

Aspiration and measurement of gastric contents

Gamma camera images

Gamma camera

Gamma camera

Gamma camera

Isotope localization monitor over fundus

Gamma camera

Subjects

? volunteers

11 patients with dumping syndrome

8 volunteers

Results

Cellulose decreased and pectin increased gastric emptying time

Guar increased gastric transit times in 8 of 11 patients

Guar delayed gastric emptying

10 volunteers Wheat bran caused faster gastric emptying; no effect of other fiber supplements

9 volunteers Larger meal slowed emptying; no effect of bran or pectin

10 volunteers Increased radiolabel detected after control or guar, but not after bran supplement; no effect on gastric emptying time

8 volunteers No effect of guar or bran on gastric emptying time

protein. Substitution of barley for wheat had a similar effect (Judd, 1982). The consumption of cellulose (16 g/d) did not affect fecal nitrogen loss in adults (Slavin and Marlett, 1980), but the addition of 21 g/d of cellulose to the diets of adoles­cents did increase nitrogen loss (Kaur et al., 1985). Consump­tion of pectin (36 g/d) also significantly increased nitrogen loss (Cummings et al., 1979a). Similarly, subjects consuming diets high in fruits and vegetables had increased fecal excretion and decreased apparent digestibility of nitrogen (Kelsay et al. 1978, 1981). Cummings (1978) has concluded that the increased fecal nitrogen is not nutritionally significant in Western industrialized communities where protein intake is usually high. These studies are limited in their ability to distinguish the source of excreted nitrogen, to assess whether protein associated with the cell wall of the fiber source is less digestible, and to determine whether dietary fiber interferes with protein digestion and absorption. Studies in which the fecal nitrogen has been fractionated indicate that most of it is in the bacterial mass, or consists of unabsorbed intestinal secretions and mucosal cell debris (Stephen and Cummings, 1980a). Studying changes in ile­ostomy contents as an indicator of in vivo digestion, Sandberg et al. (1981, 1983) found no change in fecal nitrogen after consumption of 16 g wheat bran, but a significant increase after 15 g citrus pectin.

Many studies in which the intake of dietary fiber has been increased (either by qualitative changes in diet or addi­tion of fiber supplements) have shown an increase in fecal fat in healthy subjects (Cummings et al., 1979a; Judd, 1982; Kaur et al., 1985; Kelsay et al., 1978, 1981; Slavin and Marlett, 1980; Southgate and Durnin, 1970) and patients with pancreatitis (Dutta and Hlasko, 1985). The apparent digestibility values correlated poorly with the changes in fecal fat (Southgate, 1982), suggesting that fecal fat analyses are limited in their utility for detecting effects on fat digestion and absorptiono Further, although the increase in fecal fat in normal subjects was statistically significant, fat excretion remained below levels associated with steatorrhea or with frank impairment in fat assimilation. There is a considerable literature on the overall effects (without distinctions between direct effects on absorption and other secondary effects) of high-fiber diets and fiber supplements on several aspects of lipid metabolism, includ­ing circulating lipids and lipoproteins. These studies are discussed in Section V-F.

Numerous studies have evaluated the ·relationships between dietary fiber and the fate of ingested carbohydrates. Some of these have focused on the availability of carbohydrate endogenous to various foods, as assessed mainly by glycemic response [see reviews by Jenkins and Jenkins (1985) and Jenkins et al. (1986a)J. Others have focused on the modification of plasma glucose and insulin responses by addition of dietary fiber to a test meal or diet (see Section V-G). These studies do not distinguish effects on digestion from those on absorption;

32

however, in studies of ileostomy patients, the in vivo digestion of starch was not affected by consumption of either wheat bran or pectin (Sandberg et al., 1981, 1983). Starch also appeared in the terminal ileum of individuals with intact gastrointestinal tracts following ingestion of low- and high-fiber diets; more starch was detected with the diet higher in fiber which was also more calorically dense (Stephen et al., 1983). These authors concluded that incomplete starch digestion was the normal occur­rence, although the amount of starch reaching the colon may be determined by diet composition.

The effects of various fiber sources on mineral and vitamin absorption have also been studied. These effects are discussed in Section VI-0 of this report.

2. Proposed mechanisms

A variety of mechanisms has been proposed by which dietary fiber may influence the digestion and absorption of nutrients. Some are discussed in detail in other sections of this report; the remainder will be considered in this section.

• gastric filling and emptying (Section V-A)

• gastric acidity and the pH profile of the gut

• small intestinal transit and mixing

• intraluminal digestion, including pancreatic secretion and digestive enzyme activities

• interactions with the intestinal surface and effects on surface hydrolysis

bile salt alterations, including interference with micelle formation and bile acid resorption (Sections V-E and V-F)

• bulk interference with nutrient diffusion and absorption, including effects on the unstirred water- layer

• gastrointestinal hormone release (Section V-G)

• alterations in sites of absorption and sites of maximal transport

• changes in intestinal morphology (Section V-H)

The effects of fiber on the pH or rate of change of pH in the stomach and other regions of the gastrointestinal tract have not been studied extensively. However, Lennard-Jones et al.

33

(1968) reported that gastric acidity was greater following a meal of corn flour (low protein) than unrefined corn meal (high pro­tein). Wheat an~ rice brans and unrefined grains were shown to have greater buffering capacity than did polished rice and white flour (Tovey, 1974), but the same foods were also shown to be antral stimulants. Rydning and Berstad (1986) fed healthy vol­unteers fiber supplements with a standard meal of porridge and juice and found that both guar gum and wheat bran preparations significantly lowered the peak pH elevation, but that guar gum shortened the acid-neutralizing effect of the meal while wheat bran lengthened it. The mechanism of fiber-induced changes in gastric pH is not known, but such changes may alter peptic digestion of protein and subsequent proteolytic activity in the small intestine. ·Changes in intestinal pH in response to dietary fiber have not been studied in humans.

The specific effect of dietary fiber on small intes­tinal transit has been difficult to determine as an independent variable. Only indirect methods of study can be employed in humans. Early radiological studies suggested a more rapid tran­sit with whole-meal bread than with white bread (Mccance et ale, 1953). · Another technique is the measurement of expired hydrogen arising from the production of hydrogen by cecal bacteria from an unabsorbable carbohydrate (lactulose); the appearance of breath hydrogen is used as a marker for the arrival of a meal at the terminal ileum. The mouth-to-cecum transit time includes gastric filling time, gastric emptying time, and small intestinal transit time; thus, conclusions about small intestinal motility are lim­ited with this measurement. Jenkins et al. (1978a) found that mouth-to-cecum transit times were greatly increased by guar, moderately increased by gum tragacanth and pectin, not affected by methylcellulose, and decreased by wheat bran. All the fiber sources were added to 400 ml of a glucose/xylose/lactulose test solution at doses equivalent to 12 g fiber. Viscosity was corre­lated with mouth-to-cecum transit time. Urinary excretion of xylose was reduced initially, but a compensatory increase later suggested that although absorption was delayed by several fiber sources, there was no overall impairment. Bond and Levitt (1978) used breath hydrogen· to assess mouth-to-cecum transit time and found no effect of wheat bran. Lembcke et al. (1984) found that addition of guar to a glucose load significantly prolonged mouth­to-cecum transit time in both healthy controls and partially gastrectomized patients. The changes potentially could alter absorption by reducing contact time, but little evidence exists to support or refute this hypothesis. Many of the studies .of effects of fiber on mouth-to-cecum transit time used very simple test meals atypical of a meal of mixed foods. The use of lactu­lose fermentation to detect transit of a meal assumes that the lactulose remains with the bolus of the meal with which it was ingested and does not "stream ahead" of the meal. However, it is known that portions of liquid and solid phases of a meal transit the ileum separately (Urban, 1984).

34

Another potential influence of dietary fiber on nutrient availability is its interaction with pancreatic and/or intestinal digestive enzymes and their substrates, which may lead to a decreased efficiency of digestion and limit the diffusion of absorbable products in the intestinal lumen. In vitro studies (Dunaif and Schneeman, 1981; Isaksson et al., 1982; Schneeman, 1978, 1982; Schneeman and Gallaher, 1985) indicate that in most cases the activity of human pancreatic enzymes (amylase, lipase, trypsin, and chymotrypsin) is reduced below control levels after incubation with most fiber sources. The in vivo implications of these observations are not clear. Evidence from animal studies (Calvert et al., 1985) suggests that fibers do not act directly to affect the pancreatic content of digestive enzymes or secre­tory activitj, but act to alter enzyme activity in the intestinal lumen. In vivo effects of fibers have been studied in patients with pancreatic disorders by Isaksson et al. (1984). In totally pancreatectomized patients, breath 14co2 expired after ingestion of 14c-labelled triolein was significantly reduced when a low­methoxy pectin was added to a granulated pancreatic enzyme sup­plement. Pectin also reduced trypsin, lipase, and amylase activ­ities of jejunal aspirates in the same patients after a test meal supplemented with pancreatic enzyme substitute. In patients with chronic pancreatitis, breath 14co2 excretion was reduced by wheat bran, which also caused a reduction in lipase and amylase activities of a duodenal aspirate after a test meal.

The interactions of dietary fiber with the intestinal surface and effects on surface hydrolysis have not been studied in humans. As noted by Vahouny and Cassidy (1985), the effects of various fibers (especially viscous fiber preparations) on the activities of mucosal-associated digestive enzymes in animals have not been consistent. Differences in the effects on the enzymatic digestion of disaccharides and peptides at the mucosal surface seen in these studies may be related to differences in the extent of tissue rinsing and preparation for in vitro studies and the nutritional state of the animal at the time of the study.

Another potential effect of fiber which has been stud­ied inadequately in vivo is bulk interference with nutrient dif­fusion and absorption. In vitro, viscous fibers can influence the accessibility of absorbable nutrients to the mucosal surface. Studies with various gums and carboxymethylcellulose indicate that the uptake of sugars and amino acids by rat intestinal seg­ments is inversely related to the viscosity of the incubating medium (Elsenhans et al., 1980; Johnson and Gee, 1981). That viscous fiber preparations may also alter the transport characteristics of nutrients at the mucosal surface, perhaps by increasing the resistance or thickness of the unstirred water layer, is suggested by the observation that the inhibitory effects can be reversed by rinsing the intestinal segments. Determining whether this effect occurs in vivo requires further investigation; increased motility may overcome any effects of increased unstirred layer thickness. In addition to the acute effects of fibers, Sigleo et al. (1984) have reported adaptive

35

responses which led to increases in both apparent Km and the maximal transport capacity for amino acids and glucose analogs in rinsed intestinal segments of rats fed diets containing either A

cellulose or pectin compared with rats given a fiber-free diet. Such adaptive responses are morphological as well as functional (see Section V-H).

Finally, dietary fibers may have an influence on the sites of absorption or sites of maximal transport capacity for nutrients (Leeds, 1982). If some types of dietary fiber slow nutrient absorption, then sites of absorption may shift distally; this might induce transport at distal sites and reduce transport at proximal sites. Whether these effects are seen in humans remains to be investigated.

In summary, there are a variety of demonstrated and potential influences of high dietary fiber intake on both intraluminal and mucosal intestinal physiology which may affect the rate and extent of nutrient digestion and/or absorption. Emphasis has been placed on the potentially negative influence of fiber on nutrient absorption. A range of fiber levels needs to be examined. Further studies, especially in humans using test meals more representative of a typical meal or with an appropriate animal model, are needed to establish the existence and/or significance of these effects.

E. BILE ACID EXCRETION

The ability of dietary fibers to sequester bile acids in vitro implies that they may be able to alter the enterohepatic circulation of these compounds or alter biliary bile composition, thus modifying lipid digestion and absorption. The mechanism by which these effects may occur is outlined by Story (1980) as follows: adsorption of bile acids by certain dietary fiber com­ponents may increase the excretion of bile acids in feces thereby reducing the amount returning to the liver via the enterohepatic circulation. In response, the liver may increase synthesis of primary bile acids.

In general, there is a reasonably good relationship between the ability of various fiber components to sequester bile acids in vitro and their overall effects on fecal bile acid excretion (Vahouny, 1985). Table 6 presents a summary of studies of the effects of various dietary fiber sources on fecal bile acid excretion in humans. The results of these studies indicate that pectin (15-40 g/d, mean= 30 g/d; 2-3 wk) causes a consis­tent increase in bile acid excretion (34-70%, mean= 50%), as does oat bran (40-125 g/d, mean= 91 g/d; 10-21 d; 51-111%, mean= 72.3% increase). Increases are also observed with other gel-forming fibers such as guar and psyllium seed hydrocolloid.

36

Table 6. Studies on the effect of dietary fiber on daily fecal bile acid excretion in humans.*

--

Fiber Test 24-Hr Fecal Bile Acid Reference Source Subjects Dose Duration Excretion(% Change)

Jenkins et al. (1976a) Pectin 7 normal 36 g/d 2 wk +34

Kay & Truswell (1977a) Pectin 9 normal 15 g/d 3 wk +40

Miettinen & Tarpila (1977) Pectin 2 normal 40 g/d 2 wk +75 7 hyperlipidemic

Jenkins et al. (1976a) Guar 7 normal 36 g/d 2 wk +84

Ross et al. (1983) Gum arabic 5 normal 25 g/d 3 wk 0

Forman et al. (1968) Psyllium 2 normal 9.6 g/d 6 wk +302

"-"' Stanley et al. (1973) Psyllium 22 normal 15 g/d 16 d +70 :-....J

Mathur et al. (1968) Legumes 10 normal ** 55 wk +55

Schweizer et al. (1983) Soya pulp 6 normal 21 g/d 3 wk +21 Soya fiber 6 normal 21 g/d 3 wk -11

Tsai et al. (1983) Soy polysaccha:ride 14 normal 25 g/d 2 wk 0

Anderson et al. (1984a) Beans 10 hyperlipidemic 115 g/d 3 wk -30

Kretsch et al. (1979) Oat bran 6 normal 40 g/d 15 d +111

Kirby et al. (1981) Oat bran 8 hyperlipidemic 100 g/d 10 d +51

Anderson et al. (1984a) Oat bran 10 hyperlipidemic 100 g/d 3 wk +65

* Adapted from Kay and Truswell (1980). ** "Large amount" of Cicer arietinum added to a high-fat diet.

Table 6. Studies on the effect of dietary fiber on daily fecal bile acid excretion in humans (continued).*

Fiber Test 24-Hr Fecal Bile Acid Reference Source Subjects Dose Duration Excretion(% Change)

Judd & Truswell (1981) Rolled oats 10 normal 125 g/d 3 wk +62

Shurpalekar et al. (1971) Cellulose 10 children 100 g/d 10 d +45

Stanley et al. (1973) Cellulose 22 normal 15 g/d 16 d +25

Walters et al. (1975) Bagasse 9 normal 10.5 g/d 12 wk +50

Eastwood et al. (1973) Wheat bran 8 normal 16 g/d 3 wk 0

Jenkins et al. (1975a) Wheat bran 6 normal 36 g/d 3 wk 0

Walters et al. (1975) Wheat bran 8 normal 39 g/d 3 wk 0 \.,J CX>

Cummings et al. (1976a) Wheat bran 6 normal 64 g/d 3 wk +90

Kay & Truswell (1977b) Wheat bran 6 normal 54 g/d 3 wk 0

McDougall et al. (1978) Wheat bran 5 normal 50 g/d 3 wk +240

Tarpila et al. (1978) Wheat bran 22 with 70 g/d 12 mo -49 diverticulitis

Eastwood et al. (1986) Wholemeal bread 28 normal 23 g/d 6 mo 0

Reddy et al. (1987) Whole wheat+ rye 15 normal 25 g/d 4 wk -10

* Adapted from Kay and Truswell (1980).

~ \()

Table 6. Studies on the effect of dietary fiber on daily fecal bile acid excretion in humans (continued).*

Reference

Antonis & Bersohn (1962)

Raymond et al. (1977)

Kretsch et al. (1979)

stasse-Wolthuis et al. (1979a)

Kay et al. (1985)

Fiber Source

Mixed sources

Mixed sources

Mixed sources

Mixed sources

Mixed sources

* Adapted from Kay and Truswell (1980).

Subjects

14 normal

8 normal & hyperlipidemic

6 normal

46 normal

12 normal

Test Dose

15 g/d

60 g/d

93 g/d

45 g/d

55 g/d

Duration

15-39 wk

4 wk

15 d

3 wk

5 wk

24-Hr Fecal Bile Acid Excretion(% Change)

+26

0

+111

0

0

Wheat bran, a dietary fiber source rich in insoluble components, does not show a consistent change in bile acid excretion (16-70 g/d, mean= 44 g/d; 3-72 wk, mean= 17.3 wk; -49 to +240%, mean= 35% change).

When evaluating the effects of dietary fiber on bile acid excretion, effects on both daily excretion and on concentra­tion of bile acids in feces must be consideredc Changes in daily excretion are related to the effects of a dietary fiber source on cholesterol balance. Attention must be paid to concentration of bile acids in the colon as a result of the relationship which has been observed between increased bile acid concentration and increased susceptibility to colon cancer (Hill, 1986). Various sources of fiber cause changes in one or both of these variables.

In addition to affecting the amount and concentration of fecal bile acids excreted, dietary fibers can also modify the spectrum of biliary bile acids. Several investigators have reported an increase in the chenodeoxycholic acid pool and a reduction in the deoxycholic acid pool in the bile of humans in response to wheat bran (Pomare and Heaton, 1973; Pomare et al., 1976; Watts et al., 1978). Pectin caused the opposite response in these two bile acids (Hillman et al., 1986; Miettinen and Tarpila, 1977). Studies of animals have also indicated changes in the relative amounts of bile acids excreted in feces (Reddy et al., 1980a; Story, 1985, 1986; Story and Thomas, 1982). Changes in pool size can alter several phases of cholesterol metabolism including cholesterol synthesis, cholesterol absorp­tion, cholesterol saturation in bile, and bile acid reabsorptiono The involvement of these changes in determining effects on serum cholesterol and gallstone formation will be discussed. later (Section V-F and VI-C.4).

F. SERUM LIPIDS

The effects of various sources of dietary fiber on serum lipids have been studied extensively and have been the subject of several comprehensive reviews (Judd and Truswell, 1985; Kay and Truswell, 1980; Story, 1980; Vahouny, 1982, 1985). Early studies with fiber-rich foods (Keys et al., 1960) showed that plasma total cholesterol levels could be reduced by diets containing fresh fruits, vegetables, and legumes (45 g dietary fiber). Isa­nutrient substitution of either sugar or bread by a mixture of vegetables (40 g dietary fiber) was also found to lower plasma cholesterol level, although substitution by fruit (20 g dietary fiber) had no effect (Grande et al., 1974). Other studies which have demonstrated a hypocholesterolemic effect of diets rich in fruits, vegetables, and/or legumes include those of Grande et ale (1965), Jenkins et al. (1979a), Luyken et al. (1962), and Miettinen (1980). In general, adding cereals to the diet has had little effect on circulating cholesterol levels. For exam­ple, wheat bran fed at levels from 9-38 g/d for 21-365 din 14 studies showed no change in nine studies, an increase of 7%

40

in one study, and reductions of 7-22% (mean= 15%) in four studies. Other studies which have examined the effects of a variety of high-fiber diets on serum lipids are summarized in Table 7a. Many of these studies have demonstrated the ability of such diets to lower serum cholesterol levels without affecting serum triglyceride levels. Because the proportions of other com­ponents of the diet are altered by the inclusion of fiber-rich foods, these effects cannot be attributed to dietary fiber alone. In fact, some of these studies fail to show an effect of fiber, per se, when the fat and carbohydrate contents of the experi­mental diets are manipulated. This finding is not unexpected because lowering the lipid content of the diet has a well­established hypolipidemic effect.

Studies have also been conducted with concentrated and purified sources of dietary fiber such as brans and gums. Results of studies of the influence of such fiber supplements on the serum lipid levels of subjects without hyperlipidemic disorders are summarized in Table 7b, and are discussed below.

Preparations of guar gum in doses of 6-36 g/d (mean= 19 g/d) fed for 14-66 d (mean= 30 d) have been reported to produce a modest, but consistent, decrease in serum total cho­lesterol levels of 5-17% (mean= 12.7%), without changes in serum triglyceride levels. Limited evidence available in these studies of normal volunteers, in addition to data from studies of subjects with hyperlipidemia (see Section VI-C.3), suggests that the change in total cholesterol levels is mediated mainly by a decrease in low-density lipoprotein (LDL)-cholesterol levels. In 14 studies described in Table 7b, investigators reported that pectin fed at levels of 6-36 g/d (mean= 18.8 g/d) for 14-83 d (mean= 35 d) lowered serum cholesterol levels 5-18% (mean= 8%) under a variety of conditions [see also reviews by Judd and Truswell (1985) and Kay and Truswell (1980)] in a manner similar to guaro

Many studies of wheat bran have been conducted and most have not found an effect on serum lipids, even when very high levels were fed [see Table 7b and also reviews by Judd and Truswell (1985) and Kay and Truswell (1980)]. Studies in which hypocholesterolemic effects were detected suggest that the vari­ety of wheat (soft white vs. hard red spring) and the coarseness of the bran may modulate its effects. Other dietary fiber sources (corn bran, soybean hulls, and. corn hulls) have also been studied and generally have yieldeq negative results. Cellulose added to the diet has not been found to lower plasma cholesterol levels except when fed at very high levels (100 g/d) to children. Isolated lignin has not been extensively studied, but results have been inconsistent. In pooling the results of a number of short-term metabolic studies with largely insoluble test fibers (psyllium seed fiber, rice bran, corn bran and wheat), Kies (1985) observed that variability in the cholesterol response of subjects was partly dependent on pre-study serum lipid levels and transit time.

41

Table 7a. Studies on the effects of high-fiber diets on serum lipid levels in healthy subjects.*

% Change

Cholesterol Fiber No. of

Reference Source Duration Amount Subjects Total VLDL LDL HDL TG Comments

Albrink et al. (1979) HFHC 1 wk 18 g/d CF 7 -23 - - - NS TG inc by high-CHO alone

Stasse-Wolthuis et al. HFHC 3 wk 45 g/d OF 46 -8 - - -5 - Fiber from mixed sources (1979a) HFLC II II II -7 - - -6 - in diet

Ullrich & Albrink (1982) HFHC 4 d 41 g/d NDF 8 -12 - - -16 NS Same changes seen with low-fiber, high-CHO diet

Flanagan et al. (1980) Reducing diet 3 mo 12 g/d OF 13 -8 - - +13 - Balanced, low-energy diet; subjects w/ varicose veins

~ Roughage diet 3 mo 32 g/d OF 16 NS - - +9 - 30 g All-BranID + 30 g wheat N

bran added; subjects with inflammatory bowel disease

Masarei et al. (1984) LOV 6 wk 30 g/d OF 18 men -5 - -7 -6 NS LOV diet had lower satu-II II II 18 women NS - - - - rated fat and protein

than omnivore control diet

Kay et al. (1985) LFLF 5 wk 20 g/d OF 12 -22 - NS -12 - Compared to low-fiber, HFLF II 55 g/d OF II -29 - -34 -11 - high-fat diet HFHF II 53 g/d OF II -25 - -31 NS

Albrink & Ullrich (1986) HFLS 10 d 80-86 g/d OF 6 NS - NS NS NS Compared to low-fiber, low-LFHS II 5-10 g/d OF II NS - NS dee inc+ sucrose diet; only effect of HFHS II 34-68 g/d OF II NS - NS dee inc fiber was moderate increase

in TG with high sucrose

* See code for abbreviations at end of Table 7b; amounts of fiber reported are dependent on the method of fiber analysis.

Table 7b. Studies on the effect of dietary fiber supplements on serum lipid levels in healthy subjects.*

% Change

Cholesterol Fiber Test No. of

Reference Source Duration Dose Subjects Total VLDL LDL HDL TG Comments

Fahrenbach et al. (1965) Guar 66 d 6 g/d 23 -5 " 52 d 12 g/d II -6

Jenkins et al. (1975b) Guar 2 wk 36 g/d 7 -16 - - - - Cholesterol increased on control diet

Jenkins et al. (1976a) Guar 2 wk 36 g/d 7 -16 - - - - Dietary fat also reduced

Jenkins et al. (1979a) Guar 3 wk 20 g/d 3 -13 - - NS NS

Khan et al. (1981) Guar 4 wk 9 g/d 24 -17 NS -26 NS NS Guar vs. placebo

Penagini et al. (1986) Guar 2 wk 11-14 g/d 6 -16 - - - NS

Keys et al. (1961) Citrus pectin 83 d 15 g/d 6 -5 - - - - High-legume diet; II II Ii 6 -5 - - - - High-sucrose diet;

Pectin vs. control

Fahrenbach et al. (1965) Pectin 66 d 6 g/d 23 NS II 52 d 12 g/d II II

Palmer and Dixon (1966) Pectin 4 wk 2, 4, 6, 8, 17 dee - - - - Significant decrease with & 10 g/d doses >6 g/d

Jenkins et al. ( 1975b) · Pectin 2 wk 36 g/d 7 -12

Ourrington et al. (1976) Pectin 3 wk 12 g/d 12 -8 NS -18

* See code for abbreviations at end of Table 7b.

Table 7b. Studies on the effects of dietary fiber supplements on serum lipid levels in healthy subjects (continued).*

% Change

Cholesterol Fiber Test No. of

Reference Source Duration Dose Subjects Total VLDL LDL HDL TG Comments

Jenkins et al. (1976a) Pectin 2 wk 36 g/d 7 -12

Kay & Truswell (1977a) Citrus pectin 3 wk 15 g/d 9 -15 - - - NS

Jenkins et al. (1979a) Pectin 3 wk 31 g/d 5 -13 - - NS NS

Stasse-Wolthuis et al. Citrus pectin 5 wk 28 g/d OF 15 -7 - - NS - Low-fiber control diet (1979b)

Judd & Truswell (1982) Pectin (HM) 6 wk 15 g/d 10 -18 - - NS NS Pectin (LM) II II II -16 - - NS NS

.i:::-

.i:::-Hillman et al. (1985) Pectin 4 wk 12 g/d 10 NS - - NS NS

Vargo et al. (1985) Pectin 20 d 15 g/d 10 NS

Ross et al. (1983) Gum arabic 21 d 25 g/d 5 -6 - - - NS

Behall et al. (1984) Locust bean 4 wk 0.75 g OF/ 12 -14 NS -17 NS NS Compared to low-fiber gum 100 kcal control diet without

Karaya gum II II II -10 NS -10 NS NS fiber supplements Carboxymethyl- II II II -16 NS -18 NS NS cellulose gum

Jenkins et al. (1975a) Wheat bran 3 wk 36 g/d 6 NS - - - NS

Jenkins et al. (1975b) Wheat bran 3 wk 36 g/d 5 NS

* See code for abbreviations at end of Table 7b.

Table 7b. Studies on the effects of dietary fiber supplements on serum lipid levels in healthy subjects (continued).*

% Change

Cholesterol Fiber Test No. of

Reference Source Duration Dose Subjects Total VLDL LDL HDL TG Comments

Persson et al. (1976) Wheat bran 6 wk 10 g/d 13 NS - - - - Subjects were elderly ii II 20 g/d 14 -7

Kay and Truswell (1977b) Wheat bran 3 wk 23-35 g/d 6 NS - - - NS

Mathur et al. (1977) Wheat bran 1 mo 40 g/d 20 -22 - - - - Added to usual diet

Farrell et al. (1978) Wheat bran 24 d 12 g/d 14 -21 - - - NS Diet high in saturated fat

McDougall et al. (1978) Wheat bran 4 wk 50 g/d 9 NS NS NS NS NS cereal

Tarpila et al. (1978) Wheat bran 12 mo 9 g/d CF 22 NS - - - NS Subjects had diverticular disease

Jenkins et al. (1979a) Wheat bran 3 wk 20 g/d 6 NS - - NS NS

Munoz et al. (1979) Soft white 30 d 26 g/d 6 NS - NS NS dee Metabolic ward study wheat bran Hard red 30 d 26 g/d 9 -12 - -21 NS dee

spring wheat

Stasse-Wolthuis et al. Wheat bran 5 wk 37 g/d OF 16 +7 - - NS - Low-fiber control diet (1979b)

van Berge-Henegouwen Wheat bran 4 wk 34-38 g/d 7 -10 NS NS dee -24 et al. (J.979) (coarse)

* See code for abbreviations at end of Table 7b.

Table 7b. studies on the effects of dietary fiber supplements on serum lipid levels in healthy subjects (continued).*

% Change

Cholesterol Fiber Test No. of

Reference Source Duration Dose Subjects Total VLDL LDL HDL TG Comments

Moore et al. (1985) Wheat bran 6 wk 7-12 g/d 7 NS - -25 +46 NS

Roth & Leitzmann (1985a) Bran-muesli 12 wk 29 g/d OF 18 NS - NS +19 Processed bran II 32 g/d OF II NS - NS NS

Munoz et al. (1979) Corn bran 30 d 26 g/d 6 NS - NS NS dee Metabolic ward study

Roth & Leitzmann (1985a) Cornflakes 12 wk 19 g/d OF 18 NS - NS NS

de Groot et al. (1963) Rolled oats 3 wk 140 g/d 21 -11

.J:::- Judd and Truswell (1981) Rolled oats ()\

3 wk 125 g/d 10 NS - - NS NS

Roth & Leitzmann (1985a) Rolled oats 12 wk 19 g/d OF 18 NS - NS +23

Kretsch et al. (1979) Oat bran (T) 15 d 0.6 g/kg/d 6 NS - - - NS Oat bran added to egg Oat bran (UT) II I! Ii NS - - - NS formula diet

Van Horn et al. (1986) Oat bran 6 wk 39 g/d 69 -3 NS NS NS NS Supplements added to diet Oatmeal " 35 g/d II -3 NS NS NS NS of American Heart Assoc.

Keys et al. (1961) Cellulose 42 d 15 g/d 6 NS - - - - High-starch diet; (Alphacel®) n II 6 NS - - - - High-sucrose diet;

Cellulose vs. control

Prather (1964) Cellulose 4 wk 13 g/d 5 NS (Alphacel®)

* See code for abbreviations at end of Table 7b.

Table 7b. Studies on the effects of dietary fiber supplements on serum lipid levels in healthy subjects (continued).*

% Change

Cholesterol Fiber Test No. of

Reference Source Duration Dose Subjects Total VLDL LDL HDL TG Comments

Shurpalekar et al. Cellulose 10 d 100 g/d 10 -25 - - - - Subjects were children (1971) fed high-cholesterol diet

Kaur et al. (1981) Cellulose 3 wk 21 g/d 9 NS - - - NS Compared to low-fiber control

Behall et al. (1984) Cellulose 4 wk .75 g OF/ 12 NS NS +8 NS NS 100 kcal

Hillman et al. (1985) a--Cellulose 4 wk 15 g/d 10 NS - - NS NS

Hillman et al. (1985) Lignin 4 wk 12 g/d 10 NS - - NS NS

Kies and Fox (1977) Hemicellulose 14 d 4.2 g/d 12 -8 - - - -34 Results w/ low-cholesterol (psyllium) ID 14.2 g/d va -11 - - - -20 diet; TG increased w/ in-

ii 24.2 g/d Ii -17 - - - -34 creasing hemicellulose on high-cholesterol diet

Burton & Manninen (1982) Psyllium 16 wk 25-30 g/d 12 -20 - - - NS Elderly, constipated, ( Vi-Siblin®) hospitalized patients

Munoz et al. (1979) Soybean hulls 30 d 26 g/d 5 -14 - NS NS dee Metabolic ward study Textured veg. 30 d 26 g/d 3 NS - NS NS dee

protein

Schweizer et al. (1983) Soya pulp 3 wk 21 g/d OF 6 NS - NS NS Soya fiber Ii Ii II +14 - +19 NS

* See code for abbreviations at end of Table 7bo

Table 7b. Studies on the effects of dietary fiber supplements on serum lipid levels in healthy subjects (continued).*

% Change

Cholesterol Fiber Test No. of

Reference Source Duration Dose Subjects Total VLDL LDL HDL TG Comments

Tsai et al. (1983) Soy poly- 17 d 25 g/d 14 NS -28 NS NS NS VLDL-cholesterol also saccharide reduced on placebo

Mathur et al. (1968) Bengal gram 55 wk ? 20 -22 - - - - Bengal gram replaced wheat "chana" & cereals in high-fat diet

Gormley et al. (1977) Apples 16 wk 2/d 80 -7 - - +14 - Compared to matched controls

Gormley et al. (1979) Peas 6 wk 30 g/d 56 -6 - - +13 - Compared to cornflake control group

~ <X>

Stasse-Wolthuis et al. Veg. & fruit 5 wk 43 g/d OF 15 NS - - NS - Low-fiber control diet (1979b)

Robertson et al. (1979) Carrot 3 wk 200 g/d 5 -11 - - - NS Raw carrot (6 g fiber) given at breakfast

Jenkins et al. (1979a) Carrot 3 wk 20 g/d 6 NS - - -17 NS Cabbage II II Ii NS - - NS NS Apple II II II NS - - NS NS

Fraser et al. (1981) Leafy veg. 3 wk 200 kcal/d 12 dee dee dee NS NS Isocaloric supplemented Root veg. I! 300 kcal/d IO NS NS NS NS NS diets ·compared to diet with

Grains II 400 kcal/d II dee NS dee NS NS 400 kcal/d sucrose

* See code for abbreviations at end of Table 7b.

.j::::­\0

Table 7b. Studies on the effects of dietary fiber supplements on serum lipid levels in healthy subjects (continued).

Reference Fiber Source Duration

Test Dose

No. of Subjects

% Change

Cholesterol

Total VLDL LDL HDL TG

Raymond et al. (1977) Mixed sources 4 wk 60 g/d 8 NS NS NS NS NS

CF= crude fiber

CHO= carbohydrate

dee= decrease reported, value not given

OF= dietary fiber

HDL = high-density lipoprotein

HFHC = high-fiber, high-carbohydrate diet

HFHF = high-fiber, high-fat diet

CODE FOR ABBREVIATIONS IN TABLES 7a AND 7b

HFHS = high-fiber, high-sucrose diet

HFLC = high-fiber, low­carbohydrate diet

HFLF = high-fiber, low-fat diet

HFLS = high-fiber low-sucrose diet

HM= high-methoxy pectin

inc= increase reported, value not given

LDL = low-density lipoprotein

LFLF = low-fiber, low-fat diet

LFHS = low-fiber, high-sucrose diet

LM = low-methoxy pectin

LOV = lacto-ovo-vegetarian diet

- = not determined

NDF = neutral detergent fiber

Comments

Fiber supplement added to formula diets with and without cholesterol

NS= not statistically significant (p >0.05)

T = toasted

TG = triglycerides

UT= untoasted

veg= vegetables

VLDL = very-low-density lipoprotein

Oat bran, which is high in soluble gums, is an excep­tion to the general observation of lack of effect of brans; however, its hypolipidemic effects are not as pronounced in nor­mal volunteers (Table 7b) as they are in subjects with elevated serum lipid levels (Sections VI-C.2 and VI-C.3). Psyllium seed hydrocolloid also exhibited hypocholesterolemic effects and variable effects on serum triglyceride levelso

Soy products have been tested in several studies and have shown inconsistent effects on serum lipid levels. Two studies with legumes (Bengal gram and peas) showed a hypocholes­terolemic effect. No clear patterns of serum lipid response have emerged from studies of the consumption of fiber-containing foods such as fruits, vegetables, and grains.

In summary, insoluble fiber sources have generally been found to be ineffective in lowering serum cholesterol levels in subjects without lipid disorders. Preparations with soluble, viscous fiber components have exhibited hypolipidemic effects in such studies and have been found useful in treating patients with various hyperlipidemic disorders (see Sections VI-C.2 and VI-C.3). Animal experiments have generally supported the results obtained in human studies.

A number of mechanisms have been suggested to explain the hypocholesterolemic effect of certain dietary fibers. Much attention has been paid to the ability of dietary fiber to bind bile acids, leading to decreased absorption and increased excre­tion of bile acids, thus increasing the amount of cholesterol required for bile acid synthesis. Also, adsorption of bile salts in.the small intestine could interfere with micelle formation, reducing cholesterol absorption and increasing neutral steroid excretion, and thereby reducing the entry of dietary cholesterol into body cholesterol pools. However, the changes in steroid excretion are neither consistent enough nor large enough to fully explain the lipid-lowering effects of various dietary fiber com­ponents (Story, 1985). Attention has turned to effects of fiber on regulatory enzymes of cholesterol and bile acid synthesis. Little is known about the effects of changes in the pool size of the minor bile acids on overall sterol balance. The short-chain fatty acids liberated by the bacterial metabolism of fermentable dietary fibers, especially propionate, may act to decrease hepatic synthesis of cholesterol, as demonstrated in animal studies by Chen et al. (1984).

Effects on lipoprotein secretion and metabolism may also contribute to the observed effects. With increased entry of cho­lesterol into the bile acid synthetic pathway, less is available for lipoprotein synthesis and thus, less very-low-density lipo­protein (VLDL) is available for secretion into the circulation. However, soluble fiber sources have been found to lower circu­lating LDL-cholesterol concentrations without altering VLDL-cho­lesterol concentrations. Since VLDL particles are the precursors of LDL particles, Chen and Anderson (1986) have hypothesized that

50

these observations may be explained by an increased catabolism of LDL. Soluble fibers can increase serum concentrations of acetate in fasted subjects (Cummings and Branch, 1986; Pomare et al., 1985); acetate has been shown to inhibit cholesterol synthesis in hepatocytes (Beynen et al., 1982,) which might result in an increase in peripheral LDL receptors and LDL clearance. Another possible influence is the presence of tocotrienols (which have been found to be very effective inhibitors of cholesterol syn­thesis) in some plant food sources of dietary fiber (Qureshi et al., 1986).

G. POSTPRANDIAL SERUM GLUCOSE AND HORMONE LEVELS

Numerous studies have evaluated the effects of dietary fiber on glucose tolerance and serum levels of glucose, insulin, and various gut hormones. Many of these studies were carried out in diabetic subjects and are summarized elsewhere (Section VI-C.2b). The vast majority of studies evaluating the effects of dietary fiber on glucose tolerance have been short-term, single­meal studies [summarized by Wolever and Jenkins (1986)]. Of these short-term studies, approximately 85% demonstrate a lower­ing of blood glucose or improvement in glucose tolerance with high-fiber treatment. About 75% demonstrate a decrease in serum insulin levels or an enhanced insulin sensitivity with high-fiber treatment. These short-term studies are approximately equally divided into those using high-dietary-fiber foods as compared to low-fiber foods and those using a dietary fiber supplement such as guar, pectin, ispaghula, or wheat bran. The majority of the studies using dietary fiber supplements evaluated soluble fiber and almost uniformly found enhancement of glucose tolerance and increased insulin sensitivity. The small number of studies that used insoluble fiber products such as wheat bran or ispaghula had positive results in less than 50% of the cases.

The physiologic significance of single-meal studies or short-term treatment with high levels of dietary fiber is ques­tionable since it is well known from animal studies that dietary fiber causes changes in intestinal weight, villus morphology, and intestinal motility which may take days to weeks to develop. Although long-term studies would be preferable, only a small num­ber have extended treatment with dietary fiber for 7 d or longer in normal subjects (Table 8). Long-term studies in diabetic sub­jects are more numerous and are discussed elsewhere (Section VI­C.2b). Albrink and Ullrich (1982) reviewed their experience with high-fiber diets compared to low-fiber liquid or solid diets. A high-fiber diet (60 g NDF) fed for 7 d produced a lower insulin response to a standard meal than did a low-fiber (1 g NDF) liquid diet (Albrink et al., 1979). Serum glucose levels were not different. When both high- and low-fiber diets consisted of solid food, there was no difference in glucose and insulin levels (Albrink and Ullrich, 1982). Fukagawa et al. (1984) fed 35 g/1000 kcal of dietary fiber from a mixed diet and found lower fasting glucose and insulin levels and enhanced tissue insulin

51

Table 8. Effects of dietary fiber on postprandial serum glucose and insulin.*

% Change Study Fiber Source

Reference Design (and Test Dose) Duration Subjects Insulin Serum Glucose Comments

Albrink et al. SC,R Mixed foods 7 d 7 men dee NS Liquid formula control diet (1979) (>60 g/d NOF) containing l g/d of NDF

Albrink & SC,R, Mixed foods 10 d 7-8 men NS NS Three diets containing O, 18, & 36% Ullrich (1982) co ( >68 g/d NDF) per group NS NS sucrose; only 36% sucrose had an

dee NS effect on insulin compared to control

Fukagawa SC,AS Mixed foods 4 wk 6 men dee dee Fasting glucose and insulin and et al. (1984) (35 g/1000 kcal) glucose disposal by euglycemic clamp

all decreased \J1 N

Jenkins et al. SC,AS Pectin 6 wk 3 men NS NS (1977a) (36 g/d)

Munoz et al. C,MW HRS wheat bran 30 d 5-8 men NS NS Oral glucose tolerance tests (1979) SW wheat bran per group NS NS performed after 30 don diets

Corn bran NS dee (60 min) Soy hulls NS dee (30, 60 min)

Apple powder NS dee (30, 60 min) Carrot powder NS dee (30, 60 min)

(26 g/d)

* See code for abreviations at end of table; amounts of fiber reported are dependent on the method of fiber analysis.

\JI \,-J

Table 8. Effects of dietary fiber on postprandial serum glucose and insulin (continued).

study Fiber Source Reference Design (and Test Dose) Duration

Nestel et al. SC,AS, Mixed foods 10 d (1984) MW (100 g/d)

Villaume et al. SC,AS Wheat bran 7 wk (1984) (11 g/d)

AS= alternating sequence

C = controlled

CO= crossover

dee= decreased

% Change

Subjects Insulin Serum Glucose Comments

7 men NS NS

5 men inc dee

CODE FOR ABREVIATIONS IN TABLE 8

HRS= hard red spring

inc= increased

MW= metabolic ward

NDF = neutral detergent fiber

- = not determined

Glucose production rate and glucose utilization rate measured

Test meals at 0, 1, 3, 7 wk; glucose curve progressively decreased; insulin wave rose at 7 wk

NS= not statistically significant (p >0.05)

R = random allocation

SC= self-controlled

SW = soft white

sensitivity as assessed by the insulin clamp technique. Jenkins et al. (1977a) fed three medical students 36 g of pectin daily for 6 wk and noted no change in postprandial serum glucose and insulin levels. Munoz et al. (1979) fed several different high­fiber preparations including wheat and corn brans, soy hulls, and freeze-dried apple or carrot powder to male subjects hospital­ized on a metabolic ward for 30 d. Oral glucose tolerance and insulin sensitivity improved with most of the fiber preparations. Nestel et al. (1984) fed 16 or 100 g of mixed dietary fiber in a high-carbohydrate (62%) diet for 10 d and found no changes in glucose utilization or production. Villaume et al. (1984) fed 20 g of wheat bran (about 9 g of dietary fiber) for 7 wk and evaluated glucose and insulin response to a test meal at O, 1, 3, and 7 wk. Serum glucose levels decreased at 30 and 60 min during the test meals on fiber, but insulin levels rose and by 7 wk were significantly higher during the test meal. These data imply that insulin sensitivity was decreased. Thus, of 14 trials in these studies using fiber for an extended period in normal individuals, six showed no effect. Because a greater majority of studies in diabetic patients (Section VI-C.2b) show effects of dietary fiber on glycemic control, it seems likely that subjects with abnormal glucose tolerance are more responsive to dietary fiber than are normal subjects. In addition, the differing proportions of simple and complex carbohydrates in the test diets may be a confounding factor.

Although most studies of dietary fiber have focused on glucose and insulin levels, a few studies have evaluated other hormones. Most studies of gastric inhibitory polypeptide (GIP) response to dietary fiber have shown a decrease with fiber sources including wheat bran (Beck et al., 1986), brown rice (Collier and □ 'Dea, 1982), lentils (Collier et al., 1984), psyllium (Florholmen et al., 1982), whole-grain bread (Hagander et al., 1985), and guar (Jenkins et al., 1979b; Morgan et al., 1979); however, some studies have shown no change with whole­grain bread and apples (Hagander et al., 1984), guar (Jenkins et al., 1980a), or guar and pectin (Levitt et al., 1980). Glucagon has generally been found to decrease with dietary fiber treatment including mixed foods (Crapo et al., 1980), whole-grain bread (Hagander et al., 1985), guar and pectin (Levitt et al., 1980), guar (Morgan et al., 1979), and apple or carrot powder (Munoz et al., 1979), but in one study with coarse wheat bran cereal, glucagon was found to increase with a test meal (Connell et al., 1980). Enteroglucagon decreased with a single-meal stim­ulation with guar and hemicellulose in one study (Monnier et al., 1982), but was not affected by guar in others (Jenkins et al., 1979b, 1980a). Gastrin and vasoactive intestinal peptide did not change with wheat bran (Beck et al., 1986; Connell et al., 1980). In one study, somatostatin was increased in normal subjects and in subjects with noninsulin-dependent diabetes after a meal with guar (Shimoyama et al., 1982). Most of the studies evaluating these hormone levels have been acute or single-meal studies. Further studies evaluating the chronic effects of dietary fiber on these and other gut hormones are indicated.

54

INTESTINAL MORPHOLOGY AND CELL PROLIFERATION

The morphology of the small intestine differs from that of the large bowel in that the villi protrude from the mucosal surface (Creamer, 1974). The villus cells produce enzymes (peptidases, carbohydrases) that digest nutrients and contain transport sites that assimilate digested nutrients. Like the large bowel, the small intestinal mucosa is composed of crypts, the site of epithelial cell proliferation and development. Epithelial cells do not complete maturation until they have migrated out of the crypt and onto the villus. Epithelial cell proliferation and turnover are greatest in the proximal small intestine and decline with progression down the intestine, the ileum showing lower levels than the jejunum. The large bowel has an even slower rate of cell proliferation than the ileum Factors influencing intestinal morphology and cell proliferation include diet, pancreatic and biliary secretions, hormones, enteric microflora, blood flow, luminal nutrients, intestinal resection, and intestinal mucosal cell growth factors (Sprinz, 1971; Vahouny and Cassidy, 1986; Williamson, 1978). Changes in small bowel mucosal morphology and cell proliferation, in turn, influence digestion and absorption (Vahouny and Cassidy, 1986). The intestinal microflora is also an important factor, since in germ-free rodents intestinal cell proliferation rates are decreased (Sprinz, 1971).

This section reviews the data from human and animal studies concerning the effects of dietary fibers on the small and large intestinal morphology and cell proliferation.

1 . Morphology

Variations in small intestinal morphology of humans from different geographic locations have been described. In European adults, the jejunal villi are predominantly finger­like. Africans and Asians from-Uganda have predominantly leaf­shaped villi which are shorter and broader. The reason for this difference is unknown, but enteric organisms have been suggested (Cook et al., 1969). A post-mortem study from India has shown that fetuses have finger-like villi, but that within the first month of life many of the villi are leaf-like (Chacko et al., 1969). This change increased with development, becoming maximal in older children. Other changes noted include tongue-shaped villi, ridges, and convolutions. The most marked changes were seen in the duodenum and proximal jejunum, the abnormalities diminishing distally (Chacko et al., 1969). Possible etiologies considered include infection, infestation, diet, and malnutri­tion. Small bowel biopsies from vegetarians living in San Francisco (Owen and Brandborg, 1977) showed fusion of villus tips, branching and broadening, but not shortening. Increased numbers of plasma cells and lymphocytes were present; villi were broad, irregular, and convoluted. The authors concluded that

55

broad, branched, and fused villi of normal height in vegetarians should not be interpreted as pathological.

In summary, human data show geographic variations and developmental changes in small bowel mucosal morphology. These changes may be related to environmental factors, including diete There is no direct evidence to indicate that consumption of fiber produces these changes (which could result from dietary deficien­cies or exclusions) or that these changes are pathological.

Studies in rats have shown that diets containing 15% by weight of fiber produce ultrastructural cell surface changes, detectable with electron microscopy (Cassidy et al., 1981). ·These abnormalities were present in both small intestine (jejunum) and colon, the severity being greatest with alfalfa> pectin> cellulose> bran. The authors suggested that these changes were related to the bile acid-binding capacities of the individual fibers. Another scanning electron microscopy study (Tasman-Jones et al., 1982) found no abnormalities with added cellulose while the addition of 10% pectin produced changes similar to those seen with a commercial laboratory ration when fed to newly weaned rats, namely progression from leaf-like villi to broad-leafed, long-ridged villi. Rats fed cellulose developed more villi in the jejunum and ileum than did the rats fed a com­mercial laboratory ration. These results fail to support a bile acid-binding mechanism for the fiber effect. Light microscopy studies found increased small intestinal length, weight, crypt length, and muscle thickness, but lower levels of brush-border digestive enzymes and no change in glucose absorption with 18% pectin (Brown et al., 1979). Sigleo et al. (1984) fed 10% pectin or cellulose, and found that these rats ate more·and grew faster than control animals fed no fiber. Small bowel villus height was increased with both diets as was the absorption of a hexose and an amino acid. However, because the presence of luminal nutri­ents may modify small bowel structure and function (Dworkin et al., 1976), it cannot be concluded that the fiber alone stimulat~d growth and absorption. Calvert et al. (1985) fed fiber sources at levels of 5-10% and found that rats fed fiber ate less and weighed less. Small intestinal mucosal protein levels were generally lower. Guar gum and Metamucil® increased proximal intestinal invertase and alkaline phosphatase, and distal thymidine kinase. However, in studies in which food intake and weight gain were comparable, fiber-containing diets increased small and large intestinal weights (Younoszai et al., 1978). Elsenhans et al. (1981) fed fiber sources at levels from 10 to 40% by weight. Body weight gain was slower due to reduced energy intake. Small intestinal length increased but not mucosal mass (DNA content). Degradable fibers such as guar, tragacanth, and gum arabic increased cecal weight whereas carrageenan, gum karaya, and methylcellulose increased colonic weight. Therefore, the site and degree of fiber degradation appear to influence intestinal growth. Poksay and Schneeman (1983) found that the feeding of 10% guar increased small intestinal weight despite decreases in food intake and body weight. Ecknauer et al. (1981)

56

were unable to prevent the reduction in villus size that occurs with feeding an elemental diet by supplementing the diet with 24.3% cellulose plus 10.8% petroleum jelly.

In the large intestine, wheat bran fed at a 20% level for 2 wk produced mucosal hyperplasia, deeper colonic crypts, and increased muscle thickness (Jacobs and Schneeman, 1981). In rats fed wheat bran for 4 wk, mucosal hyperplasia (DNA level) was greater in the cecum than in the colon, the distal colon being greater than the proximal (Jacobs and White, 1983). Pectin and guar, but not oat bran, produced cecal and proximal colonic hyperplasia (Jacobs and Lupton, 1984). Guar and wheat bran stimulated growth of the distal colon.

The morphology data from animal experiments are diffi­cult to interpret due to the failure of many studies to control for changes in total nutrient intake and weight gain. Studies in which food intake increased with the addition of fiber supple­ments show increased mucosal growth, but this may be the result of increased luminal nutrition and not the addition of fiber. Degradable (fermentable) fibers generally appear to increase intestinal weight and length. The effects of fiber supplements on rat colonic circular muscle cell size showed that wheat bran produced proximal and distal hypertrophy, pectin and guar had no effect, and oat bran produced a decrease in cell size in the proximal colon (Jacobs, 1985).

2. Cell proliferation

There are few pertinent studies of intestinal cell proliferation in humans. Lipkin et al. (1985) found that Seventh Day Adventist (SDA) vegetarians had lower levels of rectal mucosal cell proliferation than subjects with sporadic adenomas or familial colon cancer (previously affected or asymp­tomatic). The level of cell proliferation in the SDA was similar to that in a "health conscious-population". Colonic crypts were longer in the SDA and health conscious populatirin than in the groups at high risk of cancer. No dietary intake data were pro­vided for these groups. Addition of calcium supplements appeared to decrease rectal cell proliferation in subjects with familial colon cancer (Lipkin and Newmark, 1985). Obtaining data on usual food intake in such studies is essential before any conclusions can be reached about the effect of fiber on human large bowel cell proliferation. · Furthermore, is not clear whether changes in rectal mucosal cell proliferation reflect similar changes occurring in the proximal and distal colon.

Animal studies show that small intestinal and colonic epithelial cell proliferation decrease with fasting. Similarly, when luminal nutrients are excluded from the colon, there is decreased epithelial cell proliferation. With refeeding there is an increase in small bowel and colonic epithelial cell pro­liferation and crypt cellularity (Hagemann and Stragand, 1977).

57

The effects of dietary fibers on intestinal cell proliferation in animals have been reviewed (Jacobs, 1986a, 1987). Feeding of diets containing 10% guar increased jejunal mucosal cell mass, whereas feeding 10% pectin decreased villus height and increased crypt length in the small intestine. Both pectin and guar increased the rate of villus epithelial cell migration in the small intestine (Jacobs, 1983a). Animals fed diets containing 4% guar gum, but not 10% cellulose, had an increased small intestinal crypt cell production rate (Johnson et al., 1984). Lactase and alkaline phosphatase levels were decreased with dietary guar.

In another study, rats fed 10% wheat bran or cellulose ate more, grew more rapidly, and demonstrated increased jejunal epithelial cell proliferation and faster cell migration rates (Vahouny et al., 1985). Mucosal glycoprotein synthesis appeared to be greater in the fiber-fed groups, but goblet cell number was decreased with cellulose. Schneeman et al. (1982) found that consumption of wheat bran increased goblet cell numbers in the small intestine. Changes in the intestinal surface mucus coat could alter intestinal function and predisposition to patho­logical processes (Vahouny and Cassidy, 1986). Large bowel epithelial cell turnover was slower in the distal colon of rats fed either 10% pectin or 10% guar, but faster in the proximal colon of rats fed 20% oat bran (Jacobs and Lupton, 1984). Johnson and Gee (1986) found that 10% guar or 10% carboxymethyl­cellulose in the diet increased the small intestinal crypt cell production rate when compared with cellulose (Solkaflo~); how­ever, in the large bowel, guar only increased cecal crypt cell production rate and carboxymethylcellulose only increased colonic crypt cell production rate. A diet containing 20% wheat bran increased epithelial cell proliferation in the crypt base of the cecum and the upper part of the proximal and distal colonic crypts (Jacobs and White, 1983). This same study found a decreased rate of epithelial cell migration in the cecum but faster migration rates in the proximal colon. Goodlad and Wright (1983) added 30% kaolin or cellulose (Solkafloc®) to an elemental diet fed to mice. Cellulose but not kaolin prevented the mucosal atrophy of the colon in mice fed an elemental diet. Decreased muscle bulk accounted for most of the reduction in intestinal weight~ Changes in fecal bulk (weight) did not correlate with colonic mucosal uptake of labeled thymidine in rats fed a-cellulose (Sircar et al. 1983).

