The Journal of Nutrition Vol. 128 No. 4 April 1998, pp. 714-719

A Prospective Study of Dietary Fiber Types and Symptomatic Diverticular Disease in Men^1,2

Walid H. Aldoori^{*, 3}, Edward L. Giovannucci^{*, **, 4}, Helaine R H. Rockett^**, Laura Sampson^*, Eric B. Rimm^*, , and and Walter C. Willett^{*, , **}

Departments of ^* Nutrition and  Epidemiology, Harvard School of Public Health and ^** Channing Laboratory, Department of Medicine, Harvard Medical School and Brigham and Women's Hospital, Boston, MA 02115

	ABSTRACT

Abstract Introduction Methods Results Discussion References

To examine prospectively dietary fiber calculated from food composition values based on analytic techniques and specific dietary fiber types in relation to risk of diverticular disease, we analyzed data from a prospective cohort of 43,881 U.S. male health professionals 40-75 y of age at base line; subjects were free of diagnosed diverticular disease, colon or rectal polyps, ulcerative colitis and cancer. The insoluble component of fiber was inversely associated with risk of diverticular disease relative risk (RR) = 0.63, 95% confidence interval (CI), 0.44-0.91, P for trend = 0.02, and this association was particularly strong for cellulose (RR = 0.52, 95% CI, 0.36-0.75, P for trend = 0.002). The association between diverticular disease and total dietary fiber intake calculated from the AOACstandards method was not appreciably different from results using the Southgate or Englyst method [for AOAC method, RR = 0.60, 95% CI, 0.41-0.87; for Southgate method, RR = 0.61, 95% CI, 0.42-0.88; for Englyst method, RR = 0.60, 95% CI, 0.42-0.87, for the highest quintiles]. Our findings provide evidence for the hypothesis that a diet high in dietary fiber decreases the risk of diverticular disease, and this result was not sensitive to the use of different analytic techniques to define dietary fiber. Our findings suggest that the insoluble component of fiber was significantly associated with a decreased risk of diverticular disease, and this inverse association was particularly strong for cellulose.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Diverticular disease is one of the most common disorders of the colon among the elderly in Western societies (Walters and Smith 1990). In North America, diverticular disease is estimated to occur in one third of all persons >45 y of age, and in two thirds of all persons >85 y of age (Roberts and Veidenheimer 1990). In the great majority of cases, the condition is asymptomatic (Schwartz and Graham 1988), and only between 10 and 25% of affected individuals develop symptoms (Roberts and Veidenheimer 1990). Symptomatic diverticular disease results in 200,000 hospitalizations in the U.S. annually (Thompson and Patel 1986).

Earlier this century, diverticular disease was widely believed to be extremely rare and was regarded as a pathological curiosity. This prompted Burkitt and Painter to call it "a 20th century problem" or "a disease of Western civilization" (Painter and Burkitt 1971 and 1975), in contrast to its rarity in many developing countries (Painter 1985b). This sharp contrast has been largely attributed to dietary changes, mainly in the decline of dietary fiber intake from cereal grains (Painter and Burkitt 1975). The dietary fiber hypothesis is supported by human (Brodribb and Humphreys 1976, Findlay et al. 1974, Gear et al. 1979, Leaky et al. 1985, Manousos et al. 1985, Miettinen and Tarpila 1978) and animal studies (Fisher et al. 1985, Hogdson 1972, Jaskiewicz 1986). Previous case-control studies have consistently found that patients with diverticula consumed less fiber than did nonpatients, and we observed similar findings, as well as a beneficial effect of physical activity, in the prospective Health Professionals Follow-up Study (HPFS)⁵ (Aldoori et al. 1994, 1995a and 1995b). Agreement concerning the optimal methods for analysis of fiber subfractions is still an extremly contentious area; however, we studied prospectively dietary fiber calculated from several analytic techniques and specific fiber types in our database to explore the relative predictiveness of the various fiber fractions to the risk of symptomatic diverticular disease in the HPFS.

	SUBJECTS AND METHODS

Abstract Introduction Methods Results Discussion References

Study population.  The HPFS is a prospective study of heart disease and cancer among 51,529 U.S. male health professionals 40-75 y of age in 1986 and residing in 50 states. The study population consists of 29,683 dentists, 3745 optometrists, 2218 osteopathic physicians, 4185 pharmacists, 1600 podiatrists and 10,098 veterinarians. The study began in 1986 when cohort members completed a detailed food-frequency questionnaire and provided information about medical history, age, weight, height, smoking, alcohol consumption, physical activity and history of professionally diagnosed medical conditions. Every 2 y (1988, 1990 and 1992), follow-up questionnaires were sent to update information on potential risk factors and to identify newly diagnosed cases of various diseases.

