effects of wheat bran and olestra on objective measures of stool and subjective reports of gi...

Effects of Wheat Bran and Olestraon Objective Measures of Stool andSubjective Reports of GI SymptomsJ. McRorie, Ph.D., J. Kesler, B.S, M.B.A., L. Bishop, T. Filloon, Ph.D., G. Allgood, Ph.D.,M. Sutton, Ph.D., M.B., B.C. sir, T. Hunt, M.D., Ph.D., A. Laurent, M.D., and C. Rudolph, M.D., Ph.D.The Procter & Gamble Company, Cincinnati, Ohio; PPD Pharmaco, Austin, Texas; and Children’s HospitalMedical Center, Cincinnati, Ohio

OBJECTIVE: The aim of this study was to compare the effectsof two nondigested, nonabsorbed dietary components onobjective and subjective measures of gastrointestinal (GI)function.

METHODS: A placebo-controlled parallel study comparedthe effects of wheat bran (20 g/day or 40 g/day in cereal), awell-known dietary fiber, with those of olestra (20 g or 40g/day in potato chips), a nonabsorbed fat, on stool output,stool apparent viscosity (log peak force for extrusion [PF]),stool water content, and GI symptoms. Sixty subjects re-sided on a metabolic ward for 9 days: 3 days baseline and 6days treatment.

RESULTS: Compared with placebo, consumption of 20 g/daywheat bran for 6 days resulted in a rapid (within 38 h)increase in mean (6SE) stool output (placebo, 1506 29g/day; bran, 2466 35 g/day, p , 0.05), a directionalincrease in mean stool water content (placebo, 81.26 0.8%;bran, 83.96 0.8%), stool water output (placebo, 1596 54g/day; bran, 2386 30 g/day), and bowel movement fre-quency (BM/day) (placebo, 2.26 0.4; bran, 2.66 0.4), andno stool-softening effect (placebo log PF, 2.96 0.1 g; branlog PF, 2.96 0.1 g). Wheat bran 40 g/day results were notsignificantly different from wheat bran 20 g/day. Comparedwith placebo, consumption of olestra 20 g/day and 40 g/dayfor 6 days showed no significant difference in mean stooloutput (1516 18 g/day and 2046 28 g/day, respectively),mean BM frequency (1.86 0.2 BM/day and 2.16 0.3BM/day, respectively), and stool water output (1386 13g/day and 1846 31 g/day, respectively), a significant (p ,0.05) decrease in stool water content (75.56 1.7% and72.6 6 2.2%, respectively), and either no effect on stoolapparent viscosity (olestra 20 g/day, mean log PF, 3.060.1 g) or a gradual stool-softening effect beginning studyday 6 (olestra 40 g/day, log PF, 2.76 0.1 g). None of thetreatment groups showed a significant increase in GI symp-toms compared with placebo.

CONCLUSIONS: Consumption of wheat bran in excess oflevels in a typical Western diet significantly increased stooloutput, but did not soften normal-viscosity stool nor result in

an increase in common GI symptoms. The observed plateaueffect for wheat bran at 40 g/day suggests a maximal me-chanical stimulatory effect. Consumption of olestra in ex-cess of usual snacking conditions did not result in a signif-icant increase in stool output or common GI symptoms. Atthe highest level tested, olestra resulted in a gradual stool-softening effect after several days of consumption. (Am JGastroenterol 2000;95:1244–1252. © 2000 by Am. Coll. ofGastroenterology)

INTRODUCTION

Definitions of dietary fiber can be quite variable, but thecharacteristic feature is a resistance to hydrolysis by endog-enous enzymes of the human gastrointestinal (GI) tract(1–4). The term “dietary fiber” is typically associated withnondigestible cellulose and hemicellulose components ofplant cell walls, but has been expanded to include pectins,gums, mucilages, lignin, and synthetic polysaccharides(e.g.,methylcellulose) (2). Dietary substances that share theproperty of indigestibility and behave as dietary fiber in-clude oligosaccharides, sugar alcohols (e.g., sorbitol), hy-drocarbon waxes and oils, connective tissue proteins, chitinexoskeletons, and hair (2). A relatively new component ofthe Western diet, sucrose polyester (olestra), also shares theproperty of indigestibility in the human GI tract.