3. Relationship of changes in intestinal cell proliferation to the development of disease

As mentioned earlier, higher rates of human colonic cell proliferation are associated with greater risk for colon cancer (Lipkin et al., 1985). Factors that decrease colonic cell proliferation are thought to have antineoplastic potential (Lipkin and Newmark, 1985). The question has, therefore, been raised as to whether the stimulating effects of individual

58

dietary fiber sources on intestinal growth could enhance tumor development. Evidence to support this hypothesis may be found in animal studies in which wheat ~ran was fed concurrently with carcinogen exposure (Jacobs, 1983b; Jacobs, 1984) and with the continuous feeding of guar gum and pectin (Jacobs and Lupton, 1986). Currently, there is no evidence that fiber stimulates small or large bowel epithelial cell growth in humans. Other nutrients may also influence cell proliferation. Calciu~ in humans (Lipkin and Newmark, 1985) and vitamin C in animals (Deschner et al., 1983) have been shown to decrease cell pro­liferation; dietary fat in animals has been shown to increase proliferation of distal colonic cells (Bird and Stamp, 1986; Jacobs and Amorde, 1986).

Documentation of food intakes and cell proliferation measurements in the same individuals will be required before firm conclusions can be reached about the effects of fiber on human intestinal epithelium and any possible implications for human health. Data in animals convincingly show that individual dietary fibers can alter small and large intestinal morphology and cell proliferation. In addition, evidence suggests that these growth effects can stimulate experimental colon carcin­ogenesis.

I. COLON MICROFLORA

The human colon contains an enormous microflora within its lumen that constitutes a portion of a complex ecosystem (Moore, 1978). The components of the system are in dynamic interaction and consist of (1) secretions delivered from the small intestine, including electrolytes, water, and protein material; (2) undigested and partially digested foods passing through the small intestine into the colon; (3) metabolic reac­tions occurring actively within the colonic epithelium; and (4) over 400 species of aerobic and anaerobic bacteria, unknown populations of viruses and fungi, and occasional protozoa and helminths. When any one of these components is significantly altered, the balance of the ecosystem is changed, which may in turn change any single component or all of the other components.

The bacterial concentration in the colon is 1011 to 1012 colony-forming units/ml with anaerobes outnumbering aerobes 100-1000:l depending on the population or individual studied. The predominant organisms are Bacteroides, Bifidobacterium, Eubacterium, anaerobic gram-positive cocci, clostridia, entero­cocci, and various enterobacteriaceae (Hill and Drasar, 1975; Simon and Gorbach, 1984). Bacteria comprise 55% of the solid part of feces of subjects eating a typical Western diet (Stephen and Cummings, 1980a). However, the facts that there are over 400 species of aerobic and anaerobic bacteria in the colon and that isolation techniques are laborious have made the acquisition of needed information difficult.

59

Initial reports suggested that humans ingesting high­fiber diets had a larger aerobic than anaerobic bacterial colonic flora (Drasar and Hill, 1972). However, further studies on the relationship of dietary intake to the colonic flora have revealed inconsistent results. Some have shown that the proportion of anaerobic bacteria increased when intake of wheat bran increased (Fuchs et al., 1976), but others have not demonstrated any qual­itative change (Baird et al., 1977; Drasar et al., 1976; Finegold and Sutter, 1978). Although the bacterial flora appears to remain qualitatively stable, most investigators agree the total bacterial mass increases with increased fiber intake, especially with increasing wheat bran intake (Drasar et al., 1976; Fuchs et al., 1976).

Although the qualitative bacterial flora is constant in individuals, metabolic effects can be demonstrated which in turn may affect the ecosystem. The polysaccharide components of fiber are degraded by bacterial fermentation to carbon dioxide, hydro­gen, methane, and short-chain fatty acids (SCFA). The principal SCFA are acetic, lactic, propionic, and butyric acids (Stephen and Cummings, 1980a), and these acids are the main anions in human fecal material (Wrong et al., 1965). The rate and degree of production of SCFA depend on the type of substrate present and numerous chemical factors, such as pH and redox potential. Increasing intake of a high-fiber food, such as red kidney beans, will increase bacterial fermentation and production of SCFA (Fleming and Rodriguez, 1983; Fleming, et al., 1985). The human colonic flora is capable of digesting from 100% of fiber aspectin to less than 50% as cellulose and certain hemicelluloses (Bryant, 1978; Ehle et al., 1982; Salyers et al., 1985).

Acetate, butyrate, and propionate are absorbed from the human colon, and some evidence suggests that butyric acid may be an energy source for the colon epithelium (Dawson et al., 1964; McNeil et al., 1978; Pomare et al., 1985; Roediger, 1980, 1982). Bicarbonate appears to pass into the colonic lumen at the time of SCFA absorption (McNeil et al., 1979) and SCFA stimulate sodium resorption (Roediger and Moore, 1981). The role played by SCFA produced in the human colon and the full significance of bacte­rial fermentation in the colon remain to be elaborated. However, it is apparent that the enormous bacterial flora is capable of having significant local and metabolic effects through the fer­mentation process (Cummings, 1983; Topping and Illman, 1986).

- Gastrointestinal bacteria also play a role in maintain­ing the enterohepatic circulation of bile acids and other com­pounds. This role is fulfilled largely in -the small bowel, but the deconjugation and dehydroxylation of bile acids can occur in the colon; the species responsible are present throughout the intestines. The relationship of fiber to bile acids is discussed elsewhere (Section V-E), but it should be noted that fiber, bile acids, and colon bacteria interrelate in the ecosystem. Bac­terial enzymes degrade bile acids, and bile acids in turn can inhibit bacterial growth (Floch et al., 1971). Control of

60

bacterial populations is also accomplished by some SCFA and pH changes induced by bacterial metabolism (Simon and Gorbach, 1986). The therapeutic intake of a carbohydrate subject to bacterial metabolism such as lactulose, will reduce cecal pH to 4.8-5.0, cause diarrhea, and decrease the absorption of ammoniumion, but only minimally affect the flora quantitatively and qualitatively (Elkington, 1970).

Colonic bacteria may have a role in colon cancer. One major mechanism under study is the influence of enzymes produced by colonic flora. In colon cancer patients, 7-a-dehydroxylase and cholesterol dehydrogenase levels were elevated (Mastromarino et al., 1976). However, a study of the fecal flora of colon­cancer and noncolon-cancer patients demonstrated changes compared to normal controls, but could not differentiate the patient groups (Vargo et al., 1980). Dietary influences have been sug­gested by the demonstration of increased bacterial production of B-glucuronidase on high-meat diets (Reddy et al., 1974). Much higher levels of 7-a-dehydroxylase and nitroreductase, as well as fecal B-glucuronidase, have been shown in omnivores as com­pared to vegetarians (Goldin et al., 1980). Dietary fiber intake also appears to affect bacterial metabolism of estrogens. The bacterial enzymes B-glucuronidase and sulfatase hydrolyze conju­gated estrogen to free estriol which is reabsorbed. Vegetarian women have a marked increase in fecal excretion of estrogen and decreased estriols (Goldin et al., 1982). Clearly, the colonic flora plays a major role in bile acid and steroid metabolism through its enzyme activity. How fiber reacts in this system, and long-term effects, are not yet clear. However, the most recent evidence demonstrates varied bacterial enzyme activity between subjects on high- and low-fiber intake.

In summary, the available data on responses of the colonic microflora to dietary fiber are relatively limited. Although most reported experimental effects do not appear to be detrimental to the host, and therapeutic use of various fibers has not demonstrated significant untoward effects, a number of unanswered questions concerning the fiber-microflora relation­ships deserve increased emphasis

J. COLON FUNCTION

In addition to containing a large bacterial flora which metabolizes substrates in the ileal effluent, the colon functions to conserve water and electrolytes during the formation of feces and acts as a reservoir for feceso Effects of dietary fiber on the colonic bacteria and their fermentation products are con­sidered.in Section V-I; other influences of dietary fiber on colon function will be considered here.

61

1. Transit time

The rate of passage of ingested material from mouth to anus varies from person to person and from time to time in an individual. Methodology is a major complicating factor in the study of transit time. Measuring the time from ingestion to appearance in the feces of a nonabsorbable dye, such as carmine or Brilliant Blue, is perhaps the easiest technique, but it does not permit quantitation and is entirely visual. Further, because only the first appearance of dye can be used as a marker, such transit time measurements may not measure the bulk of the meal with which the dye was ingested. The observation that dye transit is always shorter than transit times assessed with a quantifiable marker supports this speculation {Kotb and Luckey, 1972). Use of the liquid marker polyethylene glycol allows quantitative measurements of marker concentration, as does the use of various solid markers, such as radio-opaque pellets, radioactive chromic oxide, gravel, ball bearings, radio pills, or chromium sesquioxide (Roth and Leitzmann, 1985b; Van Soest et al., 1983). The use of both liquid and solid markers, as well as the use of solid markers of appropriate size and density, may be important for following the liquid and solid phases of digesta, although this has been little studied in humans. No clear differences were demonstrated between the transit of the solid and liquid phases of digesta in subjects when polyethylene glycol and radio-opaque pellets were ingested simultaneously during low-fiber and wheat bran-supplemented mixed diets (Marlett et al., 1986)0 However, the slow bacterial degradation of poly­ethylene glycol observed in this and other studies limits its utility as a marker when transit times are 4 d or longer (Marlett et al., 1986). Differences in transit to the cecum may exist between solid and liquid phases of a meal (Urban, 1984).

Gastrointestinal transit time is frequently .expressed as the time required for 80% of quantifiable markers to be excreted (Hinton et al., 1969). Because of the high intraindividual vari­ability seen with this method of calculation, Cummings et al. (1976b) proposed the determination of mean transit time (the sum of the products of the numbers of pellets in a stool and the time of excretion of that stool which is divided by the total number of pellets recovered). The results of Marlett et al. (1981), in a comparison of the two methods, suggested that true biological variation was detected by the 80% pellet method and that calcula­tion of mean transit time was not a more reproducible method.

In addition to diet, transit time can be altered by the act of defecation, suppression of the urge to defecate, drugs, emotional stress, and exercise (Roth and Leitzmann, 1985b). Except for foods which contain specific cathartics, such as prunes and rhubarb, the dietary fiber content of foods seems to be the most important variable in decreasing gastrointestinal transit time. Effects vary depending upon the source of fiber.

62

Wheat bran has been found by many investigators to decrease transit time with a variety of doses and measurement methods (Balasubramanian et al., 1987; Cummings et al., 1976a,b, 1978, 1979b; Harvey et al., 1973; Heller et al., 1980; Kirwan et al., 1974; Payler et al., 1975; Spiller et al., 1986; Stephen and Cummings, 1980b; Wrick et al., 1983; Wyman et al., 1976). Wheat fiber in the form of whole-meal bread was also found to decrease transit time (Eastwood et al., 1986). Coarse bran was more effective than fine bran (Heller et al., 1980; Kirwan et al., 1974; Wrick et al., 1983) and raw bran more effective than cooked bran (Wyman et al., 1976). In the studies of Harvey et al. (1973) and Payler et al. (1975), wheat bran (20 or 30 g/d) decreased transit time in subjects with initial times of 3 d or more and increased transit time in subjects with initial transit times of 1 d. In comparison to other fiber sources, Cummings et al. (1978) found that wheat bran was more effective in speed­ing mean transit time than fiber from cabbage and apple, or from carrot, none of which had an effect on transit time. Wrick et al. (1983) using three different markers found that bran low­ered mean transit time more than cellulose, which had an incon­sistent effect, or cabbage fiber, which had no effect on transit time.

Pectin was found to have no effect on gastrointestinal transit time (Cummings et al., 1979a; Kay and Truswell, 1977a; Spiller et al., 1980). The observed effects of added cellulose have been variable, with some studies showing a decrease in tran­sit time (Spiller et al., 1980; Wrick et al., 1983) and others showing no effect (Fleming et al., 1983; Slavin and Marlett, 1980). Bagasse had no effect on transit time measured by radio­opaque pellets, but did decrease transit time measured by carmine (Baird et al~, 1977). Soy fiber added to a purified liquid diet decreased transit time of healthy adults in a dose-dependent man­ner (Slavin et al., 1985). Soy polysaccharide fiber in concen­trations typically present in commercially available tube feed­ings has much less effect on gastrointestinal transit of a non­ambulatory population who might be expected to consume such a diet habitually. Substitution of fruits and vegetables for their corresponding juices in a diet with no cereal fiber also decreased transit time (Kelsay et al., 1978).

Roth and Leitzmann (1985b) noted that transit time is assumed to be the result rather than the cause of colonic events. However, Stephen and Cummings (1980b) showed that when transit time was slowed pharmacologically, stool weight decreased, and conversely when transit time was shortened pharmacologically, stool weight increased. Several investigators have noted a reciprocal relationship- between transit time and fecal weight in their studies of effects of dietary fiber (Findlay et al., 1974; Spiller et al , 1982). Thus, those fibers that influence stool weight, primarily those high in insoluble fiber (see Section V-J.2), appear to affect gastrointestinal transit time. Harvey et al. (1973) first proposed that dietary fiber "normalized" transit time, i.e., decreasing the time when transit was longer

63

than 3 to 4 d and increasing the time when it was 1 to 1.5 d. This relationship has not always been apparent in reports presenting only mean values, but when data on individual responses have been published they tend to support the hypothesis.

2. Stool weight

A wide range of values for stool weight has been recorded; however; what constitutes normal stool weight is not known. In a review of the literature on dietary fiber intake and stool output in normal groups, Eastwood et al. (1980) noted values ranging from 71 to 200 g/d. Most estimates of stool · weight for persons consuming a typical Western diet are in the range 100-150 g/d (Eastwood and Brydon, 1985). Diet has a major role in determining stool weight, but the effects of dietary fiber vary by source and from individual to individual.

Feces are a complex and nonuniform mixture of micro­organisms, undigested food residues, soluble ions, organic com­pounds, and water. Various fiber sources may act to increase fecal weight by increasing bacterial proliferation and production of fermentation products, increasing the amount of undigested residue in the colon, and by increasing the water content of stool; however, water content seems to be controlled fairly tightly in the range of 73-79% (Eastwood et al., 1980). Stool moistures of less than 70% have been observed in constipated individuals (Fischer et al., 1985; Marlett et al., 1987). Cummings (1986) has provided a comprehensive review of the effects of various fibers on fecal weight and has calculated the mean change in fecal weight per gram of fiber source con­sumed (see Table 9). Some effects of specific fiber sources on stool weight are discussed below.

Wheat bran has been shown to increase stool weight significantly in a variety of forms and doses (Baird et al., 1977; Balasubramanian et al., 1987; Bright-See et al., 1985; Brodribb and Groves, 1978; Cummings et al., 1976a, 1978, 1979b; Floch and Fuchs, 1978; Heller et al., 1980; Kay and Truswell, 1977b; Kirwan et al., 1974; Spiller et al., 1986; Stephen and Cummings, 1980b; Wrick et al., 1983; Wyman et al., 1976). The particle size of bran is an important determinant of its efficacy in promoting stool bulk, with coarser brans found to be more effective than fine brans (Brodribb and Groves, 1978; Heller et al., 1980; Kirwan et al., 1974; Wrick et al., 1983). Investi­gators have hypothesized that this effect is produced mainly by the water-holding capacity of coarse bran. Wyman et al. (1976) also showed that cooking bran can reduce its effectiveness as a fecal-bulking agent.

The effects of other fiber sources on stool weight also have been examined. Robertson et al. (1979) found that consump­tion of 200 g/d raw carrot had little effect on stool weight.

64

Table -9. Average increase in fecal output per gram of fiber fed.*

Fiber Source

Wheat bran

Fruit and vegetables

Oats

Gums and mucilages

Corn

Cellulose

Soya

Pectin

Increase in Fecal Weight**

(gig Fiber Fed)

5.7 + 0.5

4 .. 9 + 0 .. 9

3.9 + LS

3.5 + 0.7

3.4 + 0.4

3.0 + 0.6

2.8 + 0.8

1 3 + 0.3

* Adapted from Cummings (1986).

**Mean+ standard error of the mean.

65

Number of Studies

31

20

3

16

4

6

4

10

Comments

Raw bran, 6.7; cooked bran, 5.3

Carrot, cabbage, peas, apple, banana, potato, bean, and mixed sources

Oat bran or rolled oats

Psyllium, 3.5; guar, 2.3; sterculia, 1.7

Corn bran or corn germ meal

Methylcellulose, 8.2; carboxymethylcellulose, 4.0 not included

Soy bean hulls, soy pulp, and soy polysaccharide

Degree of methoxylation not important

With a diet containing a wide range of fruits and vegetables, Kelsay et al. (1978) found an increase in stool weight from 89 to 208 g/d. Cummings et al. (1978), in their study of fiber con­centrates, found that fecal weight increased 127% on bran, 69% on cabbage, 59% on carrot, 40% on apple, and 20% on guar gum. The change in fecal weight correlated with increased dietary intake of pentose-containing polysaccharides, but this can be only a partial explanation for the changes in fecal weight in view of the influence of particle size. Also, potatoes, turnips, cauli­flower, and rhubarb did not increase stool weight, in contrast to bran, carrots, oranges, apples, Brussels sprouts, and spring cabbage (Eastwood, 1978; Eastwood et al., 1983).

Cellulose (Fleming et al., 1983; Slavin and Marlett, 1980; Spiller et al., 1980), bagasse (Baird et al., 1977), and ispaghula (Prynne and Southgate, 1979) have also been shown to increase stool weight. Effects of pectin on fecal weight have been variable, with some studies showing no effect (Fleming

. et al., 1983; Kay and Truswell, 1977a; Spiller et al., 1980) and others showing a slight increase (Cummings et al., 1979a; Durrington et al., 1976).

Spiller et al. (1977) and Spiller (1986) have suggested that the relationship between fecal weight and transit time may be useful for assessing appropriate dietary fiber intake. Tran­sit time has been observed to decrease with increasing fecal weight to a point at which further increases in fecal weight do not correspond with a reduction in transit time. Few studies have examined this effect in a dose-response fashion. Two such studies (Kelsay et al., 1981; Spiller et al., 1986) indicated that an intake of 19.4 g/d NDF (from fruits and vegetables) or 32.4 g/d total dietary fiber (15 g from food and 17.4 g from hard red wheat bran in bread) resulted in fecal outputs in a range which the data would suggest produced optimal transit time.

In summary, depending upon the source, dietary fiber may have a distinct impact on stool weight and individuals may vary in their responses to fiber. In general, the insoluble fiber sources have a much greater impact on stool weight than do the soluble fibers. The presence of fiber residues appears to increase total stool water. However, because most studies have not demonstrated a significant change in stool moisture with fiber supplementation in healthy individuals, fiber residues appear to be no more effective at holding water in the colon lumen than do the other constituents in excreta.

3. Stool frequency

Studies of normal subjects consuming their usual diets indicate a range in stool frequency from three stools per day to three stools per week (Connell et al., 1965; Devroede, 1978; Rendtorff and Kashgarian, 1967). The range may be narrowed to between three and 11 stools per week when diet is controlled

66

(Devroede, 1978). Stool frequency is clearly affected by factors other than diet and has been shown to be poorly correlated with transit time (Devroede, 1978). Some, but not all, studies have shown that fiber sources which increase fecal weight and decrease transit time may increase stool frequency; however, when it occurs, this change is usually small in comparison to changes in the other two parameters (Balasubramanian et al., 1987; Fleming et al., 1983; Kelsay et al., 1978; Wrick et al., 1983; Wyman et al., 1976).

4. Fecal constituents

The effects of various fiber sources on the fecal excretion of bile acids, short-chain fatty acids, electrolytes, and minerals are considered in other sections of the report. In terms of modifications of fecal constituents by fiber, it is important to note that consumption of sources of cereal fiber (such as wheat bran) can lead to a dilution of the fecal contents (Eastwood and Brydon, 1985). On the other hand, water-soluble, fermentable sources of fiber (such as pectin) do not result in a significant increase in fecal weight, and thus fecal constituents are relatively more concentrated.

67

VI. EPIDEMIOLOGICAL STUDIES AND HUMAN TRIALS OF HEALTH EFFECTS OF DIETARY FIBER

A. STRENGTHS AND WEAKNESSES OF EPIDEMIOLOGICAL STUDIES

Epidemiological studies in nutrition examine associ­ations between dietary components and health status or disease. At the population level, associations can be shown readily between intake of a dietary component, such as dietary fiber, and biochemical measures or disease incidence. The correlates of incidence of a disease in a population when such associations can be shown, however, are not necessarily the same as the causes of disease in individuals (Morgenstern, 1982; Rose, 1985).

As compared with studies of populations, the demonstra­tion of relationships between dietary components and disease in individuals has been much more difficult because diets cannot be observed directly and inferences about them must be made using indicators with large degrees of uncertainty. This has been particularly true in the case of dietary fiber which is derived from many diverse foods. Measures of health status may also be subject to varying degrees of uncertainty (Ware et al., 1981). Nonetheless, when coupled with clinical, pathological, and experimental evidence, results from epidemiological studies are an important part of the evidence necessary to infer causality. Criteria for judging the relevance of an association between diet and disease in epidemiological studies include: (1) the consistency of the association; (2) the strength of the associa­tion; (3) the specificity of the association; (4) the temporal relationship of the association; and (5) the coherence of the association (U.S. Surgeon General's Advisory Committee on Smoking and Health, 1964).

Epidemiological studies can be designed to evaluate the differences among populations, examine the changes within popula­tions over time, study retrospectively the differences in groups with and without a disease, or follow a cohort prospectively. Regardless of the study design employed, the major limitation of epidemiological research in the area of nutrition and disease has been the uncertainty in the measurement of diet (Byers and Funch, 1984). In addition to the limitations of the methods typically used to assess food intake, the availability of reliable data for the dietary fiber composition of foods has been limited. Many older studies have used crude fiber values, an unreliable esti­mate of dietary fiber content of a single food. Better data for dietary fiber composition are now becoming available, although significant limitations remain for their use. In light of the differential physiological effects of different fibers, studies which examine relationships with the type and source of dietary fiber may provide more useful information than studies in which only total dietary fiber is examined. However, older studies based on crude fiber content or on foods or food groups may

69

still be useful because crude fiber, vegetable, fruit, and cereal intake are likely to covary considerably.

Some special considerations in interpreting different types of epidemiological studies are given below.

1. Ecological studies

In ecological (or correlational) studies, disease rates or physiological measurements of defined populations are correlated with their food usage or intake. This type of study is most useful for generating hypotheses. Such studies are, how­ever, subject to the "ecological fallacy" in which even strong correlations may be spurious if the true etiological factor is strongly correlated with both diet and disease (Morgenstern, 1982). Associations between disease risk or health status and dietary factors in populations observed in ecological studies cannot be necessarily interpreted as causal because of the heterogeneity of diets within populations and because of the number of potentially confounding variables (Byers and Graham, 1984). Diet is highly correlated with many other environmental or lifestyle factors, e.g., smoking, alcohol consumption, and general socioeconomic levels, all of which may be related directly or indirectly to disease risk. In addition, there may be associations between various elements of the diet, such as fat and fiber which are inversely correlated in the international data. Attempts to detect and account for confounding variables are essential for the proper interpretation of ecological data, but information about potential confounders is frequently lacking and statistical instability for highly collinear variables often makes multivariate analysis impossible.

International studies typically compare disease inci­dence and/or mortality rates in different countries to per capita consumption of various foods or nutrients, as measured in special surveys or inferred from food production and/or "disappearance" data. Simple correlation coefficients or multiple regression analyses are usually reported. Many sources of error exist in such data, for instance, the use of food for purposes other than human consumption. Special dietary surveys are sometimes employed but are rarely conducted simultaneously in several countries. Surveys using a single diet recall or record are adequate for estimating population mean levels of intake, but are not adequate for classifying individuals with regard to intake. The accuracy .and completeness of disease data also may vary from country to country.

Comparing regions within a country is advantageous because significant variation due to the measurement of both diet and disease can be controlled. However, there may also be less variation in disease incidence and/or dietary factors in such studies as compared to international studies. Studies of a single population over time or studies of migrant groups

70

have an element of internal control not present in more typical epidemiological studies, but such studies are often constrained by the lack of good dietary data for the distant past. More­over, migrants are self-selected and often make many cultural adaptations in addition to diet. Studies of selected religious, racial, and ethnic groups are often useful for hypothesis gen­eration because these groups frequently differ significantly in dietary habits and other factors from others in the same general population.

2. Individual-based studies

In case-control (retrospective) studies, the usual diet is determined for a series of patients with a specified disease or condition and for a control group. Experimental and cpntrol subjects are drawn from the population within the same community or from the same hospital or group of hospitalse Direct comparisons of the dietary intake of cases and controls are made and the potential contribution of confounding variables can be minimized either by the careful matching of control sub­jects, or by statistical adjustment methods in analysis. This type of study is very efficient in design, but can be affected by uncertainties regarding the representativeness of both the cases and controls and the validity of the dietary measures. Persons selected as cases should be a representative sample of all such persons within a defined population. Control subjects should be representative of the general population from which the cases are drawn and should be selected in a fashion that will minimize bias by the risk factors to be measured.

Case-control studies frequently fail to confirm strong associations observed in correlational studies of diet and disease. Whether this failure is caused by the better control of confounding variables, by the inability of individual-based studies to make sufficiently valid measurements of diet, or by insufficient variation of diet within a single population (or by a combination of these factors) is not known (Byers and Graham, 1984).

In case-control studies, food frequency methods are used most commonly to estimate dietary intake. The food fre­quency method tends to give an upward bias in the estimate of usual intake. The number of foods included may vary substan­tially in different studies, but sometimes a relatively small number of selected food items is used. In some cases, useful information can be obtained because many nutrients of interest are almost entirely contained in a relatively small number of foods (Madan et al., 1981). Because dietary fiber can be obtained from many sources, a restricted list is probably not appropriate for assessing dietary fiber intake.

Although the absolute levels of nutrients may be biased either upward or downward based on food frequency measures, there

71

is evidence that the food frequency method properly ranks sub­jects according to nutrient intake (Willett et al., 1985). In the analysis and interpretation of food frequency data, the researcher is interested only in the proper classification of individuals into appropriate relative rank order within the population with regard to the nutrient of interest (Byers and Graham, 1984).

Case-control comparisons may be made according to all levels of food frequency measured. More commonly, food frequency categories are combined or the extremes of frequency are com­pared. The combining of categories of food frequency in analysis can affect statistical inferences, because arbitrary combinations may artificially inflate or deflate true underlying relative risk estimates. Stable estimates of risk depend on an adequate number of subjects in each category, especially the base (reference) category. The number of cases and controls in different exposure categories should be reported together with the relative risk estimates (Byers and Graham, 1984). When an a priori rationale for category definition is lacking, the use of tertiles, quar­tiles, or quintiles of the study population is logical (Byers and Graham, 1984).

Because diet in the past may have influenced develop­ment of the disease under study and because the disease itself may have caused alterations in the diets of cases, determination of current diet is not usually appropriate in retrospective studies. However, there has been little validation of methods for assessing retrospective dietary data. A study in Buffalo (Byers, 1986) of the reliability of a retrospective food fre­quency history has indicated misclassifications in the quintiles of dietary fiber similar to those reported by Willett et al. (1985) for dietary cholesterol. The effect of such misclassi­fication would be to bias relative risk estimates toward unity. Such misclassification does not appear to be large enough to totally obscure risks of the order suggested by international correlational studies. Little is known about the true between­person variablility of fiber intake in the U.S. population. In Western New York, however, there is a greater than two-fold difference in the midpoints of the upper and lower quartiles of dietary fiber data in a large group of population-based controls (Byers, 1986). This observation indicates that insufficient variation in dietary intake within a population may not explain the failure of case-control studies to detect associations between diet and disease. A major concern of case­control studies is biased reporting of past diet wherein the cases have a different pattern of reporting error than do the controls. This concern has not been directly or adequately investigated and therefore remains as a major problem in making inferences of true associations based on retrospective studies.

In cohort (prospective) studies, measures of diet and other potential risk factors are made on a large group of healthy individuals and subsequent disease rates are

72

then observed among those with various levels of exposure to those factors over a period of years. In such studies there is no need to attempt retrospective measures of diet•and, therefore, no concern about the possibility of biased recall. However, such studies necessitate very large numbers of subjects who must be followed over a relatively long period of time in order for a sufficient number of incident cases to be collected. Careful and reliable measurements of dietary intake, usually by food frequency methods which are used more because of their rela­tive ease and low cost, and careful follow-up are necessary for such studies. Nondietary factors, such as cigarette smoking and socioeconomic status, which may be associated with diet and also influence disease outcome also must be measured. Another major factor that should be monitored is change in diet during the study period.

B. EVALUATION OF METHODOLOGIES FOR HUMAN STUDIES

The ideal human study is the randomized controlled clinical trial in which subjects are randomly assigned to control and treatment groups and then followed to observe differences in physiological functions, biochemical measurements, or disease rates. Depending on the purpose of the study, the subjects may be healthy volunteers or individuals with specific diseases, symptoms, or complaints. Studies may be long-term intervention trials or investigations of short-term acute effects.

Intervention studies present considerable methodological difficulties, including a heavy dependence on the motivation of the subjects to comply with the treatment. The simplest nutri­tional study design, in which supplements are added to the diet in a capsule or other nonfood form, is least demanding of the subjects. In the study of dietary fiber, this design is used most frequently in therapeutic intervention trials with isolated fiber sources. Modification of the total diet pattern (for exam­ple, iA studies of the effects of a high-fiber diet on diabetic control or obesity) often presents greater difficulties in assuring compliance.

Short-term clinical trials are frequently conducted to examine the effects of various fiber sources and isolated fiber constituents on physiological functions. In such studies, inves­tigators may be able to achieve good compliance with the treat­ments, especially if conducted in a metabolic ward. Because some fiber sources contribute energy, care must be taken in designing fiber supplements or fiber-containing meals for such trials to assure that diets are isocaloric and similar in content of elements other than fiber. The acute effects of dietary fiber observed in short experiments may not resemble those seen after longer periods of ingestion because of unknown adaptive changes in the gastrointestinal tract and in the colon microflora. Thus, an adequate period of treatment, often several weeks, is necessary for interpretation of results (Painter, 1985).

73

An adequate number of subjects and a sufficiently large change in fiber intake are also needed to detect differences associated with treatment.

In both intervention trials and clinical studies, treatments should be adequately randomized to prevent selection bias. If only one treatment is used, and subjects serve as their own controls, adequate baseline data should be collected and attention paid to randomizing the timing of treatment periods. Care must be taken when randomized, crossover designs are used for studies of dietary fiber because of the carry-over effects observed with some fibers (Van Soest et al., 1978). Ideally, subjects, and investigators as well, should be blind to the treatments applied throughout the st_udy. Obviously, blinding is not possible in studies in which total dietary patterns are altered, and may not be feasible when fiber supplements are given because of their recognizable form and physiological effects. One weakness of intervention studies and clinical trials is that compliance with the treatment may not be complete and there is often little means to determine adherence to the protocol. Marlett et al. (1986) have examined some methods of determining compliance with a dietary fiber supplement (wheat bran) and concluded that incorporating markers into the supple­ment and measuring fecal fiber may be useful for this purpose. A high level of fiber supplementation was necessary to increase fecal fiber sufficiently to utilize it as a measure of compliance, and the study was short term.

In extended trials, repeated measures of dietary intake are needed to detect changes as the study progresses; however, even periodic checks on dietary intake will not prevent the bias resulting from subjects trying to please the investigators or modifying their diets independently of the instructions given. Assuming no change in dietary intake or collecting inaccurate dietary intake data can result in an inaccurate estimate of an individual subject's response to the treatment or intervention. Random misclassification of individuals lowers correlation and regression coefficients and underestimates any association between diet and health status of the experimental groups. Biased misclassification, on the other hand, can create effects where none truly exist, or can hide effects that should be seen.

C. STUDIES OF POTENTIAL BENEFICIAL EFFECTS OF DIETARY FIBER

1 . Weight reduction/control

The use of dietary fiber for weight loss has found acceptance by the lay public and by some health professionals. However, an evaluation of the available literature reveals that data to support this conclusion are scanty. Table 10 presents a summary of the effects of various fiber supplements and high­fiber diets on weight reduction. Some of the studies have

74

Table 10. Studies on the effects of dietary fiber supplements and high-fiber diets on weight loss.*

Fiber Placebo Double Test . Type of No. of Weight Reference Source Control Blind Duration Dose Diet Subjects Change Comments

Yudkin (1959) Methylcellulose No No 6 wk 10 g/d low CHO, 20 F: -4.5 kg high fat NF: -2.7 kg

Duncan et al. Methylcellulose Yes Yes 8 wk 4.5 g/d low calorie 85 NS (1960) ( 1 Celevac 1

)

Evans & Miller Methylcellulose No No 1 wk 10 g/d usual diet 11 F: -0.6 kg Weight loss greater (1975) in obese subjects

Mickelsen et al. High-fiber Yes No 8 wk 12 slices/d low calorie 16 F: -8.8 kg Lower energy intake (1979) bread (cellulose) C: -6.3 kg with fiber diet

Weinreich et al. Wheat brnn No No 5 wk 24 g/d usual diet 25 F: -0.4 kg -....j (1977) \JI

Henry et al. (1978) Bread w/ bran Yes No 5 wk ? usual diet 17 NS Subjects not obese

Dodson et al. Wheat bran No No 3 mo 20 g/d high fiber 42 F: -7.4 kg Diabetic subjects and (1981) NF: -3.9 kg nondiabetic controls

Hylander & Wheat bran No No 3 wk 19.8 g/d low calorie 43 NS Rossner (1983)

Krotkiewski (1984) Wheat bran No No 3.3 wk 20 g/d usual diet 21 F: -0.6 kg/wk Alternating sequence

* See code for abbreviations at end of table; amounts of fiber reported are dependent on the method of fiber analysis.

Table 10. Studies on the effects of dietary fiber supplements and high-fiber diets on weight loss (continued).*

Fiber Placebo Double Test Type of No. of Weight Reference Source Control Blind Duration Dose Diet Subjects Change Comments

Rossner et al. Cereal & Yes Yes 8 wk 5 g/d OF low calorie 48 NS (1985) citrus fiber

Ryttig et al. (1985) Cereal & Yes Yes 11 wk 10 g/d OF low calorie 89 F: -6.3 citrus fiber C: -4.2

(1 Dum0Vital1 )

Tuomilehto et al. Guar Yes Yes 4 mo 15 g/d usual diet 33 F: -2.5 kg Subjects with hyper-(1980) C: -0.4 kg cholesterolemia

Valle-Jones Sterculia & No No 6 wk 30 ml/d low calorie 53 F: -3.6 kg (1980) guar ( 1 Prefil 1

) (20 g/d OF) NF: -1.8 kg

-...J Dodson et al. Guar No No 3 mo 20 g/d high fiber 42 F: -4.0 kg Diabetic subjects and 0\

(1981) NF: -3.9 kg nondiabetic controls

Evans & Miller Guar No No 1 wk 16 g/d · usual diet 11 F: -0.5 kg Weight loss greater (1975) in obese subjects

Krotkiewski (1984) Guar No No 8 wk 20 g/d usual diet 9 F: -4.3 kg

Guar No No 3.3 wk 20 g/d usual diet 21 F: -0.9 kg/wk Alternating sequence

Walsh et al. (1984) Glucomannan Yes Yes 8 wk 3 g/d usual diet 20 F: -2.5 kg (konjac) C: +0.7 kg

Hylander & Ispaghula No No 3 wk 19.8 g/d low calorie 44 NS Rossner (1983) (LunelaxID)

* See code for abbreviations at end of table.

Table 10. Studies on the effects of dietary fiber supplements and high-fiber diets on weight loss (continued).

fiber Placebo Double Test Type of Nao of Weight Reference Sour·ce Control Blind Duration Dose Diet Subjects Change Comments

Krotkiewski (1985) Ispaghula No No 4 wk 15 g/d usual diet 25 F: -2.2 kg Hunger ratings lower (Metamucil®) with oat bran fiber

Oat bran No No 4 wk 5 g/d? usual diet 25 F: -2.9 kg biscuits

Oat bran No No 50 wk 20 g/d low calorie 14 F: -16.4 kg NF: -10.6 kg

Oat bran No No 56 wk 20 g/d low calorie 122 F: -15.9 kg Lower drop-out rate NF: -14.l kg with oat bran

Russ & Atkinson Oat bran, wheat Yes Yes 12 wk 40 g/d OF formula+ 60 NS (1986) bran, guar low calorie

-...J Weinsier et al. Foods No No 4-6 mo Variable low calorie 26 F: -12 kg Exercise and behavior --..J

(1982) modification

Weinsier et aL Foods No No 26 wk Variable low calorie 60 F: -8 kg Exercise and behavior (1983) modification

Stevens et al. Foods No No 6 wk Variable usual diet 52 F: -0.7 kg Subjects with non-(1985) Foods & oat bran No No 6 wk +50 g/d modified 52 F: -2.3 kg insulin-dependent

NF: -0.4 kg diabetes

Russ & Atkinson Foods Yes No 8 wk ~25 g/d OF low calorie 32 NS Exercise & behavior (1985) (single) modification; compli-

ance difficult

CODE FOR ABBREVIATIONS IN TABLE 10

C = control treatment OF= dietary fiber NF= no-fiber treatment

CHO= carbohydrate F = fiber treatment NS= not statistically significant (p >0.05)

demonstrated weight loss in subjects treated with a purified dietary fiber supplement or high-fiber diet, but few of these studies included a placebo control, and even fewer used a double­blind design. In contrast, most of the studies which did not demonstrate weight loss were better controlled. In studies which demonstrated weight loss, the actual amount of loss has been small, even though reported to be statistically significant~ Interpretation of the results of these studies is complicated by differences in the experimental protocols; some studies have examined the effects of adding fiber supplements to the usual diet, some have used fiber supplements or fiber-rich foods in a low-calorie diet, and some have included exercise and behavior modification as part of the weight reduction program.

Most of the studies showing some weight loss have used a soluble fiber supplement or a variety of high-fiber foods, but others have used sources containing insoluble fibers (see Table 10). This observation of a greater effect with soluble than insoluble fiber is consistent with data from many other studies evaluating the effects of dietary fiber on blood glucose, insulin, and lipids (see Sections V-F and V-G). Soluble fibers appear to be the most effective in altering metabolic variables, and this might be expected to influence weight loss. However, this assumption is based on the hypothesis that fiber has meta­bolic effects that are not yet fully explained. Some investiga­tors have suggested that dietary fibers may alter the secretion of gut hormones (see Section V-G) which in turn promote weight loss. The delay in the absorption of nutrients by increased amounts of fiber in the diet may partially explain the alteration of gut hormones and metabolic rate and may promote fecal energy losses. Studies of patients with impaired absorption resulting from intestinal bypass surgery or short gut syndrome have demon­strated marked changes in gut hormone secretion (Atkinson et al., 1979; Bray et al., 1979). Welch et al. (1985) have shown that· infusion of fat into the ileum delays gastric emptying and decreases food intake. This mechanism may.come into play if dietary fiber slows the absorption of fat, permitting it to travel further down the intestine (see Section V-0).

An alternate hypothesis to explain weight loss on a high-fiber diet is that the increased bulk of such a diet inter­feres with the ingestion of high-calorie foods (see Sections V-A and V-8). Ingesting very large amounts of plant foods represents a marked alteration in dietary regimen for most subjects, and the poor compliance observed with such a regimen (Russ and Atkinson, 1985) may suggest that high-fiber diets are a poor long-term treatment modality for obesity.

In summary, the limited data from clinical trials sug­gesting that dietary fiber supplements or high-fiber diets are useful for weight reduction are contradictory. When a positive effect has been found, the total weight loss is modest. Long­term follow-up studies of the ability of subjects to maintain the weight loss have not been conducted. Dietary fiber may have a

78

limited role as an adjunct in the treatment of obesity, but con­trolled, long-term clinical trials are needed before this role can be established.

2. Diabetes

ae Epidemiological studies

Many hypotheses about the etiology of diabetes have been advanced, including a number of dietary explanations of the origin of the disorder. Trowell (1973) postulated that "pro­longed consumption of fiber-depleted starch is conducive to the development of diabetes mellitus in susceptible genotypese" Although much of the epidemiologic data supports this hypothesis, these associations do not prove a causal relationship (Trowell, 1974). Mann (1985) reviewed the evidence regarding dietary fiber and starchy foods as nutrition factors linked to the etiology of diabetes. Selected studies linking dietary carbohydrate and/or dietary fiber to the development of diabetes will be reviewed briefly.

Himsworth and Marshall (1935) reported a careful case­control study comparing diets of 143 diabetic subjects prior to the onset of their diabete$ to diets of 137 and 121 nondiabetic subjects in two control groups. They concluded that the diet of diabetic subjects before the onset of the disease contained the same proportion of protein, a lower proportion of carbohydrate, and a higher proportion of fat compared to the diet of controls. Diabetic subjects reported average intakes of 50.9% of energy as carbohydrate and 37.0% as fat while nondiabetic subjects reported 55.9% of energy intake from carbohydrate and 31.8% from fat. Presumably the fiber intake of nondiabetic subjects was greater than that of diabetic subjects, but these data were not provided.

Trowell (1960) reported that diabetes was a rare disease among African hospital patients. He suggested that this observation was related to the high-carbohydrate and low fat content of African diets. Walker (1961) suggested that high levels of cereal fiber intake might prevent the development of diabetes. Subsequently, Walker et al~ (1970) documented that healthy Bantu schoolchildren in rural South Africa had lower fasting blood glucose values than urban children; blood glucose values at 1 hr after admihistration of 50 g glucose orally also were significantly lower in Bantu children than in white childrene

Wapnick et al. (1972) confirmed these observations by performing oral glucose tolerance tests in African hospital maintenance workers, African students, and European students and lecturers. The African maintenance workers had a high-fiber intake from maize meal, whereas the other two groups did not. Fasting blood glucose values were significantly lower in African maintenance workers than in the Europeans, as were values at 1 hr

79

after glucose administration. Plasma insulin responses to oral glucose also were significantly lower in African maintenance workers than in Europeans. The ·African students who consumed "Western" diets had glucose and insulin responses that were intermediate between the African maintenance workers and Europeans.

West and Kalbfleisch (1971) compared the influence of nutritional factors on the prevalence of diabetes in 11 countries of Asia, South America, and Central America, and two locations in the United States. The dietary data were obtained from detailed nutrition surveys conducted by different protocols. The authors concluded that the most impressive and consistent association in these studies was between prevalence of diabetes and body fatness (weight/height). They also noted a positive association between prevalence of diabetes and per capita intake of fat and of sugar, and a negative association between prevalence of diabetes and per capita total carbohydrate consumption. In studies of these and other selected populations, West (1974a,b) reported a strong inverse association of diabetes prevalence and the percentage of calories derived from carbohydrate. Mann (1985) noted that the highest prevalence of diabetes in these studies occurred in affluent Pennsylvania where low-fiber white flour would have con­tributed nearly 30% of energy intake while the lowest prevalence occurred in East Pakistan where whole wheat, leguminous seeds, and vegetables would have contributed nearly 70% of energy intake.

In a review of diabetes death rates in England and Wales from 1920-1970, Trowell (1974) noted that the reduction in death rates from 1941-1957 was associated with use of a high-fiber flour. During the period of lower death rates from diabetes the intake of grains was higher and high-fiber National flour was used instead of white flour.

Trowell (1975) formalized the dietary fiber hypothesis of the etiology of diabetes. In addition to the experiencie with the low prevalence of diabetes in groups consuming high-fiber diets in Africa (Trowell, 1960; Walker, 1961; Walker et al., 1970) and the reduction in diabetes death rate during wartime (Trowell, 1974), Trowell (1973) noted that the prevalence of diabetes was low in Jews residing in Yemen where large amounts of whole-meal bread were eaten. After 20 yr of residence in Israel, the prevalence of diabetes in Yemenite Jews was higher and was associated with a lower intake of dietary fiber as well as other dietary changes (Cohen et al., 1961).

Thus, available evidence provides some support for a hypothesis linking fiber-depleted diets to the pathogenesis of diabetes. Nevertheless, because of the many additional differ­ences in diet, nutrition, genetics, environment, socioeconomic status, and physical activity between populations in primitive and urban industrialized cultures, this hypothesis must await further data to emerge as more than plausible speculation

80

(Anderson, 1980). Intervention studies (summarized in Sec-tion VI-C.2b) indicating that fiber-supplemented diets or high­fiber diets improve glycemic control, lower insulin requirements, and enhance peripheral sensitivity to insulin (Anderson and Bryant, 1986) suggest that dietary fiber may reduce the risk for developing diabetes in certain individuals.

b. Clinical trials and other human studies

The effects of dietary fiber in diabetes have been studied in three major ways. First, meal studies have used various fiber-containing foods or incorporated different types of fiber into test meals and examined postprandial glycemic and hormone responses. Second, fiber supplements have been added to diabetic diets and the short- and long-term effects on glycemic control, insulin requirements, and blood lipid concentrations have been examined. Third, high-fiber diets with varying levels of complex carbohydrate have been fed to diabetic subjects and short- and long-term effects have been studied.

Interest in the use of dietary fiber in the management of diabetes was stimulated by the meal studies of Jenkins et al. (1976b, 1978b, 1979b) which demonstrated that incorporating puri­fied fiber into meals could reduce the blood glucose response to mealse Their studies showed that the water-soluble fibers such as guar and pectin could reduce the postprandial serum glucose response of diabetic subjects to either mixed meals or an oral glucose load, but that water-insoluble fibers such as cellulose and wheat bran generally did not exert these effects. These results have been confirmed by others (Monnier et al , 1978; Morgan et al., 1979; Poynard et al., 1980). Jenkins et al. (1976b, 1978a) also observed that the serum insulin response with different fibers followed the same pattern as the blood glucose concentration. In addition, serum concentrations of GIP and glucagon have been shown to be reduced by addition of guar to a test meal.

As the study of effects on postprandial serum glucose has progressed, it has become clear that dietary fiber is not the sole determinant of glycemic response (Jenkins and Jenkins, 1985). The importance of physical form of the food was illus­trated by the study of Jenkins et al. (1983a) in which the post­prandial blood glucose elevations in diabetic subjects were found to be identical after white and whole-meal bread, but markedly reduced after white spaghetti. Interactions with other compo­nents of a meal may also have an influence (Jenkins et al., 1980b)e The effectiveness of guar in lowering the glycemic response is dependent on the viscosity of the preparation and on its incorporation into the test meal; effects have been found to be greater when guar is added to the liquid rather than the solid phase of the meal (Wolever et al., 1979).

81

Results of short- and long-term studies of fiber sup­plements in patients with diabetes are summarized in Table lla. Twenty-two of the studies of fiber supplements were conducted with guar products. These studies include individuals with type I (insulin-dependent) and/or type II (noninsulin-dependent) diabetes. The duration of studies ranged from 5-182 d with a median duration of 3 wk; the guar supplement dose ranged from 5-32 g/d with a median dose of 23 g/d. Different groups measured different parameters of glucose and lipid metabolism. The effects of guar compared with control diets can be summarized as follows: insulin doses were significantly decreased in four studies and not significantly altered in three studies with a mean change of 10.1% for the seven studies reporting these values; fasting blood glucose values were significantly reduced in six of 13 studies reporting these values with a mean reduction of 7.2%; 24-hr glycosuria was significantly reduced in 14 of 17 studies reporting these values with a mean reduction of 42%; serum cholesterol concentrations were reduced significantly in seven of 11 studies reporting these values with a mean reduction 8.9%; fasting serum triglyceride levels were significantly increased in two studies, decreased in one study, and not altered in six studies. Failure to observe effects of guar in some studies may have been due to short duration, inadequate dose of guar, or poor compliance with the supplementation regimen. The poor palatability of most guar preparations may have discouraged use; many investigators have reported unpleasant side effects with guar including nausea, vomiting, and abdominal distension, especially if it is not introduced gradually into the diet.

Wheat bran supplements have also been evaluated in diabetes, but most studies have found insignificant effects on glycemic control and serum lipid levels. Other fiber sources, such as apple powder, corn bran, soy hulls, and cellulose, have also been found to have little effect. Limited investigations of fiber supplements such as hairy basal seeds, glucomannon, xanthan gum, and psyllium suggest they may improve some aspects of glycemic control, but further research is needed to confirm this possibility.

In addition to fiber supplements, a variety of high­fiber diet regimens emphasizing fiber-rich foods, such as the diets developed by Anderson and colleagues (Anderson and Chen, 1979; Anderson and Ward, 1979; Kiehm et al., 1976), have been tested in diabetic subjects. Studies with such diets are sum­marized in Tables llb and llc. Most of these diets have a high proportion of complex carbohydrate (Table llb), although some have been devised with a more "standard" (<55% of calories) carbohydrate content (Table llc). These diets differ from tra­ditional diabetic diets, not only being higher in dietary fiber, but also higher in complex carbohydrates, lower in fat, and usually lower in cholesterol.

82

Table lla. Studies on the effects of fiber supplements in patients with diabetes.*

% Change

Fiber Study Study Test Insulin Fasting Glycos- Choles-Reference Source Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Jenkins et al. (1977b) Guar SC A 1-2 wk 25 g/d 4 OM - - -46 Guar SC MW 5 d 25 g/d 5 OM - - -39

Jenkins et al. (1978b). Guar SC A&MW 5 d 23 g/d 9 OM - NS -45 Guar SC A 8 wk 23 g/d 5 OM -22 - -73

Jenkins et al. (1979c) Guar C,R A&MW 5 d 24 g/d 18 OM - - -46 Guar SC A 12 wk 18 g/d 9 OM -20 - dee

Jenkins et al. (1979d) Guar C,R,CO A&MW 5 d 14-26 g/d 9 OM - - -38 - - Urinary 3-hydroxy-butyrate 40% lower

Jenkins et al. (1980b) Guar SC A&MW 5 d 24 g/d 10 DM - - NS - - Low CHO diet OJ v-1 Guar SC A&MW 5 d 24 g/d 14 OM - - -64 - - High CHO diet

Cohen et al. (1980) Guar P,R,SB A 3 mo 15-20 g/d 22 OM - NS - - - Subjects poorly obese controlled; guar

in capsules

Aro et al. (1981) Guar CO,R,08 A 3 mo 21 g/d 9 NIODM - -17 -60 -14 NS

Botha et al. (1981) Guar SC A 3 mo 32.l g/d 9 IDDM - NS - -17

Carroll et al. (1981) Guar SC A 4 wk 6-43.2 g/d 6 IDDM NS - dee -9 - Lower mean blood glucose

Dodson et al. (1981) Guar SC A 3 mo 20 g/d 8 IDDM -21 NS - NS NS High fiber diet

Johansen (1981) Guar SC A 4 wk 18 g/d 10 NIODM - NS -36 -14

·* See code for abbreviations at end of Table llc.

Table Ila. Studies on the effects of fiber supplements in patients with diabetes (continued).*

% Change

Fiber study Study Test Insulin Fasting Glycos- Choles-Reference Source Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Koepp & Hegewisch Guar SC A 4 wk 0.45 g/kg/d 10 IDDM -8 - - - - Lower plasma (1981) (<20 g/d) children viscosity

Kyllastinen & Guar SC A 2 mo 12-24 g/d 14 NIDDM - -8 NS -13 NS Lahikainen (1981) elderly

Stokholm et al. · Guar CO,R,DB A 1 wk 12 g/d 10 NIDDM - NS -31 (1981) elderly

Smith & Holm (1982) Guar SC A 3 wk 30 g/d 17 DM dee -8 - - NS 1 Lower postprandial glucose & insulin

Smith et al. (1982) Guar SC A l wk 8.2 g/d 8 DM - -18 CP Guar SC A 2 wk 9.15 g/d 5 IDDM NS -80 p -

II II II II II 5 NIDDM - -18 -62

Kuhl et al~ (1983) Guar AS MW 1 wk 24 g/d 12 IDDM NS - -63 - - Lower mean blood pregnant glucose

Gatti et al. (1984) Guar SC MW 2 wk 10-15 g/d 5 DM - -24 dee -17 -17

Najemnik et al. (1984) Guar R,CO A 2-4 wk 15 g/d 79 NIDDM - - -16 to - - Lower postprandial -50 glucose

Mclvor et al. (1985) Guar SC A 6 mo 32 g/d 8 NIDDM obese - - - NS +73

McNaughton et al. Guar C,R A 28 d 5 g/d 28 NIDDM - - - NS NS (1985) Guar SC A 12 wk 5 g/d 13 DM - - - -14 NS

* See code for abbreviations at end of Table llc.

Table lla. Studies on the effects of fiber supplements in patients with diabetes (continued).*

% Change

Fiber Study Study Test Insulin Fasting Glycos- Choles-Reference Source Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Mcivor et al. (1986) Guar P,R A 16 wk 32 g/d 18 NIDOM - - - NS +70

Ray et al. (1983) Guar & SC A 9 wk 20 g/d 12 NIDOM - -39 -73 -15 NS Subjects poorly wheat bran 10 g/d obese controlled

Kanter et al. (1980) Guar & SC A 3 d 16 g/d 6 OM - - - - - Lower postprandial pectin 10 g/d glucose & insulin

Gardner et al. (1984) Pectin SC A 3 mo 20 g/d 17 OM NS - - - - No change in HbAlc on insulin

Mayne et al. (1982) Apple fiber SC A 7 wk 15 g/d 12 NIDDM - -8 - NS NS

0) Mahalko et al. Apple powder C,R A 4 wk 26 g/d 10 NIDOM NS NS NS \JI - NS

(1984) Apple powder C,R A 4 wk 52 g/d 8 NIDDM - NS NS +6 NS

Soy hulls C,R A 4 wk 26 g/d 10 NIDOM - NS NS NS NS Soy hulls C,R A. 4 wk 52 g/d 8 NIDDM - NS NS NS NS

Corn bran C,R A 4 wk 26 g/d 10 NIDOM - NS NS NS NS Corn bran C,R A 4 wk 52 g/d 8 NIDOM - NS NS NS -15

Cohen et al. (1980) Wheat bran P,R,SB A 3 mo 25 g/d 22 OM - NS - - - Subjects poorly obese controlled

Bosello et al. Wheat bran SC A 30 d 20 g/d 38 IGT - - - -10 -11 OGTT improved (1980) Wheat bran SC A 3 mo 20 g/d 38 IGT - - - -10 -18 OGTT improved

* See code for abbreviations at end of Table llc.

Table lla. Studies on the effects of fiber supplements in patients with diabetes (continued).*

% Change

Fiber Study Study Test Insulin Fasting Glycos- Choles-Reference Source Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Dodson et al. (1981) Wheat bran SC A 3 mo 20 g/d 8 NIDDM - NS - NS NS High-fiber diet

Monnier et al. Wheat bran co MW 10-15 d 15 g/d 17 NIDDM NS NS -36 - - 10 labile subjects; (1981) lower HbAlc

Karlstrom et al. Wheat & C,R MW 3 wk ~42 g/d 14 NIDDM - -6 -37 NS NS (1984) rye elderly

Miranda & Cellulose co A 10 d. 20 g/d 8 IDDM - NS - - - Diet also changed; Horwitz (1978) crude fiber lower mean plasma

glucose

Viseshakul et al. Hairy basal SC A l mo 30 g/d 14 NIDDM + - -31 co 0\ (1985) seeds 2 IDll\1

Doi et al. (1979) Konja Mannon SC MW 20-30 d 3.6 or 13 NIDDM - -29 - -11 (glucomannon) 7.2 g/d

Osilesi et al. Xanthan gum C,R A 6 wk 12 g/d 9 NIDDM - -38 - -13 NS Lower postprandial (1985) glucose

Capani et al. (1980) Psyllium SC MW 9 d 21 g/d 9 OM - -29 - -13

Fagerberg (1982) Metamucil ® P,R A 2 mo 3.6 or 7.2 40 NIDDM - -19 NS -5 NS Subjects elderly 2 mo + 10.8 g/d & constipated

Frati-Munari et al. Psyllium SC A 10 d 45 g/d 8 NIDDM - dee - -11 -15 Subjects lost weight (1983)

* See code for abbreviations at end of Table llc.

