5

reduction in incidence of morbid events will be considerable among .the hypertensives, but the effect might be small when the whole population is considered. If the cut-off points are low, however, the effect might be less pronounced in the hypertensive part of the population, but the impact on the whole population might be substantial. The effect of treatment in our study might be related to the type of treatment. It has been shown that betablockade can reduce the incidence of sudden death in patients with a myocardial infarction,21,22 and it is probable that similar effects might be seen in hypertensive patients. The number of end-points in subgroups on different therapy in the present trial prevented a proper analysis of mortality in patients on different types of antihypertensive therapy. A Scandinavian multicentre study which aims to do that is in progress. The lower mortality from cardiovascular causes in the treatment group was not counterbalanced by an increased mortality from other causes. In fact, there was a tendency to lower mortality from non-cardiovascular causes. If this tendency is also found in a larger series it might imply that it was due to some overall effect of the hypotensive treatment or of the general care of the hypertension clinic. Another reason might be some type of negative selection of those not being treated, but we have not found any indication for that type of bias. The lack of a placebo-treated control group makes interpretation of the findings hazardous. Although we have tried to take into account as many influencing factors as possible, this cannot entirely make up for the fact that the trial was not primarily designed to answer the question at which we have directed our analysis; yet these results suggest that antihypertensive treatment might have a substantial impact on the incidence of cor-

onary heart-diseases. This study was supported by grants from the Bank of Sweden, Tercentenary Foundation, the Swedish Medical Research Council, and the Swedish National Association against Heart and Chest Diseases.

Request for reprints should be addressed to L. W., Department of Medicine I, Sahlgren’s Hospital, S-413 45 Goteborg, Sweden. REFERENCES 1.

Hood, B, Björk, S., Sannerstedt, R., Angervall, G. Acta med. scand. 1963, 174, 393. 2. Hodge, J. V, Smirk, F. H. Am. Heart J. 1967, 73, 441. 3. Veterans Administration Cooperative Study Group, J. Am. med. Ass. 1970,

213, 1143. 4. Veterans Administration Cooperative Study Group, Circulation, 1972, 45, 991. 5. Br. med J. 1972, ii, 1. 6. Svärdsudd, K., Berglund, G., Tibblin, G. Drugs, 1976, 11, (suppl. 1) 34. 7. Wilhelmsen, L., Tibblin, G., Werkö, L. Prev. Med. 1972, 1, 153. 8. Wilhelmsen, L., Berglund, G., Werkö, L. ibid. 1973, 2, 57. 9. Wilhelmsen, L., Wedel, H., Tibblin, G. Circulation, 1973, 48, 950. 10. Jonsell, S. Acta radiol. 1969, 20, 325. 11. Rose, G., Blackburn, H. Monogr. Ser. W.H.O. 1968, 56. 12. Astrand, I., Blomqvist, G., Areskog, N. H., Bjerkelund, K., Carlsten, A., Furberg, C., Grevin, K-E., Hansen, F., Kaijser, L., Kallio, V., Malmström, G., Nordgren, L., Punsar, S., Pyörälä, K., Thulesius, O. Acta med. scand. 1967, suppl. 481, 14. 13. Berglund, G., Anderson, O., Wilhelmsen, L. Br. med. J. 1976, 2, 554. 14 The National Health Survey. U.S. Public Health Service Publications. 1966, no 1000, Ser. 11-no 13. Elmfeldt, D., Wilhelmsen, L., Tibblin, G., Vedin, J. A., Wilhelmsson, C., Bengtsson, C. J. chron. Dis. 1975, 28, 173. 16. Harmsen, P., Tibblin, G. Acta med. scand. 1972, 191, 463.

15

17. World Health Organisation. EURO 5010 (1). Copenhagen 1968. 18. World Health Organisation. CVD 71:1. Monaco 1970. 19 Armitage, P. Statistical Methods in Medical Research; p. 135. Oxford, 1974. 20. Wilhelmsen, L., Ljungberg, S., Wedel, H., Werkö, L. J. chron. Dis. 1976, 29, 331. 21. Wilhelmsson, C., Vedin, J. A., Wilhelmsen, L., Tibblin, G. Lancet, 1974, ii, 1157. 22. A Multicentre International Study Group. Br. med. J. 1975, iii, 735.