Although it is generally accepted that consumption ofincreasing levels of dietary fiber will increase stool outputand colonic transit rate (2, 5), the mechanism by whichdifferent fiber types affect colonic function is less clear.Predominantly insoluble fibers, such as wheat bran, do nothave a significant water-holding capacity (6), yet still in-crease transit rate and stool bulk (2, 6, 7). A key factor forthe laxative efficacy of insoluble fibers appears to be particlesize, with coarse particles having a greater laxative effect atthe same dose compared with fine particles (6, 7). Theimportance of particle size and the negligible effect ofwater-holding capacity associated with wheat fiber was fur-ther demonstrated in a clinical study comparing wheat branand indigestible plastic particles (8), which showed that

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 95, No. 5, 2000© 2000 by Am. Coll. of Gastroenterology ISSN 0002-9270/00/$20.00Published by Elsevier Science Inc. PII S0002-9270(00)00809-1

laxative efficacy was similar between similar-sized parti-cles. Taken together, these data suggest that wheat bran mayexert a laxative effect by mechanically stimulating bowelmucosa, independent of water-holding capacity and/or bac-terial fermentation. In contrast, a predominantly soluble,gel-forming fiber, such as psyllium, has a high water-hold-ing capacity (in vitro, psyllium$40 g water/g fiber) (9) thatis capable of resisting dehydration in the bowel (10), result-ing in high-moisture, mechanically soft, bulky stools (10,11). Olestra, a nondigested, nonabsorbed fat, also has beenshown to exert a dose-responsive stool-softening/stool-bulk-ing effect after several days of high-level consumption (12,13), but it is not clear how this effect compares to dietaryfiber. Further, wheat fiber has been shown to increase stoolbulk, but its effect on stool viscosity in humans is notknown. The aim of this study was to compare and contrastthe effects of wheat bran and olestra on objective measuresof stool and subjective reports of GI symptoms.

MATERIALS AND METHODS

Study DesignThis was a single-site (PPD Pharmaco, Inc., Austin, TX),double-blind (potato chips only), placebo-controlled paral-lel-design clinical trial. The study was open to all subjectsaged 18 yr or older who did not have conditions that mightpreclude participation in a controlled-feeding/stool-collec-tion study (e.g.,recent GI surgery, moderate/severe consti-pation). Subjects were housed on a metabolic ward for 9days and consumed mealsad libitum that conformed toAmerican Heart Association Step I diet guidelines (no morethan 30% of calories from fat). Nonalcoholic beverageswere availablead libitum.

The first 3 days (days 1–3) comprised the baseline period,during which subjects consumed placebo snacks. Duringthis period, subjects rated each bowel movement (BM) on a7-point scale (15 watery, difficult-to-control diarrhea; 45normal; 75 hard, difficult-to-pass constipation); all stoolswere collected for analysis, and GI symptoms were rated ona scale of 0 (none) to 5 (severe) and recorded daily. The final6 days (days 4–9) comprised the treatment period, duringwhich subjects consumed snacks according to their ran-domly assigned treatment group. Sixty subjects were ran-domly assigned to one of five treatment groups balanced forgender and age: placebo (5 oz triglyceride potato chips/day),olestra 20 g/day (2.5 oz olestra potato chips, 2.5 oz triglyc-eride potato chips/day), olestra 40 g/day (5oz olestra potatochips/day), bran 20 g/day (0.55 cups Bran Buds and 0.55cups Puffed Wheat), and bran 40 g/day (1.1 cups BranBuds). The low-fiber (Puffed Wheat) and high-fiber (BranBuds) cereals are quite different in appearance and texture,and could not be blinded. To minimize any bias this mightintroduce into the study, the subjects consuming bran werehoused on an adjacent but separate isolation ward from thesubjects consuming potato chips. Again, subjects rated eachBM, collected all stools for analysis, and recorded any GI

symptoms. Study snacks were purchased and repackaged inplain, food-grade bags made of white foil laminate to main-tain the blind. The protocol was reviewed and approved byan independent Institutional Review Board (Research Con-sultants Review Committee, Austin, TX).

Measurement of Stool Content,Apparent Viscosity, and Stool Viscosity RatioAccess to toilets was controlled by study staff, and BMswere collected in preweighed/labeled stool collection con-tainers and returned to study personnel. Stool samples wereweighed, then apparent viscosity was measured objectivelyby a peak force (PF) back-extrusion technique (10) using aStevens QTS25 Texture Analyzer (Michael G. Brown &Assoc., Newtown, PA). A recent study showed a correlationbetween a visual description of stool and an objective mea-sure of mean apparent viscosity: firm, 2149 g PF (log PF3.3); soft, 816 g PF (log PF 2.9); loose, 221 g PF (log PF2.3); and liquid, 59 g PF (log PF, 1.8) (13). After determi-nation of stool apparent viscosity, each BM on study days 3,8, and 9 was returned to its original container and homog-enized, and an aliquot was shipped on dry ice for determi-nation of stool water content by lyophilization to Hill TopResearch, Ltd., West Palm Beach, FL. Stool viscosity ratiois the ratio of the highest value for stool apparent viscositydivided by the lowest value for stool apparent viscosity persubject per day (13). If only one BM was produced in agiven day, the stool viscosity ratio had a value of 1.