Table.llb. Studies on the effects of high-fiber, high-carbohydrate (>55% of calories) diets in patients with diabetes.*

% Change

Study Study Fiber Insulin Fasting Glycos- Choles-Reference Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Kiehm et al. (1976) SC MW 2 wk 14.2 g/d · 13 OM -36 -26 - -24 -15 crude fiber

Anderson & Ward (1978) SC MW 2 wk 70 g/d 10 OM dee -26 - -24 -24 Subjects lost weight SC A 15 mo 40 g/d 10 OM dee -25 - -5 -43 on diets

Anderson & Ward (1979) SC MW 2 wk 65 g/d 20 OM -58 NS -17 -29 NS Subjects were lean

Simpson et al. (1979a) CO,R A 6 wk ? 14 NIDDM - -12 - -14 NS

Simpson et al. (1979b) C,R A 6 wk ? 12 IDOM -6 -36 - -10 NS

Anderson & Sieling SC MW 16 d 214-77 g/d 21 OM -78 -26 - -25 dee TG lower with weight 00 --..J (1980) obese loss

Anderson et al. (1980a) co MW 14-35 d 64 g/d 14 OM -56 NS - -32 -11 CO,AS MW 9-14 d 64 g/d 11 OM -45 -6 - -22 NS

Simpson et al. (1981a) C,R A 6 wk 96.6 g/d 9 IDOM -5 -38 -80 -15 NS Diet high in legumes C,R A 6 wk 96.6 g/d 18 NIDDM - -15 -94 -14 - and cereal fiber

Barnard et al. (1982) SC 11MW11 26 d 10-20 g/d 60 NIODM dee -28 - -19 -34 Subjects exercised crude fiber and lost weight

Pedersen et al. (1982) P,R A 4 wk 63 g/d 40 NIDOM -22 to NS NS -17 NS Lower plasma -32 ketone bodies

Ney et al. (1982) P,R A 16 wk 60-70 g/d 20 OM low - ·-85 pregnant

* See code for abbreviations at end of Table llc; amounts of fiber reported are dependent on the method of fiber analysis.

Table llb. Studies on the effects of high-fiber, high-carbohydrate (>55% of calories) diets in patients with diabetes (continued).*

% Change

Study Study Fiber Insulin Fasting Glycos- Choles-Reference Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Ward et al. (1982) C,R A 6 wk 100 g/d 7 NIDDM - -14 - - - Higher monocyte insulin binding

Anderson (1983) SC MW 2 wk N 65 g/d 12 OM -72 NS - -26 NS

Barnard et al. (1983) SC "MW" 26 d 35-40 g/d 69 NIDDM dee -26 - -25 -26 Subjects exercised and SC A 2-3 yr 35-40 g/d 52 NIDDM dee -13 - -19 -14 lost weight

Taskinen et al. (1983) SC MW 2 wk 55 g/d 10 IDDM - NS -40 -14 NS SC A 4 wk 55 g/d 10 !DOM - NS NS NS NS

Lindsay et al. (1984) SC MW 2 wk 64 g/d 12 !DOM NS NS children

CD McCulloch et al. (1985) CD R,P A 4-6 mo 31.8 g/d 40 !DOM - - - - - No change in HbA1 ; poor

l. c comp iance

Stevens et al. (1985) SC,R A 6 wk 19 g/d 12 NIDDM - -58 - -40 SC,R A 6 wk 25 g/d 13 NIDDM - -40 - -28 - Oat bran added

Story et al. (1985) SC A 48 mo 45 g/d 14 OM -72 NS - -11 -24 obese

Hollenbeck et al. (1986) CO,R MW 4 wk 26.7 g/ 6 NIDDM - NS NS NS NS High CHO control diet 1000 kcal

Pacy et al. (1986) SC A 3 mo 40-45 g/d 13 NIDDM - - - NS NS Some weight loss

* See code for abbreviations at end of Table llc; amounts of fiber reported are dependent on the method of fiber analysis.

Table llc. Studies on the effects of high-fiber, standard-carbohydrate (~55% of calories) diets in patients with diabetes.*

% Change

Study Study Fiber Insulin Fasting Glycos- Choles-Reference Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Rivellese et al. (1980) CO,R MW 10 d 5l~ g/d 8 OM -9 - - -14 -12 Lower postprandial glucose

Kay et al. (1981) CO,R 11MW11 14 d 30 g/d 5 NIDDM - - - - - Lower postprandial Nursing elderly glucose, insulin & GIP

Home

Manhire et al. (1981) CO,R A 6 wk 32.6 g/d 16 OM - - NS NS - Lower postprandial glucose

Kinmonth et al. (1982) CO,R A 6 wk 56 g/d 10 IDOM -7 -38 -76 children

0)

\0 Hj0llund et al. (1983) P,R MW 3 wk 53 g/d 18 NIDDM - -15 -77 -11 NS Increased insulin sensitivity

Rivellese et al. (1983) C,R MW 10 d 54 g/d 14 OM - - - -20 NS Lower postprandial glucose

Rosman et al. (1983) C A 12 wk 30 g/d 10 OM - - - - NS Negative correlations between fiber and mean glucose & TG

Karlstrom et al. (1984) C,R MW 3 wk ~42 g/d 14 NIDDM - -11? -37 NS NS elderly

* See code for abbreviations at end of Table llc; amounts of fiber reported are dependent on the method of fiber analysis.

\0 D

Table llc. Studies on the effects of high-fiber, standard-carbohydrate (<55% of calories) diets in patients with diabetes (continued).

% Change

Study Study Fiber Insulin Fasting Glycos- Choles-Reference Design Type Duration Dose Subjects Dose Glycemia uria terol TG Comments

Riccardi et al. (1984) C,R MW 10 d 54 g/d 14 OM dee NS - -16 NS Low-fiber, low-CHO control diet

Rivellese et al. (1985) C,R MW 10 d 65 g/d 5 IDDM w/ - NS - - - Lower postprandial renal failure glucose

CODE FOR ABBREVIATIONS IN TABLES lla, llb, AND llc

A= ambulatory

AS= alternating sequence

C = controlled

CHO= carbohydrate

GIP= gastric inhibitory polypeptide

HbAlc = glycosylated hemoglobin

IDDM = insulin-dependent diabetes mellitus

IGT = impaired glucose tolerance

CO= crossover low= insulin dose increments during pregnancy lower with high-fiber,

DB= double blind high-carbohydrate diet than with low-fiber control diet

dee= decrease reported, exact value or statistical significance not given MW= metabolic ward

OM= diabetes mellitus, mixed types or - = not determined types not specified

NIDDM = noninsulin-dependent diabetes mellitus

NS= not statistically significant (p >0.05)

OGTT = oral glucose tolerance test

P = parallel

R = random allocation

SB= single blind

SC= self controlled

TG = triglycerides

High-fiber, high-carbohydrate diets have been studied in insulin-dependent (IDDM) and noninsulin-dependent (NIDDM) diabetic subjects and in lean and obese diabetic subjects. In a majority of studies, such diets have significantly lowered insulin requirements, decreased fasting plasma glucose concen­tration, and reduced the urinary excretion of glucose. Insulin doses were decreased by a mean of 41.4% in 10 of 17 studies, fasting serum glucose was reduced by a mean of 16.2% in 15 of 24 studies, and glycosuria was reduced by a mean of 39.7% in five of eight studies. Substantial reductions in plasma choles­terol concentration (mean reduction of 18% in 20 of 23 studies) and, in some cases, plasma triglyceride level (mean reduction of 10% in eight of 20 studies) have also been reported. The use of high-fiber diets with more moderate levels of carbohydrate also has been shown to have positive effects on glycemic control and to reduce plasma lipids; however, the magnitude of the changes usually is not as great as with high-fiber, high-carbohydrate diets. In one study, Anderson et al. (1980a) found that a low­fiber, high-carbohydrate diet decreased insulin dose and serum cholesterol level, but increased serum triglyceride level.

High-fiber, high-carbohydrate diets also have been reported to have other beneficial effects for diabetic patients, such as lowering blood pressure and promoting weight loss in some instances The beneficial effects of such diets on glycemic con­trol and serum lipid levels cannot be attributed to dietary fiber alone, because weight loss and decreases in fat intake may also promote such effects. Nonetheless, these studies have had a sub­stantial impact on dietary recommendations given to persons with diabetes. The American Diabetes Association (1979) and diabetes associations in other countries now advise the consumption of a diet rich in high-fiber carbohydrate foods.

Several mechanisms to explain the possible role of dietary fibers and high-fiber, high-carbohydrate diets in modu­lating metabolic control in diabetes have been proposed, although a full explanation has not yet been advanced. The effects of the soluble fiber supplements such as guar may depend in part on their ability to slow or delay carbohydrate absorption (Blackburn et al., 1984a,b) by virtue of their viscosity. No causal links between the lower postprandial insulin levels and changes in stimulatory gut hormones (see Section V-G) have been established. High-fiber, high-carbohydrate diets do not seem to increase the secretion of endogenous insulin and may increase insulin sen­sitivity. These diets increase insulin receptor number for circulating monocytes and enhance peripheral insulin sensitivity (Anderson, 1983, 1986a; Fukagawa et al., 1984; Pedersen et al., 1982; Ward et al., 1982). Guar gum supplements also have been found to increase peripheral insulin sensitivity in noninsulin­dependent diabetic subjects (Tagliaferro et al., 1985). In addition, high-fiber, high-carbohydrate diets typically contain many foods with low glycemic indices. However, the clinical utility of foods with low glycemic indices, especially in a mixed

91

meal, has been questioned (Diabetes Care and Education Dietetic Practice Group, 1985).

3. Hyperlipidemia and cardiovascular disease

a. Epidemiological studies of coronary heart disease

There is no unequivocal diagnostic prognos­ticator of either state or severity of coronary heart disease (CHO), although three major risk factors have been identified: elevated serum cholesterol levels, hypertension, and cigarette smoking. The analysis of cholesterol has been refined to the measurement of plasma lipoproteins and, generally speaking, elevations of LDL are statistically correlated with increased risk, whereas elevations in high-density lipoprotein (HDL) have bien regarded as "protective.'' The analysis of lipoproteins has undergone further refinement to examine levels of specific apo­lipoproteins, namely, apo-B, which is the major protein component of LDL and apo-AI, which is the major protein of HDL. In humans, the only available data relating to fiber intake and CHO are retrospective or are data on current levels of serum/plasma lipids, lipoproteins, or apolipoproteins which are examined for their possible predictive values.

One group of subjects who lend themselves to this type of inquiry is the vegetarian population whose dietary lifestyle is one of low intake of cholesterol and fat and high intake of plant products. Hardinge and his colleagues (Hardinge and Stare, 1954; Hardinge et al., 1958) examined cholesterol levels in vegetarians, lacto-ovo vegetarians, and nonvegetarians and found that vegetarians had significantly lower serum cholesterol levels than the other two groups. Fiber intake (presumably crude fiber) was 24 g/d in male vegetarians compared to 16 g/d in lacto-ovo vegetarians and 11 g/d in nonvegetarians. Fiber intake as g/1000 kcal was 7.9, 5.4, and 2.9 for vegetarian, lacto-ovo vegetarian, and nonvegetarian men, respectively. The female subjects ingested (g/d and g/1000 kcal) 21 and 8.6, 13 and 5.2, and 8 and 3.1 for the vegetarians, lacto-ovo vegetarians, and nonvegetarians respectively.

Analyses of serum lipids and dietary fiber intake in these groups carried out almost 30 years later (Kritchevsky et al., 1984a) confirmed the significantly lower serum cho­lesterol (but higher triglycerides) in vegetarians. Total fiber intake (as g/1000 kcal/d) was similar for vegetarians, lacto-ovo vegetarians, and nonvegetarians, respectively. The single class of fiber which was significantly more prevalent in the vegetarian diet was pectin. Lower plasma lipid levels of vegetarians have been confirmed by other investigators (Burslem et al., 1978; Knuiman and' West, 1982; Sacks et al., 1975). The vegetarians also exhibited a higher ratio of HDL­cholesterol/total cholesterol. A study of apolipoproteins has

92

shown apo-8 and apo-AI levels to be significantly lower in vegetarians (Burslem et al., 1978). The authors concluded that vegetarianism affected both level and composition of HDL.

The largest available group of subjects of varying vegetarian status is the Seventh-Day Adventist (SDA) population in California. A survey of diet and CHO risk in this population (Phillips et al., 1978) showed that vegetarian/nonvegetarian status was strongly related to CHO risk in SDA males younger than 65 yr, but there was no relationship in males older than 65 yr, and that strict female vegetarians exhibit increased risk of CHO. The factors underlying this discrepancy have not been elucidated. A study in the United Kingdom found a significantly negative association with vegetarianism, but none with fiber intake. Other aspects of vegetarian lifestyle (such as exercise and smoking habits) may affect other recognized CHO risk factors.

The correlation of dietary fiber and CHO in nonvegetar­ian societies also has been studied. A follow-up of diet and CHO mortality in men in London and Southeast England showed that after 10 yr, men with high-energy intake or high-cereal intake had a lower rate of disease (Morris et al., 1977). These data are coincident with dietary surveys in three large ongoing American CHO studies (Puerto Rico, Hawaii, Framingham) which also showed that men with CHO had lower intakes of total energy and of complex carbohydrate (Gordon et al., 1981).

Liu et al (1982) analyzed data relating to dietary lipids, sugar, fiber, and CHO mortality in 20 developed coun­tries and found slightly higher mortality rates corresponding to lower fiber intake, but the differences were not statistically significant. Analysis of data from a prospective study involving 1001 men showed, after 20 yr, that fiber intake was significantly lower in men with CHO when considered alone, but not after adjustment for other risk factors (Kushi et al., 1985). A 10-yr follow-up study of diet and disease in the Netherlands (Kromhout et al., 1982) showed no significant difference in fiber intake between CHO victims (27.2 + 8.1 g/d; 9.5 g/1000 kcal) and survivors (30.8 + 9.7 g/d;-10.0 g/1000 kcal), but reduced survival in the lowest quintile of fiber intake.

The epidemiological data on the relationship of dietary fiber to CHO are inconclusive. However, another line of evidence is provided by the data on fiber effects on hyperlipidemia, which have been reviewed recently (Jenkins et al., 1986b; Schneeman and Lefevre, 1986). In general, these data show that soluble fibers such as pectin, guar gum, locust bean gum, or oat gum reduce significantly serum total cholesterol and LDL-cholesterol levels with little effect on HDL levels. Insoluble fibers such as bran or cellulose have essentially no effect. Studies on the effects of dietary fiber sources in hyperlipidemia are reviewed in Section VI-C.3b.

93

Animal studies have also been conducted to examine the relationship of fiber to heart disease. For example, fibers (grain residues, peat, wheat straw) were found to reduce the atherogenicity of a semipurified, high-fat, cholesterol­free diet in rabbits (Kritchevsky and Tepper, 1965; Kritchevsky et al., 1977; Moore, 1967). Pectin inhibited atherogenesis in cholesterol-fed chickens (Fisher et al.,- 1966) o Alfalfa and wheat straw were less sudanophilic for vervet monkeys than was cellulose (Kritchevsky et al., 1981). Pectin, however, did not appear to affect either cholesterolemia or sudanophila in vervet monkeys (Kritchevsky et al., l986a,b).

b. Trials of lipid-lowering effects

The effects of dietary fiber intake on elevated serum l~pid levels have been studied extensively. Results of studies on effects of various fibers on elevated cholesterol and triglyceride levels in diabetic subjects were summarized in Section VI-C.2. Table 12 presents a summary of studies evaluating the use of dietary fiber supplements and high-fiber diets to alter serum lipid levels in patients with idiopathic or genetic hyperlipidemias. The high-fiber diets are usually lower in fat and cholesterol, as well.

Supplementation with guar (8-15 g/d) has been shown to have a consistent serum cholesterol-lowering effect, with decreases of approximately 10-15%. Guar supplements tended to lower LDL-cholesterol levels without affecting HDL-cholesterol levels. Significant effects on triglyceride levels have not been observed. Effe~ts similar to those seen with guar have also been observed with pectin supplements.

Soy products have been found to have some hypocho­lesterolemic effects in subjects with various hyperlipidemic disorders. Effects on triglyceride levels have been variableo Results seem to be dependent on the composition of the diets used and may be related to components in soy products other than fiber. The limited number of studies reported suggest that dried beans have both hypocholesterolemic and hypotri­glyceridemic effects.

Oat bran has been shown to exert a substantial choles­tirol-lowering effect in patients with hypercholesterolemia, with some studies also showing a triglyceride-lowering effect. Oat bran is rich in oat gum and is well tolerated by most patients. Psyllium, another preparation rich in water-soluble fiber, has not been studied extensively, but has been found to decrease serum cholesterol levels in patients with hypercholes­terolemia and to reduce serum triglyceride levels in subjects with hypertriglyceridemia.

In contrast to supplements containing high levels of soluble. fiber, supplements containing mainly insoluble fibers

94

Table 12.· Studies on the effects of dietary fiber in patients with hyperlipidemias.*

% Change

Serum Cholesterol Fiber Study Test

Reference Source Design Duration Dose Subjects Total VLDL LDL HDL TG Comments

Jenkins et al. (1979e) Guar SC 2 wk 15 g/d 10 HL (II) -11 - - - NS Lipid-lowering drugs used

Jenkins et al. (1980c) Guar SC 8 wk 15.5 g/d 8 HL -14 - -16 NS NS (Ila, IIb)

Crispbread SC 2 wk 13 g/d 17 HL -10 - -11 NS NS Hydrated SC 2 wk 8 g/d 8 HL -8 - -11 NS NS

Semihydrated SC 2 wk 11 g/d 4 HL NS - NS NS NS

Tuomilehto et al. Guar R,P,DB 4 mo 15 g/d 33 HC NS - - NS NS Significant weight loss (1980) ( IIa, IIb)

\0 Simons et al. (1982) Guar SB,SC 3 mo 18 g/d 17 1° HC -15 - - - NS \Jl

13 1° HC -11 NS -14 NS NS

Wirth et alo (1982) Guar R,CO 2 mo 15.6 g/d 12 HLP(Ila) -6 - -14 NS NS Bezafibrate also used

Aro et al. (1984) G11ar DB,CO 6 wk 15 g/d 14 HC -12 NS -12 NS NS Guar DB,CO 12 wk 15 g/d 14 HC NS NS NS NS NS

Bosello et al. (1984) Guar SC 60 d 16 g/d 12 FCHL -11 -24 -10 NS -22

Zavoral et al •. Locust co 8 wk 10-35 g/d 18 FHC adults -10 to NS -11 to -7 to NS (1983) bean gum & children -17 -19 -9

Delbarre et al. Lemon pectin C 6 wk 6_ g/d 10 HL NS - - - NS Some patients w/ gout; (1977) Apple pectin ID 6 wk 6 g/d 13 HL NS - - - NS therapeutic HL diet

* See code for abbreviations at end of table; amounts of fiber reported are dependent on the method of fiber analysis.

Table 12. Studies on the effects of dietary fiber in patients with hyperlipidemias (continued).*

% Change

Serum Cholesterol Fiber Study Test

Reference Source Design Duration Dose Subjects Total VLDL LDL HDL TG Comments

Nakamura et al. (1982) Pectin SC 2 wk 9 g/d 12 HC -9 - - NS NS

Miettinen & Tarpila Pectin SC 2 wk 40-50 g/d 7 HL & -13 - - - NS Low-cholesterol diet (1977) 2 normal

-Palumbo et al. (1978) Cellulose sc,co 6 mo >14 g/d OF 14 HLP (Ila) NS - - - NS Cholestyramine more &/or soy hulls effective than fiber

Sirtori et al. (1979) Soybean R,CO 3 wk 60-100 g/d 42 HL -20 NS -21 - NS Combined results from protein (Ila, Ilb, protocols w/ low-lipid

(Temptein®) III) diets, varying choles-\0 terol + P/S ratio O'I

oescovich et al. (1980) Textured SC 8 wk 60-120 g/d 127 HC -19 - dee NS NS Soy products substituted veg protein for animal protein in (

1Cholsoy1) therapeutic diet

Sasaki et al. (1985) Soybean C,R,P 2 mo 11.4-15 g/d 9 HL NS -44 NS NS -28 crude fiber II II 5.7 g/d 6 HL NS NS NS NS NS

Shorey et al. (1985) Soybean DB,CO 4 wk 25 g/d 31 HC -5 to - - -9 NS polysaccharide -11

Verrillo et al. (1985) Soybean C,R,P 2 mo 60 g/d 19 HLP (II) -29 - -39 NS +12 Soy replaced animal substitution protein

Soybean II II 60 g/d 38 HLP (II) -30 - -36 NS +18 Soy added to low-fat addition low-cholesterol diet

* See code for abbreviations at end of table; amounts of fiber reported are dependent on the method of fiber analysis.

Table 12. Studies on the effects of dietary fiber in patients with hyperlipidemias (continued).*

% Change -

Serum Cholesterol Fiber Study Test

Reference Source Design Duration Dose Subjects Total VLDL LDL HDL TG Comments

Lieberthal & Martens Psyllium SC 5 wk 11.5 g/d 19 HC -10 to - - - - Patients in chronic (1975) · (Konsyl CB) -14 disease hospital

Dpnielsson et al. Psyllium SC 7.2-10.8 g/d 13 HLP: (1979) 2-23 mo 3 iCH -17 - - - NS

2-7 mo 5 iTG NS - - - -47 3-29 mo 5 iCH + iTG -17 - - - -57

Kirby et al. (1981) Oat bran C 10 d 92 g/d 8 HC -13 - -14 NS NS

Anderson et al. Oat bran R,P,SC 21 d 100 g/d 10 HC -19 - -22 NS -19 Metabolic ward; weight \0 (1984a) Beans 19 " 115 g/d 10 HC -19 - -23 NS NS loss significant -.....J

Anderson et al. Oat bran R,P,SC 21 d 100 g/d 10 HC -23 - -23 -20 -21 Metabolic ward; weight (1984b) or beans loss significant

II ii 24 wk 50 g/d ii -26 - -24 inc -25 Ambulatory

Jenkins et al .. Legumes SC 16 wk 140 g/d 7 HL (IIa, -7 - NS NS -25 (1983b) IIb, IV)

Thiffault et al. Lignin-90.5% SC 4-7 mo 1.2-4 g/d 6 HLP (II) -22 - - - - 0.5 % methylcellulose; (1970) cholestyramine similar

Linder and Moller Lignin SC 4 wk 2 g/d 7 HLP (II) +8 - - - NS (1973) (

1 Lignilin' )

* See code for abbreviations at end of table; amounts of fiber reported are dependent on the method of fiber analysis.

\0 co

Table 12. Studies on the effects of dietary fiber in patients with hyperlipidemias (continued).

Reference

Brooks et al. (1976)

Mathur et al. (1977)

Lindgarde & Larsson (1984)

Lithell et al. (1984)

Choudhury et al. (1984)

Jenkins et al. (1985)

C = controlled

CH= cholesterol

CO= crossover

DB= double blind

dee= decrease reported

OF= dietary fiber

Fiber Source

Wheat bran

Wheat bran

Wheat bran (Fiber form®)

Wheat bran (FiberfornfID)

Diet

Diet

Study Design Duration

SC 2 mo

SC 1 mo

CO,R 8 wk

CO,R,DB 6 wk

SC 7.5 mo

SC 3 mo

Test Dose

50 g/d

40 g/d

10.5 g/d

10.5 g/d

52 g OF/ 2000 kcal

60 g/d OF

Subjects

9 2°HTG

10 HC

12 HC

21 HLP (Ilb, IV)

31 HL (Ila, Ilb,III, IV)

12 HL (IV, IIb, III)

% Change

Serum Cholesterol

Total VLDL LDL HDL TG ColIVTients

NS - - - NS

-25 - - - - Added to usual diet

NS NS NS +30 -24 Moderate weight loss

NS - - NS NS Patients had low HDL-cholesterol

-22 -37 -25 NS -24 Diet high in vegetable foods; hypocaloric diet for obese subjects

-9 - -10 NS -16 Foods of low glycemic index

CODE FOR ABBREVIATIONS IN TABLE 12

FCHL = familial combined hyperlipidemia LDL = low-density lipoprotein SB= single blind

HC = hypercholesterolemia - = not determined SC= self controlled

HDL = high-density lipoprotein NS= not statistically TG = triglycerides significant (p >0.05)

HL = hyperlipidemia veg= vegetable P = parallel

HLP = hyperlipoproteinemia VLDL - very-low-density P/S = polyunsaturated/saturated lipoprotein

HTG = hypertriglyceridemia R = random allocation

have been found to exert little effect on serum lipid levels in subjects with hyperlipidemia. Most studies with wheat bran, especially those with better control, have not demonstrated an effect on elevated serum cholesterol or triglyceride levels. Results with lignin have been contradictory.

In summary, clinical trials suggest that the consumption of soluble fibers, particularly guar and oat bran, can have bene­ficial effects on serum lipids in persons with hyperlipidemia. Mechanisms proposed to explain these effects are discussed in Section V-F.

c. Hypertension

Epidemiological studies have shown that blood pressure is lower in vegetarians and other groups consuming diets with high levels of dietary fiber than in persons consuming diets with lower levels of fiber (Armstrong et al., 1977; Rouse et al., 1982; Sacks et al., 1974; Trowell, 1981; Wright et al., 1979). Typically, these diets are low in fat and refined sugars, have high amounts of complex carbohydrates, and may vary in their content of sodium, potassium, chloride, and calcium; thus, it is not clear which component of the vegetarian-type diet is res­ponsible for the blood pressure-lowering effect observed. Early studies with rabbits (Nuzum et al., 1926) indicated that animal protein caused greater elevations in blood pressure than vege­table protein. Controlled trials demonstrating a singular effect of dietary fiber are lacking.

Results of clinical studies on the effects on blood pressure of different high-fiber diets in normal and hyper­tensive subjects are summarized in Table 13. In general, these studies show a lowering of blood pressure in response to diets with increased amounts of fiber from various sources. These clinical trials suffer the same shortcoming as the epidemio­logical studies in that components of the diet other than fiber were altered. The weight loss that occurred in many studies is another confounding variable. However, a number of the studies have demonstrated a beneficial therapeutic effect of high-fiber, low-fat, and low-sodium diets which may prove an attractive first alternative to antihypertensive medication for many patients (Anderson and Tietyen-Clark, 1986).

4e Gallstones

Gallbladder disease is one of the most common disorders in economically developed populations. The principal component of most gallstones is cholesterol and the etiology of cholelithiasis is thought to include precipitation of stones from bile -which is supersaturated with cholesterol. However, super­saturated bile does not always lead to gallstone formation.

99

Table 13. Studies on the effects of dietary fiber on blood pressure.*

Change in Blood Pressure (mm Hg)

Study Study Test Reference Diet Design Type Duration Dose Subjects Systolic Diastolic Comments

Wright et al. (1979) HF SC A 4 wk 24 g/d 17 normal -3.9 -3.7

Wholemeal SC A 8 wk 29 g/d 14 normal NS -2.7 bread & bran

Wholemeal SC A 6 wk ? 12 HP NS NS High variability in bread & bran individual BP

Anderson (1983) HFHC SC MW 2 wk 65 g/d 6 normal BP -8.5 -8.0 Diabetic subjects II II II II II 6 EHP -15.2 -8.7

1--' Rouse et al. (1983) LOV C,R,CO A D 6 wk ? 59 normal -6.8 -2.7 Some meals in dining

D hall; BP changes not related to urinary Na or K excretion

Dodson et al. (1983) HFLFLS C,R A 4 wk 35-40 g/d 53 diabetes NS -6.7 BP unchanged in controls; (mild HP) diet effect greater in

whites and West Indians than in Asians

Dodson et al. (1984) HFLFLS C,R A 3 mo 40-45 g/d 50 diabetes -15.5 -8.6 No improvement in con-(mild HP) trols; weight loss on

diet; diet effect same as bendrofluazide (Pacy et al. , 1984)

* See code for abbreviations at end of table.

I-' D I-'

Table 13. Studies on the effects of dietary fiber on blood pressure (continued).

Reference

Lindahl et al. (1984)

Dodson et al. (1985)

Anderson (1986b)

A= ambulatory

BP= blood pressure

C = controlled

CO= crossover

Diet

Vegan ii

HFLFLS Ii

HFH VLC

EHP = essential hypertension

HF= high-fiber diet

Study Design

SC ii

SC II

C,R ii

Study Test Type Duration Dose Subjects

A 4 mo ? 27 EHP Ii l yr Ii II

A 3 mo 40-45 g/d 19 EHP II l yr II II

MW 21 d 32 g/d 20 obese II II 2 g/d (17 HP)

CODE FOR ABBREVIATIONS IN TABLE 13

HFH = high-fiber, hypocaloric diet

HFHC = high-fiber, high-carbohydrate diet

HFLFLS = high-fiber, low-fat, low-sodium diet

HP= hypertension

K = potassium

Na= sodium

Change in Blood Pressure (mm Hg)

Systolic Diastolic Comments

-7 -10 Weight loss on diet; sub--9 -5 jective improvement in

health

-11 0 -11.3 Good compliance; weight -12.3 -13.7 loss on diet; hyperten-

sive medication reduced

-7.2 -5.5 Weight loss similar on -5.5 -2.7 both diets; greater BP

drop with HFH

NS= not statistically significant (p >0.05)

LOV = lacto-ovo-vegetarian diet

MW= metabolic ward

R = random allocation

SC= self control

VLC = very-low-calorie (formula) diet

Several general reviews on lithogenesis have appeared recently (Heaton, 1985, 1986; Judd, 1985).

Among the conditions which have been associated with gallstones are obesity (Sarles et al., 1970), diabetes (Heaton, 1973), and hypertriglyceridemia (Ahlberg et al., 1979). The lithogenic index, which relates biliary concentrations of cho­lesterol, bile acids, and phospholipids, may be an indicator of susceptibility, but is not consistently reliable in diagnosing gallstone disease. The deoxycholic acid content of bile has also been related to lithogenicity in some studies. Deoxycholic acid is a secondary bile acid and its concentration may also be an indicator of the metabolic activity of the intestinal bacteria which dehydroxylate the primary bile acids.

Epidemiological evidence concerning high-fiber intake and gallbladder disease is indirect. Gallstone incidence is low in sub-Saharan Africa where it is assumed that the population has a high fiber intake (Burkitt, 1976). However, many aspects of their lifestyle (dietary and otherwise) differ from those of persons in the Western world.

Epidemiological data from Japan (Kameda et al., 1984) show that between 1950 and 1975 energy intake increased by 4%, animal protein intake increased by 129%, fat intake increased by 190%, carbohydrate intake decreased by 32%, and gallstone inci­dence increased by 459%. The nutrient intakes were calculated per capita per day. The authors concluded the increase in gall­stone incidence was attributable to increased fat and decreased fiber intake.

In a study of three very obese subjects on a high-fiber diet, Meyer et al. (1979) found that the primary biliary bile acid pools, but not the deoxycholic acid pool, were reduced. The biliary saturation index was increased in two of the three subjects. One study on dietary fiber and lithogenic index and bile acid metabolism showed that bran (30 g/d) decreased biliary" deoxycholic acid by 24% and, generally, reduced the saturation index of bile in normal subjects (Watts et al., 1978). Another study suggested that guar gum decreases biliary bile acid con­centration by 23% in normal subjects (Hansen et al., 1983). Oat bran (18 g/d) enlarged the biliary cholic acid pool and the chenodeoxycholate/deoxycholate pool in six normal subjects._ The differences were not significant and the lithogenic index was unchanged (Arffmann et al., 1983a).

Administration of 30 g wheat bran/d to 10 healthy male subjects did not affect biliary bile acid or phospholipid metabolism but lowered the cholesterol concentration, thus reducing the lithogenic index by 34% (Wechsler et al., 1984). In another experiment, normal subjects were fed pectin (12 g/d), cellulose (15 g/d), or lignin (12 g/d) for 4-wk periods (Hillman et al., 1986). The lithogenic index fell 11% on the pectin diet, but was unaffected by the other two fibers. The ratio of

102

primary/secondary bile acids fell when pectin was fed (32%), rose on cellulose (45%), and was relatively unaffected by lignin (+11%). The data suggest unique effects for each type of fiber.

In studies of subjects with gallstones, a refined car­bohydrate diet increased cholesterol saturation of bile and its deoxycholate pool (Thornton et al., 1983). In another· study, wheat bran (30 g/d) had no effect on the saturation index of bile of patients being treated with ursodeoxycholic acid (Frenkiel et al., 1986). Other studies have shown that 33-50 g/d of die­tary wheat bran led to significant reductions of the cholesterol saturation index of subjects with gallstones (McDougall et al., 1978; Pomare et al., 1976; Watts et al., 1978).

One epidemiological study of diet and gallstones showed that case and control subjects ingested similar amounts of fiber and that fiber emerged as a significant protective factor only on multivariate analysis when dietary sugar was excluded from logistic regression models (Scragg et al., 1984). An earlier study reported that both male and female patients ingested sig­nificantly less fiber than controls, but they also ingested fewer calories, and less protein, fat, and carbohydrate (Smith and Gee, 1979). Intake of crude fiber (g/1000 kcal) was identical in all groups. The reduced intake of fiber in the cases may have been a reflection of reduced total dietary intake.

Wheat bran (30-50 g/d) reduced deoxycholic acid content of bile (as percent of bile acids) by 30% and cholesterol satura­tion index by 30% as well (McDougall et al., 1978; Pomare et al., 1976; Watts et al., 1978); thus, wheat bran may influence one aspect of the lithogenicity complex.

Gallstones can be produced in hamsters, mice, rabbits, prairie dogs, or monkeys by using semipurified diets. In the mouse, prairie dog, or monkey the diet contains large amounts of fat and cholesterol. The rabbit model is fed a high-fat diet. The hamster model of Dam (1971) is fed a fat-free diet containing 74% carbohydrate and 20% casein. Addition of lignin (5%) or lactulose (13.6%) to the diet of hamsters led to a sig­nificant reduction in gallstone formation. The lithogenic index was reduced by 36% (p <0.05) when lignin was fed and by 18% when lactulose was fed in the diet (Rotstein et al., 1981). Pectin (4.2%) or cellulose (7%) inhibited gallstone formation in ham­sters by 76 and 64%, respectively. Pectin appeared to promote regression of gallstones, but cellulose had no effect on pre­established stones (Kritchevsky et al., 1984b). Pectin (5%) also inhibited lithogenesis in this animal model (Bergman and van der Linden, 1975).

The influence of dietary fiber on gallstone formation and regression in humans is still unresolved. The data suggest that other components of the diet (i.e., sugar, energy, and alcohol) play impprtant roles in lithogenesis and that the fiber association may be a reflection of changes in the _gross diet.

103

The question of the influence of different types of fiber is also open. The possible special effects of wheat bran and the increase in biliary deoxycholate in subjects eating fermentable fibers require explanation (Miettinen and Tarpila, 1977; Pomare, 1983) .

5. Gastrointestinal disorders

a. Peptic and duodenal ulcers

Ecological studies of effects of dietary fiber intake on incidence of duodenal ulcer are inconclusive; the geographic distribution of duodenal ulcer is not consis­tently associated with consumption of diets either high or low in fiber (Tovey, 1985). As summarized in Table 14a, clinical studies have not reported significant differences in healing of gastric or duodenal ulcers in patients consuming high- or low­fiber diets and receiving antacid therapy (Rydning and Berstad, 1985a, Rydning et al., 1986). However, ingestion of diets high in fiber was associated in one study (not yet confirmed) with a lower rate of recurrence of duodenal ulcers (Rydning et al., 1982). Results of clinical investigations of possible mechanisms for effects of fiber on gastric and duodenal ulcers are presented in Table 14b.

b. Constipation

Large variations in bowel habit, both fre­quency and weight, make it difficult to define an abnormal or "constipated" state. Approaches to the definition of consti­pation have differed in the parameters described; emphasis has been placed variously on stool frequency, stool weight, transit time, and stool consistency, but it is difficult to provide objective standards of normality for these parameters (Stephen, 1985). The definition of constipation recommended by the FDA is three or fewer bowel movements per week (Food and Drug Administration, 1985). Some workers have sought to include a subjective factor in the definition by describing constipation as the infrequent passage of small, hard stools., passed with difficulty (Roth and Leitzmann, 1985b). The subjective nature of the disorder gives rise to wide variations in the estimation of its incidence and does not permit a firm statement of the extent of the problem; however, it is widely presumed to result from consumption of diets containing inadequate fiber and to increase with increasing age (Stephen, 1985). The widespread use of laxatives is sometimes regarded as an indicator of the size of the problem, but not all persons, particularly the elderly, use laxatives to alleviate constipation (Brocklehurst, 1985; Stephen, 1985).

104

Table 14a. Studies on the effect of dietary fiber on gastric and duodenal ulcers in humans: intervention studies.

Study Test Additional Referen~e Design Duration Fiber Source Dose Treatments Subjects Results

Harju & Larmi Crossover, 4 wk Guar 5 g/d None 20 patients with Subjective reports of (1985) double-blind; wk 1: fiber duodenal ulcers beneficial effects of

self-controls or placebo guar by 15 patients; wk 2+3: Placebo-wheat one report of adverse

no treatment flour effects by a patient wk 4: fiber or w/ pyloric stenosis

placebo

Rydning & Group comparison; 4 wk Diet: high fiber 22.0 g/d mean Antacid 80 patients with 67.5% ulcers healed Berstad (1985) randomized assign- (9.1 - 44.3) treatment active duodenal

ment of patients Diet: low fiber 9.9 g/d mean ulcers 60% ulcers healed (no to diets (4.4 - 21.8) difference w/ fiber)

I-' D \JI

Rydning et al. Group comparison; 6 wk Diet: high fiber 22.6 g/d mean Antacid or 86 patients with 41% ulcers healed (1986) randomized, (9.1 - 45.6) placebo active gastric

double-blind Diet: low fiber 8.9 g/d mean treatment ulcers 50% ulcers healed assignment of (2.9 - 19.8) patients to diets

Rydning et al. Group comparison; 6 mo Diet high fiber 28.2 g/d median None 73 patients with 45% recurrence (1982) randomized (8.9 - 46.7) recently healed

assignment of Diet: low fiber 11.4 g/d median duodenal ulcers 80% recurrence patients to diets (0 - 17.7 g)

No dose-response relationship

Table 14b. Studies on the effect of dietary fiber on gastric and duodenal ulcers in humans: mechanism studies.

Study Test Additional Reference Design Duration Fiber Source Dose Treatments Subjects Results

Tovey (1974) Self-controls 3 test meals Foods not calculated None 17 adult patients Titrable total & free or analyzed with duodenal acid in gastric aspir-

ulcers ates similar following test meals of refined and unrefined grains

Rydning et al. Self-controls 3 test meals Control meal no added fiber None 10 healthy Fiber-enriched wheat (1984) subjects bran prolonged acid neu-

Fiber-enriched 10.5 g tralization, decreased wheat bran pepsin concentration

induced by test meal,

I-' Guar gum 5 g decreased bile acid

D concentration; guar gum 0\ shortened acid neutral-

ization, had no effect on pepsin concentration, decreased bile concen-tration

Rydning et al. Self-controls 3 test meals Control meal no added fiber None 8 healthy No change in gastric (1985) subjects emptying time with any

Fiber-enriched l0o5 Q of the test meals wheat bran

Guar gum 5 g

I-' D -...J

Table 14b. Studies on the effect of dietary fiber on gastric and duodenal ulcers in humans: mechanism studies (continued).

Reference

Rydning & Berstad (1985b)

Harju et al. (1984)

Study Design

Self-controls

Self-controls &

healthy controls; double-blind, randomized

Duration Fiber Source

2 test meals Control

Fiber-enriched wheat bran

2 test meals Control (wheat flour)

Guar

Test Dose

no added fiber

10.5 g

0.02 g

5 g

Additional Treatments

None

None

Subjects

6 adult patients with gastric ulcers

15 healthy subjects;

15 subjects with duodenal ulcers

Results

Postprandial bile acid concentrations signifi­cantly decreased with addition of fiber

Postprandial gastric acidity reduced by guar, not to control levels; guar delayed gastric emptying

A number of diseases and other abnormalities, such·as diabetes, hypothyroidism, uremia, neurogenic bowel disorders, and abnormalities of bowel structure in the colon, rectum, or anal canal may cause constipation. The majority of persons with this complaint, however, do not have underlying disease. Constipation also may be induced by a wide variety of drugs, endogenous hormones, lack of exercise, and other factors, as well as diet. The effects of various fiber sources on stool weight, stool frequency, and gastrointestinal transit time in normal subjects are discussed in Section V-J. Results of some trials of dietary fiber to alleviate constipation, most with elderly subjects, will be considered here.

Clark and Scott (1976) studied 25 frail, elderly, hos­pital patients with long-standing constipation who required con­tinuous management with laxatives and suppositories. Patients received 5-25 g/d of coarse wheat bran for 3 wk. Wheat bran sig­nificantly increased the number of bowel movements and decreased the number of constipated days for men but not for women, and significantly increased stool size and decreased the need for laxatives in both sexes. Subjective discomfort was reduced in some patients, but in some an increase in incontinence was also noted. The use of 10-20 g/d coarse wheat bran was compared to a bulk laxative regimen (Vi-Siblin®) for 8 wk in 10 elderly, constipated, nursing home patients (Andersson et al., 1979). Wheat bran significantly reduced transit time, and less laxative therapy was required during wheat bran treatment than during bulk laxative treatment. Smith et al. (1980) tested three regimens for 28 din 37 elderly, hospitalized subjects with constipation: one ispaghula sachet morning and night, 20 g/d wheat bran, or laxative administration when the patient experienced discomfort from constipation. Both bran and ispaghula markedly increased stool weight, in contrast to the laxative treatment, and without serious side effects or changes in colonic pressures. A 12-mo study with a high-fiber diet regimen in 30 constipated nursing home residents increased the number of bowel movements, elimi­nated the need for laxatives, and reduced nursing intervention (Hope and Down, 1986). The additional fiber was offered in a variety of ways including high-fiber breakfast cereals, high­fiber biscuits, and a bran-apple-prune supplement given with the evening meal. Dietary fiber intake was increased from 14 to 25 g/d and adequate fluid intake was also maintained.

A modest supplement of psyllium seed husk (~7 g/d) increased defecation frequency and wet and dry stool weights in an ambulatory population of habitually constipated (i.e., ~three bowel movements/wk) adults aged 18-35 yr (Marlett et al., 1987). The effects of two brans, fine wheat and corn, were studied for 5 wk in 10 constipated women aged 20-40 yr (Graham et al., 1982). Bran administration (20 g/d) was associated with a significant increase in fecal weight (157%) and bowel frequency (55%) and a significant decrease in transit time (50%). The percent moisture increased only with the wheat bran whereas subjective improvement in symptoms was noted with corn bran.

108

The addition of 20 to 22.3 g/d soy polysaccharide to the liquid formula diets of tube-fed, constipated, mentally retarded patients was also found to be beneficial (Fischer et al., 1985). Stool size increased and fecal consistency improved, but defe­cation rate and the need for elimination aids were not changed.

In summary, these and many other studies have shown that dietary fiber, particularly wheat fiber and other insoluble fiber sources, is useful in the prevention and treatment of constipation.

c. Irritable bowel syndrome

Irritable bowel syndrome is a diagnosis for a variety of disorders that are functional, i.e., not caused by organic disease states. It is thought to be a result of an inap­propriate reaction of the gastrointestinal muscles to stress and physical conditions in the gut, and the entire gastrointestinal tract may be affected (Eastwood and Passmore, 1983). Symptoms can include abdominal pain, diarrhea, or constipation, frequently accompanied by depression or anxiety. Motility disturbances,· psychiatric.disorders, and diet have been suggested as causes (Harvey, 1985). Little work has been done to examine the diets of patients with this syndrome. In one study, Hillman et al. (1982) collected weighed dietary records for 1 wk from 30 women with recently diagnosed irritable bowel syndrome (with symptoms of spastic colon and pain) and 25 healthy controls. There were no differences between patients and controls for intake of total, cereal, or fruit fiber, but vegetable fiber intake was significantly lower in patients.

Dietary fiber, especially wheat bran, has been used therapeutically in irritable bowel syndrome. A postal survey of British gastroenterologists in 1976 indicated that 84% used a high-fiber diet or wheat bran supplements to treat this disorder (Manning and Heaton, 1976). Manning et al. (1977) conducted a 6-wk trial of a high-wheat-fiber diet (containing 20 g/d fine unprocessed bran) vs. a low-wheat-fiber diet in 25 patients with irritable bowel syndrome. They reported that the frequency of pain and colon motility was reduced in patients consuming· the high-fiber diet, but that no improvement occurred with the low-fiber diet. In contrast, S0ltoft et al. (1976) did not find a significant difference in effects of wheat bran and placebo in a 6-wk trial of 52 patients with irritable bowel syndrome. The treatment group received biscuits which provided 30 g/d miller's wheat bran, and the placebo group received bis­cuits without bran; both groups received instructions about bowel habits, physical exercise, and laxative use for severe constipation. Subjective improvement was noted in 52% of the treatment group and 65% of the placebo group. Arffmann et al. (1983b) examined 18 patients in a 6-wk, double-blind, crossover study with 30 g/d wheat bran or placebo (bread crumbs). Wheat bran significantly increased fecal mass and decreased transit

109

time, but did not affect pain or other symptoms. Cann et al. (1984) studied 60 patients extensively screened for organic disease who had disturbed bowel habits and pain at least three times per week, for a duration of at least 6 mo. One portion of their study was an open trial of coarse wheat bran (gradually increased from 10 to 30 g/d) for 4 wk. Both bran and a placebo significantly reduced symptoms overall; constipation was the only symptom significantly reduced by bran, but not by placebo. The data of Hillman et al. (1984) on 2- to 3-yr follow-up in 14 patients after management of irritable bowel syndrome with a high-fiber diet (average increase of 6.7 g/d dietary fiber) showed that symptoms were greatly improved in seven, were unchanged in five, and were worse in two. Clinical course was not correlated with the amount of dietary fiber consumed, but improvement in symptoms was associated with significant decreases in anxiety.

Evidence at this time suggests that patients with irri­table bowel syndrome who are most likely to benefit from an increased intake of dietary fiber (as wheat bran) are those whose chief complaint is constipation (Fielding, 1985; Harvey, 1985)a Neither bran nor a high-fiber diet is a panacea; treatment of accompanying anxiety or depression should also be undertakene Further controlled dietary studies are desirable, particularly with selection of more homogeneous groups of patients.

d. Diverticular disease

Diverticular disease of the colon is an acquired pathologic defect of the large bowel that appears to predominate in Western, economically developed countries and to affect middle-aged and older adult men and women. Saccular herniations of the colonic mucosa averaging about 0.5 to 1.0 cm in diameter, about 95% of which occur in the sigmoid colon, are the gross anatomic features of uncompli­cated diverticulosis (Austad, 1979; Haubrich, 1985; Robbins, 1967; Simonowitz and Paloyan, 1977). The herniated mucosa penetrates the muscularis layers at vulnerable anatomic sites such as the points where nutrient arteries enter the bowel wall (LaMont and Isselbacher, 1983).

Several attempts have been made to classify forms of diverticular disease. Brodribb (1979) identified three clini­cally important forms of diverticular disease of the colon: asymptomatic, symptomatic, and complicated by bleeding or inflammation (diverticulitis). Associated sequelae of acute diverticulitis include peritonitis, abscess, and fistula. Five types of intestinal diverticular disease, based on dif­fering clinical, pathologic, and radiographic criteria, were listed by Mendeloff (1986): presolitary diverticulosis, multiple diverticulosis, generalized diverticulosis of the small intestine, generalized diverticulosis of the large intestine, and diverticulitis. From the point of view of

110

anatomic pathology, two basic forms of diverticular disease have been postulated, one associated with thickened colonic muscle, colonic narrowing, and high intracolonic pressures, and the other, in approximately equal prevalence, not associated with thickened muscle or (presumably) high intracolonic pressures (Connell, 1978). Chronic inflammatory diverticulitis results in diffuse thickening of the bowel wall, and perforation of an infected diverticulum leads to pericolic abscess, sinus tract, and sometimes generalized peritonitis. Uncommonly, massive hemorrhage and fistulous communications with other viscera may occur.

Characteristic symptoms and signs of diverticulosis ate constipation and/or diarrhea; flatulence; abdominal pain, frequently in the left iliac fossa; abdominal aching and heav­iness; colicky abdominal pain; mucus and blood in the stools; and tenesmus (Painter, 1975; Roth and Leitzmann, 1985b).

The true prevalence of diverticular disease is unknown because a great majority of cases are symptomatically silent and the diagnosis depends upon findings of colonic contrast radiog­raphy, bowel surgery, or autopsy. Painter (1985) has gathered available data from countries around the world in an effort to ascertain the prevalence of diverticular disease. He noted that it is common in affluent Western countries, but probably rare in Eastern European countries. It is reportedly rare in rural popu­lations of Africa, India, the Middle East, the Far East, Mexico, and South America, but has been found with increasing frequency in the cities of these countries.

The frequency of diverticulosis in seven ·barium enema study series ranged from 4.2% of 2090 subjects in Sweden to 40% of 500 subjects in France (Painter, 1985). A frequency of 5% in the largest series, 24,620, was reported from the Mayo Clinic in 1930. Frequency is positively associated with increasing age, amounting to approximately 20% in people over 40 and up to 70% over the age of 70 in the United Kingdom and the United States (Taylor and Duthie, 1976). In six necropsy series, frequency ranged from 4.4% of 10,167 subjects in the United Kingdom to 45% of 200 subjects in Australia.

Results of 280-unselected autopsies of Norwegian patients over 20 yr old revealed diverticular disease in 25% of the males and 43% of the females. Frequency of diverticulosis increased with age and both the frequency and numbers of diver­ticula per patient were higher in femaleso No association was apparent between the presence of diverticulosis and benign or malignant neoplasms of the colon (Eid~ and Stalsberg, 1979).

Diverticular disease is reportedly rare in native black Africans living south of the Sahara (Segal et al., 1977). During 1974 and early 1975, 16 black African residents of Johannesburg who had been eating diets containing refined carbohydrates and low levels of fiber were found to have diverticulosis. Eight of

111

the patients did not eat maize meal, the staple food of the rural black Africans, and the remainder of the patients ate refined white maize meal. In a subsequent study from January 1977 to December 1979, 42 additional cases (14 per yr) of diver­ticular disease were diagnosed from the same population group (Segal and Walker, 1982)0 The estimated mean daily dietary fiber intake of these patients, 26.5 g, contrasts with an intake of 32.5 ± 11.4 gin a sex- and age-matched urban black control group. The authors suggested that the degree and duration of westernization of the diet of black Africans in Johannesburg appeared insufficient to have evoked significant increases in the prevalence of diverticular disease when compared with the findings of the previous study (Segal and Walker, 1982).

Etiology. One hypothesis about the etiology of diverticular disease of the colon is that inadequate intake of dietary fiber over time is the major factor (Painter and Burkitt, 1971). Fiber affects such colonic f~nctions as intes­tinal transit time, intestinal motility, stool weight and consis­tency, and intracolonic pressures (Burkitt et al., 1972, 1974; Mendeloff, 1986; Painter, 1975, 1985; Srivastava et al., 1976). Many experts believe that the diverticulosis results from colonic segmentation and high intraluminal pressures that h~ve been shown to accompany low-fiber intakes and to respond favorably to increased dietary fiber consumption (Painter, 1985).

Gross estimates of food intakes and dietary patterns during the past seven or eight generations in economically developed Western nations suggest that consumption of cereal fiber has declined markedly. For example, estimated per capita wheat flour intake of 400 lb in England in 1750 contrasts with 150 lb in 1950 (Hollingsworth and Greaves, 1967). As noted above, similar changes have occurred in urban South Africa where the traditional high-fiber rural diet has shifted toward rela­tively low fiber content (Segal and Walker, 1982).

Amounts of dietary fiber in the Japanese diet between 1911 and 1980 were estimated on the basis of food consumption tables and calculation of dietary fiber and crude fiber of plant foods via published lists of analytical values (Ohi et al., 1983). In general, except during World War II, fiber content declined continuously through the period at a rate paralleling that in the United States. The prevalence of diverticular disease in _Japan increased markedly during the late 1970s, although it is considered relatively low compare·d with other industrialized nations.

Typical daily intakes of dietary fiber in rural popula­tions generally range from 50 to 70 g whereas westernized urban dwellers consume about 15 to 25 g (Mendeloff, 1986). Although the low-fiber hypothesis of the etiology of colonic diverticular disease appears to be the most generally accepted, a universal consensus does not exist. For instance, the hypothesis does not

112

explain such major exceptions as the fact that approximately one­half of westernized populations do not develop colonic diver­ticula, but maintain an intake of dietary fiber in the range of low to medium (Mendeloff, 1986).

Dietary intervention in diverticular disease. Table 15 summarizes conclusions from selected reviews of clini­cal trials of dietary fiber in the treatment of diverticular disease of the colon. In a study conducted for the Food and Drug Administration by the LSRO, another ad hoc expert panel noted that supplements of dietary fiber such as coarse wheat bran are used routinely in clinical management of uncomplicated diver­ticular disease. Dietary fiber relieved constipation and, in most available clinical reports, symptoms of diverticulosis. However, validation of its effectiveness in diverticular disease was needed (Talbot, 1980). The following paragraphs summarize the findings of clinical investigations over the past 10 or 15 years.

High-fiber crispbread eaten in addition to a normal diet relieved symptoms in 71% of 48 patients with diverticular disease manifesting abdominal pain, altered bowel habits, mucus, or com­binations of these. The high-fiber crispbread had an estimated crude fiber content of 4%, was readily accepted, and caused no changes in normal eating habits when taken at the recommended 6 slices/d (Plumley and Francis, 1973).

Findlay et al. (1974) tested the effects of 20 g of unprocessed wheat bran taken in two 10-g fractions daily in normal volunteers and patients with diverticular disease. Stool weights were increased in six normal subjects, but not in seven patients with diverticular disease; however, transit time was reduced in the patients with diverticulosis.

A comparison of the experience of 142 patients who were followed for over 9 yr after hospitalization for acute divertic­ulitis indicated that 73% of 104 treated medically and 79% of the 38 treated surgically had no further symptoms or hospitalizations as a result of diverticular disease once they had recovered from the acute episode (Larson et al., 1976). Hospital medical treat­ment in 43 patients consisted of rest, observation, and diet (presumably high-fiber supplemented) alone; another 43 patients received antibiotics alone or with intravenous fluids. The authors concluded that conservative medical management was the treatment of choice, with surgical intervention reserved for complications.