Dietary and physical activity assessment.  To assess dietary intake, we used a semiquantitative food-frequency questionnaire that was validated in this cohort (Rimm et al. 1992). The correlation between the diet records and the dietary questionnaire was 0.68 for total dietary fiber in our validation study. Also, there was a high degree of validity for foods that were the major contributors to total fiber intake (Feskanich et al. 1993). The dietary questionnaire included 131 food items plus vitamins and mineral supplements that collectively account for over 90% of the absolute intake of the major nutrients consumed by this population (Rimm et al. 1992). The intake of dietary fiber was computed by multiplying the consumption frequency of each food item by the fiber content of the specified portion. The values for dietary fiber were derived from food composition data tables based on the AOAC method (Prosky et al. 1984, 1985, and 1988, Vollendorf and Marlett 1993). We have also expanded our dietary fiber database to include dietary fiber content of foods based on the work of Southgate and colleagues (Paul and Southgate 1978, Southgate et al. 1976), as well as the Englyst method (Holland et al. 1991). Data on soluble fiber, insoluble fiber, cellulose, hemicellulose and lignin were available directly from some manufacturers, calculations from ingredients (personal communication from USDA), analysis done by Kellogg's specifically for the Channing laboratory, as well as other dietary fiber food composition data (Horvath and Robertson 1986, USDA 1993). The composition values for dietary lipids were obtained from U.S. Department of Agriculture sources (USDA 1993).

Dietary intake values were adjusted for total energy intake by using regression analysis (Willett and Stampfer 1986, Willett 1990) to focus on the composition of diets rather than on absolute intake, which is a combination of composition and overall food intake. This procedure also reduces variation caused by underreporting or overreporting of intake on the food-frequency questionnaire, thus improving the precision of nutrient measurement (Rimm et al. 1992, Willett 1990).

Physical activity was assessed by using a validated self-administered questionnaire that assesses the average weekly recreational physical activity (Chasan-Taber et al. 1996). Eight moderate or vigorous activities were listed on the questionnaire, and participants were asked to report the average time per week spent at each activity (Aldoori et al. 1995a). The contribution of each activity was based on its energy expenditure requirements in metabolic equivalents (MET) (Ainsworth et al. 1993), multiplied by the duration of the activity. The contributions from each activity were summed to give a physical activity index in total MET expressed as hours per week.

Identification of diverticular disease cases.  Follow-up questionnaires were sent in 1988, 1990 and 1992 to all study participants. On the forms mailed in 1990 and 1992, we asked whether diverticular disease had been diagnosed during the previous 2 y. Diverticular disease was not a specified end point in the 1988 follow-up questionnaire. After up to six mailings for each follow-up period, the response rate was 96% in 1990 and 94% in 1992. When diverticular disease was reported on the follow-up questionnaire, we mailed the cohort member a supplementary questionnaire to confirm the reporting and to ascertain the date of diagnosis, symptoms, procedure performed to confirm the diagnosis (e.g., barium study or endoscopy), treatment and any dietary changes before the diagnosis. We obtained 182 medical records from a sample of participants reporting diverticular disease to assess the validity of self-reporting and to ascertain the site of the diverticula. The records confirmed the self-reports in 95% of the cases. We therefore accepted the self-report of diverticular disease. In 96% of the cases, diverticula were located in the left colon (sigmoid, descending or mid-transverse colon), as expected in a predominantly Caucasian population (Parks 1969). We considered the cases ascertained during the four years of follow-up, i.e., between February 1, 1988 and January 31, 1992.

We excluded from this analysis men who did not report an average daily energy intake between 3.35 and 17.6 MJ (800 to 4200 kcal), or who left blank 70 or more food items on the dietary questionnaire (0.5% of the participants). We also excluded men who reported diverticular disease before 1988, cancer (other than nonmelanoma skin cancer), colon or rectal polyps and ulcerative colitis at base line and at subsequent follow-up. After these exclusions, the base-line population consisted of 43,881 men eligible for this analysis. In this population, we identified 362 newly diagnosed cases of symptomatic diverticular disease (Painter 1985a). Of these, 271 had abdominal pain or change in bowel habits, 50 cases had blood in stool and 41 cases had gastrointestinal bleeding as the major presenting symptom. We specifically used symptomatic diverticular disease as our primary end point, to reduce the possibility of detection bias.