Statistical MethodsData from all subjects were included in the analysis. Ob-jective and subjective measures were averaged over thebaseline and treatment periods, respectively, with a changefrom baseline being determined from each subject’s averagetreatment and baseline values. Primary analysis came fromthe (unadjusted) treatment period responses. A Wilcoxonrank-sum approach was performed for pair-wise treatmentcomparisons of the treatment phase data for all responsevariables. Treatment groups were balanced for gender, age,and race. Subjects were considered to be evaluable if theyconsumed at least 90% of the assigned active treatmentmaterial while in study.

RESULTS

DemographicsSixty subjects, ages 18 to 82 yr (mean age, 48.7 yr; 26 men,34 women, 43 whites, seven blacks, nine Hispanics, oneother) were enrolled in the study. All subjects completed atleast the first 6 days of the study, and all subjects met the$90% compliance criterion for consuming test product forthe study days completed. One subject (bran 20 g/day treat-ment group) withdrew from the study on study day 7 afterreports of nausea and vomiting. There were no seriousadverse events.

1245AJG – May, 2000 GI Effects of Wheat Bran and Olestra

Stool Weight, Stool Output, and BM FrequencyBowel movements (1131 stool samples) were collected andmeasured for stool wet weight (g/BM) and summed for dailytotal stool output (g/day). During the treatment period, meanstool weight (6SE) for the bran treatment groups (bran 20g/day, 1016 12 g/BM; 40 g/day, 1226 23 g/BM) and theolestra treatment groups (olestra 20 g/day, 1006 17 g/BM;olestra 40 g/day, 1096 14 g/BM) was not significantlydifferent from the placebo group (736 8 g/BM). Mean stooloutput (6SE) was 1506 29 g/day for the placebo group,151 6 18 g/day for the olestra 20 g/day group, 2046 28g/day for the olestra 40 g/day group, 2466 35 for the bran20 g/day group, and 2316 27 g/day for the bran 40 g/daygroup (Fig. 1, Table 1). Stool output increased rapidly(within 38 h of first dose) for the bran treatment groups (Fig.

2, study day 5), but did not increase (20 g/day) or increasedgradually (40 g/day) for the olestra treatment groups. Meandaily stool output for the bran treatment groups was signif-icantly (p , 0.05) higher than the placebo and olestra 20g/day treatment groups. Mean BM frequency was,3 BMs/day for all treatment groups. During the baseline period,mean (6SE) BM frequency was 1.886 0.13 BMs/day.During the treatment period, mean (6SE) BM frequency/day was 2.26 0.4 for the placebo group, 1.86 0.2 for theolestra 20 g/day group, 2.16 0.3 for the olestra 40 g/daygroup, 2.61 0.4 for the bran 20 g/day group, and 2.46 0.5for the bran 40 g/day group (Fig. 3). There were no signif-icant differences (p . 0.05) in BM frequency betweentreatment groups.

The detectable differences for each variable were fecal

Figure 1. Stools (n5 1131) were collected and weighed. Compared with placebo (PLA), mean stool output (g/day) was significantly (p ,0.05) greater for both the 20-g/day and 40-g/day wheat bran groups. Compared with olestra (OA) 20 g/day, mean stool output was alsosignificantly (p , 0.05) greater for both the 20-g/day and 40-g/day wheat bran groups. Note the increase in stool output between baselineand placebo (treatment phase), presumably due to adaptation to the AHA step I diet.

Table 1. Objective Measures of Stool

PlaceboOlestra

20 g/dayOlestra

40 g/dayBran

20 g/dayBran

40 g/day

Stool apparent Min 2.6 2.6 2.3 2.2 2.8viscosity Max 3.2 3.4 3.3 3.2 3.2(log force; g) Mean 2.9 3.0 2.7 2.9 2.9

Median 2.9 3.0 2.7 3.0 2.9Stool output Min 49 66 101 86 58(g/day) Max 424 236 432 557 460

Mean 150 151 204 246 231Median 135 138 170 237 211

Objective measures of stool include stool apparent viscosity (log peak force for back extrusion, g) and stool output (g/day). The minimum (Min), maximum (Max), mean, andmedian values for each treatment group are listed.