In a group of 20 patients with diverticular disease, Taylor and Duthie (1976) compared the therapeutic efficacy of a high-roughage diet supplemented with bran, a bulk laxative (Normacol®) plus an antispasmodic, and regular large amounts of bran in tablet form (Fybranta®) that provided 18 g/d bran. In this crossover study each regimen lasted 30 d. Amounts of

113

1-J 1-J .i::-

Table 15. Reviews of therapeutic effectiveness of dietary fiber in patients with diverticular disease of the colon.

Reference

Mitchell & Eastwood (1976)

Connell (1978)

Brodibb (1979)

Brodribb (1980)

Painter (1985)

Jenkins et al. (1986c)

Type of Report

Review

Review

Review of clinical trials and epidemiology

Review

Review

Review

Conclusions

Beneficial effects of bran in the treatment of diverticular disease are unquestionable.

Anecdotal testimonials reported on the merits of dietary fiber in treating irritable colon and diverticular disease; however, dependable validating data are hard to obtain. Nearly all published treatment studies fall below standards required for double-blind assessment.

High-fiber diets relieved abdominal pain, dyspepsia, and defecation symptoms in a high proportion of patients (80-90%) with uncomplicated, symptomatic diverticular disease and reduced the number of high-pressure waves in sigmoid colon. Not known whether fiber will prevent complications.

High-fiber dietary treatment of patients with symptomatic diverticulosis reduces abnormally high intraluminal pressures in the sigmoid colon and relieves symptoms. Dietary fiber is an important factor in the etiology of diverticular disease.

Reports of seven studies confirmed the ability of a high-fiber diet to relieve symptoms of diverticular disease. Criticized three reports in which no benefit was reported.

Efficacy of high-fiber diet for relieving symptoms of uncomplicated diverticular disease has been shown in most, but not all, clinical trials. More research is needed to define best methods for disease management, including identification of the most effective sources and types of fiber and possible modifications of fat and protein intake

Conments

Typical regimen was 6-10 g/d of fiber of cereal origin; however, some patients may need up to 20 g/d

A well-balanced, interpretative review of diverticular disease in relation to dietary fiber

A well-balanced, interpretative review

Examples of reports that did not demonstrate a beneficial effect are presented

supplemental, unprocessed bran consumed on the high-roughage diet were discretionary with each patient. The bran tablets were superior to the high-roughage diet or Normacol® in relieving symptoms and improving directly measured (intracolonic pressures and colonic smooth muscle electrical activity) and indirectly measured (stool weight and transit time) indices of colonic function.

Devroede et al. (1977) tested six therapeutic dietary regimens allocated at random to 80 patients with symptomatic diverticular disease: a high-residue diet with supplemental placebo or Metamucil®, low-residue diet with placebo or Meta­mucil®, and normal diet with placebo or Metamucil®. Results did not demonstrate a clear-cut advantage for the high-fiber diet in alleviating symptoms.

In a 20-mo study to measure the effects of a high-fiber diet on serum, fecal, and biliary lipids, Tarpila et al. (1978) prescribed for 10 patients with diverticular disease a daily crude fiber intake of 5.6 g, consisting of six rusks providing 3.6 g and 25 g of wheat bran providing 2.0 g crude fiber daily. A control group of 12 additional patients consumed their custom­ary diets without bran supplements. Symptoms and signs of diver­ticular disease clearly abated during the bran period and fecal wet weight, dry weight, and fat were increased in the fiber group when measured at the 6th and 12th month of the trial.

Hyland and Taylor (1980) investigated the prevalence of complications in 100 patients, 75% of whom, following hospital­ization for acute episodes of diveiticular disease, were treated conservatively with a high-fiber diet that provided approximately 40 g fiber/d. Over 90% remained symptom free when reviewed after 5-7 yr. Strict compliance with the diet was estimated at 75% of participants; among 25 patients who did not comply, the rate of complications was 20%.

In a randomized, crossover, double-blind trial, Ornstein et al. (1981) compared the effects of two forms of dietary fiber supplement and placebo in 58 patients with uncomplicated, symptomatic diverticular disease. The patients took bran crispbread, ispaghula drink, and placebo for 4 mo each. Fiber treatment resulted in increased stool weights, increased stool frequencies, and shorter transit times compared with placebo. No benefit from the fiber supplements was observed in terms of pain scores, lower bowel symptom score, and total symptom score; however, fiber supplementation relieved symptoms of constipation. The authors concluded that dietary fiber supplements in the commonly used doses do no more than relieve constipation.

The report of Ornstein et al. (1981) elicited critical comment from clinicians who firmly believed that bran was the primary treatment for uncomplicated diverticular disease (Heaton, 1981). Although acknowledging that the study of Ornstein et al. (1981) was perhaps the best that had been reported in terms of

115

numbers of subjects, duration, and an ''impeccable randomized, crossover design'', Heaton (1981) noted that the patients had mild symptoms at the outset, their stool weights while on placebo approximated those of healthy people on an ordinary British diet, and their transit times while on placebo were near normal at 50 hr. He observed that little can be expected from a treatment aimed at restoring normal colonic function when it was nearly normal at the start. The main criticism of the Ornstein et al. (1981) study was that there was too little difference in daily fiber intake between the bran and placebo phases, that is, the study did not achieve a high-fiber intake or even the "standard" daily dose of 20 g.

According to some investigators, the response to sup­plemental bran in the diet varies considerably among individuals; clinicians should not rely on the concept of a "standard dose". For example, Painter (1985) noted that patients require differing amounts of added fiber to render stools soft and easy to pass and that the correct amount of fiber should be determined by trial and error over a period of at least 3 mo. Brodribb (1980) observed that the maximum therapeutic advantage is not achieved until approximately 3 mo after starting treatment and that, during the first few weeks, patien.ts on a high-fiber diet often complain of increased abdominal distention and flatulence.

In summary, the frequency of diverticular disease of the colon has increased markedly in the past 30 yr among citizens of economically developed Western countries. In these countries, diverticulosis has been found in approximately 33% of persons over 60 yr and 66% of those 80 yr or older (Painter, 1985). Some investigators who have published reports of clinical trials of dietary fiber in diverticular disease are convinced that the etiological relationship of low-fiber intake to development of diverticular disease has been amply documented. A high-fiber diet may relieve the symptoms of uncomplicated diverticular disease and may prevent its development.

e. Inflammatory bowel disease

The term inflammatory bowel disease is used to refer to two disorders of unknown etiology, ulcerative colitis (with symptoms of urgency to defecate, abdominal cramps, blood and mucus in stools) and Crohn disease (with symptoms of abdomi­nal pain, fever, anorexia, weight loss, and an abdominal mass). Elemental diets are commonly recommended for both diseases. In fact, Eastwood and Passmore (1983) suggest that fiber is contra­indicated in Crohn disease because of the greater risk of intestinal obstruction in the presence of strictures.

Dietary fiber has not been proposed to· have an etiologic or therapeutic role in ulcerative colitis, but has been examined in relation to Crohn disease. Martini·and Brandes (1976) used a questionnaire to determine the dietary intake of 63 patients with

116

Crohn disease before they became ill (up to 4 yr earlier). Compared with 63 matched controls, the patients consumed signifi­cantly more refined sugar (in the form of sweets and pastries) and more calories. This observation suggested that refined carbohydrates might be related to the development of Crohn disease, but did not necessarily implicate the lack of dietary fiber. Kasper and Sommer (1979) also found a higher sugar and energy intake prior to illness (up to 2 yr earlier) in 35 patients with Crohn disease than in 70 control subjects. These patients also had a higher intake of starch and dietary fiber than did the controls. In contrast, Thornton et al. (1979, 1980) found that the pre-illness diets of 30 patients with newly diagnosed Crohn disease contained significantly more sugar and slightly less dietary fiber than did the diets of 30 healthy, matched controls. The intakes of cereal and cooked vegetable fiber were essentially similar, but the intake of raw fruits and vegetables was significantly lower in the patients than in controls.

Studies of the use of dietary fiber in the treatment of Crohn disease have also given equivocal results. Heaton et al. (1979) followed the course of 32 patients treated with an unrefined-carbohydrate, high-fiber diet and 32 matched patients given no dietary instruction for over 4 yr. Nine patients in each group had undergone resection. Patients in the diet group had fewer and shorter hospital stays and required less surgery than the control group. No obstructions were observed on the high-fiber diet, even in patients with strictures. Brandes and Lorenz-Meyer (1981) found that a diet in which refined sugar was excluded was beneficial to patients with active disease. Jones et al. (1985) studied 20 patients with active Crohn disease in whom remission was induced by total parenteral nutrition or an elemental diet. One-half of these subjects were assigned to an unrefined carbohydrate, high-fiber diet, and the other half to a special diet designed to exclude foods which caused intolerance. Seven patients on the exclusion diet remained in remission for 6 mo, while none of those on the high-fiber diet remained in remission. Levenstein et al. (1985) studied 70 patients with Crohn disease (without strictures): one half assigned to a low­residue diet, and the other half permitted to eat their usual diet (which contained three times as many servings/wk of fiber­rich foods as the low-residue diet). There were no differences in outcome between the two groups in terms of symptoms, hospital­ization, need for surgery, new complications, nutritional status, or postoperative recurrence.

Additional studies are required to determine the appropriate diet for, and the role of dietary fiber in, the treatment of Crohn disease. The results of available studies are equivocal.

117

6. Colon Neoplasms

a. Nutritional epidemiology

Because foods have many effects on both the contents and physiologic functions of the large bowel, it has been hypothesized that factors in the diet may influence the development of colon cancer in humans. Consequently, the colon is the most intensively studied site in the nutritional epidem­iology of cancer. There is ample evidence from experimental studies that dietary fats and/or fiber can modulate colon cancer risk in laboratory animals (see Section VI-C.6c). That diet may play an important role in cancer in humans is suggested by the dual observations that there are considerable differences in colon cancer risk among the various nations and, that when people migrate from low-risk countries to high-risk countries, they tend to acquire the colon cancer risk of their new homeland. This observation suggests that factors in the environment, rather than genetic factors, determine most of the variability in international colon cancer rates.

Nutritional epidemiologic studies of diet and colon cancer have been of three types: ecological (correlational) studies, case-control (retrospective) studies, and cohort (prospective) studies. Each of the three types of studies has its own inherent strengths and weaknesses as discussed in Section VI-A.

Ecological (correlational) studies. Table 16 summarizes the ecological (correlational) studies that have exam­ined either dietary fiber or fiber-containing foods as related to colon cancer risk in humans. In many studies, multiple nutrients or elements of the diet were studied simultaneously, but in Table 16 only factors related to dietary fiber are included. Considered as a set, these studies show a fairly consistent pat­tern of an inverse relationship between the intake of fiber or fiber-containing foods and colon cancer. Although the strength of the associations is not indicated in this table, many of the correlations, particularly in the international studies, are quite large in the negative direction. Many of these studies simultaneously presented analyses of dietary fats or fat-contain­ing foods (not presented in Table 16) that tended to show strong positive correlations which mirrored the negative correlations observed for fiber. Whether it is, in fact, fiber in the diet which is protective, and/or fat in the diet which is a risk

. factor, is not clear. Multivariate analysis has been attempted in some cases (Liu et al., 1979), but the risk estimates generated from multivariate analyses in which there is a high degree of collinearity of variables are statistically unstable, difficult to interpret biologically, and, therefore, are not · useful estimates of effects of a single nutrient.

118

Table 16. Ecological (correlational) studies of dietary fiber and fiber-containing foods as related to colon cancer risk in humans.

Study Reference Location

Drasar & Irving (1973) Inte ma tional

Irving & Drasar (1973) International

Howell (1975) International

Armstrong & Doll (1975) International

IARC Intestinal Micro- Finland, Denmark ecology Group (1977)

Knox (1977) International

Lyon & Sorenson (1978) Utah, U.S.

Hill et al. (1979) Hong-Kong

Bingham et al. (1979a) United Kingdom

Bingham et al. (1985) United Kingdom

Liu et al. (1979) International

Enstrom (1980) California, Washington

(Mormons)

Rozen et al. (1981) Israel

Englyst et al. (1982b) Finland, Denmark

McKeown-Eyssen & Bright-See (1984)

International

Number of Groups Compared

37 countries

37 countries

37 countries

32 countries

2 countries

20 countries

1 state vs. U.S.

3 socioeconomic groups

9 regions

9 regions

20 countries

Dietary Fiber Observed Measure Association*

Crude fiber

Cereals

Cereals, pulses

Cereals

Dietary fiber

Fruits, vegetables, wheat

0

Crude fiber o

Fiber-rich foods +

Dietary fiber o Pentoses Vegetables

Pentoses, Nonstarch polysaccharides

Sum of fruits, vegetables, legumes, grains

0

2 states vs. U.S. Fruits, vegetables

2 regions Crude fiber

4 regions Nonstarch

38 countries

polysaccharides

Dietary fiber from cereals

*(-)=protective; (+)=risk-enhancing; (o) = no effect

119

Table 16. Ecological (correlational) studies of dietary fiber and fiber-containing foods as related to colon cancer risk in humans (continued)o

Study Number of Groups Dietary Fiber Observed Reference Location Compared Measure Association*

Walker et al. (1986) South Africa 5 racial- Dietary fiber 0

ethnic groups

Kuratsune et al. (1986) Japan, England, 3 countries Nonstarch 0

Denmark polysaccharides

Helms et al. (1982) Denmark 2 regions Dietary fiber

*(-)=protective; (+)=risk-enhancing; (o) = no effect

120

All of the studies listed in Table 16 are not truly independent. Many of the international studies are based on essentially the same data set, i.e., international cancer mortal­ity or incidence data and Food and Agriculture Organization (FAO) food availability statistics. Although various investigators used different subsets of countries, and/or studied food intake data from different eras, their studies should still not be seen as truly independent tests of the fiber-colon cancer association. Considered together, though, these studies indicate that the inverse relationship between fiber and fiber-containing foods and colon cancer risk in international statistics is strong, consistent, and not likely due to artifacts of the analyses from particular studies.

The various studies relating colon cancer rates to diets of particular groups in a smaller number of countries, or among regions within a single country, tended to show negative correla­tions that are weaker than those found in the larger inter­national studies. The two studies by Bingham et al. (1979a, 1985) in which colon cancer mortality rates in nine regions in Great Britain were correlated with regional dietary patterns, point out the sensitivity of correlational analyses to the particular method used to determine dietary fiber. In the first analysis (Bingham et al., 1979a), vegetables and the pentose fraction of fiber were found to be "protective", whereas total dietary fiber was not. In the reanalysis (Bingham et al., 1985), based on an updated analytical technique for dietary fiber, the same colon cancer mortality pattern was found not to be associ­ated with the pentose fraction, but was inversely associated with nonstarch polysaccharides. This finding highlights the inherent instability in correlational studies in which there are small numbers of actual data points; for instance, there were only nine data points in the studies of Bingham et al. (1979a, 1985). Other studies [e.g., Helms et al. (1982); Rozen et al. (1981)] have only two data points. In some cases [e.g., Englyst et al. (1982b)], the absolute difference between fiber intake in high­and low-risk areas, though statistically significant, was modest.

Many of the older international studies are based on intake of crude fiber rather than dietary fiber. Although crude fiber greatly underestimates total dietary fiber levels, crude fiber and dietary fiber are nonetheless highly correlated, so the basic patterns of inverse correlation in international studies based on crude fiber would not likely change considerably if dietary fiber were considered.

The major problem with drawing causal inferences from ecological studies is the "ecological fallacy" in which an unmeasured confounding factor may be responsible for the association observed. In this case, dietary fats and/or other elements of the diet may be factors which could confound the relationship between fiber and colon cancer. Because dietary fats are so strongly 11inversely 11 correlated with fiber in inter­national statistics, statistical methods used to control for

121

confounding by dietary fat are not interpretable. Currently, there are not good candidates for other potential confounders because of the lack of knowledge about the etiological factors for colon cancer.

Case-control (retrospective) studies. Table 17 summarizes the-case control (retrospective) studies of dietary fiber and fiber-containing foods as related to colon cancer risk in humans. These studies have employed many dif­ferent methods for sampling cases and controls and have been conducted in a variety of locations and population subsets. All of them employed the same basic methodology for estimating food intake, the food frequency method. Considered as a set, these case-control studies tend to show inverse relationships between fiber and fiber-containing foods less consistently than the ecological studies. Most case-control studies did not compute indices of fiber intake, but rather presented results either for specific foods or for foods grouped by type (fruits, vegetables, cereals, etc.). Many other dietary factors, in particular dietary fat and fat-containing foods, were investigated in these studies, but Table 17 displays the.findings only for the fiber and fiber-containing foods. Prior to the studies by Bjelke (1974a,b), there was little support in case-control studies for a relationship between fiber-containing foods and colon cancer risk. The observation by Madan et al. (1975) of a low intake of foods containing greater than 0.5% fiber among colon cancer patients in Israel has been frequently cited as evidence that fiber protects against colon cancer in humans. I~takes of over 200 foods were ascertained in the interview, however, and over 7 O o f these foods were inc 1 u de d in ·the 1 is t o f "high - fib e r " foods. It is difficult, therefore, to determine whether the reported association was particular to fiber, or whether it was a more general association with fruits, vegetables, and grain products and/or a marker of a type of dietary pattern. The study by Graham et al. (1978) in Western New York showed an inverse association between colon cancer risk and vegetable intake (in particular cruciferous vegetables) but total dietary fiber intakes were not computed.

The most methodologically rigorous study up to that time was conducted by Jain et al. (1980). The application of validated measures of dietary intake failed to reveal any case­control differences in crude fiber intake. Unexpectedly, total cereal consumption was found to be positively associated with colon cancer. Cancer patients in this study reported retro­spective diet history less accurately than controls. Potter and McMichael (1986) showed that colon cancer patients tended to eat more dietary fiber than did controls (in both sexes, but statis­tically significant only for females). In summary, case-control studies present mixed results. Some show fiber and fiber­containing foods to be protective; some indicate they are, in fact, risk factors; and many others indicate no association.

122

Table 17. Case-control (retrospective) studies of dietary fiber and fiber-containing foods as related to colon cancer risk in humans.

Reference

Stocks & Karn (1933)

study Location

United Kingdom

Higginson (1966) Kansas

Wynder & Shigematsu (1967) New York City

Wynder et al. (1969) Japan

Haenszel et al. (1973) Hawaii

Bjelke (1974a)

Bjelke (1974b)

Madan et al. (1975)

Phillips (1975)

Graham et al. (1978)

Dales et al. (1979)

Martinez et al. (1979)

Haenszel et al. (1980)

Jain et al. (1980)

Manousos et al. (1983)

Pickle et al. (1984)

Bristol et al. (1985)

Potter & McMichael (1986)

Tuyns (1986)

Norway

Minnesota

Israel

California

Upstate New York

San Francisco (Blacks)

Puerto Rico

Japan

Canada

Greece

Nebraska

England

Australia

Belgium

Number of Cases: Controls

300:9000

340:1020

791:309

69:307

179:357

278:1394

373:1657

198:396

41:123

330:783

72:202

461:461

588:1176

348:542

100:100

86:176

50:50

419:739

1207:3521

*(-)=protective; (+)=risk-enhancing; (o) = no effect

123

Dietary Fiber Measure

Whole-grain breads, vegetables

Fruits, cereals

Fruits, cereals

Fruits

Vegetables, cereals

Vegetables

Vegetables

Foods with >0.5% fiber

High-fiber foods

Vegetables

Foods with >0.5% fiber

Cereals Crude fiber

Vegetables Cereals

Crude fiber

Vegetables Cereals, fruits

High-fiber food

Dietary fiber

Dietary fiber

Dietary fiber

Observed Association*

0

0

0

+

0

+

+

0

0

0

0

+

If the strength of the association between dietary fiber and colon cancer risk is as strong as is suggested by the inter­national correlational studies, then it is important to consider why case-control studies often fail to demonstrate the associa­tion. Three of the possible explanations are: (1) the method of measurement of diet in case-control studies may be so prone to random error that it obscures the true relationship; (2) the variability of intake of dietary fiber within a given population may be too small to demonstrate the fiber-disease relationship that is apparent in the international statistics; or (3) the ecologic fallacy is responsible for the international correlations.

There have been several studies of the validity of food frequency methods. This issue was discussed in Section VI-A. In short, it appears that the food frequency method can reliably rank individuals into at least the extremes of intake (e.g., the extreme lower and upper quintile) (Willett et al., 1985). Proper ranking in the middle of the distribution (e.g., the second to fourth quintile) is less certain. Therefore, risk differentials at the extremes of intake should be demonstrable. Because there is a small degree of misclassification, even at the extremes, true relative risks of the order of three or four may be reduced to the order of two or three in case-control studies.

As for the question of insufficient variability of diet within a population, there are few direct quantitative data that bear on this question. Surveys such as the National Health and Nutrition Examination Survey (NHANES) and U.S. Department of Agriculture household surveys are not designed to estimate usual intake of individuals, but rather to estimate central tendencies for groups. The only available data that permit estimation of interindividual variability of dietary fiber intake in North America are those derived from examining the distribution of fiber intake among subjects from studies such as those currently in progress in western New York (Byers et al., 1987), Toronto (Jain et al., 1980), or Boston (Willett et al., 1985). These studies suggest that there may be as much as a two-fold differ­ence in the levels of fiber intake between the upper and lower quintiles of the population. When considering a two-fold difference against the regression line of the fiber-colon cancer relationship demonstrated in international statistics, relative risks of the order of three or four should be expected. The problems of error in the dietary measure and low degree of vari­ability in diet within a population will both contribute, perhaps more than additively, to a bias toward the null value in case­control studies. Some methodological studies have dealt with the question of loss of power in case-control studies due to error in the measures and/or lack of dietary variability within a single population (McKeown-Eyssen and Thomas, 1985). More research should be conducted in this area so that the general lack of agreement between case control studies and ecological studies relative to colon cancer and fiber can be interpreted properly.

124

Cohort (prospective) studies. There have been only a few cohort (prospective) studies of dietary fiber and fiber-containing foods as related to colon cancer risk in humans because these studies require much larger groups of subjects and a longer interval of time between the initiation of the study and the analysis. These studies do not present the problem of recall bias, but the measure of diet is often very cursory because it must be applied to many thousands of people. The large cohort study by Hirayama (1981) failed to show any relationship between green and yellow vegetable intake and colon cancer, but there was a negative association with rice and wheat, considered together. The study by Kromhout et al. (1982) generally lacked power, as it was a very small cohort. Only five colon cancer patients were observed. An inverse association was observed for dietary fiber and lung cancer, but not for nonlung cancers (including colon cancer). The data obtained in a 21-yr follow-up for 25,493 California Seventh-Day Adventists revealed no relationship between colorectal cancer and green salad consumption (Phillips and Snowden, 1985).

There are several prospective studies in progress which will soon provide data about the fiber-colon cancer relationship. Hennekens and colleagues (Hennekens et al., 1979; Willett et al., 1983) are following over 100,000 registered nurses who have com­pleted validated dietary questionnaires that were specifically designed to assess dietary fiber intake as well as fat intake. Colon cancer incidence has been determined in this study, and a report of the findings is forthcoming. Graham determined dietary intake by questionnaire in 60,000 New York state residents in 1980 (average age 60 yr) (Graham, 1986). The questionnaire was designed to assess dietary fiber as well as fat intake. The first 5-yr follow=up should be completed soon. The cohort of participants in NHANES I has been followed in a special longi­tudinal study (Cornoni-Huntley et al., 1983). Dietary intake will be related to subsequent cancer incidence in this popula­tion. The measure of diet in NHANES I was a single 24-hr recall, so intraindividual variation will compound the interindividual variation to be assessed in this study. Relative risk estimates will therefore be considerably biased toward the null, and any negative findings from the NHANES I analysis will have to be interpreted with caution. The American Cancer Society has recently compiled data from over one-million American volunteers who completed self-administered dietary questionnaires. This cohort of one million will be followed subsequently with regard to cancer incidence and mortality. Although this cohort is not "representative" of the U.S. population, measures are nonetheless unbiased relative to outcome and, therefore, the follow-up of this cohort with regard to fiber and colon cancer should be useful.

Intervention studies. Intervention studies involving dietary fiber supplementation and/or dietary modifica­tions related to dietary fiber have not been conducted to study

125

the dietary prevention of colon cancer. However, many studies are now in progress relative ·to this issue. Such studies often target precursor lesions for colon cancer rather than colon cancer per se. These lesions include polyps, dysplasias, and abnormal cellular morphology of the colon epithelium. The advan­tage of studying such precursor lesions is that they are more common than colon cancer and, therefore, can be investigated with a smaller number of subjects followed for a shorter period of time. Whether findings regarding fiber effects on precancerous lesions can be generalized to colon cancer remains to be determined.

Summary. Many plausible mechanisms could explain a relationship between dietary fiber intake and colon cancer risk in humans. The metabolic correlates of fiber in the diet as related to stool mutagens are reviewed in this report (Section VI-C.6b) as is the animal research on this question (Section VI-C.6c). Nutritional epidemiological data from eco­logical (correlational) studies show a fairly consistent inverse relationship between intake of fiber or fiber-containing foods and colon cancer. The results of case-control studies are con­tradictory. It is increasingly evident that considering dietary fiber as a single entity will not likely lead to interpretable findings relative to human cancer risk. There is a need to reanalyze the existing data based on our evolving knowledge about types of dietary fibers (soluble, insoluble, etc.) and to gather more high quality and better validated data about individual diets in observational research.

b. Metabolic (biochemical) epidemiology

Support for the hypothesis of a protective role for certain types of dietary fiber and fiber-containing foods in the etiology of colon cancer has been obtained from some types of epidemiologic studies, but results are highly method­dependent (Bingham, 1986). However, other lines of evidence based on metabolic (biochemical) epidemiology and laboratory animal studies also suggest that certain kinds of dietary fiber play a role in the etiology of colon cancer, strengthening the hypothesis (see discussion below). Unquestionably there are areas where more research is necessary, and no single mechanism accounts for all colon cancer. The early observations of Burkitt (1971) indicated that high-fiber diets are associated with increased fecal bulk and faster rates of transit. The hypothe­sized protective effect of dietary fiber in large-bowel carcino­genesis depends on the nature and source of dietary fiber in the diet and may be due to adsorption, dilution, and/or metabolism of carcinogens, cocarcinogens, and promoters (Gulliver et al., 1983; Kritchevsky and Story, 1983; Reddy, 1986a; Smith-Barbaro et al., 1981). Thus, individuals who consume relatively moderate or large amounts of certain fibers hypothetically would have a

_greater protection against carcinogenic and cocarcinogenic

126

compounds than do those who consume lesser quantities of these fibers.

Although the causative agents in human colon cancer have not been established, the search for genotoxic carcinogens associated with the etiology of large bowel cancer has been initiated by several laboratories using the Ames Salmonella typhimurium/mammalian microsomal assay system to determine the mutagenic (or presumptive carcinogenic) activity in feces (Bruce et al., 1977; Reddy, 1986a; Wilkins et al., 1981;). The earliest indication that stools of some healthy individuals on Western diets contain ether-extractable compounds that are mutagenic in the Ames system came from studies of Bruce et al. (1977). The work of Hirai et al. (1982) [confirmed by Gupta et al. (1983)] showed that the mutagenic activity of feces of certain healthy individuals appeared to be related to the presence of a product of anaerobic bacteria termed fecapentaene, (S)-3-(l,3,5,7,9-dodecapentaenoyloxyl)-l,2-propandiol. The carcinogenic potential of this compound is the subject of further study in an animal model system. Although there is no direct evidence that fecal mutagens play a role in human colon cancer, a recent study indicated that the mutagen produced in human colonic contents can act as an alkylating agent (Plummer et al., 1986).

Because of the potential relevance of mutagens in the colonic contents to the pathogenesis of colon cancer, the fecal mutagenic activity of populations with varied risk for the devel­opment of colon cancer and with a difference in fiber intake has been determined by several investigators. Reddy et al. (1980b, 1985) examined the dietary pattern and measured fecal mutagens of healthy individuals in the New York City area, Kuopio (rural Finland), and Helsinki (urban Finland). These populations have different rates of colon cancer and levels of consumption of dietary fiber, mainly whole-grain cereal fiber. The total dietary fiber intakes in New York, Helsinki, and Kuopio were approximately 14, 19, and 26 g/d, respectively. The fecal extracts of individuals from New York consuming high-fat/low­fiber diets showed a higher mutagenic activity than did those from individuals from Kuopio consuming high-fat/high-fiber diets. In addition, the percentage of samples showing fecal mutagenic activity was higher in individuals from Helsinki than in Kuopioo Interestingly, the individuals from Kuopio and Helsinki who excreted high levels of mutagens in the feces were consuming diets low in total dietary fiber and whole-grain and cereal fiber, whereas those consuming diets high in total fiber and whole-grain cereals and bread excreted very low levels of mutagens (Reddy, 1986b). When individuals from Helsinki who showed high fecal mutagenic activity and were consuming diets moderately low in total fiber (14-17 g/d) were given 11 g/d of supplemental fiber/din the form of whole-grain (wheat and rye) bread for 4 wk, the fecal mutagenic activity was inhibited considerably (Reddy et al., 1987). Thus, increased fiber intake in. the form of whole-wheat and rye bread may reduce

127

the production and/or excretion of fecal mutagens in healthy individuals.

Currently, most knowledge about the mechanism of the relationship of dietary fiber to colon carcinogenesis is based on experiments conducted on humans (metabolic epidemiology) and laboratory animals.· The major finding of these studies is that a primary effect of dietary fiber is modulation of the concen­tration of secondary bile acids, as well as the metabolic activ­ity of gut microflora. These bacteria, in turn, can metabolize bile acids into tumorigenic compounds in the colon (Reddy, 1986a). Bile acids stimulate the proliferative activity of colonic epithelium, but the mechanism- by which this occurs is unclear (DeRubertis et al., 1984). One mechanism may include the increased activity of colonic mucosal ornithine decarboxy­lase activity which has been suggested to play a role in the phenomenon of tumor promotion (Takano et al., 1981). In addi­tion, there is evidence to suggest that activation of protein kinase C by bile acids in colonic epithelium may represent a critical intracellular event in the process by which bile acids provoke a proliferative response (DeRubertis and Craven, 1987). Thus, there may be an important relationship between colonic bile acids, colonic ornithine decarboxylase activity, protein kinase C, and colon tumor promotion.

Because the secondary bile acids in the colon may act as promoters in the development of colon cancer, several inves­tigators have determined the fecal bile acid levels in high- and low-risk populations for colon cancer development. Jensen et al. (1982) determined fecal bile acid levels and fecal bulk of 30 randomly selected men in each of four areas of Finland and Denmark with varied risk for colon cancer development; Copenhagen (a high-risk area); Helsinki and rural Them in Denmark (inter­mediate risk areas); and rural Parikkala in Finland (a low-risk area). Mean fat intake was found to be high in all four areas with no difference between the areas. Intakes of nonstarch polysaccharides, carbohydrate, and protein (mainly from milk) were higher in the low incidence area of Parikkala than in the high incidence area of Copenhagen. The fecal bile acid concen­tration was significantly higher in the high-incidence area than in the low-incidence area, while the other two areas had inter­mediate values. Fecal bulk showed an inverse association with colorectal cancer incidence.

Reddy et al. (1978) and Domellof et al. (1982) measured the dietary pattern, fecal bile acids, and the fecal bulk of three population groups with varied risk for. colon cancer devel­opment (a high-risk population in New York, an intermediate risk population in Umea, Sweden, and a low-risk population in Kuopio, Finland). Intakes of fat and protein were similar in all three areas, but the intake of· dietary fiber (whole-grain and cereal fiber) was highest in Kuopio, followed by Umea and New York. Although the daily excretion of fecal bile acids was similar in

128

/i

all three areas, the concentration of fecal secondary bile acids was lower in Umea and Kuopio than in New York.

Reddy et al. (1985, 1987) determined the fecal bile acids and fecal bulk in healthy individuals from Kuopio and Helsinki. Food frequency questionnaires completed by each vol­unteer indicated that the frequency of consumption of cereals, porridge, and whole-grain bread, and amount of fiber from bread (primarily from rye bread) was higher in Kuopio than in Helsinki. Nutrient analysis of 3-d food records indicated no significant differences in the intake of total protein, fat, and calories between two areas, but the intake of dietary fiber, especially from the whole-wheat and rye bread was higher in Kuopio than in Helsinki. The concentration of fecal bile acids was lower and the stool weight was higher in Kuopio than in Helsinki. In addi­tion, an inverse relationship was observed between the amount of total dietary fiber and the concentration of fecal deoxycholic acid and lithocholic acid and total bile acids in both males and females (Reddy et al., 1987).

The effect of 11 g/d of supplemental dietary whole-grain fiber in the form of whole-wheat and rye bread was studied in individuals consuming high-fat, moderately low-fiber diets and excreting high levels of fecal bile acids (Reddy et al., 1987). A significant decrease in the concentration of fecal deoxycholic acid, lithocholic acid, and 12-ketolithocholic acid was observed in subjects during the fiber-supplementation period ·when compared to the control-diet period.

In summary, the studies thus far conducted have revealed that prevalence of excretion of fecal mutagens and the concentra­tions of fecal bile acids are higher in populations at high risk for the development of colon cancer than for populations at low risk. High-risk populations tend to consume diets higher in fat and lower in fiber than low-risk populations. Dietary fiber, particularly whole-grain cereals and bread, may reduce the production and/or excretion of fecal mutagens and decrease the concentrations of secondary bile acids that seemingly play a role in colon carcinogenesis. Further research is needed to elucidate the role of fecal mutagens in colon cancer and the influence of various dietary fibers.

c. Colon carcinogenesis in laboratory animals

A variety of compounds, namely 1,2-dimethyl­hydrazine (DMH), azoxymethane (ADM), 3,2 1 -dimethyl-4-aminobi­phenyl (DMAB), and methylnitrosourea (MNU), that are carcinogenic in the colon have been used in a number of animal models to study the effect of types and amount of dietary fiber on tumorigenesis at this site (Shamsuddin, 1986). Administration of these com­pounds to rodents induces both benign adenomas and malignant carcinomas of the colon ..

129

Studies exam1n1ng the role of various types of dietary fiber in animal models have provided somewhat conflicting results (Jacobs, 1986b; Reddy, 1986a). These discrepancies might have been caused by many factors that varied across studies, such as the nature and dose of the carcinogen used to induce colon tumors, differences in the susceptibility of different sexes and strains to car9inogen treatment, variation in the composition of diets, quantitative and qualitative differences in the fibers and their components fed, differences in the feeding of fiber sources during the initiation and promotion stages of carcinogenesis, differences in food intake by the animals, and/or to differences in experimental design and duration of the experiment (Jacobs, 1986b; Reddy, 1986a).

Experiments with rats using the carcinogens DMH, ADM, □MAB, and MNU are summarized in Table 18. With carcinogens requiring microsomal activation, namely DMH, ADM, and DMAB, 13 of 17 studies demonstrated evidence of protection by dietary wheat bran against colon tumor development, whereas three studies showed no effect and one study showed an enhancing effect. In the study in which an enhancing effect was observed, wheat bran was fed only before and during carcinogen treatment. On the other hand, with MNU, a direct-acting carcinogen, dietary wheat bran had no effect. Dietary soybean bran and rice bran had no effect whereas oat bran had an enhancing effect on DMH-induced colon carcinogenesis in rats. Dietary corn bran either had an enhancing effect or no effect on colon carcinogenesise Results using dietary cellulose indicate that six of nine experiments showed a protective effect and three studies showed no effect. The studies examining the effect of pectin on experimental colon carcinogenesis indicate that three out of seven experiments showed no effect, three studies showed an enhancing effect, and one study showed a protective effect. In addition, dietary lignin was found to have a protective effect in □MAB-induced colon carcinogenesis.

Protective effect. The effect of a semi­purified diet containing 15% pectin or wheat bran and 20% fat on colon carcinogenesis induced by MNU or ADM was studied in Fisher 344 rats by Watanabe et al. (1979). In this study, the control and experimental diets were fed to rats before, during, and after carcinogen treatment and until termination of the experiment. The control diet contained about 5% Alphacel® as a fiber source. The addition of pectin or wheat bran to the diet greatly inhibited colon tumor incidence induced by ADM. The diets containing pectin did not protect against MNU-induced colon carcinogenesis. Thus, pectin, a fermentable fiber, inhib­ited colon carcinogenesis induced by ADM, which requires meta­bolic activation in the liver and colon, but showed no effect with the direct-acting carcinogen MNU. The mode of action dif­fered from that of wheat bran. Wheat bran protected rats against DMH-induced colon cancer, whether the carcinogen was administered orally or subcutaneously (Barbolt and Abraham, 1978; Chen et al., 1978; Wilson et al., 1977).

130

Table 18. Effects of types and amounts of dietary fiber on chemically-induced colon cancer in animal models.*

Fiber Fed During a Fiber Source in Dietb Effect on

d Reference Fiber Source Strain of Rat Sex Carcinogen C

I p Control Experimental Colon Tumors E

Wilson et al. (1977) Wheat bran Sprague-Dawley M DMH E + 20% Barbolt & Abraham (1978) Sprague-Dawley M DMH E + 20% Barbolt & Abraham (1980) Sprague-Dawley M DMH E + 20% Bauer et al. (1979) Sprague-Dawley M DMH I - 20% 0

Cruse et al. (1978) Wistar F DMHe E - 20% 0

Abraham et al. (1980) Sprague-Dawley M DMH E + 20% Jacobs (1983b) Sprague-Dawley M DMH I - 20% +

Jacobs (1983b) Sprague-Dawley M DMH p - 20% Fleiszer et al. (1978) Chester Beatty M DMH E - 28% Barnes et al. (1983) fischer 344 M DMH E + 20% f Kroes et al. (1986) Wistar M DMH E + 3.8 Nigro et al. (1979) Sprague-Dawley M ADM E + 10% 0

I-' Nigro et al. (1979) Sprague-Dawley M ADM E + 20% \.,-I Nigro et al. (1979) Sprague-Dawley M ADM E 30% I-' +

Watanabe et al. (1979) Fischer 344 F ADM E + 15% Reddy et al. (1981) Fischer 344 M ADM E + 15% Watanabe et al. (1979) Fischer 344 F MNU E + 15% 0

Reddy & Mori (1981) Fischer 344 M DMAB E + 15%

Barnes et al. (1983) Corn bran Fischer 344 M DMH E + 20% +

Freeman et al. (1984) Wistar M DMH E - 4.5% Reddy et al. (1983a) Fischer 344 M DMAB E + 10% +

Barnes et al. (1983) Rice bran Fischer 344 M DMH E + 20% 0

Barnes et al. (1983) Soybean bran Fischer 344 M DMH E + 20% 0

Jacobs & Lupton, (1986) Oat bran Sprague-Dawley M DMH E - 20% +

Reddy et al. (1983a) Lignin Fischer 3L~4 M DMAB E + 7.5%

* See footnotes at end of table.

Table 18. Effects of types and amounts of dietary fiber on chemically-induced colon cancer in animal models (continued).

Fiber Fed During a Fiber Source in Oietb Effect on

d Reference Fiber Source Strain of Rat Sex Carcinogen C I p E Control Experimental Colon Tumors

Freeman et al. (1978) Cellulose Wistar M DMH E - 4.5% Freeman et al. (1980) Wistar M DMH E - 4.5% Freeman et al. (1980) Wistar M DMH E - 9.0% Trudel et al. (1983) Sprague-Dawley M DMH E - 22% Jacobs & Lupton (1986) Sprague-Dawley M DMH E - 10% 0

Ward et al. (1973) Fischer 344 M AOM E + 20% 0

Nigro et al. (1979) Sprague-Dawley M AOM E + 10% 0

Nigro et al. (1979) Sprague-Dawley M AOM E + 20% Nigro et al. (1979) Sprague-Dawley M ADM E + 30%

Bauer et al. (1979) Pectin Sprague-Dawley M DMH I - 6.5% + Freeman et al. (1978) Wistar M DMH E - 4.5% 0

._. Freeman et al. (1978) Wistar M DMH E - 9.0% 0 \.,J Bauer et al. (1981) Sprague-Dawley M l'v DMH E - 5.0% +

Jacobs & Lupton (1986) Sprague-Dawley M DMH E - 10% + Watanabe et al. (1979) Fischer 344 F AOM E + 15% Watanabe et al. (1979) Fischer 344 F MNU E + 15% 0

Bauer et al. (1981) Guar gum Sprague-Dawley M DMH E - 5% 0

Jacobs & Lupton, (1986) Sprague-Dawley M DMH E - 10% +

Watanabe et al. (1978) Carrageenan Fischer 344 F AOM E + 15% +

a Fiber was fed during initiation (I= before and during carcinogen treatment), promotion (P = after carcinogen treatment), or the

b entire period (E = during initiation and promotion periods). Control diets contained(+) or did not contain(-) fiber.

C DMH = 1,2-dimethylhydrazine; AOM = azoxymethane; NMU = methylnitrosourea; DMAB = 3,2'-dimethyl-4-aminobiphenyle

d Effect: (-) protective, (+) enhancing, or (o) no effect. e Very high dose of carcinogen. f Expressed as g/100 kcal.

The effect of dietary wheat bran and dehydrated citrus fiber fed at the 15% level with 5% dietary fat on intestinal carcinogenesis induced by ADM and DMAB was studied in male Fisher 344 rats (Reddy and Mori, 1981; Reddy et al., 1981). Composition of the diets was adjusted so that all animals in different experimental groups consumed approximately the same amount of protein, fat, minerals, and vitamins. All experimental diets were fed before, during, and after carcinogen administra­tion and until the termination of the experiment. The animals fed the wheat bran or citrus fiber and treated with ADM had a lower incidence (number of animals with tumors) and multiplicity (number of tumors/tumor-bearing rat) of colon tumors and tumors of the small intestine than did those fed the control diet and treated with ADM. Animals fed the wheat bran and treated with DMAB had a lower incidence and multiplicity of colon tumors

Nigro et al. (1979) reported that diets containing 35% beef fat plus 10% alfalfa, wheat bran, or cellulose showed no inhibitory effect on ADM-induced colon carcinogenesis, whereas diets containing 5% fat plus either 20% or 30% wheat bran or cellulose inhibited ADM-induced colon carcinogenesis in rats& Although the fat and fiber levels were extreme in this study, the data indicate that the amount of fat in the diet is impor­tant in assessing the effect of fiber on colon cancer. Dietary fat in large amounts may be a stronger promoter of colon cancer than dietary fiber is an inhibitor.

A recent study by Kroes et al. (1986) suggested a pro­tective effect of wheat bran at 3.8 g/100 kcal on OHM-induced colon carcinogenesis in male Wistar rats. The protective effect occurred at low, medium, and high fat levels and was dose related.

Enhancing or no effects Cruse et al. (1978) reported that a diet containing 20% wheat bran had no protective effect on DMH-induced colon carcinogenesis in rats. However, the DMH dose levels were so high that they may have overpowered any protective effect of wheat bran. The data suggested that a high­wheat bran diet reduced the number of DMH-induced early deaths in rats. Administering an excessive dose of carcinogen for a pro­longed time may obscure subtle changes in carcinogenic response induced by certain dietary modifications.

Bauer et al. (1979) studied the effect of 20% wheat bran in a fiber-free, semipurified diet containing 20% peanut oil on DMH-induced colon carcinogenesis. The experimental wheat bran diet was fed to animals only before, during, and until 2 wk after carcinogen injection. Animals fed the semipurified diets were then transferred to a standard rat pellet diet for 10-12 wk before sacrifice. Wheat bran had no effect on colon tumor inci­dence compared to the fiber-free control diet. It is possible in this study that not only the high tumor yield resulting from large doses of DMH, but also the switch from a high-fiber

133

experimental diet to a standard diet during the postcarcinogen period, contributed to the failure to show any protective effect ·of wheat bran in this study.

Jacobs (1983b) reported a study on the effect of feed­ing rats a 20% wheat bran diet either during and/or after DMH treatment and found an increase in colon tumor incidence in rats fed the wheat bran diet during carcinogen treatment. In those animals fed wheat bran after DMH administration, tumor incidence was inhibited compared to the fiber-free control group. These studies suggest that dietary wheat bran protects against colon carcinogenesis when fed during the stage of promotion (post-initiation), but that it may have the opposite effect if fed during the initiation stage.

The effect of dietary corn bran (high in hemicellulose), fed before, during, and after carcinogen treatment on DMAB­induced colon carcinogenesis in rats was studied (Reddy et al., 1983b). Animals fed a semisynthetic diet containing 10% corn bran had an increased multiplicity of colon adenomas compared to those fed a control diet containing 5% Alphacel®. An impor­tant effect of corn bran observed in this study was a significant increase in the concentration of fecal deoxycholic acid, a promotor of colon tumorse

The effects of fiber fractions such as cellulose and guar gum on DMH-induced colon carcinogenesis were studied (Jacobs and Lupton, 1986). Studies with 5% guar gum and 19% fat demon­strated no effect on colon carcinogenesis (Bauer et al., 1981), whereas a diet containing 10% guar gum and 7% fat increased colon

·tumors (Jacobs and Lupton, 1986). In another study, dietary cel­lulose had no effect on colon carcinogenesis (Jacobs and Lupton, 1986). Several of these studies are difficult to interpret because other dietary components such as fat, minerals, and other nutrients were uncontrolled and several studies employed very high doses of carcinogen.

In conclusion, the results generated from studies on the effect of dietary fiber in colon carcinogenesis in animal models suggest that (1) the effects (inhibitory or enhancing) depend on the type of fiber; (2) the chemical composition, in addition to the physical properties of fiber, may be important in modulating the effects of various dietary fibers, and; (3) wheat bran appears to inhibit colon tumor development in animal models more consistently than other fiber sources. The varied conclusions on the other types of~fiber appear to be related to differences in the experimental protocols. The experimental animal models are highly sensitive to factors such as dietary change and the timing and dose of the carcinogen. The relevance of these animal models to human cancer needs to be determined.

134

7. Other neoplasms

The relationships between dietary fiber and neoplasms other than colon cancer have not been studied extensively.

In a case-control study of breast cancer, Lubin et al. (1986) compared 818 cases with both surgical and neighborhood controls. Diets high in animal protein and low in dietary fiber were found to be associated with increased risk for breast can­cer. The association with fiber was stronger for women under 50 yr than for older women. The strength of the dietary associa­tions was weakened considerably by adjustment of the analyses for nondietary risk factors.

Another line of evidence suggesting an involvement of dietary fiber in breast cancer is the finding that lignan excre­tion was significantly lower in the urine of women with breast cancer than in normal omnivorous and vegetarian women and that lignan excretion was positively correlated with dietary fiber intake (Adlercreutz et al., 1982). Endogenous human lignans (enterolactone and enterodiol) are produced by the intestinal microflora of humans from precursors in fiber-rich foods.

La Vecchia et al. (1986) conducted a case-control com­parison of 206 patients with endometrial cancer and matched hos­pital controls. Their analyses indicated a negative association between endometrial cancer and the intake of green vegetables and fruit. A negative association with the intake of whole-grain products did not persist in multivariate analyses. Other signif­icant relationships detected included a negative association with indices of retinal and carotene intake, a positive association with total dietary fat, and a strong positive association with body mass index.

Risch et al. (1985) compared the dietary intake in 246 patients with gastric cancer and an equal number of popu­lation controls and found a protective effect of dietary fiber and of "fibrous foods" (vegetables, fruits, soybeans, seeds, and nuts) and a positive association with the intake of grains. The authors noted that one shortcoming of their study was difficulty in recruiting both cases and controls, leading to concern about the representativeness of their subjects.

Research needs in the study of dietary fiber and other cancers are similar to those in the area of colon cancer: better understanding of the differences in the function of various com­ponents of dietary fiber, greater standardization of experimental protocols, and examination of the influence of other dietary factors.

135

8. Other disorders

Dietary fiber has been hypothesized to have a rolff·in the etiology of disorders such as hemorrhoids, deep-vein thrombosis, pelvic phleboliths, hiatal hernia, varicose veins, and appendicitis. Evidence on the association of dietary fiber with these disorders has been reviewed (Burkitt, 1985; Segal, 1985; Walker and Burkitt, 1985). In all cases, evidence sup­porting a role for dietary fiber in these conditions is based primarily on epidemiological observations of their relative rarity in populations in less-developed regions compared to more westernized populations with lower dietary fiber intake. As noted several times in this report, there are many other differences in both diet and lifestyle between such population groups. Additional research directed at isolating effects of dietary fiber and fiber components on these disorders is needed before definitive conclusions can be drawn.

D. STUDIES OF POTENTIAL UNDESIRABLE EFFECTS OF DIETARY FIBER

1. Effects on mineral bioavailability

a. Mineral status of vegetarians

Studies by Hardinge et al. (1958), Anderson et al. (1981), King et al. (1981), Treuherz (1982), Gibson et al. (1983), Davies et al. (1985), and Howie and Shultz (1985) indi­cated that fiber intakes of vegetarians were higher than those of omnivores. Anderson et al. (1981) reported that mean hemoglobin, serum transferrin saturation, and serum and hair zinc levels of vegetarian women were within the normal ranges for these parame­ters. King et al. (1981) observed that the lacto-ovo vegetarian diet of pregnant vegetarian women did not appear to affect their zinc status as indicated by levels of zinc in plasma, urine, and hair. Gibson et al. (1983) found that, in spite of a higher fiber intake by the vegetarians, copper levels in serum and copper and selenium levels in hair were comparable to those of omnivores; manganese levels in hair were higher than those of omnivores. Shultz and Leklem (1983) reported that there were no differences in blood selenium levels between Seventh-Day Adventist vegetarians and omnivores.

In a study by Marsh et al. (1980), the loss of bone mineral mass was found to be less in a group of postmenopausal lacto-ovo vegetarian women (aged 50 to 89 yr) than in a group of omnivorous women of the same age range. However, no differences were seen between bone mineral mass in lacto-ovo vegetarian men and omnivorous men in any decade of life examined (Marsh et al., 1983).

Treuherz (1982) compared zinc status of adolescent lacto-ovo vegetarians with a control group.· Dietary fiber intake

136

wa~ higher for the vegetarians, and dietary zinc intake was similar, but zinc intake was significantly higher per 1000 kcal for the vegetarians. Hair zinc levels of both groups were within ranges obtained by other workers, but levels were significantly lower for the vegetarians (218 vs. 249 µg/g).

Ganapathy and Dhanda (1980) studied two male and two female Punjabi students living in the United States and consuming lacto-ovo vegetarian diets similar to those consumed in their own country. Males and females were in positive iron balance over a 7-d period with iron intakes of 9.8 and 9.4 mg/d, respectively.

Kies et al. (1983) reported zinc balances of vegetarians and omnivores consuming controlled vegetarian diets. Fecal zinc excretion and zinc balances of subjects indicated better utiliza­tion of zinc from vegetarian diets by practicing vegetarians than by omnivores consuming vegetarian diets.

From the information available, it appears that the mineral status of vegetarians is essentially equivalent to that of omnivores, and that if vegetarians do have decreased bio­availability of minerals due to higher fiber intakes, they are able to adjust. Alternately, fiber intakes of vegetarians may not be as elevated as has been assumed, which is likely to be true for lacto-ovo vegetarians. Walker (1985), in a review of the occurrence of mineral deficiencies in Third World popula­tions where intakes of fiber are high, concluded that there is little evidence that these populations suffer from mineral deficiencies resulting from high fiber intakes. However, such populations consuming refined grains, e.g., rice, do not have high fiber intakes; expression of mineral consumption in terms of energy intake or body weight may reveal adequate mineral intake.

b. Blood mineral levels

Blood serum levels of minerals would not likely be altered by fiber unless a deficiency is well-advanced, due to the homeostatic maintenance of blood mineral levels. Any decreases noted are likely to remain within the normal ranges. Nonetheless, serum mineral levels have been used to assess the effect of dietary fiber on mineral nutriture.

Anderson et al. (1980b) reported that serum levels of calcium, iron, and magnesium, and total iron-binding capacity, and hemoglobin levels were normal in 15 diabetic patients fed high-fiber diets for an average of 21 mo.

Rattan et al. (1981) found that serum levels of iron, total iron-binding capacity, calcium, phosphorus, zinc, and magnesium were within the normal ranges for 68 adults taking 2 tablespoons of bran for 6 mo, for 43 adults not taking bran supplements, and for 20 adult vegetarians (eight taking bran supplements).

137

Vaaler et al. (1985) studied 33 insulin-dependent diabetic subjects i~ Norway for 3 mo on their ordinary diet with low-fiber bread (15-20 g fiber/d), for 3 mo when guar gum (29 g/d) was added, and for 3 mo when bran was added (33 g/d)e The last two diets were consumed in a crossover design. Serum concentrations of calcium, inorganic phosphate, magnesium, iron, zinc, and selenium measured at the end of each period did not differ from the control period or from each other. Urinary excretion of calcium was slightly lowered (p <0.01) during treatment with wheat bran, but inorganic phosphorus, magnesium, and zinc were unchanged.

Godara et al. (1981) fed 21 g cellulose to nine adoles­cent girls for 21 d and compared results with those observed on a control low-fiber diet. Serum calcium, inorganic phosphorus, and iron levels decreased significantly.

c. Mineral availability from test meals

Absorption of iron as measured by changes in serum iron or in total iron content as measured by whole-body counter was less in human subjects given breads made with whole­meal flour in place of white flour (Dobbs and Baird, 1977; Elwood et al., 1970; Vellar et al., 1968) and when bran was added to rolls (Bjorn-Rasmussen, 1974). Whole-wheat bread added to a meal resulted in a decrease in absorption of nonheme iron which could not be explained by the presence of phytate (Simpson et al., 1981b). When bran, pectin, or cellulose was added to muffins prepared with wheat flour, only bran significantly lowered the absorption of iron in a meal (Cook et al., 1983). A further study with a low-fiber vs. a high-fiber meal demonstrated that the iron absorption in the low-fiber meal was about double that in the high-fiber meal. The authors concluded that this decrease from 6 to 3% absorption was not a major problem in iron availability in humans.

Sandstrom et al. (1980) reported that when white and whole-meal breads contained equal amounts of zinc, the absorp­tion of zinc from a test meal was greater from white than from whole-meal bread. The addition of protein to the whole-meal bread resulted in improved iron absorption. When Sandstrom and Cederblad (1980) substituted defatted soy flour for 25% of the protein in chicken in a test meal, zinc absorption was not influenced. Feeding a soybean meal or an animal protein meal with the same zinc content resulted in similar absorptions of· zinc.

Test meals may give information about the effects of fiber on the absorption of minerals from a particular meal, but will not necessarily indicate the overall absorption of minerals in the whole diet. Other constituents in the meal and in the diet also will affect mineral availability.

138

d. Fecal excretion of minerals

Stasse-Wolthuis et al. (1980) fed each of four groups of subjects one of the following diets: a low-fiber diet (18 g dietary fiber), a diet containing fruits and vegetables (43 g dietary fiber), a diet containing citrus pectin (28 g dietary fiber), or a diet containing wheat bran (37 g dietary fiber). Results were compared with those obtained when a low fiber control diet was consumed for 2.5 wk. Subjects fed the bran diet had increased magnesium excretion, but there were no significant changes in either calcium or magnesium excretions on any of the other diets.