COLONIC RESPONSE TO DIETARY FIBRE FROM

CARROT, CABBAGE, APPLE, BRAN, AND GUAR GUM D. A. T. SOUTHGATE HELEN HOUSTON W. P. T. JAMES

H. CUMMINGS W. BRANCH D. J. A. JENKINS

J.

Medical Research Council Dunn Nutrition Unit, Cambridge, and Medical Research Council Gastroenterology Unit, Central Middlesex Hospital, London

20 g/day of concentrated from carrot, cabbage, and was added to the controlled bran, apple, guar gum basal diet of nineteen healthy volunteers. Fæcal weight increased by 127% on bran, 69% on cabbage, 59% on carrot, 40% on apple, and 20% on guar gum. These changes in fæcal weight were correlated with an increased intake of pentose-containing polysaccharides from the fibre. On the basal diet there were pronounced individual differences in fæcal weight, and from these the response of subjects to the fibre preparations could be predicted. Addition of fibre shortened mean transittime through the gut and significantly diluted an inert marker in the fæces. Diet-induced changes in colonic function may explain international differences in the prevalence of colonic disease, whilst personal variation in the response to dietary fibre may determine individual susceptibility to large-bowel disease within a com-

Summary

Approximately dietary fibre

munity. Introduction DIETARY fibre in the form of bran, wholemeal bread, other cereal-based food increases fxcal weight, shortens transit-time, and alters colonic function. These findings have been linked with epidemiological observations which show that in communities with relatively high intakes of dietary fibre the incidences of most largebowel disorders are low.1,2 In Britain fruit and vegetables, not wheat, constitute the main source of dietary fibre. Even in countries where fibre intakes are high, wheat is rarely the main source of fibre. The effects of fibre from commonly eaten vegetables or fruit have been little investigated, although attempts were made to do this in the 1930s and 40s.3,4 This omission is not surprising since wheat bran is a readily obtainable and concentrated source of fibre, whilst the extraction of fibre from non-cereal foods requires large amounts of starting material, processing on an industrial scale, and the collaboration of people with widely differing expertise. We fed concentrated fibre preparations from carrot, cabbage, bran, apple, and guar gum to healthy volunteers to compare their effects on colonic function. or

Methods Nineteen healthy male volunteers aged 20-38 who were either students or members of the hospital staff took part in the study. Each subject took a controlled basal diet for 3 weeks and a capsule containing 10 radio-opaque pellets with each meal (three times daily). The pellets were given as a marker in the metabolic studies and to allow mean transit-time (M.T.T.) through the gut to be measured as described elsewhere.5 Fseces were collected throughout the 3-week period and for 1 week afterwards, each stool being collected separately into a plastic bag suspended over a toilet. The sample was labelled, sealed, and then stored at -20 °C.