1246 McRorie et al. AJG – Vol. 95, No. 5, 2000

water, 5.67%; water output, 146 ml/day; stool viscosity (logpeak force), 0.28 g; stool output, 116 g/day; and BM fre-quency, 1.49 BMs/day. These values represent the differ-ences, in their respective absolute units, that our study had80% power for detecting using 2-sided 0.05 significancetesting.

Stool Viscosity and Stool Viscosity RatioApparent viscosity was evaluated objectively by measuringapparent viscosity as peak force for extrusion (PF). Forreference, visual comparison ratings (12, 13) are given inparentheses (e.g.,“firm”). During the baseline period, meanstool apparent viscosity was relatively constant for all treat-ment groups at a log PF of approximately 3.0 (firm stool,Fig. 4). During the treatment phase, the olestra 20 g/day,bran 20 g/day, and bran 40 g/day showed no stool-softeningeffect compared to placebo (Fig. 4, study days 4–9). Theolestra 40 g/day treatment group showed a gradual decreasein stool apparent viscosity beginning the third day of con-sumption (study day 6, Fig. 4). Mean log PF during thetreatment phase was placebo, 2.886 0.06 g (firm); olestra20 g/day, 3.006 0.06 g (firm); olestra 40 g/day, 2.7060.08 g (soft); bran 20 g/day, 2.891 0.08 g (firm); and bran40 g/day, 2.946 0.04 g (firm) (Table 1). Stool apparentviscosity for the olestra 40 g/day treatment group was sig-nificantly (p , 0.05) less than for the placebo and brantreatment groups. Mean stool viscosity ratio (6SEM) was

calculated for baseline and each treatment group: baseline,2.286 4.02; placebo, 2.366 2.90; olestra 20 g/day, 1.7861.22; dietary fiber 20 g/day, 2.336 1.88; olestra 40 g/day,1.836 1.90; and dietary fiber 40 g/day, 1.996 1.73.

Stool Water Content and Stool Water OutputStool water content was measured for each BM on studydays 3, 8, and 9. Stool water content ranged from 58.8%(olestra 40 g/day, firm stool) to 99.8% (placebo, liquidstool). During the baseline phase, mean stool water content(6SE) was 79.26 1.2 for the placebo group, 80.96 1.9%for the olestra 20 g/day group, 80.26 1.6% for the olestra40 g/day group, 80.56 1.1% for the bran 20 g/day group,and 81.26 1.1% for the bran 40 g/day group. During thetreatment phase, mean stool water content (6SE) for theolestra 20 g/day (75.56 1.7%) and olestra 40 g/day (72.662.2%) groups was significantly (p , 0.05) less than that forthe placebo group (81.26 0.8%), as well as the bran 20g/day (83.96 0.8%) and 40 g/day (83.46 0.7%) treatmentgroups (Fig. 5). During the treatment period, mean stoolwater output (6SE) was 1596 54 g/day for the placebogroup, 1386 13 g/day for the olestra 20 g/day group, 184631 g/day for the olestra 40 g/day group, 2386 30 g/day forthe bran 20 g/day group, and 2316 35 g/day for the bran 40g/day group (Fig. 6). There were no significant differences(p . 0.05) in stool water output between the placebo andolestra or bran treatment groups. Stool water output for the

Figure 2. Graph of mean stool output by study day. Study days 1–3 represent the baseline period, and study days 4–9 represent thetreatment period. Both the 20-g/day and 40-g/day wheat bran groups showed a marked increase in stool output within 38 h of the first dose(study day 5). The olestra 20 g/day treatment group did not show an increase in stool output, and the olestra 40-g/day treatment groupshowed an increase in stool output on the third day of consumption.


Figure 3. Bar graph of mean BM frequency by treatment group. Mean BM frequency was,3 BMs/day for all treatment groups. Therewere no significant differences in BM frequency between any treatment groups.

Figure 4. Graph of stool apparent viscosity (log peak force) by study day. Stool apparent viscosity remained relatively constant andcomparable during baseline, and decreased over time for all treatment groups, but only the olestra 40-g/day treatment group showed asignificant stool-softening effect beginning the third day of treatment (study day 6).


olestra 20 g/day treatment group was significantly (p ,0.05) less than for the bran 20 g/day treatment group.

Bowel Movement RatingSubjects rated each BM (n5 1131 BMs) on a 7-point scale(1 5 watery, difficult-to-control diarrhea; 45 normal; 75hard, difficult-to-pass constipation). Mean BM rating for thebaseline period was 3.966 0.08. During the treatmentphase, mean BM rating was 3.76 0.2 for the placebo group,3.5 6 0.2 for the olestra 20 g/day group, 3.06 0.2 for theolestra 40 g/day group, 3.36 0.2 for the bran 20 g/daygroup, and 3.76 0.1 for the bran 40 g/day group. Mean BMrating for the olestra 40 g/day treatment group was signif-icantly (p , 0.05) less compared to the placebo and bran 40g/day treatment groups.