Godara et al. (1981) fed nine female adolescents on a low-fiber diet and the low-fiber diet plus 21 g cellulose for 21-d periods. Fecal excretions of calcium, phosphorus, and iron were significantly increased when cellulose was added to the diet.

Tsai et al. (1983) fed a basal diet containing 2.8 g crude fiber and the same diet with 25 g soy polysaccharide in a crossover design to 14 subjects for 17-d periods. The soy polysaccharide contained 60% total dietary fiber or 30% NDF. Fecal excretions of calcium, phosphorus, iron, magnesium, zinc, and copper were not significantly different for individuals fed these two diets.

Turnlund et al. (1984) studied the effects of a-cellulose and phytate on zinc absorption using a stable isot­ope of zinc. Subjects were given 0.5 g cellulose and 2.34 g phytic acid/kg body weight for 9 to 15 do Fecal monitoring of the zinc isotope revealed that cellulose had no effect, whereas phytate decreased the absorption of zinc.

Dintzis et ale (1985) measured copper, iron, zinc, and calcium concentrations in dry-milled corn bran, wheat brans, and soybean hulls before they were baked into bread and after passage through the human gastrointestinal tract. The concentrations of all four elements increased in corn bran, calcium increased in wheat brans, and zinc increased and iron decreased in soy hulls as these fiber sources passed through the gastrointestinal tract.

If minerals are bound by fiber and are rendered unavailable to the body, increased fecal mineral excretion should result. The assumption is made that urinary excretion would not be affected. However, more complete information can be obtained when minerals are determined in both urine and feces so that balances can be calculated.

139

e. Results of mineral balance studies -supplements to self-selected diets

Guthrie and Robinson (1978) added wheat bran (14 g/d) to the diets of four females for 4 wk. Zinc balances were not affected.

Following a 2-wk control period, Schweizer et al. (1983) supplemented the normal diets of two males and four females with 21 g dietary fiber from a nonpurified soya pulp or a purified soya fiber for 3 wk each. Fecal calcium, magnesium, and iron, but not zinc, were significantly increased after addition of the fiber; however, mineral intakes were also increased when the fiber was added.

Balasubramanian et al. (1987) studied five females and two males, 59-76 yr of age, consuming self-selected diets. After a 10-d control period the subjects added 30 g wheat bran supplement to their diets for two periods of 10 d each increas­ing fiber intake from 8.5 g NDF to 20.9 g NDF/d. Dietary calcium intakes were calculated. All feces were analyzed for calcium content during the last 8 d of each 10-d period. Apparent cal­cium absorption as determined from calculated intake and analyzed feces decreased during the last 8 d of the study.

Adding a fiber supplement to the usual diets of human subjects is probably the most practical way to conduct studies of this type, but results are more uncertain because of greater possibilities of inaccuracy in reporting food intakes or in making food collections.

f. Results of mineral balance studies -controlled diets

Results of human studies in which mineral balances were determined when fiber was increased in the diet are tabulated in Table 19. In these studies the dietary intakes were controlled. Factors that may have affected the results of the studies reported in Table 19 are:

• Levels of mineral intake

Decreased absorption of minerals due to binding by fiber is not so important if mineral intakes are high. When intakes are marginally adequate, negative balances of minerals are more likely to result if minerals are bound to fiber. The level of mineral intake needed to main­tain balance also depends on the level of intake to which the subject is accustomed.

140

Table 19. Studies on the effects of fiber in the diet on mineral balances in humans.*

Study Fiber Fiber Intake Mineral Balances Reference Design Subjects Duration Source (g/d)

Ca p Mg Fe Zn Cu

Mccance & Widdowson ND 4 M, 4 F 14-28 d Brown bread 20-338

more neg NE NE (1942a) dee

14-28 d White bread 7-128

Mccance & Widdowson RR 3M, 3 F 3 wk Brown bread 25-438 neg NE NE

(1942b) dee 3 wk Dephytinized 25-43

8 neg dee dee brown bread dee

3 wk White bread 7-128

Widdowson & McCance ND 4 M, 4 F 42-91 d Brown bread 20-338 - - - dee

(1942) 42-70 d White bread 7-123

1--' Walker et al. (1948) ND 3M 3-5 wk Brown bread 398

NE NE .t::- negb negb 1--' dee dee

1-4 wk White bread 12a

Cullumbine et al. (1950) ND 12 M 3 wk Unpolished rice dee b imp dee b NE

2-3 wk Polished rice

Reinhold et al. (1973) ND 3M 32 d Unleavened 303 neg NE - - neg

wholemeal bread dee dee 16 d White bread 9a

Reinhold et al. (1976) ND 2 M 20 d Leavened 3f neg neg neg - neg wholemeal bread dee dee dee dee

20 d White bread 22c

Ismail-Beigi et al. ND 3M 20 d Cellulose in 35c neg NE neg - neg (1977b) apple compote dee dee dee

20 d Control diet 19c

* See footnotes and code for abbreviations at end of table.

Table 19. Studies on the effects of fiber in the diet on mineral balances in humans (continued).*

Study Fiber Fiber Intake Mineral Balances Reference Design Subjects Duration Source (g/d)

Ca p Mg Fe Zn Cu

Sandstead et al. (1978) RR 5 M 28-30 d SW wheat bran llf - - - NE NE imp 5 M 28-30 d Corn bran 24f - - - NE NE NE 5 M 28-30 d Control diet

Cummings et al. (1979b) ND 4 M 3 wk Wheat fiber 53d neg

2i dee

3 wk Control

Cummings et al. (1979a) ND 5 6 wk Pectin 46e NE 3 wk Control 15e

Drews et al. (1979) RR 8 M 4 d Cellulose 7+14 - - NE - NE NE (adolescents) 4 d Hemicellulose 7+14 - - more - neg more

...... neg neg ~ N dee dee dee

4 d Pectin 7+l,4 - - NE - NE NE Control 7

Kies et al. (1979) RR 12 M 14 d Hemicellulose 8f+4 14 d II 8+14 - - - - dee 14 d II 8+24 - - - - neg

Kelsay et al. (1979a,b) co 12 M 26 d Fruits & veg 2/ neg NE neg NE neg neg

/ dee dee dee dee

26 d Control

Slavin and Marlett (1980) ND 7 F 1 mo Cellulose 2/ more neg - NE 1 mo Control lOf

* See footnotes and code for abbreviations at end of table.

Table 19. Studies on the effects of fiber in the diet on mineral balances in humans (continued).*

Study Fiber Fiber Intake Mineral Balances Reference Design Subjects Duration Source (g/d)

Ca p Mg Fe Zn Cu

Kelsay et al. (1981) 4x4 12 M 21 d Fruits & veg lOf NE - NE NE NE LS 21 d Fruits & veg 19f NE - NE NE NE

21 d Fruits & veg 2/ NE - NE dee NE 21 d Control /

Van Dokkum et al. (1982) ND 12 M 20 d Medium-fiber 2/ NE - NE NE NE NE coarse-bran bread

9f 12 M 20 d White breadg 4 M 20 d High-fiber 9 35f NE - NE neg NE NE

coarse-bran bread

22f 4 M 20 d Medium-fiber9 NE - NE NE NE NE I-' fine-bran .j:::-v-1 bread

4 M 20 d Wholemeal9 22f NE - inc NE NE inc bread

Kelsay & Prather (1983) 3x3 12 M 4 wk Fruits, veg 2/ neg dur- - neg dur- - neg dur- NE LS & spinach ing wk 4 ing wk 4 ing wk 4

5f dee dee dee

4 wk Control & spinach

2/ 4 wk Fruits, veg NE - NE - NE dee w/o spinach

Andersson et al. (1983) ND 6 M 24 d Wholemeal 3i NE - NE NE NE bread

24d 24 d Brown bread NE - NE NE NE 24 d White bread 16d

* See footnotes and code for abbreviations at end of table.

1---' ~ ~

Table 19. Studies on the effects of fiber in the diet on mineral balances in humans (continued).

Study Fiber Fiber Intake Mineral Balances Reference Design Subjects Duration Source (g/d)

Ca p Mg Fe

Behall et al. (1987) RR 11 M 4 wk Cellulose 6+23 NE - NE NE 4 wk Na-carboxy- 6+23 NE - NE NE

methylcellulose 4 wk Locust bean gum 6+24 NE - NE NE 4 wk Karaya gum 6+r4 NE - NE NE 4 wk Control 6

a Estimated from Mccance and Widdowson•s food tables (Paul and Southgate, 1978). Values include fiber from bread only. b Balances improved with time. c Analyzed acid detergent fiber values which do not include hemicellulose. d Calculated from Mccance and Widdowson•s food tables (Paul.and Southgate, 1978) by investigators. ; Analyzed total dietary fiber.

Analyzed neutral detergent fiber values which include cellulose, hemicellulose, and lignin. g Compared with results for same subjects fed medium-fiber coarse-bran bread.

CO= crossover

dee= decreased

F = females

CODE FOR ABBREVIATIONS IN TABLE 19

imp= improved

inc= increased

LS= Latin square

M = males

ND= not detailed

neg= negative

NE= no effect

- = not determined

RR= random rotation

SW= soft white

veg= vegetables

Zn Cu

NE NE NE NE

NE NE NE NE

• Presence of phytate

Unrefined cereals are high in phytate which may be partially responsible for negative mineral balances observed on brown bread intakes. Removal of phytate in one study resulted in improved mineral balances.

• Presence of oxalic acid

The presence of oxalic acid in the spinach fed in the diets containing fiber from fruits and vegetables was partially responsible for negative balances originally attributed to fiber. Negative balances resulted only when spinach was fed with a diet containing higher levels of fiber.

Level of protein intake

High levels of protein intake have been shown to decrease calcium balance due to increased excretion of calcium in the urine. Detrimental effects of fiber on calcium balance may be more pronounced on high-protein diets.

• Kind of fiber

In vitro, some fibers have been shown to bind minerals more readily than others, but the in vivo and in vitro findings are sometimes contradictory. Wheat fiber, hemicellulose, and cellulose resulted in negative balances in some studies. Pectin, which is a soluble fiber, did not affect mineral balances

Level of fiber intake

Most of the negative mineral balances reported in the studies in Table 19 resulted when more than 25 g/d of insoluble fiber (NDF) was fed.

• Length of study period

Most of the study periods were from 2 to 4 wk. In two studies in which experimental periods were extended to 8 to 18 wk, negative mineral balances improved. Thus, if sufficient time is allowed, the body may adjust to any decreased mineral availability.

Some of the studies listed in Table 19 involved .only a few subjects, which makes it difficult to assume that results apply to the population in general and to interpret statistical analyses. In some of the earlier studies, statistical analyses were not performed on the data; however, individual data were reported and the decreased mineral balances due to fiber in the diet appear to be real. Some of the studies employed no

145

crossover or Latin square design and, therefore, any effects of feeding one diet before another were not eliminated.

g. Summary

Evidence that fiber has an adverse effect on mineral bioavailability is conflicting. Although some test meals containing fiber result in decreased absorption of iron and zinc, the actual amount of decrease is probably small in relation to the total daily intake of these minerals. Some reports indicate that minerals in feces were increased due to fiber in the diet, but others report no effect. Results of balance studies showed that with high intakes -of wheat fiber mineral balances were decreased or negative. In two studies, the addition of cellu­lose resulted in negative balances; in two others, there was no effect. There are many different kinds of cellulose which might be expected to elicit different responses. Hemicellulose decreased zinc balance in two studies. Consumption of diets containing about 25 g NDF/d in fruits and vegetables did not effect mineral balances unless oxalic acid in spinach was included. Locust bean gum, karaya, carboxymethylcellulose, and pectin did not affect mineral balances ..

Given the possibility that there is likely to be an adaptation to any alteration in mineral availability resulting from an increased fiber intake, a moderate level of fiber intake of 20-25 g/d of NDF (or insoluble fiber) does not appear to pose a problem.

2. Other potential deleterious effects

A sudden change from a low-fiber to a high-fiber diet can provoke symptoms of gastrointestinal distress, including abdominal distention, increased gas production, diarrhea, nausea, and vomiting. Reports of such side effects are not uncommon in clinical trials of fiber sources such as wheat bran and guar gum, especially when large amounts are introduced abruptly into the diet. There are anecdotal reports and suggestive evidence of other, more serious side effects including gastric bezoar in diabetic autonomic neuropathy (Levitt et al., 1980), small bowel volvulus (Duke and Var, 1977), and sigmoid volvulus (Ballantyne,· 1982). Esophageal and intestinal obstruction have been reported when large amounts of hydrophilic fiber sources have been consumed with inadequate fluid (Berman and Schultz, 1980; Kang and Doe, 1979; Noble and Grannis, 1984). The slowing of gastric emptying by guar gum was shown to produce esophageal reflux in a patient with pyloric stenosis (Harju and Larmi, 1985). Various fiber sources have also been found to exacerbate steatorrhea in patients with pancreatic insufficiency (Dutta and Hlasko, 1985; Isaksson et al., 1984). Wrick et al. (1983) fdund that supple­ments of pure cellulose and fine wheat bran can be constipating (i.e., cause difficult passage of hard stools).

146

Van Soest (1984) has suggested that adsorption of bacteria by isolated cellulosic compounds may lead to selec­tive depletion of the col0nic flora with potentially harmful consequences. Further investigation is needed to confirm this possibility.

Persorption, the paracellular passage of particles from the intestine into the blood or lymph circulation, has been reported with microcrystalline cellulose (Pahlke and Friedrich, 1974). A wide variety of other ingested particles (starch gran­ules, diatoms, pollen) also can be persorbed (Volkheimer, 1977) and those resistant to enzymatic degradation are eliminated in the urine. Accumulation of these particles in the kidney may pose a hazard, but this possibility has not been investigated thoroughly. In animal studies, persorbed carrageenans have been shown to be taken up by macrophages and to interfere with macrophage function (Nicklin and Miller, 1984), but these studies have not been extended to humans.

Dietary fiber may delay or impede the absorption of drugs. Brown et al. (1978) showed that a high-fiber meal decreased digoxin bioavailability. Guar gum has been shown to reduce serum digoxin levels in the early absorption period and also to reduce penicillin absorption (Huupponen et al., 1984), but to have no effect on glipizide absorption (Huupponen et al., 1985).

The effects of dietary fiber on vitamin metabolism in animals and humans have been reviewed recently (Kasper, 1986; Kelsay, 1982). Because of the effects of dietary fiber on lipid absorption, attention has focused on the fat-soluble vitamins.

In rats, Phillips and Brien (1970) found that 3% pectin in the diet had no effect on hepatic accumulation of vitamin A from ingested vitamin A or carotene. Cellulose (5 or 10% of the diet) did not alter vitamin A absorption in rats (Sina et al., 1976). In humans, Kasper et al. (1979) found mean postprandial serum vitamin A levels after test meals containing wheat bran, microcrystalline cellulose, pectin, guar, or carrageenan were equal to or greater than control values. Lignin also had no effect on postprandial serum vitamin A levels (Barnard and Heaton, 1973). However, Kelsay (1982) found greater fecal excretion of vitamin A activity with a diet of fresh fruits and vegetables than with a low-fiber diet containing added carotene. Rattan et al. (1981) found serum vitamin A levels to be the same in vegetarians and in persons consuming 2 tbs/d wheat bran for 6 mo as in control subjects.

High intake of fiber from unleavened whole-meal wheat bread is associated with the development of rickets in some population groups (Reinhold, 1976), suggesting that vitamin D metabolism may be impaired. Zoppi et al. (1982) noted that treatment of constipation in infants with wheat bran led to

147

blood levels of calcium, phosphorus, and alkaline phosphatase activity characteristic of vitamin D-dependent rickets.

Schaus et al. (1985) found that 6 and 8%, but not 3%, pectin decreased vitamin E bioavailability in rats. Graded doses of wheat bran (4, 8, 12, 16, or 20%), however, had no effect of vitamin E availability in rats (Mongeau et al., 1986)~

Studies in humans have shown that a variety of fiber sources (wheat bran, pectin, cellulose, and hemicellulose) have no significant effect on the absorption of the water-soluble vitamins, folate, ascorbic acid, vitamin 8-6, and riboflavin (Keltz et al., 1978; Leklem et al., 1980; Miller et al., 1980; Roe et al., 1978; Russell et al., 1976). ·

148

VII. ESTIMATES OF FIBER INTAKE IN THE UNITED STATES

Table 20a presents estimates of dietary fiber intake in the United States together with descriptions of the studies in which they were generated. Estimates of dietary fiber intake in other Western countries are also included in Table 20b for comparison.

A. DIETARY METHODOLOGIES USED TO ESTIMATE FIBER INTAKE

Most of the methodologies available for dietary assessment have been used in studies designed to estimate fiber intake. These include recalls of foods consumed in the preceding 24 hr or longer periods, daily dietary records of varying lengths (with a~d without weighing of food consumed), diet histories, food frequency methods, and per capita food disappearance data. The appropriate use of dietary intake data obtained with these methodologies has been reviewed (Anderson, 1986c) and some con­cerns regarding the use of these methods to assess fiber intake are considered briefly in this report (Section VI-A). In addi­tion, composite meals have been prepared for the analysis of dietary fiber content based on collection of duplicate portions, assessment of "usual diet" by surveys, institutional menus, or diets designed for metabolic studies.

Bo FOOD COMPOSITION DATABASES FOR DIETARY FIBER

Until recently, the only values for fiber avail-able in food composition databases were those obtained by the crude fiber method or the Southgate (1969) method for unavail­able carbohydrate. The Southgate fiber values in Mccance and Widdowson's food composition tables (Paul and Southgate, 1978) have been used most frequently for calculating dietary fiber intake. Crude fiber values seriously underestimate total dietary fiber content, but on daily intake or per capita basis, crude fiber values are significantly correlated (r = 0.84-0.92) with Southgate fiber values (Bright-See and McKeown-Eyssen, 1984; Kelsay and Clark, 1984; Marlett and Bokram, 1981). Southgate fiber values tend to overestimate total dietary fiber because (1) the colorimetric assays used for assessment of constitutent sugars are relatively nonspecific and subject to interference (Laine et al., 1981) and (2) starch contamina­tion can inflate values by 5 to 15% for foods containing starch (Marlett and Chesters, 1985).

Recently new databases have been compiled from fiber values derived by a variety of analytical methods. The between­laboratory variation using a single method of fiber analysis is known to be high (Theander, 1981) and the magnitude of the error

149

Table 20a. Estimates of dietary fiber intake in Western countries: United States.

Fiber Intake Reference Population g/day g/1000 kcal Intake Data Fiber Database Comments

Kahaner et al. 4 men, 2 women 3.3 1.7 3-wk records Calculated, crude fiber (1976) 22-47 yr

Ahrens & Boucher - 19.l 6.7 US diet homogenate Analyzed, Southgate Reflects food brought into (1978) 9.2 3.2 based on 1965- Analyzed, "commercial household

1966 USDA survey food laboratory" 3.5 1.2 Calculated, crude fiber

Brauer et al. - 7.4 NA*· Metabolic study Analyzed, NDF (1981) 21.6 NA diets Calculated, Southgate

Marlett & Bokram 57 men 19.9 8.3 2-d records Calculated, Southgate Dietary and crude fiber signifi-(1981) 143 women 13.4 10.2 cantly correlated, r = 0.86

~ 57 men 4.8 2.1 Calculated, crude fiber \.n D 143 women 3.8 2.9

Reddy et al. 44 adults, New 15.0 5.9 3-d dietary recall Calculated, Southgate (1983b) York City

40 adults, New 14.0 5.8 Data from Domellof et al. (1982) York City

Bright-See & food supply 5.9 NA 1972-1974 FAD food Calculated, crude fiber Assumed extraction ratio of wheat McKeown-Eyssen disappearence and maize that led to high fiber (1984) 27.4 NA data Calculated, Southgate intakes (Rutishauser, 1985)

Bright-See (1985) food_ supply 22.8 NA 1972-1974 FAQ food Calculated, Southgate Data of Bright-See & McKeown-Eyssen disappearance (1984) recalculated using differ-data ent fiber values for grain products

*NA= not available.

Table 20a. Estimates of dietary fiber intake in Western countries: United States (continued).

Fiber Intake Reference Population g/day g/1000 kcal Intake Data Fiber Database Comments

Kelsay & Clark 13 men 9.5 4.0 28-d food composite Analyzed, NDF NDF:CF, r = 0.84 (1984) 16 women 7.7 4.7 from each subject

13 men 4.2 1.7 28-d records Calculated, crude fiber 16 women 3.7 2.3

Anderson - NA* 7.5 Average US diet, Analyzed, Englyst modi-(1986d) USDA survey data fication of Southgate

NA 8.6 Metabolic study diet extraction technique NA 12.3 "Bean test diet" (lignin determined by NA 15.7 High-fiber ashing material insol-

·diabetic diet uble in 72% H2so

4)

NA 25.2 High-fiber, 70% CHO diet

I-' \Jl Johnson & Marlett - 10.8 4.8 14-d menu Analyzed, NDF Nursing home menus; CF:NDF, I-'

(1986) 4.3 2.1 28-d menu Calcluated, crude fiber r = 0.67; 7-d, 14-d, and 28-d 23.3 10.6 7-d menu Calculated, Southgate menus did not differ

Marlett & Johnson 7 adults 9.0 4.6 10-d records Analyzed, NDF (1986) >65 yr 18.8 9.6 Calculated, Southgate

6 older adults 7.4 4.5 Analyzed, NDF (institu- 18.5 11.8 Calculated, Southgate tionalized)

Lanza et al. 5509 men 12.9 5.5 24-hr recall Calculated, various Fiber values for 125 foods applied (1987) 6149 women 9.4 6.5 (NHANES II) databases to 2500 foods reported; fiber (g/d)

5509 men 15.7 NA Calculated, Southgate did not increase with age, but 6149 women 11.1 NA fiber (g/1000 kcal) did

Murphy & Calloway 996 women 13.2 8.0 24-hr recall Calculated, Southgate 235 foods in nutrient database (1986) 18-24 yr 2.5 1.5 (NHANES II) Calculated, crude fiber used for 1267 foods reported

*NA= not available.

Table 20b. Estimates of dietary fiber intake in Western countries: other countries.

Fiber Intake Reference Population g/day g/1000 kcal Intake Data Fiber Database Comments

Bingham et al. 63: 20-80 yr 19.9 9.1 7-d weighed record Calculated, Southgate From Cambridgeshire, England (1979b) 32 men 20.l 8.4

31 women 19.8 10.1

Jain et al. 542 colorectal Dietary history Calculated, crude fiber Data are means of intakes of (1980) cancer cases: colorectal disease patients

285 men 6.5 2.0 in Canada 257 women 5.6 2.4

controls: 285 men 6.8 2.3 Matched neighborhood controls 257 women 5.5 2.7

controls: 262 men 6.4 2.1 Hospitalized controls, non-

I--' 273 women 5.3 2.5 malignant GI surgery \Jl N

Kay et al. (1980) 200 men 17.8 6.9 24-hr recall Calculated, Southgate Faculty and staff, University of 35-59 yr Guelph

Anderson et al. 56 women 30.9 19.0 3-d records Calculated, Southgate Canadian Seventh-Day Adventists; (1981) 9 vegans, 47 lacto-ovo vegetarians

Bull et al. 43: 21-69 yr 4-wk dietary diary Calculated, Southgate From Isle of Westray, Scotland (1982) 16 men 16.5 6.1

27 women 15.0 7.2

Table 20b. Estimates of dietary fiber intake in Western countries: other countries (continued).

Fiber Intake Reference Population g/day g/1000 kcal Intake Data Fiber Database Comments

Englyst et al. 30 men, rural 18.4 6.9 1-d food composite Analyzed, Englyst (1982b) Finland method; does not

28 men, urban 14.5 7.3 include lignin Finland

25 men, rural 18.0 7.2 Denmark

29 men, urban 13.2 6.1 Denmark

rural Finland 21.0 6.7 4-d record Calculated, Southgate urban Finland 16.0 7.2 rural Denmark 19.9 7.3 urban Denmark 15.3 6.6

I-' Gibson et al. 38: 4.5-5.5 yr 3-d record by Calculated, Southgate Children from Guelph (Canada) \J1 vJ (1982) 20 boys 13.5 8.1 parent

18 girls 11.8 8.5 47: 4.5-5.5 yr 3-d record by Calculated, Southgate Children from Halifax (Canada)

23 boys 13.0 7.7 parent 24 girls 12.8 8.3

Gibson & Scythes 100 women 19.4 10. 7 3-d diet records Calculated, Southgate University community in Canada (1982) 30: 6 yr

Potter et al. 368 colon 20.8 9.0 Food frequency Calculated, Southgate Assumed standard serving sizes; (1982) cancer questionnaire, Australia

<50 yr 16.7 NA* self-administered >50 yr 23.l NA

732 controls 20.3 9.4 <50 yr 19.3 NA ~50 yr 20.0 NA

*NA= not available.

1--J \J1 +::-

Table 20b. Estimates of dietary fiber intake in Western countries: other countries (continued).

Fiber Intake Reference Population g/day g/1000 kcal Intake Data Fiber Database

Thomson et al. (1982)

Barasi et al. (1983)

Gibson et al. (1983)

Reddy et al. (1983b)

Bright-See & McKeown-Eyssen (1984)

97 men (40 yr) 17.5

102 women 15.l 18-75 yr 14.5

38 heal th food shoppers 25.7 31-78 yr 27.l

36 vegetarians 33.2 69 + 8 yr

30 controls 20.2 60 + 9 yr

23 from Malmo, 17.4 Sweden

21 from Umea, 26.0 Sweden

food supply of 4.4-16.l 38 countries

22.1-93.6

5.0

NA* NA

NA NA

20.6

11.7

7.5

10.3

NA

NA

Bright-See (1985) food supply of 13.9-45.1 NA 38 countries

*NA= not available.

7-d weighed record Calculated, Southgate

4-d weighed record Calculated, Southgate Questionnaire

4-d weighed record Questionnaire

3-d record Calculated, Southgate

3-d dietary recall Calculated, Southgate

1972-1974 FAD food Calculated, crude fiber disappearence data Calculated, Southgate

1972-1974 FAD food Calculated, Southgate disappearance data

Comments

From Edinburgh, Scotland

South Wales village; OF from weighed records vs. questionnaires: r = 0.51 (n = 102) and r = 0.64 (n = 38)

8 pure vegetarians, 28 ovo-lacto-vegetarians; post-menopausal women

Data from Domellof et al. (1982)

Correlation coefficient for crude & dietary fiber, r = 0.92; assumed extraction ratio of wheat and maize that led to high.fiber intakes (Rutishauser, 1985)

Data of Bright-See & McKeown-Eyssen (1984) re-calculated using differ­ent fiber values for grain products

Table 20b. Estimates of dietary fiber intake in Western countries: other countries (continued).

Fiber Intake Reference Population g/day g/1000 kcal Intake Data Fiber Database Comments

Gibson et al. 14: 21-26 yr 22.5 12.0 7-d weighed record Calculated, Southgate University students (1985a)

Gibson et al. 90 women 21.4 13.4 3-d record Calculated, Southgate Noninstitutionalized; recruited in (1985b) 58-89 yr part through university alumni

association

Health and Welfare 1-4 yr 9.1 5.5 Food intake, Calculated, Southgate Number of subjects not given Canada (1985) 5-11 yr 14.7 6.4 Nutrition Canada (some other values

12-19 yr men 19.0 5.8 also included) women 13.6 6.1

20-39 yr men 18.6 5.5 women 12.6 6.3

I-' 40-64 yr men 16.8 6.3 \J1 \J1 women 12.6 7.3

>65 yr men 15.1 7.3 women 12.2 8.0

all 14.6 6.3

Bingham (1986) vuu.K. diet" 23 7.5 Food available Calculated, Southgate for consumption

20 8.9 Household survey; 1-wk purchases

introduced by combining data from different sources is unknown. Indiscriminate collection of data obtained by a single method also raises concern. Small changes in the nutrient database can lead to large changes in the estimates of dietary fiber consumption (Bright-See, 1985; Bright-See and McKeown-Eyssen, 1984). The dietary fiber content of similar foods can vary considerably, as illustrated by Patraw and Marlett (1986) who found NDF values ranging from 0.5 to 2.2 g/slice for commer­cially available wheat and whole-wheat breads. Dietary fiber is obtained in small amounts from a wide variety of foods and_ many judgments about the equivalency of foods must be made when using databases with fiber values from a limited number of foods.

C. ESTIMATES OF INTAKE

Assessments of dietary fiber consumption of the United States are available only for adults. In studies of various population subgroups, estimates of crude fiber intake range from 2.5 to 4.8 g/d (and 1.5 to 2.9 g/1000 kcal), estimates of NDF intake range from 7.4 to 10.8 g/d (and 4.0 to 4.8 g/1000 kcal), and estimates of Southgate fiber intake range from 11.1 to 23.3 g/d (and 5.8 to 11.8 g/1000 kcal) (see Table 20a). These estimates for intake in the United States are similar to those obtained in other Western countries (Table 20b). The data sug­gest that men have a higher absolute (daily) fiber intake than women. However, when intake is expressed per kcal consumed, women have higher values. The limited data available for the elderly suggest that this group has a higher intake of dietary fiber/kcal than younger adults.

D. GROUPS AT RISK OF INADEQUATE OR EXCESSIVE INTAKE OF DIETARY FIBER

Because of the limitations of the available data on dietary fiber intake, assessment of the groups at risk of inadequate or excessive intake must be considered speculative at this time. Very young children, the elderly, and other groups consuming high-bulk, low-calorie diets may be at risk of excessive intake. Adolescents may select diets inadequate in dietary fiber. Much additional research is needed to con­firm or refute these possibilities.

E. ADDITIONAL RESEARCH NEEDED

Research needs in the assessment of dietary fiber intake include:

• determination of the accuracy and precision of measures of dietary fiber content in foods

156

• accumulation of validated data for the dietary fiber content of a large number of foods

evaluation of the effects of judgments concerning the equivalency of the fiber contents of foods

additional validation of methods for collection of dietary intake data

, compilation of fiber intake data for various age, sex, and ethnic groups

• compilation of data on dietary fiber intake in various disease states.

157

VIII. RECOMMENDATIONS FOR FIBER INTAKE IN THE UNITED STATES

A. INTRODUCTION

In making recommendations on dietary fiber intake for the U.S. population, the Expert Panel has considered several important related issues in order to ensure these recommendations are not misunderstood or misused. Recommendations of the Expert Panel can be interpreted properly only with consideration of the following points:

1.

2.

3.

4 ..

Intended audience: The Panel's recommended intakes for dietary fiber are directed to the general, healthy, adult population of the United States. The elderly and very young populations, or those on special diets, could conceivably be affected adversely by levels of dietary fiber intake well-tolerated by others. The recommen­dations are not intended to address specific involvement of dietary fiber in the prevention of diseases or in the treatment or control of existing diseases. The Expert Panel considered that infor­mation concerning use of dietary fiber sources in specific disease states (as noted in the Executive Summary) is not sufficiently complete to make quantitative recommendations at this time.

Identity of dietary fiber: · Dietary fiber is a complex group of materials found in a wide variety of grains, fruits, and vegetables in the human diet The various food sources of dietary fiber differ in composition and, as a result, often have different physiological effects.

Information on current intake: Information on current levels of dietary fiber intake in the U.S. population is limited. Data have been derived using a variety of measurement methods (diet records, food disappearance, experimental diets, etc.) and are based on a variety of ana­lytical methods. These methods may yield dietary fiber values for a single food that differ by 100%. Thus, estimates of intake are uncertain. This uncertainty does not prevent the Expert Panel from making recommendations; however, it does suggest that, as more and better information becomes available, the precision of recommendations and confidence in them will improve.

Knowledge of effects of dietary fiber: In compar­ison to the larger field of nutrition, knowledge is limited concerning the role of dietary fiber

159

5.

in the human diet. Considering the complexity of the involvement of dietary fiber in various aspects of nutrition, any current recommendations must be derived from a small base of information. As a result, these and other recommendations are likely to require modification as more information becomes available, especially information for various age groups and regarding more specific recommendations for therapeutic uses of dietary fiber. The Expert Panel does not intend that these recommendations be taken as the final word on the issue, but rather as a statement of expert scientific opinion regarding adequate levels of dietary fiber in the diet of healthy adults and a basis for continued improve­ment in the understanding of the role of dietary fiber in nutrition.

Dietary fiber is food: The Expert Panel's recom­mendations are designed to be carried out through additions and substitutions of foods in the diet. They are not meant to be met by addition of dietary fiber supplements to the diet or by the addition of isolated fiber sources or components to foods. Although supplementation with the polysaccharide components of dietary fiber may be useful in spe­cific therapeutic situations, it is not the Panel 1 s intent to recommend the use of fiber supplements to increase dietary fiber intake in the general popu­lation. Uses of specific sources of dietary fiber to treat specific disorders (such as the use of insoluble fiber in the treatment of constipation and diverticular disease and the use of soluble fiber in diabetes and hyperlipidemia) are outlined in individual sections of this report. Such specific therapeutic effects should not form the basis for_ decisions concerning recommendations of level or type of dietary fiber for the general population. The Expert Panel has concluded that dietary fiber is an integral part of a healthy diet; however, the available evidence is not sufficient to support specific, quantitative recommendations on the role of dietary fiber for the prevention of specific diseases in the general, healthy population.

B. FOOD SOURCES OF DIETARY FIBER

As a result of the diversity of effects of various com­ponents of dietary fiber, it is important that the dietary fiber consumed come from a wide variety of foods which provide a broad spectrum of these components. The members of the Expert Panel concurred that dietary fiber should be obtained from a variety of whole-grain products, fruits, and vegetables, including legumes.

160

The changes recommended to achieve such a diet in the general population include increased consumption of plant foods and/or substitution of plant foods higher in dietary fiber. For exam­ple, whole-grain products (breads, cereals) could be substituted for refined flour products, whole fruits could be substituted for fruit juices, and greater use could be made of plant sources of proteins (beans vs. meats). The Panel is aware that consuming several servings of fruits and vegetables daily and using whole­grain products may represent a considerable change in diet for many persons in the U.S. population whose intake of these foods is currently low. The emphasis in applying these recommendations should be on inclusion of a wide variety of foods high in dietary fiber in the diet.

Based on their investigational experience with a variety of self=selected and experimental diets, the Panel estimated that the dietary fiber in the recommended diet would comprise approx­imately 70-75% insoluble fibers and 25-30% soluble fibers. Whether this ratio is ideal is not known and may vary from indi­vidual to individual, but it is approximately the ratio found in a diet containing a wide variety of foods. Whether an elevation in the intake of soluble fiber may exert detrimental effects is also unknown. Results of some, but not all, animal studies have indicated that feeding large amounts of some purified, soluble fibers such as pectin and guar gum can increase the development of chemically-induced colon neoplasms. However, evidence has not suggested that increased intake of soluble fiber, accompanied by a proportionate increase in insoluble fiber, from commonly avail­able foods has detrimental effects in humans aside from increased production of intestinal gas. On the basis of the scientific evidence, it is not yet possible to state whether health risks would be involved if soluble fibers were to constitute a larger proportion of the total intake.

Changes targeted to dietary fiber must also be considered in light of other nutrients in the overall diet. Changes in dietary fiber consumption will probably result in other dietary changes, including fat level, protein source, complex carbohydrate content, and caloric density. From this perspective, the Expert Panel's recommendations are similar to those made by other groups. Effects of changes in dietary fiber consumption cannot be considered apart from those of other diet components. The intent of the Panel is to encourage dietary fiber consumption as part of a balanced diet. Employing supple­ments of isolated polysaccharides to achieve an increase in dietary fiber intake in the general population should be dis­couraged. Effects of these polysaccharides in isolated form are likely to be different than their effects in native foods. This concept should be extended to foods with added isolated poly­saccharides for similar reasons. Consumers should be educated and encouraged to use dietary fiber-containing foods as they occur naturally to achieve the desired dietary fiber intake.

161

C. RECOMMENDED LEVEL OF INTAKE

The Expert Panel concluded that it is not possible tq recommend a single optimal level of dietary fiber for the general U.S. population at this time. The inability to assess accurately the current intake of the population in question is a factor but, by itself, does not prevent the formulation of recommendations.

The larger problem lies in selecting an appropriate health-related outcome to which the dietary fiber intake recom­mendations can be tied. The Panel concluded that in making rec­ommendations for the general population, the health outcome to be considered should be related to normal physiological function rather than to a disease state or adverse health condition. Ideally, data on the effect of dietary fiber on the chosen outcome to be used as the basis for recommendations should be derived from studies of subjects of the same age and sex as the population to which the recommendations will be applied, fed the types of fiber-containing foods recommended at various levels. Such investigations are analogous to those conducted to determine nutrient requirements. Such studies have not been conducted for dietary fiber and should be performed in the immediate future.

The Expert Panel concluded that information now available suggests that adequate intake might be estimated by assessing large intestinal function, the most obvious and best documented physiological effect of dietary fiber intake. Intestinal transit time and fecal weight are among the primary variables used to determine intestinal function. Spiller et al. (1977) have suggested that the relationship between fecal weight and transit time may be useful in determining optimal dietary fiber intake. Transit time decreases with increasing fecal weight to a point at which further increases in fecal weight do not correspond with a reduction in transit time. Although what constitutes ideal transit time for humans is not known, Spiller (1986) notes that it is reasonable to assume that extremely long transit times are not desirable_ and that a more predictable colon function is preferable. There are few published studies which examine the effects of fiber on intestinal function in a dose­response fashion and even fewer which employ foods rather than isolated fiber sources. Two such studies indicate that an intake of approximately 32 g/d total dietary fiber (from wheat bran in bread and other foods) (Spiller et al, 1986) or approximately 19 g/d NDF (from fruits and vegetables) (Kelsay et al., 1981) resulted in fecal outputs in a range which the data suggest would produce reasonable transit time. These studies and others con­sidered by the Panel (including Balasubramanian et al., 1987; Eastwood et al., 1986; Slavin et al., 1985; Spiller et al., 1982; Wyman et al., 1976) suggest that consumption of diets containing at least 20 g/d of dietary fiber frequently results in fecal weights in a range not associated with a significant further reduction in transit time. Most of the subjects in these studies were healthy young or middle-aged adults.

162

Based on the data described above and the Panel 1 s conclusions on the potential beneficial and adverse consequences of dietary fiber consumption (see Executive Summary), as well as the collective experience and judgment of the members of the Expert Panel, a majority agreed on a recommended intake range of 20-35 g/d total dietary fiber from foods for the healthy, adult population of the United States. The minimum value was selected with special consideration of the effects of dietary fiber on intestinal function, and the maximum value was chosen with consideration of the possible deleterious effects of large amounts of dietary fiber on mineral balance. This range was also selected with consideration for the variability in the analytical methods available for quantifying dietary fiber. The Panel emphasized that this range of intakes is appropriate for normal adults, in whom the recommended levels will yield a daily intake of approximately 10-13 g dietary fiber/1000 kcal.

The recommended range may not be applicable to special groups, such as children, the elderly, or persons consuming special diets. In addition, the range of intakes and the types of dietary fiber sources recommended are not intended to serve as guidelines for the prevention, alleviation, or treatment of specific diseases, such as diabetes, coronary heart disease, or cancer. The amounts of dietary fiber needed in relation to effects on carbohydrate and lipid metabolism and other functions have not been established and should not be suggested for the general population at this time.

There was agreement that any specified values for dietary fiber intake are meaningless outside the context of a diet containing whole grains, fruits, and vegetables, as recom­mended by the Panel. However, a minority of the Panel concluded that, because of the limitations in knowledge about dietary fiber, no numerical recommendations should be offered at this time. They concluded that the available evidence is not suffi­cient to support recommendations of an optimal level or range of dietary fiber intake for the general population, but did agree that eating a diet with a greater variety and quantity of whole fruits, vegetables, and grain products is likely to be beneficial to the health of the general population. Such a diet would result in the intake of amounts of fiber within the recommended range, but such fiber levels should be regarded as a natural consequence of eating a healthy diet rather than as a primary goal of increasing the quantity of dietary fiber in the diet.

The Panel concluded that achieving the recommended level of dietary fiber intake is feasible by selection of ordinary foods that are currently available. For persons whose consump­tion of whole grain products, fruits, and vegetables, including legumes, is very low, a change in diet will be necessary to meet the recommendations. The experience of some Panel members indi­cates that many healthy adults can and do consume levels of dietary fiber within the recommended range.

163

IX. LITERATURE CITED

AACC Committee on Dietary Fiber. 1981. Collaborative study of an analytical method for insoluble dietary fiber in cereals. Cereal Foods World 26:295-297.

Abraham, R.; Barbolt, T.A.; Rodgers, J.B. 1980. Inhibition by bran of the colonic cocarcinogenicity of bile salts in rats given dimethylhydrazine. Exp. Mal. Pathol. 33:133-143.

Adlercreutz, H.; Fotsis, T.; Heikkinen, R.; Dwyer, J.T.; Woods, M.; Goldin, B.R.; Gorbach, S.L. 1982. Excretion of the lignans enterolactone and enterodiol and of equal in omnivorous and vegetarian postmenopausal women and in women with breast cancer. Lancet 2:1295-1299.

Ahlberg, J.; Angelin, B.; Einarsson, K.; Hellstrom, K.; Leijd, B. 1979. Prevalence of gallbladder disease in hyperlipoproteinemia. Dig. Dis. Sci. 24:459-464.

Ahrens, E.H., Jr.; Boucher, C.A. 1978. The composition of a simulated American diet. J. Am. Diet. Assoc. 73:613-620.

Albrink, M.J.; Newman, T.; Davidson, P.C. 1979. Effect of high­and low-fiber diets on plasma lipids and insulin. Am. J. Clin. Nutr. 32:1486-1491.

Albrink, M.J.; Ullrich, I.H. 1982. Effect of dietary fiber on lipids and glucose tolerance of healthy young men. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber in health and disease. New York: Plenum Press. p.169-181.

Albrink, M.J.; Ullrich, I.H. 1986. Interaction of dietary sucrose and fiber on serum lipids in healthy young men fed high carbohydrate diets. Am. J. Clin. Nutr. 43:419-428.

American Diabetes Association. 1979. Principles of nutrition and dietary recommendations for individuals with diabetes mellitus: 1979. Diabetes Care 2:520-523.

Anand, B.K. 1974. Neurological mechanisms regulating appetite. In: Burland, W.L.; Samuel, P.O.; Yudkin, J., eds. Obesity symposium. New York: Churchill Livingstone. p.116-145.

Anderson, B.M.; Gibson, R.S.; Sabry, J.H. 1981. The iron and zinc status of long-term vegetarian women. Am. J. Clin. Nutr. 34:1042-1048.

Anderson, J.W. 1980. The role of dietary carbohydrate and fiber in the control of diabetes. Adv. Intern. Med. 26:67-95.

Anderson, J.W. 1983. Plant fiber and blood pressure. Ann. Intern. Med. 98:842-846.

165

Anderson, J.W. 1986a. Dietary fiber in nutrition management of diabetes. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.343-360.

Anderson, J.W. 1986b. High-fiber, hypocaloric vs. very-low­calorie diet effects on blood pressure of obese men. Am. J. Clin. Nutr. 43:695 (Abstract). ""'

Anderson, S. A., editor. 1986c. ..1Guidelines for the use of dietary intake data. Prepared for the Food and Drug Administra­tion under Contract FDA No. 223-84-2059 by the Life Sciences Research Office, Federation of American Societies for Experi­mental Biology. 89p. Available from: FASEB Special Publica­tions Office, Bethesda, MD.

Anderson, J.W. 1986d. VA Medical Center, Lexington, KY. Letter, dated October 7, to S.M. Pilch, Federation of American Societies for Experimental Biology, Bethesda, MD.

Anderson, J.W.; Bryant, C.A. 1986. Dietary fiber: diabetes and obesity. Am. J. Gastroenterol. 81:898-906.

Anderson, J.W.; Chen, W.-J.L. 1979. Plant fiber: carbohydrate and lipid metabolism. Am. J. Clin. Nutr. 32:346-363.

Anderson, J.W.; Chen, W.-J.L.; Sieling, B. 1980a. Hypolipidemic effects of high-carbohydrate, high-fiber diets. Metab. Clin. Exp. 29:551-558.

Anderson, J.W.; Ferguson, S.K.; Karounos, D.; □ 'Malley, L.; Sieling, B.; Chen, W.-J.L. 1980b. Mineral and vitamin status on high-fiber diets: long-term studies of diabetic patients. Diabetes Care 3:38-40.

Anderson, J.W.; Sieling, B. 1980. High fiber diets for obese diabetic patients. Obesity/Bariatric Med. 9:109-113.

Anderson, J.W.; Story, L.; Sieling, B.; Chen, W.-J.L.; Petro, M.S.; Story, J. 1984a. Hypocholesterolemic effects of oat-bran or bean intake for hypercholesterolemic men. Am. J. Clin. Nutr. 40:1146-1155.

Anderson, J.W.; Story, L.; Sieling, B.; Chen, W.-J.L. 1984b. Hypocholesterolemic effects of high-fibre diets rich in water­soluble plant fibres. J. Can. Diet. Assoc. 45:140-148.

Anderson, J.W.; Tietyen-Clark, J. 1986. Dietary fiber: hyperlipidemia, hypertension, and coronary heart disease. Am. J. Gastroenterol. 81:907-919.

Anderson, J.W.; Ward, K. 1978. Long-term effects of high­carbohydrate, ·high-fiber diets on glucose and lipid metabolism: a preliminary report on patients with diabetes. Diabetes Care 1:77-82.

166 _

Anderson, J.W.; Ward, K. 1979. High-carbohydrate, high-fiber diets for insulin-treated men with diabetes mellitus. Am. J. Clin. Nutr. 32:2312-2321.

Andersson, H.; Bosaeus, I.; Falkheden,~T.; Melkersson, M. 1979. Transit time in constipated geriatric patients during treatment with a bulk laxative and bran: a comparison. Scand. J. Gastro­enterol. 14:821-826.

Andersson, H.; Navert, B.; Bingham, S.A.; Englyst, H.N.; Cummings, J.H. 1983. The effects of breads containing similar amounts of phytate but different amounts of wheat bran on calcium, zinc and iron balance in man. Br. J. Nutr. 50:503-510.

Antonis, A.; Bersohn, I. 1962. The influence of diet on fecal lipids in South African white and Bantu prisoners. Am. J. Clin. Nutr. 11:142-155.

Arffmann, S.; H0jgaard, L.; Giese, Bo; Krag, E. 1983a. Effect of oat bran on lithogenic index of bile and bile acid metabolism. Digestion 28:197-200.

Arffmann, S.; Andersen, J.R.; Hegnh0j, J.; Schaffalitzky de Muckadell, O.B.; Morgensen, N.B.; Krag, E. 1983b. Irritable bowel syndrome treated with wheat bran: a controlled double blind trial. Scand. J. Gastroenterol. 18(Suppl. 86):3 (Abstract).

Armstrong, B.; Doll, R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int. J. Cancer 15:617-631.

Armstrong, B.; Van Merwyk, A.J.; Coates, H 1977 Blood pressure in Seventh-Day Adventist vegetarians. Amo Jo Epidemiol. 105:444-449.

Aro, A.; Uusitupa, M.; Voutilainen, E.; Hersio, K.; Korhonen, T.; Siitonen, O. 1981. Improved diabetic control and hypocholester­olaemic effect induced by long-term dietary supplementation with guar gum in type 2 (insulin-independent) diabetes. Diabetologia 21:29-33.

Aro, A.; Uusitupa, M.; Voutilainen, E.; Korhonen, T. 1984. Effects of guar gum in male subjects with hypercholesterolemia. Am. J. Clin. Nutr. 39:911-916.

Arrigoni, Ee; Caprez, A.; Amado, R.; Neukom, H. 1984. Gravi­metric method for the determination of insoluble and soluble dietary fibres. Z. Lebensm. Unters. Forsch. 178:195-198.

Asp, Ne~Ge; Johansson, C.-Ge 1984. Dietary fibre analysis. Nutr. Abstr. Rev. Clin. Nutr. 54:735-752.

167

Asp, N.-G.; Johansson, C.-G.; Hallmer, H.; Siljestrom, M. 1983. Rapid enzymatic assay of insoluble and soluble dietary fiber. J. Agric. Food Chem. 31:476-482.

Association of Official Analytical Chemists. 1984. Official methods of analysis of the Association of Official Analytical Chemists. 14th ed. Arlington, VA: Association of Official Analytical Chemists, Inc. 114lp.

Atkinson, R.L.; Dahms, W.T.; Bray, G.A.; Lemmi, C.,; Schwartz, A.A. 1979. Gastrin secretion after weight loss by dieting and intestinal bypass surgery. Gastroenterology 77:696-699.

Austad, W.I. 1979. Diverti~ular disease of the colon. N. Z. Med. J. 90:160-162.

Baird, I.M.; Walters, R.L.; Davies, P.Se; Hill, M.J.; Drasar, B.S.; Southgate, D.A.T. 1977. The effects of two dietary fiber supplements on gastrointestinal transit, stool weight and frequency, and bacterial flora, and fecal bile acids in normal subjects. Metab. Clin. Exp. 26:117-128.

Balasubramanian, R.; Johnson, E~J.; Marlett, J.A. 1987. Effect of wheat bran on bowel function and fecal calcium in older adults. J. Am. Coll. Nutr. 6:199-208.

Ballantyne, G.H. 1982. Review of sigmoid volvulus: clinical patterns and pathogenesis. Dis. Colon Rectum 25:823-830.

Barasi, M.E.; Burr, M.L.; Sweetnam, P.M. 1983. A comparison of dietary fibre intakes in South Wales estimated from a questionnaire and weighed dietary records. Nutr. Res. 3:249-255~

Barbolt, T.A.; Abraham, R. 1978. The effect of bran on dimeth­ylhydrazine-induced colon carcinogenesis in the rat. Proc. Soc. Exp. Biol. Med. 157:656-659.

Barbolt, T.A.; Abraham, R. 1980. Dose-response, sex difference, and the effect of bran in dimethylhydrazine-induced intestinal tumorigenesis in rats. Toxicol. Appl. Pharmacol. 55:417-422.

Barnard, D.L.; Heaton, K.W. 1973. Bile acids and vitamin A absorption in man: the effects of two bile acid-binding agents, cholestyramine and lignin. Gut 14:316-318.

Barnard, R.J.; Lattimore, L.; Holly, R.G.; Cherny, S.; Pritikin, N. 1982. Response of non-insulin-dependent diabetic patients to an intensive program of diet and exercise. Diabetes Care 5:370-374.

Barnard, R.J.; Massey, M.R.; Cherny, S.; O'Brien, L.T.; Pritikin, N. 1983. Long-term use of a high-complex-carbohydrate, high­fiber, low-fat diet and exercise in the treatment of NIDDM patients. Diabetes Care 6:268-273.

168

Barnes, D.S.; Clapp, N.K.; Scott, D.A.; Oberst, D.L.; Berry, S.G. 1983. Effects of wheat, rice, corn, and soybean bran on 1,2-dimethylhydrazine-induced large bowel tumorigenesis in F344 rats. Nutr. Cancer 5:1-9.

Bauer, H.G.; Asp, N.-G.; Dahlqvist, A.; Fredlund, P.E.; Nyman, M.; Oste, R. 1981. Effect of two kinds of pectin and guar gum on 1,2-dimethylhydrazine initiation of colon tumors and on fecal B-glucuronidase activity in the rat. Cancer Res. 41:2518-2523.

. . Bauer, H.G.; Asp, N.-G.; Oste, R.; Dahlqvist, A.; Fredlund, P.E. 1979. Effect of dietary fiber on the induction of colorectal tumors and fecal B-glucuronidase activity in the rat. Cancer Res. 39:3752-3756.

Beck, B.; Villaume, C.; Bau, H.M.; Gariot, P.; Chayvialle, J.A.; Desalme, A.; Debry, G. 1986. Long-term influence of a wheat­bran supplemented diet on secretion of gastrointestinal hormones and on nutrient absorption in healthy man. Hum. Nutr. Clin. Nutr. 40C:25-33.

Behal!, K.M.; Lee, K.H.; Moser, P.B. 1984. Blood lipids and lipoproteins in adult men fed four refined fibers. Am. J. Clin. Nutr. 39:209-214.

Behall, K.M.; Scholfield, D.J.; Lee, K.; Powell, A.S.; Moser, P.B. [1987]. Effect of four refined fibers on apparent mineral balance in adult men. Am. J. Clin. Nutr. In press.

Bergman, F.; van der Linden, W. gallstone formation in hamsters.

1975. Effect of dietary fibre on Z. Ernahrungswiss. 14:218-224.

Berman, J.I.; Schultz, M.J. 1980. Bulk laxative ileus. J. Am. Geriatr. Soc. 28:224-226.

Beynen, A.C.; Buechler, K.F.; Van der Molen, A.J.; Geelen, M.J.H. 1982. The effects of lactate and acetate on fatty acid and cholesterol biosynthesis by isolated rat hepatocytes. Int. J. Biochem. 14:165-169.

Bingham, S.; Williams, D.R.R.; Cole, T.J.; James, W.P.T. 1979a. Dietary fibre and regional large-bowel cancer mortality in Britain. Br. J. Cancer 40:456-463.

Bingham, S.; Cummings, J.H.; McNeil, N.I. 1979b. Intakes and sources of dietary fiber in the British population. Am. J. Clin. Nutr. 32:1313-1319.

Bingham, S.A. 1986. Epidemiology of dietary fiber and cola­rectal cancer: current· status of the hypothesis. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.523-542.

169

Bingham, S.A.; Williams, □ .R.R.; Cummings, J.H. 1985. Dietary fibre consumption in Britain: new estimates and their relation to large bowel cancer mortality. Br. J. Cancer 52:399-402.

Bird, R.P.; Stamp, D. 1986. Effect of a high fat diet on the proliferative indices of murine colinic epithelium. Cancer Lett. 31:61-67.

Birkner, H.J.; Kern, F., Jr. 1974; In vitro adsorption of bile salts to food residues, salicylazosulfapyridine, and hemicellu­lose. Gastroenterology 67:237-244.

Bjelke, E. 1974a. Case-control study in Norway. Scand. J. Gastroenterol. 9(Suppl. 31):42-48.

Bjelke, E. 1974b. Case-control study in Minnesota. Scand. J. Gastroenterol. 9(Suppl. 31):49-53.

Bjorn-Rasmussen, E. 1974. Iron absorption from wheat bread: influence of various amounts of bran. Nutr. Metab. 16:101-110 . . Blackburn, N.A.; Holgate, A.M.; Read, N.W. 1984a. Does guar gum improve post-prandial hyperglycaemia in humans by reducing small intestinal contact area? Br. J. Nutr. 52:197-204.

Blackburn, N.A.; Redfern, J.S.; Jarjis, H.; Holgate, A.M.; Hanning, I.; Scarpello, J.H.B.; Johnson, I.T.; Read, N.W. 1984b. The mechanism of action of guar gum in improving glucose toler­ance in man. Clin. Sci. 66:329-336.