Statistical analysis.  Participants were categorized according to quintiles of fiber intake (soluble, insoluble, cellulose, hemicellulose and lignin). Follow-up time was calculated as the number of months between February 1, 1988 and the date of diagnosing diverticular disease, death or to January 31, 1992, whichever came first. The relative risk (RR), the incidence among the men in different exposure categories divided by the corresponding rate in the reference category, was used as the measure of association (Rothman 1986). Age-adjusted rates were calculated with the use of five-year categories (Rothman 1986). The Mantel extension test (Kleinbaum et al. 1988) was used to evaluate linear trends across categories of different dietary fiber variables. In addition to the dietary variables mentioned, other potentially confounding variables were modeled with multiple logistic regression, which included age, physical activity and total fat (Aldoori et al. 1994, 1995a and 1995b). The P-values are all two tailed, and for all relative risks, we calculated the 95% confidence interval (CI) (Miettinen 1976).

	RESULTS

Abstract Introduction Methods Results Discussion References

During 160,825 person-years of follow-up over a period of 4 y, 362 cases of symptomatic diverticular disease were documented in this cohort. Age and energy-adjusted intake of soluble fiber and insoluble fiber were inversely associated with risk of diverticular disease (Table 1). This association was stronger for insoluble fiber. When we analyzed soluble and insoluble fiber separately in multivariate models including age, energy-adjusted total fat and physical activity, our findings for both sources did not change appreciably from those in the age and energy-adjusted model (Table 1). When we included soluble and insoluble fiber simultaneously in the multivariate model, the inverse association with insoluble fiber was slightly stronger, but that with soluble fiber was eliminated (RR for insoluble fiber = 0.53, 95% CI, 0.33-0.85, RR for soluble fiber = 1.34, 95% CI, 0.83-2.18, for the highest quintile).

We further investigated specific dietary fiber types that were available in our dietary database. Age-and energy-adjusted intake of cellulose, hemicellulose and lignin were inversely associated with risk of diverticular disease, and this association was strongest for cellulose and lignin (Table 2). When we analyzed these sources of dietary fiber separately in multivariate models that included age, energy-adjusted total fat and physical activity, our findings for each source of dietary fiber did not change appreciably from those in the age- and energy-adjusted model (Table 2). However, when we included hemicellulose, cellulose and lignin in the multivariate model, the association with cellulose did not change appreciably, but those with hemicellulose and lignin were attenuated (RR for hemicellulose = 1.26, 95% CI, 0.81-1.96, RR for cellulose = 0.57, 95% CI, 0.34-0.94, RR for lignin = 0.80, 95% CI, 0.47-1.38 for the highest quintile). Several foods contributed to estimating total dietary fiber and specific fiber components in this cohort; however, there were specific foods that were the leading contributors to the dietary fiber variables. In (Table 3), we have listed such food sources, their percentage contribution by serving size to total dietary fiber, soluble fiber, insoluble fiber, hemicellulose, cellulose and lignin in this cohort.

We further explored the association between dietary fiber and the risk of diverticular disease by specific presenting symptoms. We observed a decreased risk of diverticular disease among cases in which symptoms were mainly abdominal pain and change in bowel habits (RR = 0.63, 95% CI, 0.40-0.99). Similar inverse trends between dietary fiber and diverticular disease were observed among the patients whose presenting symptoms were mainly bleeding or a positive fecal occult blood (RR for bleeding = 0.61, 95% CI, 0.20-1.87; RR for positive fecal occult blood = 0.43, 95% CI, 0.17-1.10).

It has been suggested that the interpretation of epidemiologic studies of fiber-disease association is limited because different analytic techniques have been used to define dietary fiber. However, in our data, results using total fiber intake calculated from the AOAC method were not appreciably different from those using the Southgate or Englyst methods (for Southgate method, RR = 0.61, 95% CI, 0.42-0.88; for Englyst method, RR = 0.60, 95% CI, 0.42-0.87, for the highest quintiles) (Table 4).