Gatrointestinal SymptomsSelf-assessment of GI experiences were rated on a scale of0 (none) to 5 (severe) and recorded by subjects in a dailydiary. Compared with placebo, there were no significantdifferences in the severity of GI symptoms in any of thetreatment groups (Table 2). There were also no episodes offecal incontinence in any of the treatment groups.

DISCUSSION

Wheat bran, a predominantly insoluble dietary fiber, is notdigested by endogenous enzymes of the human GI tract andis known to increase fecal weight and colonic transit rate.

The mechanism by which it exerts its effects on colonicfunction, and its dose-responsive effects on stool viscosity,were not clear. Olestra, a nondigested, nonabsorbed dietaryfat, also is not digested by endogenous enzymes of thehuman GI tract and has been shown to increase stool output1 g for each gram consumed and soften stool in a dose-responsive manner (12, 13). The aim of this study was tocompare and contrast the effects of wheat bran and olestraon objective measures of stool and subjective reports of sixcommon GI symptoms.

Stool output is highly dependent on dietary fiber intake.The United States has one of the lowest recorded stooloutputs in the world (approximately 100 g per day) (14),presumably due to calorie-dense, low-fiber diets. In coun-tries with less calorie-dense, high-fiber diets, mean stooloutput among healthy individuals is much higher: India (311g/day), Peru (325 g/day), Malaysia (465 g/day), and Uganda(470 g/day) (14). Compared with the U.S. average stooloutput of only 100 g/day, the AHA Step I diet (#30% ofcalories from fat) consumed in this study resulted in arelatively high stool output in the placebo group (150g/day). During the treatment phase, consumption of wheatbran 20 g/day resulted in a stool output more than twice thatof baseline within 38 h of the first dose, an increase of 9 gof stool per gram of bran. It is interesting to note that branconsumed at 40 g/day did not increase stool output abovethat observed with 20 g/day. The laxative efficacy of wheatbran is dependent on the final particle size consumed, with

Figure 5. Graph of stool water content (%). Stool water content for the olestra treatment groups was significantly (p , 0.05) less than thatfor the placebo and bran treatment groups. There was no significant difference between the placebo and bran treatment groups.


coarse particles having a greater laxative effect at the samedose compared with fine particles (6, 7). The importance ofparticle size and the negligible effect of water-holding ca-pacity (6) in the laxative efficacy of wheat bran was clearlydemonstrated in a clinical study (8) comparing wheat branground to varying degrees to plastic particles cut to match

the size of the wheat particles. In that study, plastic particles,which are not fermented and have no appreciable water-holding capacity, exerted a similar laxative efficacy to wheatfiber: coarse particles had a greater laxative effect than fineparticles. This work has been supported by a more recentstudy that demonstrated that the shape of the plastic particles

Figure 6. Graph of mean stool water output by treatment group. There were no significant differences (p . 0.05) in stool water outputbetween the placebo and olestra or bran treatment groups. Stool water output for the olestra 20-g/day treatment group was significantly (p , 0.05)less than for the bran 20-g/day treatment group. Stool water content was determined for stools collected during the last two days of the treatmentphase, resulting in a reported stool water output value directionally greater than the reported stool output value for the placebo group.

Table 2. Gastrointestinal Symptom Scores (Mean6 SEM)

PlaceboOlestra

20 g/dayOlestra

40 g/dayBran

20 g/dayBran

40 g/day

BloatingBaseline 0.506 0.29 0.946 0.36 0.676 0.33 0.396 0.18 0.146 0.36Treatment 0.716 0.27 0.976 0.30 0.866 0.33 0.926 0.27 0.366 0.13

CrampingBaseline 0.146 0.06 0.696 0.24 0.286 0.18 0.286 0.16 0.066 0.06Treatment 0.226 0.07 0.796 0.30 0.386 0.13 0.476 0.17 0.216 0.07

Passing gasBaseline 1.506 0.27 1.336 0.33 1.196 0.31 1.316 0.14 1.286 0.21Treatment 1.646 0.26 1.366 0.36 1.536 0.42 1.966 0.24 1.246 0.26

HeartburnBaseline 0.446 0.29 0.196 0.14 0.536 0.21 0.226 0.13 0.196 0.11Treatment 0.496 0.22 0.356 0.22 0.506 0.25 0.406 0.17 0.226 0.11