Bolton, R.P.; Heaton, K.W.; Burroughs, L.F. 1981. The role of dietary fiber in satiety, glucose, and insulin: studies with fruit and fruit juice. Am. J. Clin. Nutr. 34:211-217.

Bond, J.H.; Levitt, M.O. 1978. Effect of dietary fiber on intestinal gas production and small bowel transit time in man. Am. J. Clin. Nutr. 31:Sl69-Sl74.

Bosello, O.; Cominacini, L.; Zocca, I.; Garbin, U.; Ferrari, F.; Davoli, A. 1984. Effects of guar gum on plasma lipoproteins and apolipoproteins C-II and C-III in patients affected by familial combined hyperlipoproteinemia. Am. J. Clin. Nutr. 40:1165-1174.

Bosello, O.; Ostuzzi, R.; Armellini, F.; Micciolo, R.; Scuro, L.A. 1980. Glucose tolerance and blood lipids in bran-fed patients with impaired glucose tolerance. Diabetes Care 3:46-49.

Botha, A.P.J.; Steyn, A.F.; Esterhuysen, A.J.; Slabbert, M. 1981. Glycosylated haemoglobin, blood glucose and serum cholesterol levels in diabetics treated with guar gum. S. Afr. Med. J. 59:333-334.

170

Brandes, J.-W.; Lorenz-Meyer, H. 1981. Zuckerfreie Diat: eine neue Perspektive zur Behandlung des Marbus Crohn: eine randomisierte, kontrollierte Studie. Z. Gastroenterol. 19:1-12.

Brauer, P.M.; Slavin; J.L.; Marlett, J.A. 1981. Apparent digestibility of neutral detergent fiber in elderly and young adults. Am. J. Clin. Nutr. 34:1061-1070.

Bray, G.A.; Dahms, W.T.; Atkinson, R.L.; Rodin, J.; Taylor, I.; Frame, C.; Schwartz, A. 1979. Metabolic and behavioral differences between dieting and intestinal bypass. Hormo Metab. Res. 11:648-654.

Bright-See, E. 1985. Reply to letter by Rutishauser. Am. J. Clin. Nutre 41:825-826.

Bright-See, E.; McKeown-Eyssen, G.E. 1984. Estimation of per capita crude and dietary fiber supply in 38 countries. Am. J. Clin. Nutr. 39:821-829.

Bright-See, E.; McKeown-Eyssen, G.; Jacobson, E.A.; Newmark, H.W.; Mathews, R.; Morson, L.; Jazmaji, V.; Bruce, W.R. 1985. Dietary fiber and cancer: a supplement for intervention studies. Nutr. Cancer 7:211-220.

Bristol, J.B.; Emmett, P.Me; Heaton, K.W.; Williamson, R.C.N. 1985. Sugar, fat, and the risk of colorectal cancere Br. Med. J. 291:1467-1470.

Brocklehurst, J.C. 1985. Colonic disease in the elderly. Clin. Gastroenterol. 14:725-747.

Brodribb, A.J.M 1979 The treatment of diverticular disease with dietary fibre. In: Heaton, K.W., ede Dietary fibre: current developments of importance to health. Westport, CT: Technomic Publishing Company, Inc. p.63-73.

Brodribb, AoJoM. 1980. Dietary fiber in diverticular disease of the colon. In: Spiller, G.A.; Kay, RoM., eds. Medical aspects of dietary fiber. New York: Plenum Medical Book Company. p.43-66.

Brodribb, A.J.M.; Groves, C. 1978. Effect of bran particle size on stool weight. Gut 19:60-63.

Brooks, P.M.; Bremner, W.F.; Third 1 J.H.L.C. 1976. Bran, hypertriglyceridaemia and urate clearance. Med. J. Aust. 2:753, 756.

Brown, D.D.; Juhl, R.P.; Warner, S.L. 1978. Decreased bio­availability of digoxin due to hypocholesterolemic interventions. Circulation 58:164-172.

171

Brown, R.C.; Kelleher, J.; Losowsky, M.S. 1979. The effect of pectin on the structure and function of the rat small intestine. Br. J. Nutr. 42:357-365.

Bruce, W.R.; Varghese, A.J.; Furrer, R.; Land, P.C. 1977. A mutagen in the feces of normal humans. In: Hiatt, H.H.; Watson, J.D.; Winsten, J.A., eds. Origins of human cancer. Book C. Human risk assessment. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. p.1641-1646.

Bryant, M.P. 1978. Cellulose digesting bacteria from human feces. Am. J. Clin. Nutr. 31:Sll3-Sll5.

Bryson, E.; Dore, C.; Garrow, J.S. 1980. Wholemeal bread and satiety. J. Hum. Nutr. 34:113-116.

Bull, N.L.; Smart, G.A.; Judson, H. 1982. Food and nutrient intakes on Westray in the Orkney Islands. Ecol. Food Nutr. 12:97-101.

Burkitt, D. 1985. Varicose veins, haemorrhoids, deep-vein thrombosis and pelvic phleboliths. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.317-329.

Burkitt, D.P. 1971. Epidemiology of cancer of the colon and rectum. Cancer 28:3-13.

Burkitt, D.P. 1973a. Epidemiology of large bowel disease: the role of fibre. Proc. Nutr. Soc. 32:145-149.

Burkitt, D.P. 1973b. Some diseases characteristic of modern Western civilization. Br. Med. J. 1:274-278.

Burkitt, D.P. 1976. Economic development -- not all bonus. Nutr. Today 11:6-13.

Burkitt, D.P.; Walker, A.R.P.; Painter, N.S. 1972. Effect of dietary fibre on stools and transit-times, and its role in the causation of disease. Lancet 2:1408-1412.

Burkitt, D.P.; Walker, A.R.P.; Painter, N.S. 1974. Dietary fiber and disease. J. Am. Med. Assoc. 229:1068-1074.

Burslem, J.; Schonfeld, G.; Howald, M.A.; Weidman, s.w.; Miller, J.P. 1978. Plasma apoprotein and lipoprotein lipid levels in vegetarians. Metabolism 27:711-719.

Burton, R.; Manninen, V. 1982. Influence of a psyllium-based fibre preparation on faecal and serum parameters. Acta Med. Scand. 668(Suppl.):9l-94.

172

Byers, T. 1986. Fiber in relation to cancer. Presentation to the National Research Council Committee on Diet and Health, Working Group IV, July 11, Seattle.

Byers, T.; Funch, D. 1984. Towards the dietary prevention of cancer: contributions of epidemiology. Cancer Detect. Prev. 7:135-146.

Byers, T.; Graham, S. 1984. The epidemiology of diet and cancer. Adv. Cancer Res. 41:1-69.

Byers, T.E.; Graham, S.; Haughey, B.P.; Marshall, J.R.; Swanson, M.K. 1987. Diet and lung cancer risk: findings from the Western New York diet study. Am. J. Epidemiol. 125:351-363.

Calvert, R.; Schneeman, B.O.; Satchithanandam, S.; Cassidy, M.M.; Vahouny, G.V. 1985. Dietary fiber and intestinal adaptation: effects on intestinal and pancreatic digestive enzyme activities. Am. J. Clin. Nutr. 41:1249-1256.

Cann, P.A.; Read, N.W.; Holdsworth, C.D. 1984. What is the benefit of coarse wheat bran in patients with irritable bowel syndrome? Gut 25:168-173.

Capani, F.; Consoli, A.; Del Ponte, A.; Lalli, G.; Sensi, S. 1980. A new dietary fibre for use in diabetes. IRCS Med. Sci. 6:661.

Carroll, D.G.; Dykes, V.; Hodgson, W. 1981. Guar gum is not a panacea in diabetes management. N. Z. Med. J. 93:292-294.

Carryer, P.W.; Brown, M.L.; Malagelada, J.-R.; Carlson, G.L.; McCall, J.T. 1982. Quantification of the fate of dietary fiber in humans by a newly developed radiolabeled fiber marker. Gastroenterology 82:1389-1394.

Cassidy, M.M.; Lightfoot, F.G.; Grau, L.E.; Story, J.A.; Kritchevsky, D.; Vahouny, G.V. 1981. Effect of chronic intake of dietary fibers on the ultrastructural topography of rat jejunum and colon: a scanning electron microscopy study. Am. J. Clin. Nutr. 34:218-228.

Chacko, C.J.G.; Paulson, K.A.; Mathan, V.I.; Baker, S.J. 1969. The villus architecture of the small intestine in the tropics: a necropsy study. J. Pathole 98:146-151.

Chen, W.-F.; Patchefsky, A.S.; Goldsmith, H.S. 1978. Colonic protection from dimethylhydrazine by a high fiber diet. Surg. Gynecol. Obstet. 147:503-506.

Chen, W.J.L.; Anderson, J.W. 1986. Hypocholesterolemic effects of soluble fibers. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.275-286.

173

Chen, W.-J.L.; Anderson, J.W.; Jennings, D. 1984. Propionate may mediate the hypocholesterolemic effects of certain soluble plant fibers in cholesterol-fed rats. Proc. Soc. Exp. Biol. Med. 175:215-218.

Choudhury, S.; Jackson, P.; Katan, M.B.; Marenah, C.B.; Cortese, C.; Miller, N.E.; Lewis, B. 1984. A multifactorial diet in the management of hyperlipidaemia. Atherosclerosis 50:93-103.

Clark, A.N.G.; Scott, J.F. 1976. Wheat bran in dyschezia in the aged. Age Ageing 5:149-154.

Cleave, T.L. 1956. The neglect of natural principles in current medical practice. J. R. Nav. Med. Serv. 42:55-83.

Cohen, A.M.; Bavly, S.; Poznanski, R. 1961. Change of diet of Yemenite Jews in relation to diabetes and ischaemic heart­disease. Lancet 2:1399-1401.

Cohen, M.; Leong, V.W.; Salmon, E.; Martin, F.I.R. 1980. Role of guar and dietary fibre in the management of diabetes mellitus. Med. J. Aust. 1:59-61.

Collier, G.; McLean, A.; □ 'Dea, K. 1984. Effect of co-ingestion of fat on the metabolic responses to slowly and rapidly absorbed carbohydrates. Diabetologia 26:50-54.

Collier, G.; □ 'Dea, K. 1982. Effect of physical form of carbohydrate on the postprandial glucose, insulin, and gastric inhibitory polypeptide responses in type 2 diabetes. Am. J. Clin. Nutr. 36:10-14.

Connell, A.M. 1978. The effects of dietary fiber on gastro­intestinal motor function. Am. J. Clin. Nutr. 31:Sl52-S156.

Connell, A.M.; Hilton, C.; Irvine, G.; Lennard-Jones, J.E.; Misiewicz, J.J. 1965. Variation in bowel habit in two popula­tion samples. Br. Med. J. 2:1095-1099.

Connell, A.M.; Mccarroll, A.M.; Chen, M.H. 1980. Effect of fibre from bran cereal on gastroenteropancreatic hormone responses of normal adults. Ir. J. Med. Sci. 149:49-52.

Cook, G.C.; Kajubi, S.K.; Lee, F.D. 1969. Jejunal morphology of the African in Uganda. J. Pathol. 98:157-169.

Cook, J.D.; Noble, N.L.; Morck, T.A.; Lynch, S.R.; Petersburg, S.J. 1983. Effect of fiber on nonheme iron absorptipn. Gastro­enterology 85:1354-1358.

Cornoni-Huntley, J.; Barbano, H.E.; Brody, J.A.; Cohen, B.; Feldm~n, J.J.; Kleinman, J.C.; Madans, J. 1983. National Health and Nutrition Examination I: epidemiologic followup survey. Public Health Rep. 98:245-251.

174

Crapo, P.A.; Kolterman, O.G.; Waldeck, N.; Reaven, G.M.; Olefsky, J.M. 1980. Postprandial hormonal responses to different types of complex carbohydrate in individuals with impaired glucose tolerance. Am. J. Clin. Nutr~ 33:1723-1728.

Creamer, B. 1974. Intestinal structure in relation to absorp­tion. Biomembranes 4A:l-42.

Cruse, J.P.; Lewin, M.R.; Clark, C.G. 1978. Failure of bran to protect against experimental colon cancer in rats. Lancet 2:1278-1280.

Cullumbine, H.; Basnayake, V.; Lemottee, J.; Wickramanayake, T.W. 1950. Mineral metabolism on rice diets. Br. Med J. 4:101-111.

Cummings, J.H. 1978. Nutritional implications of dietary fiber. Am. J. Clin. Nutr. 31:S21-S29.

Cummings, J.H. 1982. Consequences of the metabolism of fiber in the human large intestine. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber in health and disease. New York: Plenum Press. p.9~22.

Cummings, J.H. 1983. Fermentation in the human large intestine: evidence and implications for health. Lancet 1:1206-1209.

Cummings, J.H. 1986. The effect of dietary fiber on fecal weight and composition. In: Spiller, G.A., ed. CRC handbook of dietary fiber in human nutrition. Boca Raton, FL: CRC Press, Inc. p.211-280.

Cummings, J.H.; Branch, W.J. 1986. Fermentation and the pro­duction of short-chain fatty acids in the human large intestine. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.131-149.

Cummings, J.H.; Hill, M.J.; Jenkins, D.J.A.; Pearson, J.R.; Wiggins, H.S. 1976a. Changes in fecal composition and colonic function due to cereal fiber. Am. J. Clin. Nutr. 29:1468-1473.

Cummings, J.H.; Jenkins, D.J.A.; Wiggins, H.S. 1976b. Measure­ment of the mean transit time of dietary residue through the human gut. Gut 17:210-218.

Cummings, J.H.; Southgate, D.A.T.; Branch, W.; Houston, H.; Jenkins, D.J.A.; James, W.P.T. 1978. Colonic response to dietary fibre from carrot, cabbage, apple, bran, and guar gum. Lancet 1:5-9.

Cummings, J.H.; Southgate, D.A.T.; Branch, W.J.; Wiggins, H.S.; Houston, H.; Jenkins, D.J.A.; Jivraj, T.; Hill, M.J. 1979a. The digestion of pectin in the human gut and its effect on calcium absorption and large bowel function. Br. J. Nutr. 41:477-485.

175

Cummings, J.H.; Hill, M.J.; Jivraj, T.; Houston, H.; Branch, W.J.; Jenkins, D.J.A. 1979b. The effect of meat protein and dietary fiber on colonic function and metabolism. I. Changes in bowel habit, bile acid excretion, and calcium absorption. Am. J. Clin. Nutr. 32:2086-2093.

Dales, L.G.; Friedman, G.D.; Ury, H.K.; Grossman, S.; Williams, S.R. 1979. A case-control study of relationships of diet and other traits to colorectal cancer in American blacks. Am. J. Epidemiol. 109:132-144.

Dam, H. 1971. Determinants of cholesterol cholelithiasis in man and animals. Am. J. Med. 51:596-613.

Danielsson, A.; Ek, B.; Nyhlin, H.; Steen, L. 1979. Effect of long term treatment with hydrophilic colloid on serum lipids. Acta Hepato-Gastroenterol. 26:148-153.

Davies, G.J.; Crowder, M.; Dickerson, J.W.T. 1985. Dietary fibre intakes of individuals with different eating patterns. Hum. Nutr. Appl. Nutr. 39A:139-148.

Dawson, A.M.; Holdsworth, C.D.; Webb, J. 1964. Absorption of short chain fatty acids in man. Proc. Soc. Exp. Biol. Med. 117:97-100.

de Groot, A.P.; Luyken, R.; Pikaar, N.A. 1963. Cholesterol­lowering effect of rolled oats. Lancet 2:303-304.

Delbarre, F.; Randier, J.; de Gery, A. 1977. Lack of effect of two pectins in idiopathic or gout-associated hyperdyslipidemia hypercholesterolemia. Am. J. Clin. Nutr. 30:463-465.

DeRubertis, F.R.; Craven, P.A. 1987. Relationship of bile salt stimulation of colonic epithelial phospholipid turnover and proliferative activity: role of activation of protein kinase C. Prev. Med. 16:573-580.

DeRubertis, F.R.; Craven, P.A.; Saito, R. 1984. Bile salt stimulation of colonic epithelial proliferation: evidence for involvement of lipoxygenase products. J. Clin. Invest. 74:1614-1624.

Deschner, E.E.; Alcock, N.; Okamura, T.; DeCosse, J.J.; Sherlock, P. 1983. Tissue concentrations and proliferative effects of massive doses of ascorbic acid in the mouse. Nutr. Cancer 4:241-246.

Descovich, G.C.; Ceredi, C.; Gaddi, A.; Benassi, M.S.; Mannino, G.; Colombo, L.; Cattin, L.; Fontana, G.; Senin, U.; Mannarino, E.; Caruzzo, C.; Bertelli, E.; Fragiacomo, C.; Noseda, G.; Sirtori, M.; Sirtori, ·c.R. 1980. Multicentre study of soybean protein diet for outpatient hypercholesterolaemic patients. Lancet 2:709-712.

176

Devroede, G. 1978. Dietary fiber, bowel habits, and colonic function. Am. J. Clin. Nutr. 31:Sl57-Sl60.

Devroede, G.; Vobecky, J.S.; Vobecky, J.M.; Beaudry, R.; Haddad, H.; Navert, H.; Perey, B.; Poisson, J. 1977. Medical management of diverticular disease: a random trial. Gastroenterology 72:1157 (Abstract).

Diabetes Care and Education Dietetic Practice Group. 1985. Diabetes mellitus and glycemic responses to different foods: a summary and annotated bibliography. 1985 Supplement. Avail­able from: American Dietetic Association, Chicago.

Dintzis, F.R.; Watson, P.R.; Sandstead, H.H. 1985. Mineral contents of brans passed through the human GI tract. Am. J. Cline Nutr. 41:901=908.

Dobbs, R.J.; Baird, I.M. 1977. Effect of wholemeal and white bread on iron absorption in normal people. Br. Med. J. 2:1641-1642.

Dodson, P.M.; Pacy, P.J.; Bal, P.; Kubicki, A.J.; Fletcher, R.F.; Taylor, K.G. 1984. A controlled trial of a high fibre, low fat and low sodium diet for mild hypertension in type 2 (non-insulin­dependent) diabetic patients. Diabetologia 27:522-526.

Dodson, P.M .. ; Pacy, P.J.; Beevers, M.; Bal, P.; Fletcher, R.F.; Taylor, K.G. 1983. The effects of a high fibre, low fat and low sodium dietary regime on diabetic hypertensive patients of different ethnic groups. Postgrad. Med. J. 59:641-644.

Dodson, P.M.; Pacy, PoJ.; Cox, E.V. 1985. Long-term follow-up of the treatment of essential hypertension with a high-fibre, low-fat and low-sodium dietary regimen. Hum. Nutr. Clin. Nutr. 39C:213-220.

Dodson, P.M.; Stocks, J.; Holdsworth, G.; Galton, D.J. 1981. High-fibre and low-fat diets in diabetes mellitus. Br. J. Nutr. 46:289-294.

Doi, K.; Matsuura, M.; Kawara, A.; Baba, S. 1979. Treatment of diabetes with glucomannan (Konjac mannan). Lancet 1:987-988.

Domellof, L.; Darby, Le; Hanson, D.; Mathews, L.; Simi, B.; Reddy, B.S. 1982. Fecal sterols and bacterial S-glucuronidase activity: a preliminary metabolic epidemiology study of healthy volunteers from Umea, Sweden, and metropolitan New York. Nutr. Cancer 4:120-127.

Drasar, BeS ; Hill, M.J. 1972. Intestinal bacteria and cancer. Am. J. Clin. Nutr. 25:1399-1404.

Drasar, B.S.; Irving, D. 1973 .. Environmental factors and cancer of the colon and breast. Br. J. Cancer 27:167-172.

177

Drasar, B.S.; Jenkins, D.J.A.; Cummings, J.H. 1976. The influence of a diet rich in wheat fibre on the human faecal flora. J. Med. Microbial. 9:423-431.

Drews, L.M.; Kies, C.; Fox, H.M. 1979. Effect of dietary fiber on copper, zinc, and magnesium utilization by adolescent boys. Am. J. Clin. Nutr. 32:1893-1897.

Duke, J.H., Jr.; Yar, M.S. 1977. Primary small bowel volvulus: cause and management. Arch. Surg. 112:685-688.

Dunaif, G.; Schneeman, 8.0. 1981. The effect of dietary fiber on human pancreatic enzyme activity in vitro. Am. J. Clin. Nutr. 34:1034-1035.

Duncan, K.H.; Bacon, J.A.; Weinsier, R.L. 1983. The effects of high and low energy density diets on satiety, energy intake, and eating time of obese and nonobese subjects. Am. J. Clin. Nutr. 37:763-767.

Duncan, L.J.P.; Rose, K.; Meiklejohn, A.P. 1960. Phenmetrazine hydrochloride and methylcellulose in the treatment of "refractory" obesity. Lancet 1:1262-1265.

Durrington, P.N.; Manning, A.P.; Bolton, C.H.; Hartog, M. 1976. Effect of pectin on serum lipids and lipoproteins, whole-gut transit-time, and stool weight. Lancet 2:394-396.

Dutta, S.K.; Hlasko, J. 1985. Dietary fiber in pancreatic disease: effect of high fiber diet on fat malabsorption in pancreatic insufficiency and in vitro study of the interaction of dietary fiber with pancreatic enzymes. Am. J. Clin. Nutr. 41:517-525.

Dworkin, L.D.; Levine, G.M.; Farber, N.J.; Spector, M.H. 1976. Small intestinal mass of the rat is partially determined by indirect effects of intraluminal nutrition. Gastroenterology 71:626-630.

Eastwood, M.; Brydon, W.G. 1985. Physiological effects of dietary fibre on the alimentary tract. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.105-131.

Eastwood, M.A. 1973. Vegetable fibre: its physical properties. Proc. Nutr. Soc. 32:137-143.

Eastwood, M.A. 1978. Fiber in the gastrointestinal tract. Am. J. Clin. Nutr. 31:S30-S32.

Eastwood, M.A.; Brydon, W.G.; Tadesse, K. 1980. Effect of fiber on colon function. In: Spiller, G.A.; Kay, R.M., eds. Medical aspects of dietary fiber. New York: Plenum Medical Book Company. p.l-26.

17A

Eastwood, M.A.; Elton, R.A.; Smith, JeH. 1986. Long-term effect of wholemeal bread on stool weight, transit time, fecal bile acids, fats, and neutral sterols. Am. J. Clin. Nutr. 43:343-349.

Eastwood, M.A.; Hamilton, D. 1968. Studies on the adsorption of bile salts to non-absorbed components of diet. Biochim. Biophys. Acta 152:165-173.

Eastwood, M.A.; Kirkpatrick, J.R.; Mitchell, W.D.; Bone, A.; Hamilton, T. 1973. Effects of dietary supplements of wheat bran and cellulose on faeces and bowel function. Br. Med. Jo 4:392-394.

Eastwood, M.A.; Mowbray, L. 1976. The binding of the components of mixed micelles to dietary fiberse Am. J. Clin. Nutr 29:1461-14670

Eastwood, M.A.; Passmore, R. 1983. Dietary fibre. Lancet 2:202-206.

Eastwood, M.A.; Robertson, J.A.; Brydon, W.G.; MacDonald, D. 1983. Measurement of water-holding properties of fibre and their faecal bulking ability in man. Br. J. Nutr. 50:539-547.

Ecknauer, R.; Sircar, B.; Johnson, L.R. 1981. Effect of dietary bulk on small intestinal morphology and cell renewal in the rat. Gastroenterology 81:781-786.

Ehle, F.R.; Robertson, J.B.; Van Soest, P.J. 1982. Influence of dietary fibers on fermentation in the human large intestine. J. Nutr. 112:158-166.

Eide, T.J.; Stalsberg, H. 1979. Diverticular disease of the large intestine in Northern Norway. Gut 20:609-615.

Elkington, S.G. 1970. Lactulose. Gut 11:1043-1048.

Ellis, P.Re 1985 Fibre and food products. In: Leeds, A.R.; Avenell, A., eds. Dietary fibre perspectives: reviews and bibliography. London: John Libbey and Company Ltd. p.83-105.

Elsenhans, B.; Blume, R.; Caspary, W.F. 1981. Long-term feed­ing of unavailable carbohydrate gelling agents: influence of dietary concentration and microbiological degradation on adaptive responses in the rat. Am. J. Clin. Nutr. 34:1837-1848.

Elsenhans, B.; SUfke, U.; Blume, R.; Caspary, W.F. 1980. The influence of carbohydrate gelling agents on rat intestinal transport of monosaccharides and neutral amino acids in vitro. Clin. Sci. 59:373-380.

Elwood, P.C.; Benjamin, I.T.; Fry, F.A.; Eakins, J.D.; Brown, D.A.; De Kock, P.C.; Shah, J.U. 197 □~ Absorption of iron from chapatti made from wheat flour. Am. J. Clin. Nutr. 23:1267-1271.

179

Englyst, H. 1981. Determination of carbohydrate and its com­position in plant materials. In: James, W.P.T.; Theander, O., eds. The analysis of dietary fiber in food. New York: Marcel Dekker, Inc. p.71-93.

Englyst, H.; Wiggins, H.S.; Cummings, J.H. 1982a. Determination of the non-starch polysaccharides in plant foods by gas-liquid chromatography of constituent sugars as alditol acetates. Analyst 107:307-318.

Englyst, H.N.; Bingham, S.A.; Wiggins, H.S.; Southgate, D.A.T.; Seppanen, R.; Helms, P.; Anderson, V.; Day, K.C.; Choolun, R.; Collinson, E.; Cummings, J.H. 1982b. Nonstarch polysaccharide consumption in four Scandinavian populations. Nutrc Cancer 4:50-60.

Enstrom, J.E. 1980. Health and dietary practices and cancer mortality among California Mormons. In: Cairns, J.; Lyon, J.L.; Skolnick, M., eds. Banbury Report 4. Cancer incidence in defined populations. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. p.69-92.

Evans, E.; Miller, D.S. 1975. Bulking agents in the treatment of obesity. Nutr. Metab. 18:199-203.

Fagerberg, S.-E. 1982. The effects of a bulk laxative (Metamucil®) on fasting blood glucose, serum lipids and other variables in constipated patients with non-insulin dependent adult diabetes. Curr. Ther. Res. 31:166-172.

Fahrenbach, M.J.; Riccardi, B.A.; Saunders, J.C.; Lourie, I.N.; Heider, J.G. 1965. Comparative effects of guar gum and pectin on human serum cholesterol levels. Circulation 32(Suppl. II): 11-12 (Abstract).

Farrell, D.J.; Girle, L.; Arthur, J. 1978. Effects of dietary fibre on the apparent digestibility of major food components and on blood lipids in men. Aust. J. Exp. Biol. Med. Sci. 56:469-479.

Fernandez, R.; Phillips, S.F. 1982. Components of fiber bind iron in vitro. Am. J. Clin. Nutr. 35:100-106.

Fielding, J.F. 1985. The irritable bowel controversy. Proc. Nutr. Soc. 44:139-140.

Findlay, J.M.; Smith, A.N.; Mitchell, W.D., Anderson, A.J.B.; Eastwood, M.A. 1974. Effects of unprocessed bran on colon function in normal subjects and in diverticular disease. Lancet 1:146-149.

Finegold, S.M.; Sutter, V.L. 1978. Fecal flora in different populations, with special reference to diet. Am. J. Clin. Nutr. 31:Sll6-Sl22.

180

Fischer, M.; Adkins, W.; Hall, L.; Scaman, P.; Hsi, S.; Marlett, J. 1985. The effects of dietary fibre in a liquid diet on bowel function of mentally retarded individuals. J. Ment. Defic. Res. 29:373-381.

Fisher, H.; Siller, W.G.; Griminger, P. 1966. The retardation by pectin of cholesterol-induced atherosclerosis in the fowl. J. Atheroscler. Res~ 6:292-298.

Flanagan, M.; Little, C.; Milliken, J.; Wright, E.; McGill, A.R.; Weir, D.G.; □ 'Moore, R.R. 1980. The effects of diet on high density lipoprotein cholesterol. J. Hum. Nutr. 34:43-45.

Fleiszer, D.; Murray, D.; MacFarlane, J.; Brown, R.A. 1978. Protective effect of dietary fibre against chemically induced bowel tumours in rats. Lancet 2:552-553.

Fleming, S.E.; Marthinsen, D.; Kuhnlein, H. 1983. Colonic function and fermentation in men consuming high fiber diets. Je Nutr. 113:2535-25440

Fleming, S.E.; 0 1 Donnell, A.U.; Perman, J.A. 1985. Influence of frequent and long-term bean consumption on colonic function and fermentation. Am. J. Clin. Nutr. 41:909-918.

Fleming, S.E.; Rodriguez, M.A. 1983. Influence of dietary fiber on fecal excretion of volatile fatty acids by human adults. J. Nutr. 113:1613-1625.

Floch, M.H.; Fuchs, H.-M. 1978. Modification of stool content by increased bran intake. Am. J. Clin. Nutr. 31.Sl85-Sl89o

Floch, M.H.; Gershengoren, W.; Elliott, S.; Spiro, H.M. 1971. Bile acid inhibition of the intestinal microflora: a function for simple bile acids? Gastroenterology 61:228-233.

Florholmen, J.; Arvidsson-Lenner, R.; Jorde, R.; Burhol, P.G. 1982. The effect of Metamucil on postprandial blood glucose and plasma gastric inhibitory peptide in insulin-dependent diabetics. Acta Med. Scand. 212:237-239.

Floyd, R.Ao 1978. Digoxin interaction with bran and high fiber foods. Am. J. Hosp. Pharm. 35:660.

Food and Drug Administration. 1985. Laxative drug products for over-the-counter human use: tentative final monograph. Fed. Regist. 50:2124-2158.

Forman, DeT.; Garvin, J.E.; Forestner, J.E.; Taylor, C.B. 1968. Increased excretion of fecal bile acids by an oral hydrophilic colloid. Proc. Soc. Exp. Biol. Med. 127:1060-1063.

181

Fraser, G.E.; Jacobs, D.R., Jr.; Anderson, J.T.; Foster, N.; Palta, M.; Blackburn, H. 1981. The effect of various vegetable­supplements on serum cholesterol. Am. J. Clin. Nutr. 34:1272-1277.

Frati-Munari, A.C.; Fernandez-Harp, J.A.; Becerril, M.; Chavez­Negrete, A.; Banales-Ham, M& 1983. Decrease in serum lipids, glycemia and body weight by plantago psyllium in obese and diabetic patients. Arch. Invest. Med. 14:259-268.

Freeman, H.J.; Spiller, G.A.; Kim, Y.S. 1978. A double-blind study on the effect of purified cellulose dietary fiber on 1,2-dimethylhydrazine-induced rat colonic neoplasia. Cancer Res. 38:2912-2917.

Freeman, H.J.; Spiller, G.A.; Kim, Y.S. 1980. A double-blind study on the effects of differing purified cellulose and pectin fiber diets on 1,2-dimethylhydrazine-induced rat colonic neo-plasia. Cancer Res. 40:2661-2665. ·

Freeman, H.J.; Spiller, G.A.; Kim, Y.S. 1984. Effect of high hemicellulose corn bran in 1,2-dimethylhydrazine-induced rat intestinal neoplasia. Carcinogenesis 5:261-264.

Frenkiel, P.G.; Lee, D.W.T.; Cohen, H.; Gilmore, C.J.; Resser, K.; Bonorris, G.G.; Marks, J.W.; Schoenfield, L.J. 1986. The effect of diet on bile acid kinetics and biliary lipid secretion in gallstone patients treated with ursodeoxycholic acid. Am. J. Clin. Nutr. 43:239-250.

Fuchs, H.-M.; Dorfman, S.; Floch, M.H. 1976. The effect of dietary fiber supplementation in man. II. Alteration in fecal physiology and bacterial flora. Am. J. Clin. Nutr. 29:1443-1447.

Fukagawa, N.K.; Minaker, K.L.; Hageman, G.; Young, VaR.; Anderson, J.W. 1984. High-carbohydrate, high-fiber diets increase peripheral insulin sensitivity of healthy young men. Clin. Res. 32:796A (Abstract).

Furda, I. 1981. Simultaneous analysis of soluble and insoluble dietary fiber. In: James, W.P.T.; Theander, O., eds. The analysis of dietary fiber in food. New York: Marcel Dekker, Inc. p.163-172.

Ganapathy, S.; Dhanda, R. 1980. Protein and iron nutrition in lacto-ovo-vegetarian Inda-Aryan United States residents. Indian J. Nutr. Diet. 17:45-52.

Gardner, D.F.; Schwartz, L.; Krista, M.; Merimee, T.J. 1984. Dietary pectin and glycemic control in diabetes. Diabetes Care 7:143-146.

Gatti, E.; Catenazzo, G.; Camisasca, E.; Torri, A.; Denegri, E.; Sirtori, C.R. 1984. Effects of guar-enricherj pasta in th~ treat­ment of diabetes and hyperlipidemia. Ann. Nutr. Metab. 28:1-10.

182

Gibson, R.S.; Anderson, B.M.; Sabry, J.H. 1983. status of a group of post-menopausal vegetarians. Assoc. 82:246-250.

The trace metal J. Am. Diet.

Gibson, R.S.; Anderson, B.M.; Scythes, C.A. 1982. Regional differences in hair zinc concentrations: a possible effect of water hardness. Am. J. Clin. Nutr. 37:37-42.

Gibson, R.S.; Gibson, I.L.; Kitching, J. 1985a. A study of inter- and intrasubject variability in seven-day weighed dietary intakes with particular emphasis on trace elements. Biol. Trace Elem. Res. 8:79-91.

Gibson, R.S.; Martinez, O.B.; MacDonald, A.C. 1985b. The zinc, copper, and selenium status of a selected sample of Canadian elderly women. J. Gerontol. 40:296-302.

Gibson, R.S.; Scythes, C.A. 19820 Trace element intakes of women. Br. J. Nutr. 48:241-248.

Godara, R.; Kaur, A.P.; Bhat, C.M. 1981. Effect of cellulose incorporation in a low fiber diet on fecal excretion and serum levels of calcium, phosphorus, and iron in adolescent girls. Am. J. Clin. Nutr. 34:1083-1086.

Goldin, B.R.; Adlercreutz, H.; Gorbach, S.L.; Warram, J.H.; Dwyer, J.T.; Swenson, L.; Woods, M.N. 1982. Estrogen excretion patterns and plasma levels in vegetarian and omnivorous women. N. Engl. J. Med. 307:1542-1547.

Goldin, B.R.; Swenson, L.; Dwyer, J.; Sexton, M.; Gorbach, S.L. 1980. Effect of diet and Lactobacillus acidophilus supplements on human fecal bacterial enzymes. J. Natl. Cancer Inst. 64:255-261.

Goodlad, R.A.; Wright, N.A. 1983. Effects of addition of kaolin or cellulose to an elemental diet on intestinal cell prolifera­tion in the mouse. Br. J. Nutr. 50:91-98.

Gordon, T.; Kagan, A.; Garcia-Palmieri, M.; Kannel, W.B.; Zukel, W.J.; Tillotson, J.; Sorlie, P.; Hjortland, M. 1981. Diet and its relation to coronary heart disease and death in three popula­tions. Circulation 63:500-515.

Gormley, T.Ro; Kevany, J.; Egan, J.P.; McFarlane, R~ Effect of apples on serum cholesterol levels in humans Food Sci. Technol. 1:117-128.

1977. Ir. J.

Gormley, T.R.; Kevany, J.; O1 Donnell, B.; McFarlane, R. 1979. Effect of peas on serum cholesterol levels in humans. Ir. J. Food Sci. Technol. 3:101-109.

Graham, D.Y.; Moser, S.E.; Estes, M.K. 1982. The effect of bran on bowel function in constipation. Am. J. Gastroenterol. 77:599-603.

183

Graham, S. 1986. State University of New York at Buffalo, Buffalo, NY. Personal communication with T. Byers, State University of New York at Buffalo, Buffalo, NY.

Graham, S.; Dayal, H.; Swanson, M.; Mittelman, A.; Wilkinson, G. 1978. Diet in the epidemiology of cancer of the colon and rectum. J. Natl. Cancer Inst. 61:709-714.

Grande, Fe; Anderson, J.T.; Keys, A. 1965. Effect of carbo­hydrates of leguminous seeds, wheat and potatoes on serum cholesterol concentration in man. J. Nutr. 86:313-317.

Grande, F.; Anderson, J.T.; Keys, A. 1974. Sucrose and various carbohydrate-containing foods and serum lipids in man. Am. J. Clin. Nutr. 27:1043-1051.

Grimes, D.S.; Gordon, C. 1978. Satiety value of wholemeal and white bread. Lancet 2:106.

Gulliver, W.P.; Kutty, K.P.; Laher, J.M.; Barrowman, J.A. 1983. In vitro interaction of 7,12-dimethylbenz[a]anthracene and its biliary metabolites with dietary fibers. J. Natl. Cancer Inst. 71:207-210.

Gupta, I.; Baptista, J.; Bruce, W.R.; Che, C.T.; Furrer, R.; Gingerich, J.S.; Grey, A.A.; Marai, L.; Yates, P.; Krepinsky, J.J. 1983. Structures of fetapentaenes, the mutagens of bacterial origin isolated from human feces. Biochemistry 22:241-245.

Guthrie, B.E.; Robinson, M.F. 1978. Zinc balance studies during wheat bran supplementation. Fed. Proc. Fed. Am. Soc. Exp. Biol. 37:254 (Abstract).

Haber, G.B.; Heaton, K.W.; Murphy, D.; Burroughs, L.F. 1977. Depletion and disruption of dietary fibre: effects on satiety, plasma-glucose, and serum-insulin. Lancet 2:679-682.

Haenszel, W.; Berg, J.W.; Segi, M.; Kurihara, M.; Locke, F.B. 1973. Large-bowel cancer in Hawaiian Japanese. J. Natl. Cancer Inst. 51:1765-1779.

Haenszel, W.; Loche, F.B.; Segi, M. 1980. A case-control study of large bowel cancer in Japan. J. Natl. CaAcer Inst. 64:17-22.

Hagander, B.; BjHrck, I.; Asp, N.-G.; Lundquist, I~; Nilsson­Ehle, P.; Schrezenmeir, J.; Schersten, B. 1985. Hormonal and metabolic responses to breakfast meals in NIDDM: comparison of white and whole-grain wheat bread and corresponding extruded products. Hum. Nutr. Appl. Nutr. 39A:114-123.

184

Hagander, B.; Schersten, B.; Asp, N.-G.; Sartor, G.; Agardh, C.-D.; Schrezenmeir, J.; Kasper, H.; Ahren, B.; Lundquist, I. 1984. Effect of dietary fibre on blood glucose, plasma immuno­reactive insulin, C-peptide and GIP responses in non insulin dependent (type 2) diabetics and controls. Acta Med. Scand. 215:205-213.

Hagemann, R.F.; Stragand, J.J. 1977. Fasting and refeeding: cell kinetic response of jejunum, ileum and colon. Cell Tissue Kinet. 10:3-14.

Hansen, W.E.; Maurer, H.; Vollmar, J. Brauning, C. 1983. Guar gum and bile: effects on postprandial gallbladder contraction and on serum bile acids in man. Hepatogastro­enterology 30:131-1330

Hardinge, M.G.; Chambers, A.C.; Crooks, H.; Stare, F.J. 1958. Nutritional studies of vegetarians. III. Dietary levels of fiber. Am. J. Clin. Nutr. 6:523-525.

Hardinge, M.G.; Stare, F.J. 1954. Nutritional studies of vegetarians. 2. Dietary and serum levels of cholesterol. Am. J. Clin. Nutr. 2:83-88.

Harju, E.; Heikkila, J.; Larmi, T.K. 1984. Effect of guar gum on gastric emptying after gastric resection. J. Parentere Enter. Nutr 8:18-20.

Harju, E.J.; Larmi, T.K. 1985. Effect of guar gum added to the diet of patients with duodenal ulcero J Parenter. Enter. Nutr. 9:496-500.

Harvey, R.F. 1985. Functional gastrointestinal disorders: irritable bowel and other syndromes. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Presso p.217-228.

Harvey, R.F.; Pomare, E.W.; Heaton, K.W. 1973. Effects of increased dietary fibre on intestinal transit. Lancet 1:1278-1280.

Haubrich, W.S. 1985. Diverticula and diverticular disease of the colon. In: Berk, J.E.; Haubrich, W.S.; Kalser, N.Ho; Roth, J.L.A.,; Schaffner, F., eds. Backus' gastroenterology. 4th ed. Philadelphia: WoB. Saunders Company. Vol. 4. p.2445-2473.

Heading, R.C.; Tothill, P.; Maclaughlin, G.P.; Shearman, D.J.C. 1976. Gastric emptying rate measurement in man: a double isotope scanning technique for simultaneous study of liquid and solid components of a meal. Gastroenterology 71:45-50.

Health and Welfare Canada. 1985. Report of the Expert Advisory Committee on Dietary Fibre. [58]p. Available from: Health Protection Branch, Health and Welfare Canada, Ottawa.

185

Heaton, K.W. 1973. The epidemiology of gallstones and suggested aetiology. Clin. Gastroenterol. 2:67-83.

Heaton, K.W. 1980. Food intake regulation and fiber. In: Spiller, G.A.; Kay, R.M., eds. Medical aspects of dietary fiber. New York: Plenum Medical Book Company. p.223-238.

Heaton, K.W. 1981. Is bran useful in diverticular disease? Br. Med. J. 283:1523-1524.

Heaton, K. 1985. Gallstones. In: Trowell, H.; Burkitt, D., Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.289-304.

Heaton, K.W. 1986. Gallstones. In: Spiller, G.A., ed. CRC handbook of dietary fiber in human nutrition. Boca Raton, FL: CRC Press, Inc. p.361-367.

Heaton, K.W.; Thornton, J.R.; Emmett, P.M. 1979. Treatment of Crohn's disease with an unrefined-carbohydrate, fibre-rich diet. Br. Med. J. 2:764-766.

Heller, S.N.; Hackler, L.R.; Rivers, J.M.; Van Soest, P.J.; Roe, D.A.; Lewis, B.A.; Robertson, J. 1980. Dietary fiber: the effect of particle size of wheat bran on colonic function in young adult men. Am. J. Clin. Nutr. 33:1734-1744.

Helms, P.; J~rgensen, I.M.; Paerregaard, A.; Bjerrum, L.; Poulsen, L.; Mosbech, J. 1982. Dietary patterns in Them and Copenhagen, Denmark. Nutr. Cancer 4:34-40.

Hennekens, C.H.; Speizer, F.E.; Rosner, B.; Bain, C.J.; Belanger, C.; Peto, R. 1979. Use of permanent hair dyes and cancer among registered nurses. Lancet 1:1390-1393.

Henry, R.W.; Stout, R.W.; Love, A.H.G. 1978. Lack of effect of bran enriched bread on plasma lipids, calcium, glucose and body weight. Ir. J. Med. Sci. 147:249-251.

Higginson, J. 1966. Etiological factors in gastrointestinal cancer in man. J. Natl. Cancer Inst. 37:527-545.

Hill, M.; Maclennan, R.; Newcombe, K. 1979. Diet and large­bowel cancer in three socioeconomic groups in Hong Kong. Lancet 1:436.

Hill, M.J. 1986. Bile acids and colorectal cancer in humans. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.497-513.

Hill, M.J.; Drasar, B.S. 1975. The normal colonic bacterial flora. Gut 16:318-323.

186

Hillman, L.C.; Peters, S.G.; Fisher, C.A.; Pomare, E.W. 1985. The effects of the fiber components pectin, cellulose and lignin on serum cholesterol levels. Am. J. Clin. Nutr. 42:207-213.

Hillman, L.C.; Peters, S.G.; Fisher, C.A.; Pomare, E.W. 1986. Effects of the fibre components pectin, cellulose, and lignin on bile salt metabolism and biliary lipid composition in man. Gut 27:29-36.

Hillman, L.C.; Stace, N.H.; Fisher, A.; Pomare, E.W. 1982. Dietary intakes and stool characteristics of patients with the irritable bowel syndrome. Am. J. Clin. Nutr. 36:626-629.

Hillman, L.C.; Stace, N.H.; Pomare, E.W. 1984. Irritable bowel patients and their long-term response to a high fiber diet. Am. J. Gastroenterol. 79:1-7.

Himsworth, H.P.; Marshall, E.M. 1935. The diet of diabetics prior to the onset of the disease. Clin. Sci. 2:95-115.

Hinton, J.M.; Lennard=Jones, J.E.; Young, A.C. 1969. A new method for studying gut transit times using radioopaque markers. Gut 10:842-8470

Hirai, N.; Kingston, D.G.I.; Van Tassell, Rel.; Wilkins, T.D. 19820 Structure elucidation of a potent mutagen from human feces. J. Am. Chem. Soc. 104:6149-6150.

Hirayama, T. 19810 A large-scale cohort study on the rela­tionship between diet and selected cancers of digestive organs. In: Bruce, W.R.; Correa, P.; Lipkin, M.; Tannenbaum, S.Re; Wilkins, T.D., eds. Banbury Report 7. Gastrointestinal cancer: endogenous factors. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. p.4O9-429.

Hj~llund, E.; Pedersen, O.; Richelsen, B.; Beck-Nielsen, H.; S0rensen, N.S. 1983. Increased insulin binding to adipocytes and monocytes and increased insulin sensitivity of glucose trans­port and metabolism in adipocytes from non-insulin-dependent diabetics after a low-fat/high-starch/high-fiber diet. Metabolism 32:1067-1075*

Hollenbeck, C.8.; Coulston, A.M.; Reaven, G.M. 1986. To what extent does increased dietary fiber improve glucose and lipid metabolism in patients with noninsulin-dependent diabetes mellitus (NIDDM)? Am. J. Clin. Nutr. 43:16-24.

Hollingsworth, D.F.; Greaves, J.P. 1967. Consumption of carbohydrates in the United Kingdom. Am. J. Clin. Nutr. 20:65-72.

Holt, S.; Heading, R.C.; Carter, D.C.; Prescott, L.F.; Tothill, P. 1979. Effect of gel fibre on gastric emptying and absorption of glucose and paracetamol. Lancet 1:636-639.

187

Hope, A.K.; Down, E.C. 1986. Dietary fibre and fluid in the control of constipation in a nursing home populatione Med. J. Aust. 144:306-307.

Howard, P.; Mahoney, R.R.; Wilder, T. 1986. amino acids by dietary fibres and wheat bran. 34:135-140.

Binding of Nutr. Rep. Inte

Howell, M.A. 1975. Diet as an etiological factor in the development of cancers in the colon and rectum. J. Chronic Dis. 28:67-80.

Howie, B.J.; Shultz, T.D. 1985. Dietary and hormonal inter­relationships among vegetarian Seventh-Day Adventists and nonvegetaria~ men. Am. J. Clin. Nutr. 42:127-134.

Hunt, J.N. 1954. The viscosity of a test-meal: its influence on gastric secretion and emptying. Lancet 1:17-18.

Huupponen, R.; Karhuvaara, S.; Sepp~l~, P. 1985. Effect of guar gum on glipizide absorption in man. Eur. J. Clin. Pharmacol. 28:717-719.

Huupponen, R.; Seppala, P.; Iisalo, E. 1984. Effect of guar gum, a fibre preparation, on digoxin and penicillin absorption in man. Eur. J. Clin. Pharmacol. 26:279-281.

Hyland, J.M.P.; Taylor, I. 1980. Does a high fibre diet prevent the complications of diverticular disease? Br. J. Surg. 67:77-79.

Hylander, B.; Rossner, S. 1983. Effects of dietary fiber intake before meals on weight loss and hunger in a weight-reducing club. Acta Med. Scand. 213:217-220.

IARC Intestinal Microecology Group. 1977. Dietary fibre, transit-time, faecal bacteria, steroids, and colon cancer in two Scandinavian populations. Lancet 2:207-211.

Irving, D.; Drasar, B.S. 1973. Fibre and cancer of the colon. Br. J. Cancer 28:462-463.

Isaksson, G.; Lundquist, I.; Akesson, B.; Ihse, I. 1984. Effects of pectin and wheat bran on intraluminal pancreatic enzyme activities and on fat absorption as examined with the triolein breath test in patients with pancreatic insufficiency. Scand. J. Gastroenterol. 19:467-472.

Isaksson, G.; Lundquist, I.; Ihse, I. 1982. fiber on pancreatic enzyme activity in vitro. 82:918-924.

Effect of dietary Gastroenterology

Ismail-Beigi, F.; Faraji, B.; Reinhold, J.G. 1977a. Binding of zinc and iron to wheat bread, wheat bran, and their components. Am. J. Clin. Nutr. 30:1721-1725.

188

Ismail-Beigi, F.; Reinhold, J.G.; Faraji, B.; Abadi, P. 1977b. Effects of cellulose added to diets of low and high fiber content upon the metabolism of calcium, magnesium, zinc and phosphorus by man. J. Nutr. 107:510-518.

Jacobs, LeR. 1983a. Effects of dietary fiber on mucosa! growth and cell proliferation in the small intestine of the rat: a comparison of oat bran, pectin, and guar with total fiber deprivation. Amo J. Clin. Nutr. 37:954-960.

Jacobs, L.R. 1983b. Enhancement of rat colon carcinogenesis by wheat bran consumption during the stage of 1,2-dimethylhydrazine administration. Cancer Res. 43:4057-4061.

Jacobs, L.Re 1984e Stimulation of rat colonic crypt cell proliferative activity by wheat bran consumption during the stage of 1,2-dimethylhydrazine administration. Cancer Res. 44:2458-2463.

Jacobs, L.R. 1985. Differential effects of dietary fibers on rat intestinal circular muscle cell size. Dig. Dis. Sci. 30:247-252.

Jacobs, L.R. 1986a. Dietary fiber and gastrointestinal epithelial cell proliferation. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.211-228.

Jacobs, L.R. 1986b. Relationship between dietary fiber and cancer: metabolic, physiologic, and cellular mechanisms. Proc Soc Exp Biol. Med. 183·290-311.

Jacobs, L.R. 1987. Effect of dietary fiber on colonic cell proliferation and its relationship to colon carcinogenesis. Prev. Med. 16:567-572.

Jacobs, L.R.; Amorde, K. 1986. Source and saturation of die­tary fats determine regional alterations in rat colonic cell proliferation. Gastroenterology 90:1473 (Abstract).

Jacobs, L.R.; Lupton, J.R. 1984. Effect of dietary fibers on rat large bowel mucosal growth and cell proliferation. Am. J. Physiol. 246:G378-G385.

Jacobs, L.R.; Lupton, J.R. 1986. Relationship between colonic luminal pH, cell proliferation, and colon carcinogenesis in 1,2-dimethylhydrazine treated rats fed high fiber diets. Cancer Res. 46:1727-1734.

Jacobs, L.R.; Schneeman, BeO. 1981 Effects of dietary wheat bran on rat colonic structure and mucosal cell growth. J. Nutr. 111:798-803.

189

Jacobs, L.R.; White, F.A. 1983. Modulation of mucosal cell proliferation in the intestine of rats fed a wheat bran diet. Am. J. Clin. Nutr. 37:945-953.

Jain, M.; Cook, G.M.; Davis, F.Gv; Grace, M.G.; Howe, G.R.; Miller, A.B. 1980. A case-control study of diet and cola­rectal cancer. Int. J. Cancer 26:757-768.

James, W.P.T.; Branch, W.J.; Southgate, D.A.T. 1978. Calcium binding by dietary fibre. Lancet 1:638-639.

James, W.P.T.; Theander, O. 1981. The analysis of dietary fiber. New York: Marcel Dekker, Inc. 276p.

Jenkins, D.J.A.; Hill, M.S.; Cummings, J.H. 1975a. Effect of wheat fiber on blood lipids, fecal steroid excretion and serum iron. Am. J. Clin. Nutr. 28:1408-1411.

Jenkins, D.J.A.; Leeds, A.R.; Newton, C.; Cummings, J.H. 1975b. Effect of pectin, guar gum, and wheat fibre on serum-cholesterol. Lancet 1:1116-1117.

Jenkins, D.J.A.; Jenkins, A.L. 1985. Dietary fiber and the glycemic response. Proc. Soc. Exp. Biol. Med. 180:422-431.

Jenkins, D.J.A.; Leeds, A.R.; Gassull, M.A.; Houston, H.; Goff, D.V.; Hill, M.J. 1976a. The cholesterol lowering properties of guar and pectin. Clin. Sci. Mal. Med. 51:8p-9p.

Jenkins, O.J.A.; Goff, D.V.; Leeds, A.R.; Alberti, K.G.M.M.; Wolever, T.M.S.; Gassull, M.A.; Hockaday, T.D.R. 1976b. Unabsorbable carbohydrates and diabetes: decreased post­prandial hyperglycaemia. Lancet 2:172-174.

Jenkins, D.J.A.; Leeds, A.R.; Houston, H.; Hinks, L.; Alberti, K.G.M.M.; Cummings, J.H. 1977a. Carbohydrate tolerance in man after six weeks of pectin administration. Proc. Nutr. Soc. 36:60A (Abstract).

Jenkins, D.J.A.; Wolever, T.M.S.; Hockaday, T.D.R.; Leeds, A.R.; Howarth, R.; Bacon, S.; Apling, E.C.; Dilawari, J. 1977b. Treatment of diabetes with guar gum: reduction of urinary glucose loss in diabetics. Lancet 2:779-780.

Jenkins, D.J.A.; Reynolds, D.; Leeds, A.R.; Waller, A.L.; Cummings, J.H. 1979a. Hypocholesterolemic action of dietary fiber unrelated to fecal bulking effect. Am. J. Clin. Nutr. 32:2430-2435.

Jenkins, D.J.A.; Taylor, R.H.; Nineham, R.; Goff, D.V.; Bloom, S.R.; Sarson, D.; Alberti, K.G.M.M. 1979b. Combined use of guar and acarbose in reduction of postprandial glycaemia. Lancet 2:924-927.

1an

Jenkins, D.J.A.; Wolever, T.M.S.; Nineham, R.; Bacon, S.; Smith, R.; Hockaday, T.D.R. 1979c. Dietary fiber and diabetic therapy: a progressive effect with time. Adv. Exp. Med. Biol. 119:275-279.

Jenkins, D.J.A.; Hockaday, T.D.R.; Wolever, T.M.S.; Nineham, R.; Goff, D.V.; Haisman, P.; Charnock, R.; Taylor, R.H.; Bacon, S. 1979d. Dietary fibre and ketone bodies: reduced urinary 3-hydroxybutyrate excretion in diabetics on guar. Br. Med. J. 2:1555.

Jenkins, D.J.A.; Leeds, A.R.; Slavin, B.; Mann, J.; Jepson, EaM. l979e. Dietary fiber and blood lipids: reduction of serum cholesterol in type II hyperlipidemia by guar gum. Am. J. Clin. Nutr. 32:16-18.

Jenkins, D.JaA.; Thorne, M.J.; Camelon, K.; Jenkins, A.; Rao, A.V.; Taylor, R.H.; Thompson, L.U.; Kalmusky, J.; Reichert, R.; Francis, T. 1982. Effect of processing on digestibility and the blood glucose response: a study of lentils. Am. J. Clin. Nutr. 36:1093-1101.