	DISCUSSION

Abstract Introduction Methods Results Discussion References

These prospective data support the hypothesis that high intake of dietary fiber reduces the risk of diverticular disease and suggest that the inverse relation is particularly strong for the insoluble component of fiber and most notably for cellulose. To our knowledge, different analytic definitions of dietary fiber or the assessment of specific fiber types has not been considered in previous studies of diverticular disease.

Earlier, we reported that fiber from fruit and vegetable sources, and not from cereal sources, was associated with a reduced risk of diverticular disease (Aldoori et al. 1994). Dietary fiber is often a mixture of soluble and insoluble fiber (Bennett and Creda 1996). In our data, the insoluble component of fiber was significantly associated with a decreased risk of diverticular disease, and this association was particularly strong for cellulose. Insoluble fiber is the major dietary fiber fraction, and overall, fruits and vegetables tend to be higher in cellulose than cereals (Marlett 1992, Slavin 1987). This might partially explain the observed beneficial effect of fiber from fruits and vegetables in our study. Cellulose comprises approximately one third or less of the total fiber in most foods except for legumes, in which it is about half (Marlett 1992). In fruits and vegetables, cellulose represents an average of 30⁺ and 50⁺% of the insoluble fraction, respectively (Marlett 1992).

Soluble fiber is metabolized by colonic bacteria more than insoluble fiber, and hence has little effect on stool weight (Bennett and Creda 1996, Slavin 1987). However, it has been suggested that fermentable fiber can increase fecal output by stimulating microbial growth, with the production of short-chain fatty acids (SCFA) and other products (Stephen and Cummings 1980). Luminal SCFA are recognized now as an important fuel source for the colon (Rabassa and Rogers 1992), particularly in the distal colon, which is the most common site for diverticular disease. However, this phenomenon is not exclusive to soluble fiber, because some 60⁺% of ingested cellulose appears in human feces (Topping and Illman 1986). This indicates that insoluble fiber is fermented to some degree because SCFA are not abolished when the diet contains only insoluble fiber such as cellulose or wheat bran (Topping 1991). Also, in an animal study, it was observed that insoluble fiber may modify the production of SCFA (Cheng et al. 1987).

Agreement concerning an optimal method for analysis of fiber is an extremely contentious area. We recognize that for some types of fiber such as hemicellulose and lignin, there may be no current satisfactory data. However, in our analysis, we made the best attempt possible to use all available sources of fiber data. Intakes calculated with these data do provide probably the best biological test of the value of fiber; it is particularly useful to know the relative predictiveness of the various fiber fractions in relation to the risk of symptomatic diverticular disease. We acknowledge that it is possible that further improvements in analyses of fiber subfractions may provide different results in the future.

Fiber intake is correlated with consumption of several micronutrients that may also be associated with lower risk of diverticular disease. Indeed, several such nutrients were inversely related to the risk of diverticular disease. However, these inverse associations were no longer significant when each of these nutrients was adjusted for other risk factors (Aldoori et al. 1994). This attenuation was due primarily to the influence of dietary fiber and to a lesser extent to physical activity, whereas the inverse association between dietary fiber and diverticular disease was not appreciably altered. This emphasizes the importance of dietary fiber in the etiology of diverticular disease and suggests that this inverse association with diverticular disease is not explained by other nutrients in dietary fiber sources. Our findings for dietary fiber and diverticular disease were not materially different using analytic techniques to ascertain dietary fiber. This argues that our dietary questionnaire discriminated and ranked participants according to their dietary fiber intake reliably, regardless of the method of fiber analysis used. However, it is expected that such techniques are highly correlated because correlations reflects association between foods high and low in fiber. Indeed the correlations among the Southgate, Englyst and AOAC methods ranged from 0.88 to 0.94.

It has been argued that diverticular disease may be an asymptomatic condition and that the symptomatic presentation is due to existing irritable bowel syndrome (IBS)(Thompson et al. 1982). However, the prevailing understanding is that diverticula can cause symptoms and that because of the abnormal motility of the colonic muscle and the associated pain, diverticular disease may cause symptoms whether or not diverticula are inflamed (Painter 1985a). Moreover, it was shown that intraluminal pressure was significantly higher in diverticular disease patients than in controls, unlike patients with IBS who had lower pressure than controls (Trotman and Misiewicz 1988). Indeed, in our study, the beneficial effect of fiber was not confined to the cases who presented with abdominal pain or change in bowel habits, but was evident even among the cases who presented with bleeding or with fecal occult blood.