NauseaBaseline 0.036 0.03 0.086 0.08 0.086 0.08 0.176 0.14 0.006 0.00Treatment 0.136 0.07 0.116 0.07 0.046 0.03 0.466 0.36 0.016 0.01

UrgencyBaseline 0.696 0.26 0.316 0.10 0.316 0.12 0.286 0.12 0.316 0.10Treatment 0.566 0.29 0.406 0.15 0.606 0.23 0.406 0.14 0.406 0.16

Gastrointestinal symptoms were rated by subjects on a scale of 0 (none) to 5 (severe) and recorded in a daily diary. With the exception of passing gas, all mean GI symptomseverity scores were,1 (mildest rating). Mean severity for passing gas was#2. Compared with placebo, none of the treatment groups showed a significant increase in the severityof GI symptoms.


may also be a determinant of laxation efficacy (15). Takentogether, these data, and the observed plateau effect ofwheat bran at 40 g/day on stool output, suggest that there isa maximal mechanical stimulatory effect for wheat bran.

In contrast to the effects of wheat bran, consumption ofolestra did not significantly affect stool output, but direc-tionally increased stool output approximately 1 g for eachgram consumed. The 20 g/day olestra dose (2.5 ounces ofpotato chips daily for 6 consecutive days) exceeded thechronic daily intake of the highest consuming age/genderpopulation subgroup of savory snacks (e.g., potato chips)based on Market Research Corporation of America(MRCA) data (16), and represents the highest level ofchronic consumption predicted by the Food and Drug Ad-ministration (17). The 40 g/day olestra dose (5 ounces ofpotato chips daily for 6 consecutive days) exceeded thechronic daily intake of the highest consuming subgroup ofsavory snacks (e.g.,potato chips) by several fold based onMRCA data (16) and is twice the highest level of chronicconsumption predicted by the Food and Drug Administra-tion (17). These data show that consumption of olestra undernormal snacking conditions, which are well below the levelsused in this study, would have a negligible effect on stooloutput. Neither wheat bran nor olestra exhibited a significanteffect on stool weight (g/BM) or BM frequency.

Consumption of a soluble gel-forming dietary fiber, suchas psyllium, results in a modest, dose-responsive increase instool output (3.7 g of stool per gram consumed) (2), anincrease in stool water content (10, 11), and a decrease instool apparent viscosity (10), resulting in soft bulky stools.Insoluble dietary fibers, such as wheat bran, are known toincrease stool bulk, but their effects on stool viscosity hadnot been determined. In this study, consumption of wheatfiber at 20 g/day rapidly increased stool output and direc-tionally increased stool water content compared to placebo.A recent study demonstrated a close association of stoolwater content and stool apparent viscosity (10), yet con-sumption of large amounts of dietary fiber in this studyresulted in a directional increase in stool water content withno concomitant decrease in stool apparent viscosity. This islikely due to the nonlinear relationship between stool ap-parent viscosity and stool water content (10). For stool watercontent#75%, a small change in stool water content resultsin a relatively large change in stool apparent viscosity. Incontrast, for stool water content$80%, a large increase instool water content results in a relatively small change instool apparent viscosity (10). The bran treatment groups hada baseline stool water content of approximately 81%, and atreatment mean stool water content of 83–84%, resulting inlittle or no change in stool apparent viscosity. This obser-vation is supported by data from three individual subjects inthe bran treatment groups who exhibited very firm/hardstools (,75% stool water content) in the baseline period(mean log PF, 3.4 g, 3.3 g, and 3.7 g) and markedly softerstools in the treatment period (mean log PF, 3.1 g, 2.9 g, and3.2 g, respectively, a 50–65% reduction in stool apparent

viscosity). The observation is further supported by twostudies in which wheat fiber was fed to pigs that exhibitedfirm/hard stool in baseline (,75% water content) and sig-nificantly (p , 0.01) softer stool during treatment withwheat fiber (18, 19). Taken together, these data suggest thatconsumption of wheat fiber will only result in a stool-softening effect if subjects are experiencing firm/hard stoolsbefore treatment.

Consumption of olestra 20 g/day resulted in a decrease instool water content and did not affect stool apparent viscos-ity compared with placebo. Consumption of olestra at 40g/day also resulted in a decrease in stool water contentcompared with placebo, and a gradual stool-softening effectafter several days of consumption, similar to the effects ofsoluble dietary fiber (10, 11). Olestra is an inert semisolidthat is not fermented by bacteria, has no osmotic activity,and is not a mechanically coarse fiber, so it would not beexpected to increase stool water content. The decrease instool water content observed with olestra consumption inthis study is caused by the increase in the solid componentof stool (olestra remains with the solid component duringlyophilization), resulting in a calculated smaller percentwater content. The stool-softening effect of olestra, a semi-solid at body temperature, appears to be dependent upon theconcentration of olestra in stool.