Jenkins, D.J.A.; Wolever, T.M.S.; Jenkins, A.L.; Lee, R.; Wong, G.S.; Josse, R. 1983a. Glycemic response to wheat products: reduced response to pasta but no effect of fiber. Diabetes Care 6:155-159.

Jenkins, DJ A.; Wong, G.S.; Patten, R.; Bird, J.; Hall, M.; Buckley, G.C.; McGuire, V.; Reichert, R.; Little, J.A. 1983b. Leguminous seeds in the dietary management of hyperlipidemia. Am. J. Clin. Nutr. 38:567-573.

Jenkins, D.J.A.; Wolever, T.M.S.; Jenkins, A.L.; Thompson, L.U.; Rao, A.V.; Francis, T. 1986a. The glycemic index: blood glucose response to foods. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.167-179.

Jenkins, D.J.A.; Rainey-Macdonald, C.G.; Jenkins, A.L.; Benn, G. 1986be Fiber in the treatment of hyperlipidemia. In: Spiller, G.A , ed. CRC handbook of dietary fiber in human nutritiono Boca Raton, FL: CRC Press, Inc. p.327-344.

Jenkins, D.J.A.; Jenkins, A.L.; Wolever, T.M.S.; Rao, A.V.; Thompson, L.U. 1986c. Fiber and starchy foods: gut function and implications in disease. Am. J. Gastroenterol. 81:920-930.

Jenkins, D.J.A.; Wolever, T.M.S.; Kalmusky, J.; Giudici, S.; Giordano, C.; Wong, G.S.; Bird, J.N.; Patten, R.; Hall, M.; Buckley, G.; Little, J.A. 1985. Low glycemic index carbohydrate foods in the management of hyperlipidemia. Am. J. Clin. Nutr. 42:604-617.

191

Jenkins, D.J.A.; Wolever, T.M.S.; Leeds, A.R.; Gassull, M.A.; Haisman, P.; Dilawari, J.; Goff, D.V.; Metz, G.L.; Alberti, K.G.M.M. 1978a. Dietary fibres, fibre analogues, and glucose tolerance: importance of viscosity. Br. Med. J. 1:1392-1394~

Jenkins, D.J.A.; Wolever, T.M.S.; Nineham, R.; Taylor, R.; Metz, G.L.; Bacon, S.; Hockaday, T.D.R. 1978b. Guar crispbread in the diabetic diet. Br. Med. J. 2:1744-1746.

Jenkins, D.J.A.; Wolever, T.M.S.; Nineham, R.; Sarson, D.L.; Bloom, S.R.; Ahern, J.; Alberti, K.G.M.M.; Hockaday, T.D.R. 1980a. Improved glucose tolerance four hours after taking guar with glucose. Diabetologia 19:21-24.

Jenkins, D.J.A.; Wolever, T.M.S.; Bacon, S.; Nineham, R.; Lees, R.; Rowden, R.; Love, M.; Hockaday, T.D.R. 1980b. Diabetic diets: high carbohydrate combined with high fiber. Am. J. Clin. Nutr. 33:1729-1733.

Jenkins, D.J.A.; Reynolds, D.; Slavin, B.; Leeds, A.R.; Jenkins, A.L.; Jepson, E.M. 1980c. Dietary fiber and blood lipids: treatment of hypercholesterolemia with guar crispbread. Am. J. Clin. Nutr. 33:575-581.

Jensen, O.M.; Maclennan, R.; Wahrendorf, J. 1982. Diet, bowel function, fecal characteristics, and large bowel cancer in Denmark and Finland. Nutr. Cancer 4:5-19.

Johansen, K. 1981. Decreased urinary glucose excretion and plasma cholesterol level in non-insulin dependent diabetic patients with guar. Diabete Metab. (Paris) 7:87-90.

Johnson, E.J.; Marlett, J.A. 1986. A simple method to estimate neutral detergent fiber content of typical daily menus. Am. J. Clin. Nutr. 44:127-134.

Johnson, I.T.; Gee, J.M. 1981. Effect of gel-forming gums on the intestinal unstirred layer and sugar transport in vitro. Gut 22:398-403. -

Johnson, I.T.; Gee, J.M. 1986. Gastrointestinal adaptation in response to soluble non-available polysaccharides in the rat. Br. J. Nutr. 55:497-505.

Johnson, I.T.; Gee, J.M.; Mahoney, R.R. 1984. Effect of die­tary supplements of guar gum and cellulose on intestinal cell proliferation, enzyme levels and sugar transport in the rat. Br. J. Nutr. 52:477-487.

Jones, V.A.; Dickinson, R.J.; Workman, E.; Wilson, A.J.; Freeman, A.H.; Hunter, J.O. 1985. Crohn's disease: maintenance of remission by diet. Lancet 2:177-180.

lQ?

Judd, P.A. 1982. The effects of high intakes of barley on gastrointestinal function and apparent digestibilities of dry matter, nitrogen and fat in human volunteers. J. Plant Foods 4:79-88.

Judd, P.A. 1985. Dietary fibre and gallstones. In: Leeds, A.R.; Avenell, A., eds. Dietary fibre perspectives: reviews and bibliography. London: John Libbey and Company Ltd. p.40-46.

Judd, Pe A.; Truswell, A.S. 1981. The effect of rolled oats on blood lipids and fecal steroid excretion in man. Am. J. Clin. Nutr. 34:2061-2067.

Judd, P.A.; Truswell, A.S. 1982. Comparison of the effects of high- and low-methoxyl pectins on blood and faecal lipids in man. Br. J. Nutr. 48:451-458.

Judd, P9A9; Truswell, A.Se 1985. Dietary fibre and blood lipids in man. In: Leeds, A.R.; Avenell, A., eds. Dietary fibre perspectives: reviews and bibliographye London: John Libbey and Company Ltd. p.23-39.

Kahaner, N.; Fuchs, H.-M.; Floch, M.He 1976. The effect of dietary fiber supplementation in man. I. Modification of eating habits. Am. Jo Clin. Nutr. 29:1437-1442.

Kameda, H.; Ishihara, F.; Shibata, K.; Tsukie, E 1984. Clinical and nutritional study on gallstone disease in Japan. Jpn J Med 23:109-113

Kang, J.Y.; Doe., W.F. 1979. Unprocessed bran causing intestinal obstruction. Br. Med. J. 1:1249-1250.

Kanter, Y.; Eitan, N.; Brook, G.; Barzilai, D. 1980. Improved glucose tolerance and insulin response in obese and diabetic patients on a fiber-enriched diet. Isr. J. Med. Sci. 16:1-6.

Karlstrom, B.; Vessby, B.; Asp, N.-G.; Boberg, M.; Gustafsson, I.-B.; Lithell, H.; Werner, I. 19840 Effects of an increased content of cereal fibre in the diet of Type 2 (non-insulin­dependent) diabetic patients. Diabetologia 26:272=277.

Kasper, H. 1986. Effects of dietary fiber on vitamin metabo­lism. In: Spiller, G.A., ed. CRC handbook of dietary fiber in human nutritiono Boca Raton, FL: CRC Press, Inc. p.201-208.

Kasper, H.; Eilles, C.; Reiners, C.; Schrezenmeir, J. 1985. The influence of dietary fiber on gastric transit time. Hepatogastroenterology 32:69-71.

Kasper, H.; Rabast, U.; Fassl, H.; Fehle, F. 1979. The effect of dietary fiber on the postprandial serum vitamin A concentra­tion in man. Am. J. Clin. Nutr. 32:1847-1849.

193

Kasper, H.; Sommer, H. 1979. Dietary fiber and nutrient intake in Crohn's disease. Am. J. Clin. Nutr. 32:1898-1901.

Kaur, A.P.; Bhat, C.M.; Godara, R.B. 1981. Effect of cellulose on serum lipids in adolescent girls. J. Hum. Nutr. 35:456-458.

Kaur, AaP.; Bhat, C.M.; Grewal, R.B. 1985. Effect of cellu­lose incorporation in a low fiber diet on fecal excretion and digestibility of nutrients in adolescent girls. Nutr. Rep. Int. 32:383-388.

Kay, R.M. 1982. Dietary fiber. J. Lipid Res. 23:221-242.

Kay, R.M.; Grobin, W.; Track, N.S. 1981. Diets rich in natural fibre improve carbohydrate tolerance in maturity-onset, non­insulin dependent diabetics. Diabetologia 20:18-21.

Kay, R.M.; Jacobs, M.; Katan, M.B.; Lewis, B. 1985. Relation­ship between changes in plasma lipoprotein concentrations and fecal steroid excretion in man during consumption of four experimental diets. Atherosclerosis 55:15-23.

Kay, R.M.; Sabry, Z.I.; Csima, A. 1980. Multivariate analysis of diet and serum lipids in normal men. Am. J. Clin. Nutr. 33:2566-2572.

Kay, R.M.; Stitt, S. 1978. Food form, postprandial glycemia, and satiety. Am. J. Clin. Nutr. 31:738-739.

Kay, R.M.; Strasberg, S.M.; Petrunka, C.N.; WaymBn, M. 1979. Differential adsorption of bile acids by lignins. In: Inglett, G.E.; Falkehag, S.I., eds. Dietary fibers: chemistry and nutri­tion. New York: Academic Press. p.57-65.

Kay, R.M.; Truswell, A.S. 1977a. Effect of citrus pectin on blood lipids and fecal steroid excretion in man. Am. J. Clin. Nutr. 30:171-175.

Kay, R.M.; Truswell, A.S. 1977b. Effect of wheat fibre on plasma lipids and faecal steroid excretion in man. Br. J. Nutr. 37:227-235.

Kay, R.M.; Truswell, A.S. 1980. Dietary fiber: effects on plasma and biliary lipids in man. In: Spiller, G.A.; Kay, R.M., eds. Medical aspects of dietary fiber. New York: Plenum Medical Book Company. p.153-173.

Kelsay, J.L. 1982. Effects of fiber on mineral and vitamin bioavailability. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber in health and disease. New York: Plenum Press. p.91-103.

194

Kelsay, J.L. 1986. Update on fiber and mineral availability. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.361-372.

Kelsay, J.L.; Behall, K.M.; Prather, E.S. 1978. Effect of fiber from fruits and vegetables on metabolic responses of human subjects. I. Bowel transit time, number of defecations, fecal weight, urinary excretions of energy and nitrogen and apparent dig~stibilities of energy, nitrogen, and fat. Am. J. Clin. Nutr. 31:1149-1153.

Kelsay, J.L.; Behall, K.M.; Prather, E.S. 1979a. Effect of fiber from fruits and vegetables on metabolic responses of human subjects. II. Calcium, magnesium, iron, and silicon balances. Am. J. Clin. Nutr. 32:1876-1880.

Kelsay, J.L.; Jacob, R.A.; Prather, E.S. 1979b. Effect of fiber from fruits and vegetables on metabolic responses of human sub­jects. III. Zinc, copper, and phosphorus balances. Am. J. Clin. Nutr. 32:2307-2311.

Kelsay, J.L.; Clark, W.M. 1984. Fiber intakes, stool frequency, and stool weights of subjects consuming self-selected diets. Am. J. Clin. Nutr. 40:1357-1360.

Kelsay, J.L.; Clark, WoM.; Herbst, B.J.; Prather, E.S. 1981. Nutrient utilization by human subjects consuming fruits and vegetables as sources of fibero Jo Agric. Food Chern. 29:461-4650

Kelsay, J.L.; Prather, E.S. 1983. Mineral balances of human subjects consuming spinach in a low-fiber diet and in a diet containing fruits and vegetables. Am. J. Clin. Nutr. 38:12-19.

Keltz, F.R.; Kies, C.; Fox, H.M. 1978. Urinary ascorbic acid excretion in the human as affected by dietary fiber and zinc. Am. Jo Clin. Nutr. 31:1167-1171.

Keys, A.; Anderson, JoT.; Grande, Fo 1960. constant) and blood lipids in man. J. Nutr.

Diet-type (fats 70:257-266.

Keys, A.; Grande, F.; Anderson, J.T. 1961. Fiber and pectin in the diet and serum cholesterol concentration in man. Proc. Soc. Exp. Biol. Med. 106:555-558.

Khan, A.R.; Khan, G.Y.; Mitchel, A.; Qadeer, M.Ao 1981. Effect of guar gum on blood lipids~ Am. J. Clin. Nutr. 34:2446-2449.

Kiehm, T.G.; Anderson, JoW.; Ward, K. 1976. Beneficial effects of a high carbohydrate, high fiber diet on hyperglycemic diabetic men. Am. J. Clin. Nutr. 29:895-899.

195

Kies, C. 1985. Non-soluble dietary fiber effects on lipid absorption and blood serum lipid patterns. Lipids 20:802-807.

Kies, C.; Fox, H.M. 1977. Dietary hemicellulose interactions influencing serum lipid patterns and protein nutritional status of adult men. J. Food Sci. 42:440-443.

Kies, C.; Fox, H.M.; Beshgetoor, D. 1979. Effect of various levels of dietary hemicellulose on zinc nutritional status of men. Cereal Chem. 56:133-136.

Kies, C.; Young, E.; McEndree, L. 1983. Zinc bioavailability from vegetarian diets: influence of dietary fiber, ascorbic acid, and past dietary practices. In: Inglett, G.E., ed. Nutritional bioavailability of zinc. ACS Series No. 210. Washington, DC: American Chemical Society. p.115-126.

King, J.C.; Stein, T.; Doyle, M. 1981. Effect of vegetarian~sm on the zinc status of pregnant women. Am. J. Clin. Nutr. 34:1049-1055.

Kinmonth, A.-L.; Angus, R.M.; Jenkins, P.A.; Smith, M.A.; Baum, J.D. 1982. Whole foods and increased dietary fibre improve blood glucose control in diabetic children. Arch. Dis. Child. 57:187-194.

Kirby, R.W.; Anderson, J.W.; Sieling, B.; Rees, E.D.; Chen, W.-J. L.; Miller, R.E.; Kay, R.M. 1981. Oat-bran intake selectively lowers serum low-density lipoprotein cholesterol concentrations of hypercholesterolemic men. Am. J. Clin. Nutr. 34:824-829.

Kirwan, W.O.; Smith, A.N.; McConnell, A.A.; Mitchell, W.D.; Eastwood, M.A. 1974. Action of different bran preparations on colonic function. Br. Med. J. 4:187-189.

Knox, E.G. 1977. Foods and diseases. Br. J. Prev. Soc. Med. 31:71-80.

Knuiman, J.T.; West, C.E. 1982. The concentration of choles­terol in serum and in various serum lipoproteins in macrobiotic, vegetarian and non-vegetarian men and boys. Atherosclerosis 43:71-82.

Koepp, P.; Hegewisch, S. 1981. Effects of guar on plasma viscosity and related parameters in diabetic children. Eur. J. Pediatr. 137:31-33.

Kojima, N.; Wallace, D.; Bates, G.W. 1981. The effect of chemical agents, beverages, and spinach on the in vitro solubilization of iron from cooked pinto beans. Am. J. Clin. Nutr. 34:1392-1401.

,o~

Kotb, A.R.; Luckey, T.D. 1972. Markers in nutrition. Nutr. Abstr. Rev. 42:813-845.

Kretsch, M.J.; Crawford, L.; Calloway, D.H. 1979. Some aspects of bile acid and urobilinogen excretion and fecal elimination in men given a rural Guatemalan diet and egg formulas with and without added oat bran. Am. J. Clin. Nutr. 32:1492-1496.

Kritchevsky, D.; Davidson, L.M.; Goodman, G.T.; Tepper, S.A.; Mendelsohn, D. 1986a. Influence of dietary fiber on lipids and aortic composition of vervet monkeys. Lipids 21:338-341.

Kritchevsky, D.; Davidson, L.M.; Van der Watt, J.J.; Mendelsohn, D. 1986b. Influence of dietary fiber on aortic sudanophilia in vervet monkeys fed "Western" diet. In: Atherosclerosis& VII. Proceedings of the Seventh International Symposium on Atherosclerosis. Amsterdam: Elsevier Science Publishers. p.663-666.

Kritchevsky, D.; Davidson, L.Mo; Krendel, O.A.; Van der Watt, J.J.; Russell, D.; Friedland, S.; Mendelsohn, D. 1981. Influence of dietary fiber on aortic sudanophilia in vervet monkeys. Ann. Nutr. Metab. 25:125-136.

Kritchevsky, Do; Story, JoA. 1974. Binding of bile salts in vitro by nonnutritive fiber. J. Nutr. 104:458-462.

Kritchevsky, D.; Story, J A. 1983 Binding of bile salts, but not neutral steroids, and tumor promoters in the colon in man and in rodents. Environ. Health Perspect. 50:101-107.

Kritchevsky, D.· Tepper, S.A 19650 Factors affecting atherosclerosis in rabbits fed cholesterol-free diets. Life Sci. 4:1467-14710

Kritchevsky, D.; Tepper, S.A.; Goodman, G. 1984a. Diet, nutri­tion intake, and metabolism in populations at high and low risk for colon cancer: relationship of diet to serum lipids. Am. J. Clin. Nutr. 40:921-926.

Kritchevsky, D.; Tepper, S.A.; Klurfeld, D.M. 1984b. Effect of pectin and cellulose on formation and regression of gallstones in hamsters. Experientia 40:350-351.

Kritchevsky, D.; Tepper, S.A.; Williams, D.E.; Story, J.A~ 1977. Experimental atherosclerosis in rabbits fed cholesterol-free diets. Part 7. Interaction of animal or vegetable protein with fiber. Atherosclerosis 26:397-403.

Kroes, R.; Seems, R.B.; Bosland, M.C.; Bunnik, G.S.J.; Sinkeldam, EoJ. 1986. Nutritional factors in lung, colon, and prostate carcinogenesis in animal models. Fed. Proc. Fed. Am. Soc. Exp. Biol. 45:136-141.

197

Kromhout, D.; Bosschieter, E.B.; de Lezenne Coulander, C. 1982. Dietary fibre and 10-year mortality from coronary heart disease, cancer, and all causes: the Zutphen study. Lancet 2:518-522.

Krotkiewski, M. 1984. Effect of guar gum on body-weight, hunger ratings and metabolism in obese subjects. Br. J. Nutr. 52:97-105.

Krotkiewski, M. 1985. Use of fibres reduction programs. In: Bjoerntorp, Kritchevsky, D., eds. Current topics Vol. 14. Dietary fiber and obesity. Inc. p.85-109.

in different weight P.; Vahouny, G.V.; in nutrition and disease. New York: Alan R. Liss,

KUhl, C.; M0lsted-Pedersen, L.; Hornnes, P.J. 1983. Guar gum and glycemic control of pregnant insulin-dependent diabetic patients. Diabetes Care 6:152-154. ·

Kuratsune, M.; Honda, T.; Englyst, H.N.; Cummings, J.H. 1986. Dietary fiber in the Japanese diet as investigated in connection with colon cancer risk. Jpn. J. Cancer Res. 77:736-738.

Kushi, L.H.; Lew, R.A.; Stare, F.J.; Ellison, C.R.; el Lazy, M.; Bourke, G.; Daly, L.; Graham, I.; Hickey, N.; Mulcahy, R.; Kevaney, J. 1985. Diet and 20-year mortality from coronary heart disease: the Ireland-Boston Diet-Heart Study. N. Engl. J. Med. 312:811-818.

Kyllastinen, M.; Lahikainen, T. 1981. Long-term dietary supple­mentation with a fiber product (guar gum) in elderly diabetics. Curr. Ther. Res. 30:872-879.

La Vecchia, C.; Decarli, A.; Fasoli, M.; Gentile, A. 1986. Nutrition and diet in the etiology of endometrial cancer. Cancer 57:1248-1253.

Laine, R.A.; Vara, P.; Koivistoinen, P.E. 1981. Observations on the analysis of dietary fiber. In: James, W.P.T.; Theander, O., eds. The analysis of dietary fiber in food. New York: Marcel Dekker, Inc. p.21-27.

LaMont, J.T.; Isselbacher, K.J. 1983. Diseases of the small and large intestine. In: Petersdorf, R.G.; Adams, R.D.; Braunwald, E.; Isselbacher, K.J.; Martin, J.B.; Wilson, J.D., eds. Harrison's principles of internal medicine. 9th ed. New York: McGraw-Hill Book Company. p.1420.

Lanza, E.; Jones, D.Y.; Block, G.; Kessler, L. [1987]. fiber intake in the US population. Am. J. Clin. Nutr.

Dietary In press.

Larson, D.M.; Masters, S.S.; Spiro, H.M. surgical therapy in diverticular disease: Gastroenterology 71:734-737.

198

1976. Medical and a comparative study.

Leatherdale, B.A.; Green, D.J.; Harding, L.K.; Griffen, D.; Bailey, C.J. 1982. Guar and gastric emptying in non-insulin dependent diabetes. Acta Diabetol. Lat. 19:339-343.

Leeds, A.R. 1982. Modification of intestinal absorption by dietary fiber and fiber components. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber in health and disease. New York: Plenum Press. p.53-71.

Leeds, A.R.; Ralphs, D.N.L.; Ebied, F.; Metz, G.; Dilawari, J.B. 1981. Pectin in the dumping syndrome: reduction of symptoms and plasma volume changes. Lancet 1:1075-1078.

Leigh, M.J.; Miller, D.D. 1983. Effects of pH and chelating agents on iron binding by dietary fiber: implications for iron availability. Am. J. Clin. Nutr. 38:202-213.

Leklem, J.E.; Miller, L.T.; Perera, A.O.; Peffers, D.E. 1980. Bioavailability of vitamin 8-6 from wheat bread in humans. J. Nutr. 110:1819-1828.

Lembcke, B.; Ebert, R.; Ptak, M.; Caspary, W.F.; Creutzfeldt, W.; Schicha, H.; Emrich, D. 1984. Role of gastrointestinal transit in the delay of absorption by viscous fibre (guar). Hepato­gastroenterology 31:183-186.

Lennard-Jones, J.E.; Fletcher, J.; Shaw, D.G. 1968. Effect of different foods on the acidity of the gastric contents in patients with duodenal ulcer. Part III. Effect of altering the proportions of protein and carbohydrate. Gut 9:177-182.

Levenstein, S., Prantera, C.; Luzi, C., D1 Ubaldi, A. 1985. Low residue or normal diet in Crohn's disease: a prospective controlled study in Italian patients. Gut 26:989-993.

Levitt, N.S.; Vinik, A.L; Sive, A.A.; Child, P.T.; Jackson, W.P.U. 1980. The effect of dietary fiber on glucose and hormone responses to a mixed meal in normal subjects and in diabetic subjects with and without autonomic neuropathy. Diabetes Care 3:515-519.

Lieberthal, M.M.; Martens, ReA. 1975. Lowered serum cholesterol following the ingestion of a hydrophilic colloid. Am. J. Dig. Dis. 20:469-474.

.. Lindahl, O.; Lindwall, L.; Spangberg, A.; Stenram, A.; Ockerman, P.A. 1984. A vegan regimen with reduced medication in the treatment of hypertension. Br. Je Nutr. 52:11-20.

Linder, P.; Moller, Be 1973. Lignin: a cholesterol-lowering agent? lancet 2.1259-1260.

199

Lindgarde, .F.; Larsson, L. 1984. Effects of a concentrated bran fibre preparation on HDL-cholesterol in hypercholesterolaemic men. Hum. Nutr. Clin. Nutr. 38C:39-45.

Lindsay, A.N.; Hardy, S.; Jarrett, L.; Rallison, M.L. l984e High-carbohydrate, high-fiber diet in children with type 1 diabetes mellitus. Diabetes Care 7:63-67.

Lipkin, M.; Newmark, H. 1985. Effect of added dietary calcium on colonic epithelial-cell proliferation in subjects at high risk for familial colonic cancer. N. Engl. J. Med. 313:1381-1384.

Lipkin, M.; Uehara, K.; Winawer, S.; Sanchez, A.; Bauer, C.; Phillips, R.; Lynch, H. T.; Blattner, W.A.; Fraumeni, J.F., Jr. 1985. Seventh-Day Adventist vegetarians have a quiescent pro­liferative activity in colonic mucosa. Cancer Lett. 26:139-144.

Lithell, H.; Selinus, I.; Vessby, 8. 1984. Lack of effect of a purified bran preparation in men with low HDL cholesterol. Hum. Nutr. Clin. Nutr. 38C:309-313.

Liu, K.; Stamler, J.; Moss, D.; Garside, D.; Persky, Vo; Soltero, I. 1979. Dietary cholesterol, fat, and fibre, and colon-cancer mortality. Lancet 2:782-785.

Liu, K.; Stamler, J.; Trevisan, M.; Moss, D. 1982. Dietary lipids, sugar, fiber, and mortality from coronary heart disease: bivariate analysis of international data. Arteriosclerosis 2:221-227.

Lubin, F.; Wax, Y.; Madan, 8. 1986. Role of fat, animal protein, and dietary fiber in breast cancer etiology: a case-control study. J. Natl. Cancer Inst. 77:605-612.

Luyken, R.; Pikaar, N.A.; Pelman, H.; Schippers, F.A. 1962. The influence of legumes on the serum cholesterol level. Voeding 23:447-453.

Lykken, G.I.; Mahalko, J.; Johnson, P.E.; Milne, o.; Sandstead, H.H.; Garcia, W.J.; Ointzis, F.R.; Inglett, G.E. 1986. Effect of browned and unbrowned corn products intrinsically labeled with 65Zn on absorption of 65Zn in humans. J. Nutr. 116:795-801.

Lyon, D.S. 1984. Studies on the solubility of Ca, Mg, Zn, and Cu in cereal products. Am. J. Clin. Nutr. 39:190-195.

Lyon, J.L.; Sorenson, A.W. 1978. Colon cancer in a low-risk population. Am. J. Clin. Nutr. 31:S227-S230.

MacGregor, I.L.; Martin, P.; Meyer, J.H. 1977. Gastric emptying of solid food in normal man and after subtotal gastrectomy and truncal vagotomy with pyloroplasty. Gastroenterology 72:206-211.

200

Mahalko, J.R.; Sandstead, H.H.; Johnson, L.K.; Inman, L.F.; Milne, D.B.; Warner, R.C.; Haunz, E.A. 1984. Effect of consuming fiber from corn bran, soy hulls, or apple powder on glucose tolerance and plasma lipids in type II diabetes. Am. J. Clin. Nutr. 39:25-34.

Manhire, A.; Henry, C.L.; Hartog, M.; Heaton, K.W. 1981. Unrefined carbohydrate and dietary fibre in treatment of diabetes mellitus. J. Hum. Nutr. 35:99-101.

Mann, J. 1985. Diabetes mellitus: some aspects of aetiology and management of non-insulin-dependent diabetes. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.263-287.

Manning, A.P.; Heaton, K.W. 19760 Bran and the irritable bowel. Lancet 1:588.

Manning, A.P.; Heaton, K.W.; Harvey, R.F.; Uglow, P. 1977e Wheat fibre and irritable bowel syndrome: a controlled trial. Lancet 2:417-418.

Manousos, O.; Day, N.E.; Trichopoulos, D.; Gerovassilis, F.; Tzonou, A.; Polychronopoulou, A. 1983. Diet and colorectal cancer: a case-control study in Greece. Int. J. Cancer 32:1-5.

Marlett, J.A.; Balasubramanian, R.; Johnson, E.J.; Draper, N.R. 1986. Determining compliance with a dietary fiber supplement. J. Natl. Cancer Inst. 76:1065-1070.

Marlett, J.A.; Bokram, R.L. 1981. Relationship between calculated dietary and crude fiber intakes of 200 college students. Am. J. Clin. Nutr. 34:335-342.

Marlett, J.A.; Chesters, J.G. 1985. Measuring dietary fiber in human foodso J. Food Sci. 50:410-414, 423.

Marlett, J.A.; Johnson, E.J. 1986. University of Wisconsin, Madison. Personal communication with S.M. Pilch, Federation of American Societies for Experimental Biology, Bethesda, MD.

Marlett, J.A.; Lee, S.Co 19800 Dietary fiber, lignocellulose and hemicellulose contents of selected foods determined by modified and unmodified Van Soest procedures. J. Food Sci. 45:1688-1693.

Marlett, J.A.; Li, B.U.K.; Patraw, C.J.; Bass, P. 1987. Comparative laxation of psyllium with and without senna in an ambulatory constipated population. Am. J. Gastroenterol. 82:333-337.

201

Marlett, J.A.; Slavin, J.L.; Brauer, P.M. 1981. Comparison of dye and pellet gastrointestinal transit time during controlled diets differing in protein and fiber levels. Dig. Dis. Sci. 26:208-213.

Marsh, A.G.; Sanchez, T.V.; Chaffee, F.L.; Mayor, G.H.; Mickelsen, 0. 1983. Bone mineral mass in adult lacto­ovo-vegetarian and omnivorous males. Am. J. Clin. Nutr. 37:453-456.

Marsh, A.G.; Sanchez, T.V.; Mickelsen, O.; Keiser, J.; Mayor, G. 1980. Cortical bone density of adult lacto-ovo-vegetarian and omnivorous women. J. Arn. Diet. Assoc. 76:148-151.

Martinez, I.; Torres, R.; Frias, Z.; Colon, J.R.; Fernandez, N. 1979. Factors associated with adenocarcinomas of the large bowel in Puerto Rico. In: Birch, J.M., ed. Advances in medical oncology, research and education. Vol. 3. Epidemiology. New York: Pergamon Press. p.45-52.

Martini, G.A.; Brandes, J.W. 1976. Increased consumption of refined carbohydrates in patients with Crohn's disease. Klin. Wochenschr. 54:367-371. ·

Masarei, J.R.L.; Rouse, I.L.; Lynch, W.J.; Robertson, K.; Vandongen, R.; Beilin, L.J. 1984. Effects of a lacto-ovo vegetarian diet on serum concentrations of cholesterol, triglyceride, HDL-C, HDL2-C, HDL3-C, apoprotein-B, and Lp(a). Am. J. Clin. Nutr. 40:468-479.

Mastromarino, A.; Reddy, B.S.; Wynder, E.L. 1976. Metabolic epidemiology of colon cancer: enzymic activity of fecal flora. Am. J. Clin. Nutr. 29:1455-1460.

Mathur, K.S.; Khan, M.A.; Sharma, R.D. 1968. Hypocholester­olaemic effect of Bengal gram: a long-term study in man. Br. Med. J. 1:30-31.

Mathur, M.S.; Singh, F.; Chadda, V.S. 1977. Effect of bran on blood lipids. J. Assoc. Phys. India 25:275-278.

Matthee, V.; Appledorf, H. 1978. Effect of cookin~ on vegetable fiber. J. Food Sci. 43:1344-1345.

Mayne, P.O.; McGill, A.R.; Gormley, T .R.; Tomkin, G.H.; Julian, T.R.; □ 'Moore, R.R. 1982. The effect of apple fibre on diabetic control and plasma lipids. Ir. J. Med. Sci. 151:36-41.

McBurney, M.I.; Horvath, P.J.; Jeraci, J.L.; Van Soest, P.J. 1985. Effect of in vitro fermentation using human faecal inoculum on the water-holding capacity of dietary fibre. Br. J. Nutr. 53:17-24.

McBurney, M.I.; Van Soest, P.J.; Chase, L.E. 1983. Cation exchange capacity and buffering capacity of neutral-detergent fibres. J. Sci. Food Agric. 34:910-916.

Mccance, R.A.; Prior, K.M.; Widdowson, E.M. 1953. A radio­logical study on the rate of passage of brown and white bread through the digestive tract of man. Br. J. Nutr. 7:98-104.

Mccance, R.A.; Widdowson, E.M. 1942a. Mineral metabolism of healthy adults on white and brown bread dietaries. J. Physiol. 101:44-85.

Mccance, R.A.; Widdowson, E.M. 1942b. Mineral metabolism on dephytinized bread. Jo Physiol. 101:304-313.

McConnell, A.Ao; Eastwood, M.A.; Mitchell, W.D. 1974. Physical characteristics of vegetable foodstuffs that could influence bowel function. J. Sci. Food Agric. 25:1457-1464.

McCulloch, DoKo; Mitchell, R.Do; Ambler, J.; Tattersall, R.B. 1985. A prospective comparison of 'conventional' and high carbohydrate/high fibre/low fat diets in adults with established type 1 (insulin-dependent) diabetes. Diabetologia 28:208-212.

McDougall, R.M.; Yakymyshyn, L.; Walker, K.; Thurston, O.G. 1978. Effect of wheat bran on serum lipoproteins and biliary lipids. Can. J. Surg. 21:433-435e

Mcivor, M.E.; Cummings, C.C.; Mendeloff, A.I. 1985. Long-term ingestion of guar gum is not toxic in patients with noninsulin-dependent diabetes mellitus Am. J. Clin. Nutr. 41:891-8940

Mcivor, M.E., Cummings, C.C.; Van Duyn, M.A.; Leo, T.A.; Margolis, S.; Behall, K.M.; Michnowski, J.E.; Mendeloff, A.I. 1986. Long-term effects of guar gum on blood lipids. Athero­sclerosis 60:7-13.

McKeown-Eyssen, G.E.; Bright-See, E. 1984. Dietary factors in colon cancer: international relationships. Nutr. Cancer 6:160-170.

McKeown-Eyssen, GeEe; Thomas, D.C. 1985e Sample size determination in case-control studies: the influence of the distribution of exposure. Jo Chronic Dis. 38:559-568.

McNaughton, J.P.; Morrison, D.D.; Huhner, LoJ.; Earnest, M.M.; Ellis, M.A.; Howell, G.L. 1985. Changes in total serum cholesterol levels of diabetics fed five grams guar gum daily. Nutr. Rep. Int. 31:505-520.

McNeil, N.I.; Cummings, J.H.; James, W.P.T. 1978. Short chain fatty acid absorption by the human large intestine. Gut 19:819-822.

203

McNeil, N.I.; Cummings, J.H.; James, W.P.T. 1979. Rectal absorption of short chain fatty acids in the absence of chloride. Gut 20:400-403.

Mendeloff, A.I. 1986. Thoughts on the epidemiology of diverticular disease. Clin. Gastroenterol. 15:855-877.

Meyer, P.O.; DenBesten, L.; Mason, E.E. 1979. The effects of a high-fiber diet on bile acid pool size, bile acid kinetics, and biliary lipid secretory rates in the morbidly obese. Surgery 85:311-316.

Mickelsen, O.; Makdani, D.D.; Cotton, R.H.; Titcomb, S.T.; Colmey, J.C.; Gatty, R. 1979. Effects of a high fiber bread diet on weight loss in college-age males. Am. J. Clin. Nutr. 32:1703-1709.

Miettinen, T.A. 1980. Effects of dietary fiber on serum lipids and cholesterol metabolism in man. In: Gotto, A.M., Jr.; Smith, L.C.; Allen, B., eds. Atherosclerosis V: proceedings of the Fifth International Symposium. New York: Springer-Verlag. p.311-315.

Miettinen, T.A.; Tarpila, S. 1977. Effect of pectin on serum cholesterol, fecal bile acids and biliary lipids in normo­lipidemic and hyperlipidemic individuals. Clin. Chim. Acta 79:471-477.

Miller, L.T.; Shultz, T.D.; Leklem, J.E. 1980. Influence of citrus pectin on the bioavailability of vitamin B6 in men. Fed. Proc. Fed. Am. Soc. Exp. Biol. 39:797 (Abstract).

Miranda, P.M.; Horwitz, D.L. 1978. High-fiber diets in the treatment of diabetes mellitus. Ann. Intern. Med. 88:482-486.

Mitchell, W.D.; Eastwood, M.A. 1976. Dietary fiber and colon function. In: Spiller, G.A.; Amen, R.J., eds. Fiber in human nutrition. New York: Plenum Press. p.185-206.

Mod, R.R.; Ory, R.L.; Morris, N.M.; Normand, F.L. 1982. In vitro interaction of rice hemicellulose with trace minerals and their release by digestive enzymes. Cereal Chem. 59:538-542.

Madan, B.; Barell, V.; Lubin, F.; Madan, M.; Greenberg, R.A.; Graham, S. 1975. Low-fiber intake as an etiologic factor in cancer of the colon. J. Natl. Cancer Inst. 55:15-18.

Madan, B.; Cuckle, H.; Lubin, J. 1981. A note on the role of dietary retinal and carotene in human gastrointestinal cancer. Int. J. Cancer 28:421-424.

Mongeau, R.; Behrens, W.A.; Madere, R.; Brassard, R. 1986. Effects of dietary fiber on vitamin E status in rats: dose­response to wheat bran. Nutr. Res. 6:215-224.

Mongeau, R.; Brassard, R. 1982. Insoluble dietary fiber from breakfast cereals and brans: bile salt binding and water-holding capacity in relation to particle size. Cereal Chem. 59:413-4170

Mongeau, R.; Brassard, R. 1986. A rapid method for 1 the deter­mination of soluble and insoluble dietary fiber: comparison with AOAC total dietary fiber procedure and Englyst•s method. J. Food Scio 51:1333-1336.

Monnier, L.; Pham, T.C.; Aguirre, L.; Orsetti, A.; Mirouze, J. 1978. Influence of indigestible fibers on glucose tolerance. Diabetes Care 1:83-88.

Monnier, L.H.; Blotman, M.J.; Colette, C.; Monnier, M.P.; Mirouze, J. 1981. Effects of dietary fibre supplementation in stable and labile insulin dependent diabetics. Diabetologia 20:12-17.

Monnier, L.H.; Colette, C.; Aguirre, L.; Orsetti, A.; Combeaux, D. 1982. Restored synergistic entero-hormonal response after addition of dietary fibre to patients with impaired glucose tolerance and reactive hypoglycaemia. Diabete Metab. (Paris) 8:217-222.

Moore, D.J.; White, F.J.; Flatt, P.R.; Parke, D.V. 1985. Bene­ficial short-term effects of unprocessed wheat bran on lipid and glucose metabolism in man. Hum. Nutr. Clin. Nutr. 39C:63-67.

Moore, J.H. 1967. The effect of the type of roughage in the diet on plasma cholesterol levels and aortic atherosis in rabbits. Br. J. Nutr. 21:207-215.

Moore, W.E.C. 1978. Workshop II -- fiber and bacteria in the gut: summary and recommendations. Am. J. Clin. Nutr. 31:Slll­S112.

Morgan, L.M.; Goulder, T.J.; Tsiolakis, D.; Marks, V.; Alberti, K.G.M.M. 1979. The effect of unabsorbable carbohydrate on gut hormones: modification of post-prandial GIP secretion by guar. Diabetologia 17:85-89.

Morgenstern, H. 1982. Uses of ecologic analysis in epidemio­logic research. Am. J. Public Health 72:1336-1344.

Morris, J.N.; Marr, J.We; Clayton, D.G. 1977. Diet and heart: a postscript. Br. Med. J. 2:1307-1314

Munoz, J.M.; Sandstead, H.H.; Jacob, R.A. 1979. Effects of dietary fiber on glucose tolerance of normal men. Diabetes 28:496-502Q

Murphy, S.Pe; Calloway, D.H. 1986. Nutrient intakes of women in NHANES II, emphasizing trace minerals, fiber, and phytate. J. Am. Diet. Assoc. 86:1366-1372.

205

Najemnik, C.; Kritz, H.; Irsigler, K.~; Laube, H.; Knick, B.; Klimm, H.D.; Wahl, P.; Vollmar, J.; Brauning, C. 1984. Guar and its effects on metabolic control in type II diabetic sub­jects. Diabetes Care 7:215-220.

Nakamura, H.; Ishikawa, T.; Tada, N.; Kagami, Ao; Kondo, K.; Miyazima, E.; Takeyama, S. 1982. Effect of several kinds of dietary fibres on serum and lipoprotein lipids. Nutr. Rep. Int. 26:215-221.

National Advisory Committee on Nutrition Education. 1983. Proposals for nutritional guidelines for health education in Britain. London: Health Education Council.

National Cancer Institute. 1984. Diet, nutrition and cancer prevention: a guide to food choices. NIH Publication No. 85-2711. Available from: Office of Cancer Communications, National Cancer Institute, Bethesda, MD.

Neilson, M.J.; Marlett, J.A. 1983. A comparison between detergent and nondetergent analyses of dietary fiber in human foodstuffs, using high-performance liquid chromatography to measure neutral sugar composition. J. Agric. Food Chem. 31:1342-1347.

Nestel, P.J.; Nolan, C.; Bazelmans, J.; Cook, R. 1984. Effects of a high-starch diet with low or high fiber content on post­absorptive glucose utilization and glucose production in normal subjects. Diabetes Care 7:207-210.

Ney, D.; Hollingsworth, D.R.; Cousins, L. 1982. Decreased insulin requirement and improved control of diabetes in pregnant women given a high-carbohydrate, high-fiber, low-fat diet. Diabetes Care 5:529-533.

Nicklin, S.; Miller, K. 1984. Effects of orally administered food-grade carrageenans on antibody-mediated and cell-mediated immunity in the inbred rat. Food Chem. Toxicol. 22:615-621.

Nigro, N.D.; Bull, A.W.; Klopfer, B.A.; Pak, M.S.; Campbell, R.L. 1979. Effect of dietary fiber on azoxymethane-induced intestinal carcinogenesis in rats. J. Natl. Cancer Inst. 62:1097-1102.

Noble, J.A.; Grannis, W.F., Jr. 1984. Acute esophageal obstruction by a psyllium-based bulk laxative. Chest 86:800.

Nuzum, F.R.; Seegal, B.; Garland, R.; Osborne, M. 1926. Arteriosclerosis and increased blood pressure. Arch. Intern. Med. 37:733-744.

206

□ hi, G.; Minowa, K.; Oyama, T.; Nagahashi, M.; Yamazaki, N.; Yamamoto, S.-I.; Nagasako, K.; Hayakawa, K.; Kimura, K.; Mori, B. 1983. Changes in dietary fiber intake among Japanese in the 20th century: a relationship to the prevalence of diverticular disease. Am. J. Clin. Nutr. 37:115-121.

Ornstein, M.H.; Littlewood, E.R.; Baird, I.M.; Fowler, J.; North, W.R.S.; Cox, A.G. 1981. Are fibre supplements really necessary in diverticular disease of the colon: a controlled clinical trial. Br. Med. J. 282:1353-1356.

Osilesi, O.; Trout, Dal.; Glover, E.E.; Harper, S.M.; Koh, E.T.; Behall, K.M.; 0 1 Dorisio, T.M.; Tartt, J. 1985. Use of xanthan gum in dietary management qf diabetes mellitus. Am. J. Clin. Nutr. 42:597-603.

Owen, R.L.; Brandborg, L.L. 1977. Jejunal morphologic consequences of vegetarian diet in humans. Gastroenterology 72:A-88 (Abstract).

Pacy, P.J.; Dodson, P.M.; Fletcher, R.F. 1986. Effect of a high carbohydrate, low sodium and low fat diet in type 2 diabetics with moderate hypertension. Int. Jo Obesity 10:43-52.

Pacy, P.J.; Dodson, P.M.; Kubicki, A.Jo; Fletcher, R.F.; Taylor, K.G. 1984. Comparison of the hypotensive and metabolic effects of bendrofluazide therapy and a high fibre, low fat, low sodium diet in diabetic subjects with mild hypertension. J. Hyperten. 2:215-220.

Pahlke, G.; Friedrich, R. 1974. Persorption van mikrokris­talliner Cellulose. Naturwissenschaften 61:350

Painter, N. 1985. Diverticular disease of the colon. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.145-160.

Painter, N.S. 1975. Diverticular disease of the colon: a deficiency disease of Western civilization. London: Heinemann Medical Books.

Painter, N.S.; Almeida, A.Z.; Colebourne, K.W. 19720 Unprocessed bran in treatment of diverticular disease of the colon. Br. Medo J. 2:137-140.

Painter, N.S.; Burkitt, D.P. 1971. Diverticular disease of the colon: a deficiency disease of Western civilisation. Br. Med. J. 2:450-454.

Palmer, G.H.; Dixon, D.G. 1966. Effect of pectin dose on serum cholesterol levels. Am. J. Clin. Nutr. 18:437-442.

207

Palumbo, P.J.; Briones, E.R.; Nelson, R.A. 1978. High fiber diet in hyperlipemia: comparison with cholestyramine treatment in type IIa hyperlipoproteinemia. J. Am. Med. Assoc. 240:223-227.

Patraw, C.J.; Marlett, J.A. 1986. Variability in the dietary fiber content of wheat and mixed-grain commercial breads. Jo Am. Diet. Assoc. 86:794-796.

Paul, A.A.; Southgate, D.A. 1978. Mccance and Widdowson's the composition of foods. Amsterdam: Elsevier/North Holland Biomedical Press. 418p.

Payler, D.K.; Pomare, E.W.; Heaton, K.W.; Harvey, R.F. 1975. The effect of wheat bran on intestinal transit. Gut 16:209-213.

Pedersen, O.; Hj~llund, E.~ Lindskov, H.O.; Helms, P.; S~rensen, N.S.; Ditzel, J. 1982. Increased insulin receptor binding to monocytes from insulin-dependent diabetic patients after a low­fat, high-starch, high-fiber diet. Diabetes Care 5:284-291.

Penagini, R.; Velia, P.; Vigorelli, R.; Bozzani, A.; Castagnone, D.; Ranzi, T.; Bianchi, P.A. 1986. The effect of dietary guar on serum cholesterol, intestinal transit, and fecal output in man. Am. J. Gastroenterol. 81:123-125.

Persson, I.; Raby, K.; F~nss-Bech, P.; Jensen, E. 1976. Effect of prolonged bran administration on serum levels of cholesterol, ionized calcium and iron in the elderly. J. Am. Geriatr. Soc. 24:334-335.

Phillips, R.L. 1975. Role of life-style and dietary habits in risk of cancer among Seventh-Day Adventists. Cancer Res. 35:3513-3522.

Phillips,. R.L.; Lemon, F .R.; Beeson, W.L.; Kuzma, J.W. 1978. Coronary heart disease mortality among Seventh-Day Adventists with differing dietary habits: a preliminary report. Am. J. Clin. Nutr. 31:Sl91-Sl98.

Phillips, R.L.; Snowden, D.A. 1985. Dietary relationships with fatal colorectal cancer among Seventh-Day Adventists. J. Natl. Cancer Inst. 74:307-317.

Phillips, W.E.J.; Brien, R.L. 1970. Effect of pectin, a hypocholesterolemic polysaccharide, on vitamin A utilization in the rat. J. Nutr. 100:289-292 .

. Pickle, L.W.; Greene, M.H.; Ziegler, R.G.; Toledo, A.; Hoover, R.; Lynch, H.T.; Fraumeni, J.F., Jr. 1984. Colorectal cancer in rural Nebraska. Cancer Res. 44:363-369.

208

Plumley, P.F.; Francis, Ba 1973. Dietary management of diverticular disease. J. Am. Diet. Assoc. 63:527-530.

Plummer, S.M.; Kakefuda, T.; Harris, C.C. 1986. Further studies with fecapentaene 12. Prag. Clin. Biol. Res. 206:338 (Abstract).

Poksay, K.S.; Schneeman, B.O. 1983. Pancreatic and intestinal response to dietary guar gum in rats. J. Nutr. 113:1544-1549.

Pomare, E.W. 1983. Fibre and bile acid metabolism. Royal Soc. N. Z. Bull. 20:177-182.

Pomare, E.W.; Branch, W.J.; Cummings, J.H. 1985. Carbohydrate fermentation in the human colon and its relation to acetate concentrations in venous blood. J. Clin. Invest. 75:1448-1454.

Pomare, E.W.; Heaton, K.W. 1973. Alteration of bile salt metabolism by dietary fibre (bran). Br. Med. J. 4:262-264.

Pomare, E.W.; Heaton, K.W.; Low-Beer, TS.; Espiner, H.J 1976. The effect of wheat bran upon bile salt metabolism and upon the lipid composition of bile in gallstone patients. Am. J. Dig. Dis. 21:521-536.

Potter, J.D.; McMichael, A.J. 1986. Diet and cancer of the colon and rectum: a case-control study. J. Natl. Cancer Inst. 76:557-569

Potter, J.D.; McMichael, A.J.; Bonett, A.Z. 1982. Diet, alcohol and large-bowel 9ancer: a case control study. Proc. Nutr. Soc. Aust. 7:123-126.

Poynard, T.; Slama, G.; Delage, A.; Tchobroutsky, G 1980 Pectin efficacy in insulin-treated diabetics assessed by the artificial pancreas. Lancet 1:158.

Prather, E.S. 1964. Effect of cellulose on serum lipids in young women. J .. Am. Diet. Assoc. 45:230-233.

Prosky, L.; Asp, N.-G.; Furda, I.; Devries, J.W.; Schweizer, T.F.; Harland, B.F. 1984. Determination of total dietary fiber in foods, food products, and total diets: interlaboratory study. J. Assoc. Off. Anal. Chem. 67:1044-1052.

Prosky, L.; Asp, N.-G.; Furda, I.; DeVries, J.W.; Schweizer, T.F.; Harland, B.F. 1985. Determination of total dietary fiber in foods and food products: collaborative study. J. Assoce Off. Anal. Chem. 68:677-679.

Prynne, C.J.; Southgate, O.A.T. 1979e The effects of a supplement of dietary fibre on faecal excretion by human subjects. Br. J. Nutr. 41:495-503.

209

Qureshi, A.A.; Burger, W.C.; Peterson, D.M.; Elson, C.E. 1986. The structure of an inhibitor of cholesterol biosynthesis isolated from barley. J. Biol. Chem. 261:10544-10550.

Rattan, J.; Levin, N.; Graff, E.; Weizer, N.; Gilat, T. 1981. A high-fiber diet does not cause mineral and nutrient deficiencies. J. Clin. Gastroenterol. 3:389-393.

Ray, T.K.; Mansell, K.M.; Knight, L.C.; Malmud, L.S.; Owen, O.E.; Boden, G. 1983. Long-term effects of dietary fiber on glucose tolerance and gastric emptying in noninsulin-dependent diabetic patients. Am. J. Clin. Nutr. 37:376-381.

Raymond, T.L.; Connor, W.E.; Lin, D.S.; Warner, S.; Fry, M.M.; Connor, S.L. 1977. The interaction of dietary fibers and cholesterol upon the plasma lipids and lipoproteins, sterol balance, and bowel function in human subjects. J. Clin. Invest. 60:1429-1437.

Reddy, B.S. 1986a. Diet and colon cancer: evidence from human and animal model studies. In: Reddy, B.S.; Cohen, L.A., eds. Diet, nutrition, and cancer: a critical evaluation. Vol. 1. Macronutrients and cancer. Boca Raton, FL: CRC Press. p.47-65.

Reddy, B.S. 1986b. Fecal mutagens as a function of diet. In: Knudsen, I., ed. Genetic toxicology of the diet. New York: Alan R. Liss, Inc. p.213-224.

Reddy, B.S.; Hedges, A.R.; Laakso, K.; Wynder, E.L. 1978. Metabolic epidemiology of large bowel cancer: fecal bulk and constituents of high-risk North American and low-risk Finnish population. Cancer 42:2832-2838.

Reddy, B.S.; Maeura, Y.; Wayman, M. 1983a. Effect of dietary corn bran and autohydrolyzed lignin on 3,2'-dimethyl-4-amino­biphenyl-induced intestinal carcinogenesis in male F344 rats. J. Natl. Cancer Inst. 71:419-423.

Reddy, B.S.; Ekelund, G.; Bahe, M.; Engle, A.; Domellof, L. 1983b. Metabolic epidemiology of colon cancer: dietary pattern and fecal sterol concentrations of three populations. Nutr. Cancer 5:34-40.

Reddy, B.S.; Mori, H. 1981. Effect of dietary wheat bran and dehydrated citrus fiber on 3,2'-dimethyl-4-aminobiphenyl-induced intestinal carcinogenesis in F344 rats. Carcinogenesis 2:21-25.

Reddy, B.S.; Mori, H.; Nicolais, M. 1981. Effect of dietary wheat bran and dehydrated citrus fiber on azoxymethane-induced intestinal carcinogenesis in Fischer 344 rats. J. Natl. Cancer Inst. 66:553-557.

Reddy, B.S.; Sharma, C.; Mathews, L.; Engle, A.; Laasko, K.; Choi, K.; Puska, P.; Korpella, R. 1985. Metabolic epidemiology of colon cancer: fecal mutagens in healthy subjects from rural Kuopio .and urban Helsinki, Finland a Mutat. Res. 152:97-105.

Reddy, B.S.; Sharma, C.; Simi, Be; Engle, A.; Laakso, K.; Puska, P.; Korpella, R. 1987. Metabolic epidemiology of colon cancer: effect of dietary fiber on fecal mutagens and bile acids in healthy subjects. Cancer Res. 47:644-648.

Reddy, B.S4; Watanabe, K.; Sheinfil, A. 1980a. Effect of dietary wheat bran, alfalfa, pectin and carrageenan on plasma cholesterol and fecal bile acid and neutral sterol excretion in rats. J. Nutr. 110:1247-1254.

Reddy, Be Se; Sharma, C.; Darby, L.; Laakso, ~C; Wynder, E L. 1980b. Metabolic epidemiology of large bowel cancer: fecal mutagens in high- and low-risk population for colon cancer. A preliminary report. Mutat. Res. 72:511-522.

Reddy, B.S.; Weisburger, J.H.; Wynder, E.L. 1974. Fecal bacterial S-glucuronidase: control by diet. Science 183:416-417.

Reinhold, J.G. 1976. Rickets in Asian immigrants. Lancet 2:1132-1133.

Reinhold, J.G.; Faradji, B.; Abadi, P.; Ismail-Beigi, F. 1976. Decreased absorption of calcium, magnesium, zinc and phosphorus by humans due to increased fiber and phosphorus consumption as wheat breads J Nutr 106:493-503

Reinhold, J.G.; Garcia, J.S.; Garzon, P. 1981. Binding of iron by fiber of wheat and maize. Am. J. Clin. Nutr. 34:1384-1391.

Reinhold, J.G.; Nasr, K.; Lahimgarzadeh, A.; Heydayati, H. 1973. Effect of purified phytate and phytate-rich bread upon metabolism of zinc, calcium, phosphorus, and nitrogen in man. Lancet 1:283-288.

Rendtorff, R.C.; Kashgarian, M. 1967. Stool patterns of healthy adult males. Dis. Colon Rectum 10:222-228.

Riccardi, G.; Rivellese, A.; Pacioni, D.; Genovese, S.; Mastranzo, P.; Mancini, M. 1984. Separate influence of dietary carbohydrate and fibre on the metabolic control in diabetes. Diabetologia 26:116-121 ..

Risch, H.A.; Jain, M.; Choi, N.W.; Fodor, J.G.; Pfeiffer, C.J.; Howe, G.R; Harrison, L.W.; Craib, K J.P.; Miller, A.B. 1985 Dietary factors and the incidence of cancer of the stomach. Am. J. Epidemiol. 122:947-959.

211

Rivellese, A.; Parillo, M.; Giacco, A.; de Marco, F.; Riccardi, G. 1985. A fiber-rich diet for the treatment of diabetic patients with chronic renal failure. Diabetes Care 8:620-621.