It could be argued that our findings are explained by detection bias, because diverticular disease is incidently diagnosed when individuals develop symptoms for other reasons. For example, people with low dietary fiber intake might have more abdominal symptoms, and hence more endoscopies, in which diverticulosis is diagnosed. However, the opposite was found; those participants in the HPFS who reported having endoscopies for abdominal symptoms were in fact eating more dietary fiber and less total fat and were more physically active. We therefore conclude that detection bias is unlikely to have explained our results (Aldoori et al. 1994, 1995a and 1995b).

Because we relied mainly on self-report rather than obtaining complete medical records for all of the positive respondents, we could not exclude right-sided diverticulosis from our cases. However, among the 108 cases for whom we obtained medical records, exclusive right-sided diverticulosis was present in <4%, which is expected in a U.S. population (Parks 1969). This is in contrast to the Orient, where diverticular disease is predominantly right sided (Chia et al. 1991, Lee 1986, Vajrabukka et al. 1980).

Biased recall of diet is unlikely, because the dietary data were collected before the diagnosis of diverticular disease. As in almost every study, some degree of misclassification is inevitable. However, the prospective design of this study means that any misclassification would be random with regard to case status and hence would tend to attenuate any association. Our findings are most directly generalizable to nonvegetarian U.S. men 40 y of age. Apart from reports of a higher preponderance of diverticular disease in women (Roberts and Veidenheimer 1990), we have no reason to believe that the association we observed in men would be different for women. However, future studies should explore any possible differences in the relations between total dietary fiber or its specific types and the occurrence of diverticular disease due to gender.

Our findings provide evidence for the hypothesis that a diet high in dietary fiber decreases the risk of diverticular disease, and this result was not sensitive to the use of different analytic techniques to define dietary fiber. Our findings suggest that the insoluble component of fiber was significantly associated with a decreased risk of diverticular disease, and this inverse association was particularly strong for cellulose.

	ACKNOWLEDGMENTS

We thank the participants of the Health Professionals Follow-up Study. We are also indebted to Graham Colditz, Gary Curhan, Meir Stampfer, Alberto Ascherio, and to Al Wing, Jill Arnold, Elizabeth Frost-Hawes, Mildred Wolff, Cindy Dyer, Jan Vomacka, Kerry Pillsworth-Demmers and Mira Koyfman for their expert help.