In contrast to anecdotal reports of cramping and other GIsymptoms associated with a marked increase in dietary fiberconsumption, this study showed no significant increase inGI symptoms compared with placebo. This apparent lack ofincrease in GI symptoms, combined with the observationthat the wheat bran did not cause a marked decrease in stoolviscosity, is consistent with a recent study that proposed anew objective measure: stool viscosity ratio (13). The studyshowed that severe abdominal cramping/pain was associ-ated with the passage of formed stool followed by one ormore loose/liquid stools, resulting in a high stool viscosityratio (highest stool viscosity4 lowest stool viscosity withina given day) (13). Reports of no or mild cramping wereassociated with a mean stool viscosity ratio,5, whereasreports of severe abdominal cramping were associated witha mean stool viscosity ratio of 25. Subjects consumingsorbitol 40 g/day reported severe abdominal cramping andtended to pass formed stools followed by liquid/rice-waterstools, leading to a mean stool viscosity ratio of 19.5 (13).In contrast, stool viscosity ratio for baseline and placebowere 2.0 and 1.9, respectively. In the current study, stoolviscosity ratios were,2.5 for baseline, placebo, and alltreatment groups, predicting no/mild symptoms and no dif-ference from placebo.

The term “abdominal cramping” appears to be a misno-mer, in that it implies spastic contraction of the bowel wall.Lower intestinal symptoms can be replicated by passivestretch of the bowel wallvia stepwise inflation of a balloonin the colon or rectum, generating sensations from vagueawareness to severe cramping pain with increasing intralu-minal pressure (20–22). Similar sensations, from slight dis-


comfort and the urge to defecate to cramping pain, are alsostrongly correlated with high amplitude propagating con-tractions (HAPCs) (23–26), suggesting that physiologicalcolonic motor events can give rise to these sensations. Themodel proposed by the authors (13) showed that readilydeformable loose/liquid digesta propelled by HAPCsagainst formed digesta/stool in the distal bowel could causeacute bowel wall tension, resulting in cramping pain. Ifsubjects had begun the present study with hard stool (pre-sumably like many of the patients who report a low toler-ance for increases in dietary fiber consumption), the largevolume of soft digesta/stool created by high-dose wheatbran ingestion might have generated a sufficiently high ratioas to cause lower abdominal symptoms. Chronic consump-tion of olestra, at doses far exceeding normal snackingconditions, also resulted in a stool viscosity ratio of,2.5and did not increase GI symptoms. This lack of increase in GIsymptoms with olestra consumption is consistent with otherrecent placebo-controlled clinical trials (27–29) in which therewere no differences in GI symptoms between groups consum-ing potato chips prepared in olestra or triglycerides.

In summary, these data show that consumption of wheatbran in excess of levels in a typical Western diet signifi-cantly increased stool output, but did not soften normal-vis-cosity stool nor result in an increase in common GI symptoms.The observed plateau effect for wheat bran at 40 g/day suggestsa maximal mechanical stimulatory effect. Consumption ofolestra in excess of usual snacking conditions did not result ina significant increase in stool output or common GI symptoms.At the highest level tested, olestra resulted in a gradual stool-softening effect after several days of consumption.

ACKNOWLEDGMENT

This study was sponsored by The Procter & Gamble Company.

Reprint requests and correspondence:John McRorie, Ph.D.,The Procter & Gamble Company, P.O. Box 8006 ML 505, Mason,OH 45040.

Received Aug. 6, 1999; accepted Jan. 7, 2000.

REFERENCES

1. Schaller D. Fiber content and structure in foods. Am J ClinNutr 1978;31:S99-S102.

2. Spiller GA. Handbook of dietary fiber in human nutrition, 2nded. Boca Raton, FL: CRC Press Inc., 1993.

3. Trowell H. Ischemic heart disease and dietary fiber. Am J ClinNutr 1972;25:926.

4. Southgate D. The chemistry of fiber. In: Spiller G, Amen R,eds. Fiber in human nutrition. New York: Plenum Press, 1976.

5. Stevens J, VanSoest P, Robertson J, et al. Comparison of effectsof psyllium and wheat bran on gastrointestinal transit time andstool characteristics. J Am Diet Assoc 1988;88:323–6.