Rivellese, A.; Riccardi, G.; Giacco, A.; Pacioni, D.; Genovese, S.; Mattioli, P.L.; Mancini, M. 1980. Effect of dietary fibre on glucose control and serum lipoproteins in diabetic patients. Lancet 2:447-450.

Rivellese, A.; Riccardi, G.; Giacco, A.; Postiglione, A.; Mastranzo, P.; Mattioli, P.L. 1983. Reduction of risk factors for atherosclerosis in diabetic patients treated with a high­fiber diet. Prev. Med. 12:128-132.

Robbins, S.L. 1967. Pathology. 3rd ed. Philadelphia: W.B. Saunders Company. 1434p.

Robertson, J.; Brydon, W.G.; Tadesse, K.; Wenham, P.; Walls, A.; Eastwood, M.A. 1979. The effect of raw carrot on serum lipids and colon function. Am. J. Clin. Nutr. 32:1889-1892.

Robertson, J.A.; Eastwood, M.A. 1981. An examination of factors which may affect the water holding capacity of dietary fibre. Br. J. Nutr. 45:83-88.

Robertson, J.A.; Eastwood, M.A.; Yeoman, M.M. 1980. An inves­tigation into the physical properties of fibre prepared from several carrot varieties at different stages of development. J. Sci. Food Agric. 31:633-638.

Robertson, J.B.; Van Soest, P.J. 1977. Dietary fiber estimation in concentrate feed-stuffs. J. Anim. Sci. 45(Suppl. 1):254-255.

Robertson, J.B.; Van Soest, P.J. 1981. The detergent system of analysis and its application to human foods. In: James, W.P.T.; Theander, O., eds. The analysis of dietary fiber in food. New York: Marcel Dekker, Inc. p.123-158.

Roe, D.A.; Wrick, K.; McLain, D.; Van Soest, P. 1978. Effects of dietary fiber sources on riboflavin absorption. Fed. Proc. Fed. Am. Soc. Exp. Biol. 37:756 (Abstract).

Roediger, W.E.W. 1980. Role of anaerobic bacteria in the metabolic welfare of colonic mucosa in man. Gut 21:793-798.

Roediger, W.E.W. 1982. Utilization of nutrients by isolated epithelial cells of the rat colon. Gasteroenterology 83:424-429.

Roediger, W.E.W.; Moore, A. 1981. Effect of short-chain fatty acid on sodium absorption in isolated human colon perfused through the vascular bed. Dig. Dis. Sci. 26:100-106.

?1?

Rose, G. 1985. Sick individuals and sick populations. Int. J. Epidemiolo 14:32-38.

Rosman, M.S.; Smith, C.J.; Jackson, W.P.U. 1983. The effect of long-term high-fibre diets in diabetic outpatients. S. Afr. Med. J. 63:310-313.

Ross, A.H.M.; Eastwood, M.A.; Brydon, W.G.; Anderson, J.R.; Anderson, D.M.W. 1983. A study of the effects of dietary gum arabic in humans. Am. J. Clin. Nutr. 37:368-375.

Rossner, S.; van Zweigbergk, D.; Ohlin, A. 1985. Effects of dietary fiber in treatment of overweight out-patients. In: Bjoerntorp, P.; Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber and obesity. Current topics in nutrition and disease. Volu 14. New York: Alan R. Liss, Inc. p.69-76.

Roth, G.; Leitzmann, C. 1985a. Langzeiteinfluss ballaststoff­reicher FrUhstUckscerealien auf die Blutlipide beim Menschen. Aktuel. Ernahr. 10:106-109.

Roth, G.; Leitzmann, C. 1985b. Fibre and the large gut. In: Leeds, A.R.; Avenell, A., eds. Dietary fibre perspectives: reviews and bibliography. London: John Libbey and Company Ltd. po3-22.

Rotstein, O.D.; Kay, R.M.; Wayman, M.; Strasberg, S.M. 1980. Effect of lignin (LIG) and lactulose (LAC) on bile composition in hamsters receiving ethinyl estradiol (EE)e Gastroenterology 78:1247 (Abstract).

Rotstein, O.De; Kay, R.M.; Wayman, M.; Strasberg, S.M. 1981. Prevention of cholesterol gallstones by lignin and lactulose in the hamster. Gastroenterology 81:1098-1103.

Rouse, I.L.; Armstrong, B.K.; Beilin, L.Jo 1982. Vegetarian diet, lifestyle and blood pressure in two religious populations. Clin. Exp. Pharmacol. Physiol. 9:327-330.

Rouse, I.L.; Beilin, L.J.; Armstrong, B.K.; Vandongen, R. 1983. Blood-pressure-lowering effect of a vegetarian diet: controlled trial in normotensive subjects. Lancet 1:5-10.

Rozen, P.; Hellerstein, S.M.; Horwitz, C. 1981. ·The low incidence of colorectal cancer in a "high-risk" population: its correlation with dietary habits. Cancer 48:2692-2695.

Rubio, M.A.; Pethica, B.A.; interactions of carcinogens other components of dietary S.I., eds. Dietary fibers: Academic Press. p.251-271.

Zuman, P .. ; Falkehag, S. I. 1979. The and co-carcinogens with lignin and fiber¢ In: Inglett, G.E.; Falkehag,

chemistry and nutrition. New York:

213

Russ, C.S.; Atkinson, R.L. 1985. Use of high fiber diets for the outpatient treatment of obesity. Nutre Rep. Int. 32:193-198.

Russ, C.S.; Atkinson, R~L. 1986. No effect of dietary fiber on weight loss in obesity. Am. J. Clin. Nutr. 43:136 (Abstract).

Russell, R.M.; Ismail-Beigi, F.; Reinhold, J.G. 1976. Folate content of Iranian breads and the effect of their fiber content on the intestinal absorption of folic acid. Am. J. Clin. Nutr. 29:799-802.

Rutishauser, I.H.E. 1985. Estimation of dietary fiber supply. Am. J. Clin. Nutr. 41:824-825.

Rydning, A.; Berstad, A. 1985a. Fiber diet and antacids in the short-term treatment of duodenal ulcer. Scand. J. Gastroenterol. 20:1078-1082.

Rydning, A.; Berstad, A. 1985b. Intragastric bile acid con­centrations in healthy subjects and in patients with gastric and duodenal ulcer and the influence of fiber-enriched wheat bran in patients with gastric ulcer. Scand. J. Gastroenterol. 20:801-804.

Rydning, A.; Berstad, A. 1986. Fiber and duodenal ulcer. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.395-409.

Rydning, A.; Berstad, A.; Aadland, E.; 0degaard, B. 1982. Prophylactic effect of dietary fibre in duodenal ulcer disease. Lancet 2:736-739.

Rydning, A.; Berstad, A.; Berstad, T.; Hertzenberg, L. 1985. The effect of guar gum and fiber-enriched wheat bran on gastric emptying of a semisolid meal in healthy subjects. Scand. J. Gastroenterol. 20:330-334.

Rydning, A.; Nesland, A.; Berstad, A. 1984. Influence of fiber on postprandial intragastric juice acidity, pepsin, and bile acids in healthy subjects. Scand. J. Gastroenterol. 19:1039-1044.

Rydning, A.; Weberg, R.; Lange, O.; Berstad, A. 1986. Healing of benign gastric ulcer with low-dose antacids and fiber diet. Gastroenterology 91:56-61.

Ryttig, K.R.; Larsen, S.; Haegh, L. 1985. Treatment of slightly to moderately overweight persons. In: Bjoerntorp, P.; Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber and obesity. Current topics in nutrition and disease. Vol. 14. New York: Alan R. Liss, Inc. p.77-84.

214

Sacks, F.M.; Castelli, W.P.; Donner, A.; Kass, E.H. 1975. Plasma lipids and lipoproteins in vegetarians and controls. N. Engl. J. Med. 292:1148-1151.

Sacks, F.M.; Rosner, B.; Kass, E.H. 1974. Blood pressure in vegetarians. Am. J. Epidemiol. 100:390-398 .

. Salyers, A.A.; Kuritza, A.P.; McCarthy, R.E. 1985. Influence of dietary fiber on the intestinal environment. Proc. Soc. Exp. Biol. Med. 180:415-421.

Sandberg, A.-S.; Ahderinne, R.; Andersson, H.; Hallgren, B.; Hulten, L. 1983. The effect of citrus pectin on the absorption of nutrients in the small intestine. Hum. Nutr. Clin. Nutr. 37C:171-183.

Sandberg, A.-S.; Andersson, H.; Hallgren, B.; Hasselblad, K.; Isaksson, B. 1981. Experimental model for in vivo determination of dietary fibre and its effect on the absorption of nutrients in the small intestine. Br. J. Nutr. 45:283-294.

Sandberg, A.-S.; Andersson, H.; Kivisto, B.; Sandstrom, B. 1986. Extrusion cooking of a high-fibre cereal product. 1. Effects on digestibility and absorption of protein, fat, starch, dietary fibre and phytate in the small intestine. Br. J. Nutr. 55:245-254.

Sandstead, H.H.; Munoz, J.M.; Jacob, ReA.; Klevay, L.M.; Reck, S.J.; Logan, G.M., Jr.; Dintzis, F.R.; Inglett, G.E.; Shuey, W.C. 1978. Influence of dietary fiber on trace element balance. Am. J. Clin. Nutr. 31:S180-S184.

Sandstrom, B.; Arvidsson, B. Cederblad, A.; Bjorn-Rasmussen, E. 1980. Zinc absorption from composite mealse le The significance of wheat extraction rate, zinc, calcium, and protein content in meals based on breade Am. J. Clin. Nutr. 33:739-745.

Sandstrom, B.; Cederblad, A. 1980. Zinc absorption from composite meals. II. Influence of the main protein source. Am. J. Cline Nutr. 33:1778-1783.

Sarles, H.; Hautan, J.; Planche, N.E.; Lafont, H.; Gerolami, A. 1970. Diet, cholesterol gallstones, and composition of the bile. Am. Jo Dig. Dis. 15:251-260.

Sasaki, J.; Funakoshi, M.; Arakawa, K. 1985. Effect of soybean crude fiber on the concentrations of serum lipids and apolipo­proteins in hyperlipemic subjects. Ann. Nutr. Metab. 29:274-278.

Schaller, D 1976. Analysis of cereal products and ingredients Paper presented at the 61st annual meeting of the American Association of Cereal Chemists, October 6, New Orleans.

215

Schaus, E.E.; de Lumen, B.O.; Chow, F.I.; Reyes, P.; Omaye, S.T. 1985. Bioavailability of vitamin E in rats fed graded levels of pectin. J. Nutr. 115:263-270.

Schneeman, B.O. 1978. Effect of plant fiber on lipase, trypsin and chymotrypsin activity. J. Food Sci. 43:634-635.

Schneeman, 8.0. 1982. Pancreatic and digestive functiono In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber in health and disease. New York: Plenum Press. p.73-83.

Schneeman,- 8.0.; Gallaher, D. 1985. Effects of dietary fiber on digestive enzyme activity and bile acids in the small intestine. Proc. Soc. Exp. Biol. Med. 180:409-414.

Schneeman, B.O.; Lefevre, M. 1986. Effects of fiber on plasma lipoprotein composition. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.309-321.

Schneeman, B.Oo; Richter, B.D.; Jacobs, L.R. 1982. Response to dietary wheat bran in the exocrine pancreas and intestine of rats. J. Nutr. 112:283-286.

Schwartz, S.E.; Levine, R.A.; Singh, A.; Scheidecker, J.R.; Track, N.S. 1982. Sustained pectin ingestion delays gastric emptying. Gastroenterology 83:812-817.

Schweizer, T.F.; Bekhechi, A.R.; Koellreutter, B.; Reimann, S.; Pometta, D.; Bron, B.A. 1983. Metabolic effects of dietary fiber from dehulled soybeans in humans. Am. J. Clin. Nutr. 38:1-11.

Schweizer, T.F.; WUrsch, P. 1981. Analysis of dietary fiber. In: James, W.P.T.; Theander, O., eds. The analysis of dietary fiber in food. New York: Marcel Dekker, Inc. p.203-216.

Scragg, R.K.R.; McMichael, A.J.; Baghurst, P.A. 1984. Diet, alcohol, and relative weight in gall stone disease: a case­control study. Br. Med. J. 288:1113-1119.

Segal, I. 1985. Hiatal hernia and gastro-oesophageal reflux. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.241-248.

Segal, I.; Solomon, A.; Hunt, J.A. 1977. Emergence of divertic­ular disease in the urban South African black. Gastroenterology 72:215-219.

Segal, I.; Walker, A.R.P. 1982. Diverticular disease in urban Africans in South Africa. Digestion 24:42-46.

216

Selvendran, R.R. 1984. The plant cell wall as a source of dietary fiber: chemistry and structure. Am. J. Clin. Nutre 39:320-337.

Shamsuddin, A.K.M. 1986~ Carcinoma of the large intestine: animal models and huma~ disease. Hum. Pathol. 17:451-453.

Shimoyama, R.; Uehara, S.; Itagaki, Y.; Izumiyama, S.; Hirayama, A. 1982. Effect of guar intake on plasma somatostatin-like immunoreactivity in diabetic patieFlts. Hokkaido Igaku Zasshi 57:727-733.

Shorey, R.L.; Day, P.J.; Willis, R .. A.; Lo, G.S.; Steinke, F .H. 1985. Effects of soybean polysaccharide on plasma lipids. J. Am. Diet. Assoc. 85:1461-1465.

Shultz, T.D.; Howie, B.J. 1986. In vitro binding of steroid hormones by natural and purified fibers. Nutr. Cancer 8:141-1470

Shultz, TaD.; Leklem, J.E. 1983. Selenium status of vege­tarians, nonvegetarians, and hormone-dependent cancer subjects. Am. J. Clin. Nutr. 37:114-118.

Shurpalekar, K.S.; Doraiswamy, T.R.; Sundaravalli, O.E.; Rao, M.N. 1971. Effect of inclusion of cellulose in an "atherogenic" diet on the blood lipids of children. Nature (London) 232:554-555.

Siegal, S.M. 1968. Biochemistry of the plant cell wall. In: Florkin, M.; Stotz, E.H., eds. Comprehensive biochemistry. Vol. 26A. Amsterdam: Elsevier Publishing Company. p.1-51.

Sigleo, S.; Jackson, M.J.; Vahouny, G.V. 1984. Effects of dietary fiber constituents on intestinal morphology and nutrient transport. Am. J. Physiolo 246:G34-G39.

Simon, G.L.; Gorbach, S.L. 1984. Intestinal flora in health and disease. Gastroenterology 86:174-193.

Simon, G.L.; Gorbach, S.L. 1986. The human intestinal micro­flora. Dig. Dis. Sci. ·31(Suppl.):147S-162Se

Simons, L.A.; Gayst, S.; Balasubramaniam, S.; Ruys, J. 1982. Long-term treatment of hypercholesterolaemia with a new palatable formulation of guar gum. Atherosclerosis 45:101-108.

Simonowitz, D.; Paloyan, Do 1977. Diverticular disease of the colon in patients under 40 years of age. Am. J. Gastroenterol. 67:69-72.

'J 1 7

Simpson, H.C.R.; Simpson, R.W.; Lousley, S.; Carter, R.D.; Geekie, M.; Hockaday, T.D.R.; Mann, J.I. 1981a. A high carbo­hydrate leguminous fibre diet improves all aspectsof diabetic control. Lancet 1:1-5.

Simpson, K.M.; Morris, E.R.; Cook, JoD. 1981b. The inhibitory effect of bran on iron absorption in man. Ame J. Clin. Nutr. 34:1469-1478.

Simpson, R.W.; Mann, J.I.; Eaton, J.; Moore, R.A.; Carter, R.; Hockaday, T.D.R. 1979a. Improved glucose control in maturity­onset diabetes treated with high-carbohydrate-modified fat diet. Br. Med. J. 1:1753-1756.

Simpson, R.W.; Mann, J.I.; Eaton, J.; Carter, R.D.; Hockaday, T.D.R. 1979b. High-carbohydrate diets and insulin-dependent diabetics. Br. Med. J. 2:523-525.

Sina, P.A.; Provenghi, R.R.; Cantone, A. 1976. Dietary cellulose and intestinal absorption of nutrients in rats. Biochem. Exp. Biol. 12:321-327.

Sircar, B.; Johnson, L.R.; Lichtenberger, L.M. 1983. Effect of synthetic diets on gastrointestinal mucosal DNA synthesis in rats. Am. J. Physiol. 244:G327-335.

Sirtori, C.R.; Gatti, E.; Mantero, O.; Conti, F.; Agradi, E.; Tremoli, E.; Sirtori, M.; Fraterrigo, L.; Tavazzi, L.; Kritchevsky, D. 1979. Clinical experience with the soybean protein diet in the treatment of hypercholesterolemia. Am. J. Clin. Nutr. 32:1645-1658.

Slavin, J.L.; Marlett, J.A. 1980. Influence of refined cellulose on human bowel function and calcium and magnesium balance. Am. J. Clin. Nutr. 33:1932-1939.

Slavin, J.L.; Nelson, N.L.; McNamara, E.A.; Cashmere, K. 1985. Bowel function of healthy men consuming liquid diets with and without dietary fiber. J. Parenter. Enter. Nutr.- 9:317-321.

Smith, C.J.; Rosman, M.S.; Levitt, N.S.; Jackson, W.P.U. 1982. Guar biscuits in the diabetic diet. S. Afr. Med. J. 61:196-198.

Smith, D.A.; Gee, M.I. 1979. A dietary survey to determine the relationship between diet and cholelithiasis. Am. J. Clin. Nutr. 32:1519-1526.

Smith, R.G.; Rowe, M.J.; Smith, A.N.; Eastwood, M.A.; Drummond, E.; Brydon, W.G. 1980. A study of bulking agents in elderly patients. Age Ageing 9:267-271.

Smith, U.; Holm, G. 1982. Effect of a modified guar gum prep­aration on glucose and lipid levels in diabetics and healthy volunteers. Atherosclerosis 45:1-10.

218

Smith-Barbaro, P.; Hanson, D.; Reddy, B.S. 1981. Carcinogen binding to various types of dietary fiber. J. Natl. Cancer Inst. 67:495-497.

S~ltoft, J.; Gudmand-H~yer, E.; Krag, B.; Kristensen, E.; Wulff, H.R. 1976. A double-blind trial of the effect of wheat bran on symptoms of irritable bowel syndrome. Lancet 1:270-272.

Southgate, D.; Englyst, H. 1985. Dietary fibre: chemistry, physical properties and analysis. In: Trowell, H.; Burkitt, D~; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.31-55.

Southgate, DeAeTo 1969e Determination of carbohydrate in foods. II. Unavailable carbohydrates. J. Sci. Food Agric. 20:331-335.

Southgate, D.A.T. 1982. Definitions and terminology of dietary fiber. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber in health and disease. New York: Plenum Press. p.1-7.

Southgate, D.A.T. 1986. The relation between composition and properties of dietary fiber and physiological effects. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.35-48.

Southgate, D.A.T.; Durnin, J.V.G.A. 1970. Calorie conversion factors: an experimental reassessment of the factors used in the calculation of the energy value of human diets. Br. J. Nutr. 24:517-535.

Spiller, G.A. 1986. Suggestions for a basis on which to determine a desirable intake of dietary fiber In: Spiller G.A., ed. CRC handbook of dietary fiber in human nutrition. Boca Raton, FL: CRC Press, Inc. p.281-283.

Spiller, G.A.; Chernoff, M.C.; Hill, R.Ar; Gates, J.E.; Nassar, J.J.; Shipley, E.A. 1980. Effect of purified cellulose, pectin, and a low-residue diet on fecal volatile fatty acids, transit time, and fecal weight in humans. Am. J. Clin. Nutr. 33:754-759.

Spiller, G.A.; Chernoff, M.C.; Shipley, E.Ae; Beigler, M.A.; Briggs, G.M. 1977. Can fecal weight be used to establish a recommended intake of dietary fiber (plantix)? Am. J. Clin. Nutr. 30:659-661.

Spiller, G.A.; Story, J.A.; Wong, L.G.; Nunes, J.D.; Alton, M.; Petro, M.S.; Furumoto, E.J.; Whittam, J.H.; Scala, J. 1986. Effect of increasing levels of hard wheat fiber on fecal weight, minerals and steroids and gastrointestinal transit time in healthy young women. J. Nutr. 116:778-785.

219

Spiller, G.A.; Wong, L.G.; Whittam, J.H.; Scala, J. 1982. Correlation of gastrointestinal transit time to fecal weight in adult humans at two levels of fiber intake. Nutr. Rep. Int. 25: 23-30. •'

Sprinz, H. 1971. Factors influencing intestinal cell renewal. Cancer 28:71-74.

Srivastava, G.S.; Smith, A.N.; Painter, N.S. 1976. Sterculia bulk-forming agent with smooth-muscle relaxant versus bran in diverticular disease. Br. Med. J. 1:315-318.

Stanley, M.M.; Paul, D.; Gacke, D.; Murphy, J. 1973. Effects of cholestyramine, Metamucil, and cellulose on fecal bile salt excretion in man. Gastroenterology 65:889-894.

Stasse-Wolthuis, M.; Albers, H.F.F.; van Jeveren, J.G.C.; Wil de Jong, J.; Hautvast, J.G.A.J.; Hermus, R.J.J·.; Katan, M.B.; Brydon, W.G.; Eastwood, M.A. 1980. Influence of dietary fiber from vegetables and fruits, bran or citrus pectin on serum lipids, fecal lipids, and colonic function. Am. J. Clin. Nutr. 33:1745-1756.

Stasse-Wolthuis, M.; Hautvast, J.G.A.J.; Hermus, R.J.J.; Katan, M.B.; Bausch, J.E.; Rietberg-Brussaard, J.H.; Velema, J.P.; Zondervan, J.H.; Eastwood, M.A.; Brydon, W.G. 1979a. The effect of a natural high-fiber diet on serum lipids, fecal lipids, and colonic function. Am. J. Clin. Nutr. 32:1881-1888.

Stasse-Wolthuis, M.; Katan, M.B.; Hermus, R.J.J.; Hautvast, J.G.A.J. 1979b. Increase of serum cholesterol in man fed a bran diet. Atherosclerosis 34:87-91.

Stephen, A. 1985. Constipation. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.133-144.

Stephen, A.M.; Cummings, J.H. 1979. Water-holding by dietary fibre in vitro and its relationship to faecal output in man. Gut 20:722-729.

Stephen, A.M.; Cummings, J.H. 1980a. The microbial contribution to human faecal mass. J. Med. Microbial. 13:45-56.

Stephen, A.M.; Cummings, J.H. 1980b. Mechanism of action of dietary fibre in the human colon. Nature (London) 284:283-284.

Stephen, A.M.; Haddad, A.C.; Phillips, S.F. 1983. Passage of carbohydrate into the colon: direct measurements in humans. Gastroenterology 85:589-595.

,..,,..,n

Stevens, J.; Burgess, M.B.; Kaiser, D.L.; Sheppa, C.M. 1985. Outpatient management of diabetes mellitus with patient education to increase dietary carbohydrate and fiber. Diabetes Care 8:359-366.

Stocks, P.; Karn, MeN. 1933. A co-operative study of the habits, home life, dietary and family histories of 450 cancer patients and of an equal number of control patients. Ann. Eugen. 5:237-280.

Stokholm, K.H.; Lauritsen, K.B.; Larsen, S. 1981. Reduced glycosuria during guar gum supplementation in non-insulin­dependent diabetics: a double-blind, randomised cross-over study. Dan. Medo Bull. 28:41-42.

Story, JoA. 1980 Dietary fiber and lipid metabolism: an update. In: Spiller, G.A.; Kay, R.M., eds. Medical aspects of dietary fiber. New York: Plenum Medical Book Company. p.137-152.

Story, J.A. 1985. Dietary fiber and lipid metabolism. Proc. Soc. Exp. Biol. Med. 180:447-452.

Story, J.A. 1986. Modification of steroid excretion in response to dietary fiber. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Presso p.253-264.

Story, J.A.; Kritchevsky, D. 1976. Comparison of the binding of various bile acids and bile salts in vitro by several types of fiber. J. Nutr. 106:1292-1294.

Story, J.Ao; Thomas, J.N. 1982. Modification of bile acid spectrum by dietary fiber. In: Vahouny, G.V.; Kritchevsky, D., edso Dietary fiber in health and disease. New York: Plenum Press. p.193-201.

Story, J.A.; White, A.; West, L.G. 1982. Adsorption of bile acids by components of alfalfa and wheat bran in vitro. J. Food Sci. 47:1276-1279.

Story~ L.; Anderson, J.W.; Chen, W.-J.L.; Karounos, D.; Jefferson, B. 1985. Adherence to high-carbohydrate, high-fiber diets: long-term studies of non-obese diabetic men. J. Am. Diet. Assoc. 85:1105-1110.

Tadesse, K. 1982. The effect of dietary fibre on gastric secretion and emptying in man. J. Physiol. (London) 332:102P-103P (Abstract).

221

Tagliaferro, V.; Cassader, M.; Bozzo, C.; Pisu, E.; Bruno, A.; Marena, S.; Cavallo-Perin, P.; Cravero, L.; Pagano, G. 1985. Moderate guar-gum addition to usual diet improves peripheral sensitivity to insulin and lipaemic profile in NIDDM. Diabete Metab. (Paris) 11:380-385.

Takano, S.; Matsushima, M.; ErtUrk, Ee; Bryan, G.T. 1981. Early · induction of rat colonic epithelial ornithine and S-adenosyl-L­methionine decarboxylase activities by N-methyl-N'nitro-N­nitrosoguanidine or bile salts. Cancer Res. 41:624-628.

Talbot, J.M. 1980. The role of dietary fiber in diverticular disease and colon cancer. Prepared for the Food and Drug Administration, Washington, DC, under Contract No. 223-79-2275 by the Life Sciences Research Office, Federation of American Societies for Experimental Biology, Bethesda, MD. 77p. Avail­able from: NTIS, Springfield, VA; PB81-127847.

Tarpila, S.; Miettinen, T.A.; Metsaranta, L. 1978. Effects of bran on serum cholesterol, faecal mass, fat, bile acids and neutral sterols, and biliary lipids in patients with diverticular disease of the colon. Gut 19:137-145.

Taskinen, M.-R.; Nikkila, E.A.; Ollus, Ae 1983. Serum lipids and lipoproteins in insulin-dependent diabetic subjects during high-carbohydrate, high-fiber diet. Diabetes Care 6:224-230.

Tasman-Jones, C.; Owen, R.L.; Jones, A.L. 1982. dietary fiber and small-bowel morphology in rats. 27:519-524.

Semipurified Dig. Dis. Sci.

Taylor, I.; Duthie, H.L. 1976. Bran tablets and diverticular disease. Br. Med. J. 1:988-990.

Theander, O. 1981. Review of the different analytical methods and remaining problems. In: James, W.P.T.; Theander, O., eds. The analysis of dietary fiber in food. New York: Marcel Dekker, Inc. p.263~276. ·

Theander, O. 1983. Advances in the chemical characterisation and analytical determination of dietary fibre components. In: Birch, G.G.; Parker, K.J., eds. Dietary fibre. New York: Applied Science Publishers. p.77-93.

Theander, O.; Westerlund, E.A. 1986. Studies on dietary fiber. 3. Improved procedures for analysis of dietary fiber. J. Agric. Food Chem. 34:330-336.

Thiffault, C.; Belanger, M.; Pouliot, M. 1970. Traitement de l'hyperlipoproteinemie essentielle de type II par un nouvel agent therapeutique, la celluline. Can. Med. Assoc. J. 103:165-166.

222

Thomson, M.; Logan, R.L.; Sharman, M.; Lockerbie, L.; Riemersma, R.A.; Oliver, M.F. 1982. Dietary survey in 40-year-old Edinburgh men. Hum. Nutr. Appl. Nutr. 36A:272-280.

Thornton, J.R.; Emmett, P.M.; Heaton, K.W. 1979. Diet and Crohn's disease: characteristics of the pre-illness diet. Br. Med. J. 2:762-764.

Thornton, J.R.; Emmett, P.M.; Heaton, K.W. 1980. Diet and ulcerative colitis. Br. Med. J. 280:293-294.

Thornton, J.R.; Emmett, P.M.; Heaton, K.Wo 1983. Diet and gall stones: effects of refined and unrefined carbohydrate diets on bile cholesterol saturation and bile acid metabolism. Gut 24:2-6.

Topping, D.L.; Illman, R.J. 1986. Bacterial fermentation in the human large bowel: time to change from the roughage model of dietary fibre? Med. J. Aust. 144:307-309.

Tovey, F. 1985. Duodenal ulcer. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and diseaseo New York: Academic Press. p.229-240.

Tovey, F.I. 1974. Aetiology of duodenal ulcer: an investiga­tion into the buffering action and effect on pepsin of bran and unrefined carbohydrate foods. Postgrad. Med. J. 50:683-688.

Treuherz, J 1982. Possible inter-relationship between zinc and dietary fibre in a group of lacto-ovo vegetarian adolescents. J. Plant Foods 4:89-93.

Trowell, H. 1973. Dietary fibre, ischaemic heart disease and diabetes mellitus. Proc. Nutr. Soc. 32:151-157.

Trowell, H. 1974. Diabetes mellitus death-rates in England and Wales 1920-70 and food supplies. Lancet 2:998-1002.

Trowell, H. 1978. The development of the concept of dietary fiber in human nutrition. Am. J. Clin. Nutr. 3l(Suppl.):S3-Sll.

Trowell, H. 1981. Hypertension, obesity, diabetes mellitus and coronary heart disease. In: Trowell, H.C.; Burkitt, D.P., eds. Western diseases: their emergence and prevention. Cambridge, MA: Harvard University Press~ p.3-32.

Trowell, H.C. 1960. Non-infective disease in Africa. London: Edward Arnold Publishers Ltd. 48lp.

Trowell, H.C. 1975. Dietary-fiber hypothesis of the etiology of diabetes mellituso Diabetes 24:762-765.

223

Trudel, J.L.; Senterman, M.K.; Brown, R.Ao 1983. The fat/ fiber antagonism in experimental colon carcinogenesis. Surgery 94:691-696.

Tsai, A.C.; Mott, E.L.; Owen, G.M.; Bennick, M.R.; Lo, G.S.; Steinke, F.H. 1983. Effects of soy polysaccharide on gastro­intestinal functions, nutrient balance, steroid excretions, glucose tolerance, serum lipids, and other parameters in humans. Am. J. Clin. Nutr. 38:504-511.

Tuomilehto, J.; Voutilainen, E.; Huttunen, J.; Vinni, S.; Homan, K. 1980. Effect of guar gum on body weight and serum lipids in hypercholesterolemic females. Acta Med. Scand. 208:45-48.

Turnlund, J.R.; King, J.C.; Keyes, W.R.; Gong, B.; Michel, M.C. 1984. A stable isotope study of zinc absorption in young men: effects of phytate and a-cellulose. Am. J. Clin. Nutr. 40:1071-1077.

Tuyns, A.J. 1986. A case-control study on colorectal cancer in Belgium: preliminary results. Soz. Praventiumed. 31:81-82.

Ullrich, I.H.; Albrink, M.J. 1982. Lack of effect of dietary fiber on serum lipids, glucose, and insulin in healthy young men fed high starch diets. Am. J. Clin. Nutr. 36:1-9.

Urban, E. 1984. Beans, gas, and science. Gastroenterology 86:994-995.

U.S. Department of Agriculture/U.S. Department of Health and Human Se~vices. 1985. Dietary guidelines for Americans. 2nd ed. USDA Home and Garden Bulletin No. 232. 24p. Available from: U.S Government Printing Office, Washington, DC.

U.S. Surgeon General's Advisory Committee on Smoking and Health. 1964. Smoking and health. Public Health Service Publication No. 1103. 387p. Available from: U.S. Government Printing Office, Washington, DC.

Vaaler, S.; Aaseth, J.; Hanssen, K.F.; Dahl-Jorgensen, K.; Fr0lich, W.; 0degaard, B.; Agenaes, 0. 1985. Trace elements in serum and urine of diabetic patients given bread enriched with wheat bran or guar gum. In: International Symposium on Trace Element Metabolism in Man and Animals, TEMA-5. Edinburgh: Churchill Livingstone. p.446-449.

Vahouny, G.V. 1982. Dietary fiber, lipid metabolism, and atherosclerosis. Fed. Proc. Fed. Am. Soc. Exp. Biol. 41:2801-2806.

Vahouny, G.V. 1985. Dietary fibers: aspects of nutrition, pharmacology, and pathology. In: Sidransky, H., ed. Nutri­tional pathology: pathobiochemistry of dietary imbalances. New York: Marcel Dekker, Inc. p.207-277.

224

Vahouny, G.V.; Cassidy, M.M. 1985. Dietary fibers and absorp­tion of nutrients. Proc. Soc. Exp. Biol. Med. 180:432-446.

Vahouny, G.V.; Cassidy, M.M. 1986. Dietary fiber and intestinal adaptation. In: Vahouny, G.V.; Kritchevsky, D., eds. Dietary fiber: basic and clinical aspects. New York: Plenum Press. p.181-209.

Vahouny, G.V.; Le, T.; Ifrim, I.; Satchithanandam, S.; Cassidy, M.M. 1985. Stimulation of intestinal cytokinetics and mucin turnover in rats fed wheat bran or cellulose. Am. J. Clin. Nutr. 41:895-900.

Vahouny, G.V.; Tombes, R.; Cassidy, M.M.; Kritchevsky, D.; Gallo, L.L. 1980. Dietary fibers. V. Binding of bile salts, phospholipids and cholesterol from mixed micelles by bile acid sequestrants and dietary fibers. Lipids 15:1012-1018.

Vahouny, G.V.; Tombes, R.; Cassidy, M.M.; Kritchevsky, D.; Gallo, L.L. 1981. Dietary fibers. VI. Binding of fatty acids and monolein from mixed micelles containing bile salts and lecithin. Proc. Soc. Exp. Biol. Med. 166:12-16.

Valle-Jones, J.C. 1980. The evaluation of a new appetite­reducing agent (Prefil) in the management of obesity. Sr. J. Cline Pract. 34:72-74.

van Berge-Henegouwen, GP ; Huybregts, A.W ; van de Werf, S.; Damacker, P ; Schade, R W 1979. Effect of a standardized wheat bran preparation on serum lipids in young healthy males. Am. J. Clin. Nutr. 32:794-798.

Van Dokkum, W.; Wesstra, A.; Schippers, F.A. 1982. Physio­logical effects of fibre-rich types of bread. 1. The effect of dietary fibre from bread on the mineral balance of young men. Br. J. Nutr. 47:451-460.

Van Horn, L.V.; Liu, K.; Parker, D.; Emidy, L.; Liao, Ye; Pan, W.H.; Giumetti, D.; Hewitt, J.; Stamler, J. 1986. Serum lipid response to oat product intake with a fat-modified diet. J. Am. Diet. Assoc. 86:759-764.

Van Soest, P.J. 1963. Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fiber and lignin. J. Assoc. Off. Anal. Chem. 46:829-835.

Van Soest, P.J. 1984. Some physical characteristics of dietary fibres and their influence on the microbial ecology of the human colona Proc. Nutr. Soc. 43:25-33.

225

Van Soest, P.J.; Robertson, J.B.; Roe, D.A.; Rivers, J.; Lewis, B.A.; Hackler, L.R. 1978. The role of dietary fiber in human nutrition. Proceedings of the Cornell Nutrition Conference for Feed Manufacturers, October 31-November 2, Syracuse, NY. p.5-12. Available from: Cornell University, Ithaca, NY.

Van Soest, P.J.; Uden, P.; Wrick, K.F. 1983. Critique and evaluation of markers for use in nutrition of humans and farm and laboratory animals. Nutr. Rep. Int. 27:17-28.

Van Soest, P.J.; Wine, R.H. 1967. Use of detergents in the analysis of fibrous feeds. IV. Determination of plant cell-wall constituents. J. Assoc. Off. Anal. Chem. 50:50-55.

Vargo, D.; Doyle, R.; Floch, M.H. 1985. Colonic bacterial flora and serum cholesterol: alterations induced by dietary citrus pectin. Am. J. Gastroenterol. 80:361-364.

Vargo, D.; Moskovitz, M.; Floch, M.H. 1980. Faecal bacterial flora in cancer of the colon. Gut 21:701-705.

Vellar, O.D.; Borchgrevink, C.; Natvig, H. 1968. Iron-fortified bread: absorption and utilization studies. Acta Med. Scand. 183:251-256.

Verrillo, A.; de Teresa, A.; Giarrusso, P.C.; La Rocca, S. 1985. Soybean protein diets in management of type II hyperlipoprotein­aemia. Atherosclerosis 54:321-331.

Villaume, C.; Beck, B.; Gariot, P.; Desalme, A.; Debry, G. 1984. Long-term evolution of the effect of bran ingestion on meal­induced glucose and insulin responses in healthy men. Am. J. Clin. Nutr. 40:1023-1026. ·

Viseshakul, D.; Premvatana, P.; Chularojmontri, V.; Kewsiri, D.; Tinnarat, P. 1985. Improved glucose tolerance induced by long term dietary supplementation with hairy basal seeds (Ocimum canum sim) in diabetics. J. Med. Assoc. Thailand 68:408-411.

Volkheimer, G. 1977. Persorption of particles: physiology and pharmacology. Adv. Pharmacol. Chemother. 14:163-187.

van Thaer, A. 1809. GrUndsatze der rationellen Landwirtschaft. Vol. 1. Berlin: Die Verlag der Realschulbuchhandlung. p.275.

Walker, A. 1985. Mineral metabolism. In: Trowell, H.; Burkitt, D.; Heaton, K., eds. Dietary fibre, fibre-depleted foods and disease. New York: Academic Press. p.361-375.

Walker, A.; Burkitt, D. 1985. Burkitt, D.; Heaton, K., eds. foods and disease. New York:

Appendicitis. In: Trowell, H.; Dietary fibre, fibre-depleted Academic Press. p.191-203.

226

Walker, A.R.P. 1961. Crude fibre, bowel motility, and pattern of diet. S. Afr. Med. J. 35:114-115.

Walker, A.R.P.; Fox, F.W.; Irving, J.T. 1948. Studies in human mineral metabolism. 1. The effect of bread rich in phytate phosphorus on the metabolism of certain mineral salts with special reference to calcium. Biochem. J. 42:452-462.

Walker, A.R.P.; Walker, B.F.; Richardson, B.D. 1970. Glucose and fat tolerances in Bantu children. Lancet 2:51-52.

Walker, A.R.P.; Walker, B.F.; Walker, A.J. 1986. Faecal pH, dietary fibre intake, and proneness to colon cancer in four South African populations. Br. J. Cancer 53:489-495.

Walsh, D.F.; Yaghoubian, V.; Behforooz, A. 1984. Effect of glucomannan on obese patients: a clinical study. Int. J. Obesity 8:289-293.

Walters, R.L.; Baird, LM.; Davies, P.S.; Hill, M.J.; Draser, B.S.; Southgate, D.AeT.; Green, J.; Morgan, B. 1975. Effects of two types of dietary fibre on faecal steroid and lipid excretion. Br. Med. J. 2:536-538.

Wapnick, S.; Wicks, A.C.B.; Kanengoni, E.; Jones, J.J. 1972. Can diet be responsible for the initial lesion in diabetes? Lancet 2:300-302.

Ward, G.M.; Simpson, R.W.; Simpson, H.C.R.; Naylor, B.Ae; Mann, J.I.; Turner, R.C. 1982. Insulin receptor binding increased by high carbohydrate low fat diet in non-insulin-dependent diabetics. Eur. J. Clino Invest. 12:93-96.

Ward, J.M.; Yamamoto, R.S.; Weisburger, J.H. 1973. Cellulose dietary bulk and azoxymethane-induced intestinal cancer. J. Natl. Cancer Inst. 51:713-715.

Ware, J.E., Jr.; Brook, R.H.; Davis, A.R.; Lohr, K.N. 1981. Choosing measures of health status for individuals in general populationse Ame J. Public Health 71:620-625.

Watanabe, K.; Reddy, B.S.; Weisburger, J.H.; Kritchevsky, D. 1979. Effect of dietary alfalfa, pectin, and wheat bran on azoxymethane- or methylnitrosourea-induced colon carcinogenesis in F344 rats. J. Natl. Cancer Inst. 63:141-145.

Watanabe, K.; Reddy, B.S.; Wong, C.Q.; Weisburger, J.H. 1978. Effect of dietary undegraded carrageenan on colon carcinogenesis in F344 rats treated with azoxymethane or methylnitrosourea. Cancer Res. 38:4427-4430.

Watts, J.M., Jablonski, P.; Toouli, J. 1978. The effect of added bran to the diet in the saturation of bile in people without gallstones. Am. J. Surg. 135:321-324.

227

Wechsler, J.G.; Swobodnik, W.; Wenzel, H.; Heuchemer, T.; Nebelung, W.; Hutt, V.; Ditschuneit, H. 1984. Ballaststoffe vom Typ Weizenkleie senken LithogenitMt der Galle. Dtsch. Med. Wochenschr. 109:1284-1288.

Weinreich, J.; Pedersen, O.; Dinesen, K. 1977. Role of bran in normals: serum levels of cholesterols, triglyceride, calcium and total 3a-hydroxycholanic acid, and intestinal transit time. Acta Med. Scand. 202:125-130.

Weinsier, R.L.; Bacon, J.A.; Birch, R. 1983. Time-calorie displacement diet for weight control: a prospective evaluation of its adequacy for maintaining normal nutritional status. Int. J. Obesity 7:539-548.

Weinsier, R.L.; Johnston, M.H.; Doleys, D.M.; Bacon, J.A. 1982. Dietary management of obesity: evaluation of the time-energy displacement diet in terms of its efficacy and nutritional adequacy for long-term weight control. Br. J. Nutr. 47:367-379.

Welch, I.; Saunders, K.; Read, N.W. 1985. Effect of ileal and intravenous infusions of fat emulsions on feeding and satiety in human volunteers. Gastroenterology 89:1293-1297.

West, K.M. 1974a. Epidemiolgic (sic) observations on thirteen populations of Asia and the Western hemisphere. In: Hillebrand, S.S., ed. Is the risk of becoming diabetic affected by sugar consumption? Proceedings of the Eighth International Sugar Research Symposium, March 14, Washington, D.C. p.34-43. Available from: International Sugar Research Foundation, Bethesda, MD.

West, K.M. 1974b. Diabetes in American Indians and other native populations of the New World. Diabetes 23:841-855.

West, K.M.; Kalbfleisch, J.M. 1971. Influence of nutritional factors on prevalence of diabetes. Diabetes 20:99-108.

Widdowson, E.M.; Mccance, R.A. 1942. Iron exchanges of adults ,, on white and brown bread diets. Lancet 1:588-591.

Wilkins, T.D.; Lederman, M.; Van Tassell, R.L. 1981. Isolation of mutagen produced in the human colon by bacterial action. In: Bruce, W.R.; Correa, P.; Lipkin, M.; Tannenbaum, S.R.; Wilkins, T.D., eds. Banbury Report 7. Gastrointestinal cancer: endogenous factors. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. p.205-214.

Willett, W.; Stampfer, M.J.; Bain, C.; Lipnick, R.; Speizer, F.E.; Rosner, B.; Cramer, D.; Hennekens, C.H. 1983. Cigarette smoking, relative weight, and menopause. Am. J. Epidemiol. 117:651-658.

Willett, W.C.; Sampson, L.; Stampfer, M.J.; Rosner, B.; Bain, C.; Witschi, J.; Hennekens, C.H.; Speizer, F.E. 1985. Reproduci­bility and validity of a semiquantitative food frequency questionnaire. Am. J. Epidemiol. 122:51-65.

Williamson, R.C.N. 1978. Intestinal adaptation (second of two parts): mechanisms of control. N. Engl. J. Med. 298:1444-1450.

Wilmshurst, P.; Crawley, J.C.W. 1980. The measurement of gastric transit time in obese subjects using 24Na and the effects of energy content and guar gum on gastric emptying and satiety. Br. J. Nutr. 44:1-6.

Wilson, R.B.; Hutcheson, D~P.; Wideman, L. 1977. Dimethyl­hydrazine-induced colon tumors in rats fed diets containing beef fat or corn oil with and without wheat bran. Am Jo Clin. Nutr. 30:176-181.

Wirth, A.; Middelhoff, G.; Braeuning, C.; Schlierf, G. 1982. Treatment of familial hypercholesterolemia with a combination of bezafibrate and guar. Artherosclerosis 45:291-297.

Wolever, T.M.S.; Jenkins, D.J.A. 1986. Effect of dietary fiber and foods on carbohydrate metabolism. In: Spiller, G.A., ed. CRC handbook of dietary fiber in human nutrition. Boca Raton, FL: CRC Press, Inc. p.87-1190

Wolever, T.M S ; Jenkins, DJ.A.; Nineham, R.; Alberti, K.G.M.M. 1979. Guar gum and reduction of post-prandial glycaemia: effect of incorporation into solid food, liquid food, and both. Br. J. Nutr. 41:505-510.

Wong, S.; O1 Dea, K. 1983. Importance of physical form rather than viscosity in determining the rate of starch hydrolysis in legumes. Am. J. Clin. Nutr. 37:66-700

Wrick, K.L.; Robertson, J.B.; Van Soest, P.J.; Lewis, B.A.; Rivers, J.M.; Roe, D.A.; Hackler, L.R. 1983. The influence of dietary fiber source on human intestinal transit and stool output. J. Nutr. 113:1464-1479.

Wright, Ao; Burstyn, P.G.; Gibney, M.J. 1979. Dietary fibre and blood pressure. Br. Med. J. 2:1541-1543.

Wrong, O.; Metcalfe-Gibson, A.; Morrison, R.B.I.; Ng, S.To; Howard, A.V. 1965. In vivo dialysis of faeces as a method of stool analysis. Io echnique and results in normal subjects. Clin. Sci. 28:357-375.

Wyman, J.B.; Heaton, K.W.; Manning, A.Po; Wicks, A.C.B. 19760 The effect on intestinal transit and the feces of raw and cooked bran in different doses. Am. J. Clino Nutr. 29:1474-1479.

2

Wynder, E.L.; Kajitani, T.; Ishikawa, S.; Dodo, H.; Takano, A. 1969. Environmental factors of cancer of the colon and rectum. II. Japanese epidemiological data. Cancer 23:1210-1220.

Wynder, E.L.; Shigematsu, T. 1967. Environmental factors of cancer of the colon and rectum. Cancer 20:1520-1561.

Younoszai, M.K.; Adedoyin, M.; Ranshaw, J. 1978. Dietary components and gastrointestinal growth in rats. J. Nutro 108:341-350.

Yudkin, J. 1959. The causes and cure of obesity. Lancet 2:1135-1138.

Zavoral, J.H.; Hannan, P .. ; Fields, D.J.; Hanson, M.N.; Frantz, I.D.; Kuba, K.; Elmer, P.; Jacobs, D.R., Jr. 1983. The hypo­lipidemic effect of locust bean gum food products in familial hypercholesterolemic adults and children. Am. J. Clin. Nutr. 38:285-294.

Zoppi, G.; Gobio-Casali, L.; Deganello, A.; Astolfi, R.; Saccomani, F.; Cecchettin,·M. 1982. Potential complica­tions in the use of wheat bran for constipation in infanriy. J. Pediatr. Gastroenterol. Nutr. 1:91-95.

X. GLOSSARY

Acid detergent fiber: the residue produced by heating food­stuffs with hot, dilute sulfuric acid containing cetyl trimethylammonium bromide; taken as a measure of cellu­lose and lignin

Agar: cell wall constituent of red marine algae that is extracted without water; a viscous mixture of poly­saccharides of which the main component is a9arose

Bagasse: the residue of fibrous stalks remaining after Juice has been pressed from sugar cane; a commercial source of cellulose

Bengal gram: chickpeas; also called chana

Bran: the husks or outermost layer of a cereal grain, such as wheat, corn, oats, rice

Carrageenan: an algal polysaccharide extracted from an edible seaweed (carragheen, Irish moss); contains, among other constituents, polymerized sulfated D-galactopyranose units

Cellulose: main structural polysaccharide of the cell walls of plants; molecules consist of long chains of B-(1,4)-linked­glucopyranoside units

Crude fiber: the residue remaining after boiling defatted foodstuffs in dilute base and dilute acid; "classical" method does not allow the complete recovery of plant cell wall constituents and is no longer an acceptable method for determining dietary fiber

Cutin: a lipid component of the covering and cuticle of the outer cellulose walls of plants

Dietary fiber: the endogenous components of plant materials in the diet which are resistant to digestion by enzymes produced by humans

Fermentable: capable of being used as a substrate by micro­organisms in anaerobic catabolism

Galactomannan: a polysaccharide containing galactose and mannose; usually found in the hemicellulose fraction of plant cell walls

Gel: a phase that is largely liquid, but unable to flow because it is held rigid by molecular chains, usually cross-linked, that pass through it

Glucan: generic name for a polysaccharide composed of glucose

231

Glucomannan: a gel-forming fiber extracted from konjac tuber; a proprietary name

Glycan: generic name for a polysaccharide

Guar gum: a neutral polysaccharide isolated from the ground endosperm of Indian cluster bean Cyamopis tetragonolobus; used as a stabilizer and thickener in processed foods

Gum: a diverse group of complex carbohydrate derivatives that are amorphous, translucent, and water-soluble; produced by plants following mechanical injury

Gum arabic: the dried, gummy exudate from the stem and branches of Acacia senegal and other A. species; consists mainly of calcium, potassium, and magnesium salts of arabic acid; used as an emollient and thickening agent in food products; also called acacia

Gum tragacanth: gummy exudate obtained from several species of Astragalus; used as an emollient and emulsifier

Hemicellulose: any of several alkali-soluble polysaccharides found in the walls of plants; includes polymers of xylose, mannose, L-arabinose, glucuronic acid, and galacturonic acid

Insoluble fiber: fiber not soluble in water, depending upon the method of extraction; includes cellulose, lignin, some hemicelluloses, and sometimes pectin

Ispaghula: mucilaginous seeds of Plantago ovata

Karaya gum: see sterculia gum

Klasen lignin: insoluble residue produced by the hydrolysis of cellulose with 72% H2so4

Lectin: a hemagglutinating protein substance in the saline extracts of the seeds of certain plants

Lignin: constituent of the secondary cell wall of plants; a polymer of phenylpropanoid units

Locust bean gum: carob gum; obtained from the leguminous tree Ceratonia siligua

Metamucil®: a proprietary name for psyllium seed hydrocolloid

Microcrystalline cellulose: partially depolymerized form of cellulose prepared from purified wood cellulose by acid hydrolysis

232

Mucilage: a gum-like plant cell product composed of sulfate esters of complex polysaccharides

Neutral detergent fiber: the residue remaining after plant materials are boiled with neutral sodium lauryl sulfate and EDTA; considered a measure of lignin, cellulose, and hemicellulose

Noncellulosic polysaccharides: plant cell wall constituents including hemicelluloses and pectic substances

Nonstarch polysaccharides: plant cell wall constituents ~ncluding cellulose and noncellulosic polysaccharides

Pectin: a constituent of plant cell walls; consists of methyl esters of a complex uronic acid composed of galactose and arabinose

Pentosan: any of a group of polysaccharides able to form pentoses by hydrolysis; includes araban (from gum arabic) and xylan (from wood and straw)

Phytate: the calcium and mixed calcium and magnesium salt of inositol hexaphosphoric acid

Plant fiber: synonymous with dietary fiber.

Psyllium: dried, ripe seed of the herb Plantago psyllium, Pe arenana, and Pe ovata; contains a mucilage useful as a bulk laxative -

Psyllium hydrocolloid or mucilloid: white or cream-colored, nearly odorless, granular powder composed of the mucilag­inous portion of psyllium seeds or husks; used in treatment of chronic constipation

Saponin: any plant glycoside with detergent action that can be hydrolyzed to yield a carbohydrate and steroid component that is a sapogenin

Soluble fiber: fiber soluble in water, depending upon the method of extraction; includes gums, pectins, mucilages, and some hemicelluloses

Sterculia gum: exudate obtained from several species of Sterculia; used as a bulk cathartic; also called karaya gum, Indian gum

Viscous: offering resistance to flow or alteration of shape; the result of molecular cohesion of materials dissolved or suspended in fluid

233

XI. STUDY PARTICIPANTS

ad hoc EXPERT PANEL ON DIETARY FIBER

James W. Anderson, M.D. Chief Endocrine-Metabolism Section Veterans Administration Hospital Lexington, Kentucky 40511

Ri~hard L. Atkinson, Jr., M.D. Chief, Division of Clinical

Nutrition Eastern Virginia Medical School VA Medical Center Hampton, Virginia 23667

Tim Byers, M.D., M.P.H.* Department of Social and

Preventive Medicine State University of New York

at Buffalo Buffalo, New York 14214

Martin H. Floch, M.D. Chief of Medicine Norwalk Hospital Yale University School

of Medicine Norwalk, Connecticut 06856

Saxon Graham, PhoD.* Department of Social and

Preventive Medicine State University of New York

at Buffalo Buffalo, New York 14214

Jon A. Story, Ph.D. Professor Depaitment of Foods and Nutrition Purdue University West Lafayette, Indiana 47907

* Alternates ** Deceased August 1, 1986

235

Lucien R. Jacobs, M.D. Chief of Nutrition Division of Gastroenterology Cedars-Sinai Medical Center Los Angeles, California 90048

June L. Kelsay, Ph.D. Research Nutritionist ARS, USDA Beltsville Human Nutrition

Research Center Beltsville, Maryland 20705

David Kritchevsky, Ph.D. Associate Director Wistar Institute of Anatomy

and Biology Philadelphia,

Pennsylvania 19104

Judith A. Marlett, Ph.D. Professor Department of Nutritional

Scie·nce University of Wisconsin Madison, Wisconsin 53706

Bandaru S. Reddy, D.V.M., Ph.D. Department of Nutritional

Biochemistry Naylor Dana Institute for

Disease Prevention American Health Foundation Valhalla, New York 10595

George V. Vahouny, Ph.D.** Professor George Washington University

Medical Center Washington, D.C. 20037

CENTER FOR FOOD SAFETY AND APPLIED NUTRITION, FDA 200 C Street, s.w.

Washington, D.C. 20204

James P. Harwood, Ph.D. Supervisory Pharmacologist Chief Experimental Clinical

Research Section

Margaret McDowell, M.PoH. Nutritionist Clinical Nutrition Assessment

Section Division of Nutrition

Elizabeth A, Yetley, Ph.D. Chief Clinical Nutrition Assessment

Section Division of Nutrition

LIFE SCIENCES RESEARCH OFFICE

Sue Ann Anderson, Ph.D. Senior Staff Scientist

Beverly R. Lea Technical Services Manager

Barbara L. Durant Project Secretary

Louise S. Erlick Librarian

Kenneth D. Fisher, Ph.D. Director

J. Elaine Huey Technical Literature Specialist

Judith F. Miller Administrative Assistant

Susan M. Pilch, Ph.D. Senior Staff Scientist

Richard W. Leukroth, Jr., M.S. Staff Scientist

Stephen H. Simpson Technical Literature Assistant

John M. Talbot, M.D. Senior Medical Consultant

236