	FOOTNOTES

Manuscript received 27 May 1997. Initial reviews completed 2 July 1997. Revision accepted 19 December 1997.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Ainsworth B. E., Haskell W. L., Leon A. S., Jacobs D. R., Montoye H. J., Sallis J. F., Paffenbarger R. S. Compendium of physical activities: classification of energy costs of human physical activities. Med. Sci. Sports Exer. 1993; 25:71-80 [Medline]
• Aldoori W. H., Giovannucci E. L., Rimm E. B., Wing A. L., Trichopoulos D. V., Willett W. C. A prospective study of diet and the risk of symptomatic diverticular disease in men. Am. J. Clin. Nutr. 1994; 60:757-764 [Abstract/Free Full Text]
• Aldoori W. H., Giovannucci E. L., Rimm E. B., Ascherio A., Stampfer M. J., Colditz G. A., Wing A. L., Trichopoulos D., Willett W. C. A prospective study of physical activity and the risk of symptomatic diverticular disease in men. Gut 1995a; 36:276-282 [Abstract/Free Full Text]
• Aldoori W. H., Giovannucci E. L., Rimm E. B., Wing A. L., Trichopoulos D. V., Willett W. C. A prospective study of alcohol, smoking, caffeine, and the risk of symptomatic diverticular disease in men. Ann. Epidemiol. 1995b; 5:221-228 [Medline]
• Bennett W. G., Creda J. J. Benefits of dietary fiber. Myth or medicine? Postgrad. Med. 1996; 99:153-175
• Brodribb A.J.M., Humphreys D. M. Diverticular disease: three studies. I. Relation to other disorders and fiber intake. Br. Med. J. 1976; 1:424-425
• Chasan-Taber S., Rimm E. B., Stampfer M. J., Spiegelman D., Colditz G. A., Giovannucci E. L., Ascherio A., Willett W. C. Reproducibility and validity of a self-administered physical activity questionnaire for male health professionals. Epidemiology 1996; 7:81-86 [Medline]
• Cheng B. Q., Trimble R. P., Illman R. J., Stone B. A., Topping D. L. Comparative effects of dietary wheat bran and its morphological components (aleurone and pericarp-seed coat) on volatile fatty acid concentrations in rat. Br. J. Nutr. 1987; 57:69-76 ^[Medline]
• Chia J. G., Chiantana C. W., Ngoi S. S., Goh P. M., Ong C. L. Trends of diverticular disease of the large bowel in a newly developed country. Dis. Colon Rectum 1991; 34:498-501 [Medline]
• Feskanich D., Rimm E. B., Giovannucci E. L., Colditz G. A., Stampfer M. J., Litin L. B., Willett W. C. Reproducibility and validity of food intake measurements from a semiquantitative food frequency questionnaire. J. Am. Diet. Assoc. 1993; 93:790-796 [Medline]
• Findlay J. M., Smith A. N., Mitchell W. D., Anderson A.J.B., Eastwood M. A. Effects of unprocessed bran on colon function in normal subjects and in diverticular disease. Lancet 1974; 1:146-149 [Medline]
• Fisher N., Berry C. S., Fearn T., Gregory J. A., Hardy J. Cereal dietary fiber consumption and diverticular disease: a lifespan study in rats. Am. J. Clin. Nutr. 1985; 42:788-804 [Abstract/Free Full Text]
• Gear J.S.S., Ware A., Fursdon P., Mann J. I., Nolan D. J., Brodribb A.J.M. Symptomless diverticular disease and intake of dietary fiber. Lancet 1979; 1:511-514 ^[Medline]
• Hodgson W.J.B. An interim report on the production of colonic diverticula in the rabbit. Gut 1972; 13:802-804 ^{[Abstract/Free Full Text]}
• Holland, G., Welch, A. A., Unwin, D. D., Buss, D. H., Paul, A. A. & Southgate, D.A.T. (1991) In: McCance and Widdowson's The Composition of Foods, 5th ed. The Royal Society of Chemistry and Ministry of Agriculture, Fisheries, and Food, Cambridge, UK.
• Horvath, P. J. & Robertson, J. B. (1986) Detergent analysis of foods. In: CRC Handbook of Dietary Fiber in Human Nutrition (Spiller, G. A., ed.). CRC Press, Boca Raton, FL.
• Jaskiewicz K., Rossouw J. E., Kritchevsky D., Van Resburg S. J., Fincham J. E., Woodroof C. W. The influence of diet and diemethyhydrazine on the small and large intestine of Vervet monkeys. Br. J. Exp. Pathol. 1986; 67:361-369 [Medline]
• Kleinbaum, D. G., Kupper, L. L. & Muller, K.E. (1988) Applied Regression Analysis and Other Multivariable Methods. PWS-KENT Publishing, Boston, MA.
• Leaky A. L., Ellis R. M., Quill D. S., Peel A.L.G. High fiber diet in symptomatic diverticular disease of the colon. Ann. R. Coll. Surg. Engl. 1985; 67:173-174 ^[Medline]
• Lee Y. S. Diverticular disease of the large bowel in Singapore: an autopsy study. Dis. Colon Rectum 1986; 29:330-335 [Medline]
• Manousos O., Day N. E., Tzonou A., Papadimitriou C., Kapetanakis A., Polychronopoulou-Trichopoulou A., Trichopoulos D. Diet and other factors in the aetiology of diverticulosis: an epidemiological study in Greece. Gut 1985; 26:544-549 [Abstract/Free Full Text]