6. Brodribb A, Groves C. Effect of bran particle size on stoolweight. Gut 1978;19:60–3.

7. Heller S, Hackler L, Rivers J, et al. Dietary fiber: The effect ofparticle size of wheat bran on colonic function in healthyyoung adult men. Am J Clin Nutr 1980;33:1734–44.

8. Tomlin J, Read N. Laxative properties of indigestible particles.Br Med J 1988;297:1175–6.

9. Struthers BJ. Warning—Feeding animals hydrophyllic fibersources in dry diets. J Nutr 1986;116:47–9.

10. McRorie J, Pepple S, Rudolph C. Effects of fiber laxatives,and calcium docusate on regional water content, and viscosityof digesta in the large intestine of the pig. Dig Dis Sci 1998;43:738–45.

11. McRorie J, Daggy B, Morel J, et al. Psyllium is superior todocusate sodium for treatment of chronic constipation. Ali-ment Pharmacol Ther 1998;12:491–7.

12. Giannella R, Zorich N, Riccardi K, et al. Reports of diarrhea:When is stool water output clinically meaningful? A random-ized, controlled trial of olestra and sorbitol. Gastroenterology1998;114:A372.

13. McRorie J, Zorich N, Riccardi K, et al. Effects of olestra andsorbitol consumption on objective measures of diarrhea: Im-pact of stool viscosity on common gastrointestinal symptoms.Regul Toxicol Pharmacol 2000;31:59–67.

14. Cummings JH, Bingham SA, Heaton KW, et al. Fecal weight,colon cancer risk, and dietary intake of nonstarch polysaccha-rides (dietary fiber). Am J Gastroenterol 1992;103:1783–9.

15. Lewis S, Heaton K. Roughage revisited: The effect on intes-tinal function of inert plastic particles of different sizes andshapes. Dig Dis Sci 1999;44:744–8.

16. Webb DR, Harrison GG, Lee M-J, et al. Estimated consump-tion and eating frequency of olestra from savory snacks usingmenu census data. J Nutr 1997;127:1547S–54S.

17. Federal Register, Part III, Department of Health, and HumanServices, Food, and Drug Administration, 21 CFR Part 172,January 30, 1996.

18. Greenwood-Van Meerveld B, Neeley D, Tyler K, et al. Com-parison of effects on colonic motility and stool characteristicsassociated with feeding olestra and wheat bran to ambulatorymini-pigs. Dig Dis Sci 1999;44:1282–7.

19. Boring G, Cooper R, Brown S, et al. Effects of bran and olestraon digesta/stool viscosity and water content in porcine largeintestine. Gastroenterology 1998;114:G2993.

20. Lembo T, Munakata J, Naliboff B, et al. Sigmoid afferentmechanisms in patients with irritable bowel syndrome. DigDis Sci 1997;42:1112–20.

21. Serra J, Azpiroz F, Malagelda J. Perception, and reflex re-sponses to intestinal distention in humans are modified bysimultaneous or previous stimulation. Gastroenterology 1995;109:1742–9.

22. Zighboim J, Talley N, Phillips S, et al. Visceral perception inirritable bowel syndrome. Dig Dis Sci 1995;40:819–27.

23. Kraus M, Muller-Lissner S. Motility, and lower gut symptoms.Eur J Gastroenterol 1993;5:990–8.

24. Narducci F, Bassotti G, Gaburri M, et al Twenty four hourmanometric recording of colonic motor activity in healthyman. Gut 1987;28:17–25.

25. Bassotti G, Gaburri M. Manometric investigation of highamplitude propagated contractile activity of the human colon.Am J Physiol 1988;255:G660–G664.

26. Lemann M, Flourie B, Picon L, et al. Motor activity recorded inthe unprepared colon of healthy humans. Gut 1995;37:649–53.

27. Sandler R, Zorich N, Filloon T, et al. Gastrointestinal symp-toms of 3181 volunteers ingesting snack foods containingolestra or triglycerides. A 6-week randomized, placebo-con-trolled trial. Ann Intern Med 1999;130:253–61.

28. Cheskin L, Miday R, Zorich N, et al. Gastrointestinal symptomsfollowing consumption of olestra or regular triglyceride potatochips. A controlled comparison. JAMA 1998;279:150–2.

29. Zorich N, Biederman D, Riccardi K, et al. Randomized, dou-ble-blind, placebo-controlled, consumer rechallenge test ofolean salted snacks. Regul Toxicol Pharmacol 1997;26:200–9.


effects of wheat bran and olestra on objective measures of stool and subjective reports of gi...

Documents