6 The subjects were asked to participate in two 3-week dietary trials during one of which additional fibre was added to the basal diet. At least 3 weeks elapsed between the two dietary periods to ensure that one dietary period had no residual effect on the next. Six men took fibre from more than one source. The men lived either in a hostel in the grounds of the Central Middlesex Hospital or in accommodation nearby. All subjects were encouraged to lead as normal a life as possible during the study, and no medication of any sort was taken either before or during the diets. The control diet included ’Weetabix’, milk, orange juice, sugar, jam, biscuits, meat, vegetables, fruit, tea, coffee, white bread, and butter. Its daily composition (calculated from food tables6) was 11.6 MJ (2784 kcal), 108 g fat, 385 g carbohydrate, 85 g protein, and 22 g dietary fibre. Three 1-day menus were designed with this composition and fed in rotation. Only deionised water was allowed. All food was prepared in the metabolic kitchen. The carrot, cabbage, bran, and apple were prepared as follows : 2500 kg of diced carrot and 6000 kg of shredded cabbage were obtained and air dried on fluidised bed driers for between 15 and 20 h at 55°C by Batchelors Foods Ltd in Worksop. This approximately 250 kg carrot and 430 kg cabbage. The material was transferred to Proprietory Perfumes Ltd in Kent where 25 kg batches were extracted for 3 h under reflux with 500 1 of 85% v/v aqueous methanol, the resulting material was filtered, and this process was repeated. Next the material was extracted with 4261 acetone for 1 h, filtered, then stirred with 100 1 cold acetone and filtered. This last cycle was repeated five times for carrot and three times for cabbage. The resultant material was air dried. Commercially dried apple pulp was bought and 33 kg lots extracted in a similar way at Unilever Research Laboratory, Port Sunlight, Wirral. Bran with a low endosperm content was obtained from Rank Hovis McDougall Ltd. Because of its higher dietary fibre content it was possible to extract this material with modified laboratory equipment in Cambridge. The extraction procedure was designed to remove free sugars, lipid, pigments, &c. and was based on the initial steps of the Southgate method’ for measuring dietary fibre. Acetone was used instead of ether because of the danger of ether in industrial plant. The apple, carrot, and cabbage preparations were ground in a kitchen grinder and then mechanically sieved for 30 min. Material retained between the 0.5 and 1-4 mm sieves was used for the feeding studies. Guar gum was obtained as clear gum from Hercules Powder Co., Erith, Kent. It was a fine powder which passed through an 0-5mm sieve and was fed in the form in which it was obtained. The particle size for the food preparations was controlled because it is thought to affect the response of the large bowel to fibre.8 The composition of the resulting materials is given in table i together with the amounts fed daily. Dietary fibre was measured by the method of Southgate,’ free sugars and starch by the method of Hudson et al.,9 and nitrogen by an autoanalyser technique. The subjects were given three equal amounts of the dry material each day and incorporated it in their food or drink. The data presented here were obtained in the 3rd week of the dietary periods. Faecal weight has been corrected for marker output. Student’s t test was used for statistical comparisons and all mean values are given±11 S.D. Because the groups of subjects received slightly different amounts of dietary fibre from each preparation (17-2—21-9g/day) fsecal weights have been corrected to a 20 g fibre intake to compare the relative effects of the different fibres (fig. 3 only). This correction assumes that response to fibre is directly proportional to intake over this narrow range.

produced

TABLE I-COMPOSITION PER

100 g

i

*Expressed

as

monosaccharides; thus

OF MATERIALS FED

i

sum

I

of all components exceeds

100%.

tN x 6 25 (orx 5.7for bran). $Starch values expressed as glucose. also to obtain,

unaltered, the whole of the fibre trom the

starting material whilst removing as much as possible of other constituents. Carrot, cabbage, and apple normally contain 1-4% dietary fibre’O. The drying and extraction process concentrated fibre 20-40 times (fibre 61-85%) (table i). The resultant materials were colourless, almost tasteless, and the fibre had a similar chemical composition to that of untreated carrot, cabbage, and apple fibre.1O This suggested that processing had not altered the fibre to any great extent. Processing bran proved much simpler since it normally contains 27-44% fibre. In bran the proportion of fibre was increased to 60% by processing. Groups of

six subjects took apple, carrot, cabbage, and bran whilst three took guar gum. The preparations were introduced gradually into the diet with 10 g the first day, 20 g the second day, and so on until the specified dose for each material was reached (table I), Reactions to its inclusion in the diet varied. The most commonly noted effect was an increase in flatus which, with apple, carrot, and cabbage, was associated with feeling fuller after meals. Bran was the only material which caused the subject to comment on a change in bowel

Results These materials were prepared to produce sufficiently concentrated fibre from each source to allow consumption of a reasonable amount by the men. The aim was

Fig. 1-Mean daily faEcat weight during third week of each dietary period in all men investigated. Mean values for groups are given below the figure. Open circles represent data during basal diet and closed circles are data with fibre.

7 TABLE II-MEAN TRANSIT-TIME IN HOURS

*Number of

obtained. t Based on

subjects in whom technically satisfactory data

were

loge data.

habit. Some found the change acceptable but others experienced urgency when wanting to defaecate. Each fibre preparation increased fsecal output in every subject except in one person taking guar gum (fig. 1). Increases in faecal weight were all significant (P