• Marlett J. A. Content and composition of dietary fiber in 117 frequently consumed foods. J. Am. Diet. Assoc. 1992; 92:175-186 [Medline]
• Miettinen O. Estimability and estimation in case-referent studies. Am. J. Epidemiol. 1976; 103:226-235 [Abstract/Free Full Text]
• Miettinen T. A., Tarpila S. Fecal beta-sitosterol in patients with diverticular disease of the colon and in vegetarians. Scand. J. Gastroenterol. 1978; 13:573-576 [Medline]
• Painter, N. S. (1985a) The cause of diverticular disease of the colon; its symptoms and its complications. J. R. Coll. Edin. 30 : 118-122.
• Painter, N. S. (1985b) Diverticular disease of the colon. In: Dietary Fibre, Fibre-Depleted Foods and Diseases (Trowell, H., Burkitt, D. & Heaton, K.,eds.), pp. 145-160. Academic Press, London, UK.
• Painter N. S., Burkitt D. P. Diverticular disease of the colon: a deficiency disease of western civilization. Br. Med. J. 1971; 2:450-454
• Painter N. S., Burkitt D. P. Diverticular disease of the colon: a 20th century problem. Clin. Gastroenterol. 1975; 4:3-21 [Medline]
• Parks T. G. Natural history of diverticular disease of the colon; a review of 521 cases. Br. Med. J. 1969; 4:639-645
• Paul, A. A. & Southgate, D. A. (1978) McCance and Widdowson's The Composition of Foods, 4th rev. ed. Her Majesty's Stationery Office, London, UK.
• Prosky L., Asp N., Furda I., DeVries J. W., Schweizer T. F., Harland B. F. Determination of total dietary fiber in foods, food products and total diets: interlaboratory study. J. Assoc. Off. Anal. Chem. 1984; 67:1044-1052
• Prosky L., Asp N., Furda I., DeVries J. W., Schweizer T. F., Harland B. F. Determination of total dietary and food products: collaborative study. J. Assoc. Off. Anal. Chem. 1985; 68:677-679 [Medline]
• Prosky L., Asp. N., Schweizer T., DeVries J., Furda I. Determination of insoluble, soluble and total dietary fiber in foods and food products: interlaboratory study. J. Assoc. Off. Anal. Chem. 1988; 71:1017-1021 [Medline]
• Rabassa A. A., Rogers A. I. The role of short chain fatty acid metabolism in colonic disorders. Am. J. Gastroenterol. 1992; 87:419-423 [Medline]
• Rimm E. B., Giovannucci E. L., Stampfer M. J., Colditz G. A., Litin L. B., Willett W. C. Reproducibility and validity of an expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am. J. Epidemiol. 1992; 135:1114-1126 [Abstract/Free Full Text]
• Roberts, P. L. & Veidenheimer, M. C. (1990) Diverticular disease of the colon. In: Current Therapy in Gastroenterology and Liver Diseases, Vol. 3 (Bayless, T. M., ed.), pp. 416-419. B. C. Decker, Toronto, Canada.
• Rothman, K. J. (1986) Modern Epidemiology. Little, Brown and Company, Boston, MA.
• Schwartz, J. T. & Graham, D. Y. (1988) Diverticular disease of the large intestine. In: Diseases of the Colon, Rectum, and Anal Canal (Kirsner, J. & Shorter, R., eds.), pp. 519-536. Williams and Wilkins, Baltimore, MD.
• Slavin J. L. Dietary fiber: classification, chemical analyses, and food sources. J. Am. Diet. Assoc. 1987; 87:1164-1171 [Medline]
• Southgate D. A., Bailey B., Collinson E., Walker A. F. A guide to calculating intakes of dietary fiber. J. Hum. Nutr. 1976; 30:303-313 [Medline]
• Stephen A. M., Cummings J. H. Mechanism of action of dietary fiber in the human colon. Nature (Lond.) 1980; 284:283-284 [Medline]
• Thompson W. G., Patel D. G. Clinical picture of diverticular disease of the colon. Clin. Gastroenterol. 1986; 15:903-916 [Medline]
• Thompson W. G., Patel D. G., Tao H., Nair R. C. Does uncomplicated diverticular disease produce symptoms? Dig. Dis. Sci. 1982; 27:605-608 [Medline]
• Topping D. L. Soluble fiber polysaccharides: effects on plasma cholesterol and colonic fermentation. Nutr. Rev. 1991; 49:195-203 [Medline]
• Topping D. L., Illman R. J. Bacterial fermentation in the human large bowel: time to change from the roughage model of dietary fibre? Med. J. Aust. 1986; 144:307-309 [Medline]
• Trotman I. F., Misiewicz J. J. Sigmoid motility in diverticular disease and the irritable bowel syndrome. Gut 1988; 29:218-222 [Abstract/Free Full Text]
• U.S. Department of Agriculture (1993) Nutrient Database for Standard Reference, release 10. U.S. Government Printing Office,Washington, DC.
• Vajrabukka T., Saksornchai K., Jimakorn P. Diverticular disease of the colon in a far-eastern community. Dis. Colon Rectum 1980; 23:151-154 [Medline]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences