dietary fiber intake was 19.9 ± 5.3 g/day compared with the calculated value of 19.7 g/day from the 1976 National. Food Survey. There was a 4-fold range in fiber ...
nutrition
Intakes and sources the British Sheila
Bingham,
of dietary
random
Intakes
sample
of the
fiber in
2
B. Sc., J. H. Cummings,
ABSTRACT
survey
M. R. C.P.,
of dietary
population
fiber
and
and N. I. McNeil,
its different
in Cambridgeshire,
M.R.
components
England
and
have compared
C.P.
been with
measured data
in a
from
the
British National Food Survey. Sixty-three men and women ages 20 to 80 were included in the sample. Total dietary fiber intake was 19.9 ± 5.3 g/day compared with the calculated value of 19.7 g/day from the 1976 National Food Survey. There was a 4-fold range in fiber intake from 8 to 32 g/day; no significant trends with age or between men and women were detected. Vegetables supplied the majority of the fiber (4 1.3%); cereals 30.5%, and fruit and mixed sources 28.2%. Of the components of dietary fiber noncellulosic polysaccharide, cellulose, and lignin intakes were 13.8, 4.7, and 1.4 g/day, respectively. In the noncellulosic fraction, hexoses contributed 7.4 g, pentoses and uronic acids 3.3 and 3.0 g. Vegetables and unrefined cereals were the main sources of pentose. These intakes are low in comparison with limited international data from developing countries and of a similar order to those known in dietary experiments to produce low stool weights, slow transit time, and concentrated feces. They could readily be increased by simple dietary changes. Am. J. Clin. Nuir. 32: 1313-1319, 1979.
The hypothesis that diet, in particular a lack of fiber, is important in heart disease, diabetes, obesity, and many large bowel disorders has been widely discussed and gained a measure of acceptance. This hypothesis is based on epidemiological observations in which
disease
patterns
in
different
the world are compared with (1, 2). One such observation bowel disorders like diverticular cancer in rural Africa when that of the West (3, 4). Dietary are said to be much higher in This personal
dietary information observation by
parts
of
dietary intakes is the rarity of disease and compared with fiber intakes Africa.
is based either the investigators
on or
derived from measurements of the crude fiber content (5, 6) of the diet. Crude fiber is a totally inadequate estimate of fiber (7, 8) and in fact there is currently only one publication that contains actual dietary fiber intake measurements in any population in the world (9). The inadequacy of dietary fiber intake data is because only recently has its measurement in a wide range of foodstuffs become possible The American
Journal
of Clinical
Nutrition
32: JUNE
(10). It is now apparent that dietary fiber is a mixture of substances and current studies show that both the level of intake and the type of fiber are important in determining physiological events ( 1 1). We have, therefore, measured dietary fiber in a randomly selected sample ofa population in Cambridgeshire, England and using recently published data (10) calculated consumption of the various fiber constituents and their dietary sources. Methods
and subjects
A village (population 4500) 6 miles north of Cambridge was chosen as the site of the study. This village was chosen because it was believed to be a manageable size, contained a broad selection of social classes, was compact
developed
and
clearly
defined
local organizations,
geographically,
had
was largely
the list of one
‘From the Medical Research Council, Dunn tion Unit, Milton Road, Cambridge, England. 2 Address reprint requests to: Sheila Bingham, Clinical Nutrition Trumpington Street,
1979, pp. 13 13-1319.
Printed
Unit, Addenbrooke’s Cambridge, England.
in U.S.A.
well
NutriDunn Hospital,
1313
1314
BINGHAM
family doctor and was suitably close to the Dunn Nutrition Unit. Originally a market garden center, it had been developed since 1960 by light industry and new housing estates. The majority of the population found local employment or in offices and professions in Cambridge.
The or over)
electoral of 2886
for random local village
register voters
(which lists all residents age 18 was used as the sampling frame of 103 people. After contacting
selection officials a publicity
campaign
was mounted
the village. Each household was sent a letter explaining the study and posters were put up round the village. The in
selected subjects were divided into four groups geographically and individuals were then recruited. In turn over a period of 4 months each person was sent a personal
letter
inviting
him
to join
in,
followed
by a visit
to the
house
by one of the investigators. The subjects were asked to weigh and record all individual items offood and drink and to weigh anything left over, for one week. Those who agreed to take part were visited by S. B. the day before they were due to start weighing their food when an explanation and demonstration of the technique was given. Food was served out onto dishes or plates in the normal way, subjects recording the weight of the plate and cumulative weights of the food or drink added. Scales oflarge capacity were necessary to accommodate both the plate and the food. Scales (Chatillon 0265) were obtained and the charts modified by CMS Weighing Limited of London, England, to weigh 4 pounds by #{188} ounces. After the subjects had repeated the demonstration successfully, the scales, notebook, written instructions, and examples were left in the home. The subjects were visited on the 1st day of weighing to see that the technique had been understood. Further visits in the week were made if necessary. The day after completion of the record, the subjects were revisited and the records thoroughly checked to ensure that all the necessary information had been given. In order not to interfere with normal eating habits food bought away from home was not weighed but subjects were asked to describe the food eaten, its price, and place of purchase. Similar samples of food were then bought and weighed. An average of four subjects per week was studied
from May to August 1977. how well the subjects who took part in the represented the population as a whole, informa-
To see survey
tion on age and occupation was obtained. Using information published by the Office of Population, Censuses and Surveys (12), occupations were then arranged into six broad social classes; I, professional; II, intermediate; III, skilled occupations, subdivided into manual (hIM) and nonmanual (IIIN); IV, partly skilled; and V. unskilled occupations. Housewives, and subjects who were retired but did not state their former occupations, were not classified. The numbers in each class were compared with those that would have been obtained if the social class proportions were the same as those found by the latest census in 1971 (13). Subjects were also classified by age and sex, and the numbers in each 10 year age group compared with those that would have been obtained if the proportions were the same as those published in population statistics for 1974 (14). The hypothetical populations based on these published figures of social class, sex, and age for East Anglia were called the expected populations.
Total
daily energy
and nutrient
intakes
were caicu-
ET AL. lated by computer, using food tables compiled for use in individual weighed surveys by The Department of Health and Social Security (DHSS). In these tables, average nutrient contents of cooked items of food and of dishes containing more than one ingredient for instance pies, stews, puddings, etc. were calculated from food tables (15, 16) and from standard recipe books. Cooking losses were obtained mainly by experiment. These tables have been modified (S. Bingham and C. Ravenscroft, unpublished data) to include dietary fiber and its components,
measured
by
the
method
of
Southgate
et al.
(10). The subjects own recipes were used where these differed from those in the DHSS tables. Fiber intakes from the individual survey have been compared with those calculated from the National Food Survey of 1976. In this survey of6000 to 7500 households carried out every year under the auspices of the Ministry of Agriculture, Fisheries and Food, food is not weighed. The housewife is asked to keep a record of her purchases for I week and from this the quantity of food eaten by the average person per week is calculated. The published figures (17) of food consumption were used and a computer file for dietary fiber and its components accumulated by S. B. Average factors for inedible waste, and average weightings for foods within groups for 1976, as used in the National Food Survey for the file were supplied by the Ministry of Agriculture, Fisheries and Food as part of a collaborative project (18). Provisional energy intakes (19) for April to September 1977 recorded by the National Food Survey are compared with the mean energy intake from the individual survey. The figures in the paper are given ± I SD unless otherwise indicated.
Results Ofthe 103 people chosen from the electoral role 1 1 had either died or moved away. Thirty-two (82%) of the 39 men and 3 1 (58%) of the 53 women agreed to take part in the study. Of the 29 who declined to participate, two were elderly infirm women. Thus overall cooperation rates were 70% for those physically able to do so. Since disease incidence statistics
are
drawn
from
the
whole
popula-
tion no exclusions on medical grounds were made and three subjects whose diets were modified for therapeutic purposes (diabetes and gall stones) were included. All subjects who agreed to take part found the technique of weighing easy to understand, and no subject was unable to complete the survey. Most of the men were helped by the person responsible for food preparation, for instance the wife or mother. In Table 1 the distribution by age and sex of the 63 subjects who cooperated is shown. There was no significant difference between the subjects who agreed to cooperate and the
DIETARY TABLE
Age
Sex
20-29
M F M F M F M F M F M F M F
30-39 40-49 50-59 60-69 70-79 80+
not
1315
POPULATION
class of subjects
Observed
Total based
IN BRITISH
I
Age, sex, and social who participated
a
FIBER
Expected
numbers
based
8 4 5 6
7 6 6 5
7
5
5 5 9 3
4 4 3 0 0 63 on proportions
on classification of occupations economically active.
Socia
Expected
EXPCCCJ
I
M F
4 0
1 0
II
M
7
6
F
6
4
IIIN
M
5
3
5 5 5
IIIM
F M F
6 8 1
8 11 2
5
IV
M
4
7
F M F M F
6 2 1 2 11 63
7 2 2 2 8 63
5 3 3 1 2 63 in total
( 12); details
V V Total Anglian
East
in “Methods
population
and subjects”.
expected East Anglian population (X212 = 9.5 1, P > 0.5). The social class distribution of the sample is also shown in Table 1. As in the expected population, the majority were in social class III, which included clerical workcr5, bus and lorry drivers, shop assistants, and cooks, with a minority in social classes I (e.g., architects, chartered accountants) and V (e.g., laborers, cleaners). Compared with the cxpected population, there were rather more in class II, which included teachers, managers, and farmers, and less in IV, which included factory workers and ward orderlies. Overall, however, there was no significant difference between the observed and expected populations (X212 16.37, P > 0.20). Mean dietary intakes for the group were carbohydrate 256.1 ± 8 1 .6 g/day, fat 97. 1 ± 28.0 g/day; protein 72.3 ± 18.7 g/day; energy 9. 1 ± 2.4 MJ (2 180 kcal)/day and dietary fiber 19.9 ± 5.3 g/day. Of the total fiber intake, vegetables contributed 41.3%, cereals 30.5%, fruit 1 1.7%, and nuts 0.6%. The remainder was from miscellaneous mixed
TABLE
sources
g (21.7%),
such
as pies.
Intakes of the components of dietary fiber are shown in Table 2 in conjunction with average national intakes calculated from food consumption data from the National Food Survey for 1976, the latest year for which food consumption figures are published. Energy intakes for 1977 from the two surveys
(13, ‘
14).
b
Social
Unclassified,
class
includes
housewives
2
Mean daily intakes of energy dietary fiber and its components
and of
National Food Survey (1976)
Present study 977
Total Total
dietary fiber (g) noncellulosic polysaccharides (g) Hexoses(g) Pentoses (g) Uronic acids (g) Cellulose (g)
19.9 13.8
± ±
7.4±2.1 3.3 ± 1.3 3.0 ± 1.0 4.7 ± 1.2
(g)
1.4 ± 0.9
Energy
(MJ)
9.1
Provisional
results
are also shown polysaccharides intake, cellulose
19.7 13.9
5.3 3.7
Lignin
a
gradings
from
April
±
8.2 2.5 3.1
4.8 1.0 935
2.4
to September
1977.
in Table 2. The noncellulosic supplied 69.3% of the total 23.6%, and substances ana7.3%. The major part 7.4 g
lysing (53.6%) fraction
as lignin of the noncellulosic polysaccharides measured as hexose, pentose,
uronic
acid
contributing respectively.
3.3 (23.9%) The
pentose
and
and
3.0
contain-
ing polysaccharides in the noncellulosic polysaccharide fraction came mainly from vegetables (1.2 g) and whole wheat cereal (0.7 g) with
white
bread
and
other
low
extraction
flour products contributing 0.2 g. Dietary fiber intakes by sex and age are shown in Figure 1. There was no difference
BINGHAM
1316
ET AL.
24 22
:::::
20
:::::
. ...
:::::
18
.
)
‘a 16 .
14
.
A .
b
b’
12
b’
bb’
10
%
a, ..
b’
+++
+++ +++
+++
+++
+++
+++
+++
+++
+++
+++
+++ +++
+++ +++
+++ +++
8
+++ +++ +++ +++
6
+++
::
+++
t
+++
+++
+++
4 2
Group
20-29
Energy
9.6±2,0
Total.
30-39 jo#{149}cy±1.8
40-49
50-59
60-69
70-79
Men
9.93.2
8.31.6
7.52.6
8.8t3.l
10.0±2.4
Uronic
Acid
S.E.
Non-cellulos
El FIG.
total
dietary
1. Mean
fiber
daily
intakes
consumption
of dietary
between
the two sexes (19.8 ± 5.3 g/day women and 20. 1 ± 5.4 g/day men). Pentose intakes however were significantly higher for women, (3.7 ± 1 .5 g/day) than men (3.0 ± 1 1 g/day) (t = 2.09, P < 0.02). These higher intakes of pentose were derived from unrefmed wheat cereal, for example wholemeal bread, Weetabix, or All-Bran. Total dietary fiber intake from these sources was higher in the women (3. 1 ± 3.5 g) than in the men ( 1 .3 ± 1.9 g), (t = 2.58, P < 0.01). Distribution of pentose intakes in women was not normal. Of the 12 women at the top end of the distribution, i.e., those taking 4 to 7 g daily, all but one regu.
ic
polysaccharide
Hexose
in
8.221
El Lignin EII:J Cellulose
dietary
fibre
Women
fiber
and its components
by sex and age.
larly took some form of unrefined wheat cereal. In men this effect of high extraction cereal consumption was not seen. Energy intakes declined gradually with age (Fig.
1) but
were
higher
than
expected
in the
70 to 79 group. This fmding was largely due to high fat intakes in three subjects in the group of seven. Total dietary fiber intakes were highest in the 40 to 49 age group and lowest in the 70 to 79 age group but there was no significant progressive decline in fiber intakes with age. Figure 2 shows the distribution of dietary fiber intakes in the group studied. Intakes ranged from 8 to 32 g/day with 76% of the
DIETARY
FIBER ‘ean
IN BRITISH
19.9
POPULATION
1317
#{149}5.3g
11 10
9 8 7 Numbers
6 5 4 3
2
8
10
12
14
16
FIG.
population six subjects
18
2. Distribution
20
22
14 to 26 g/day. Of the than 26 g daily, in three could these high intakes be attribto consumption of whole wheat foods.
only
uted
more
Discussion This vidual
study intakes
to measure mdifiber in a random
was designed of dietary
sample of the British population. Subjects were asked to weigh their food, since this method is thought to give the most valid assessment of food intake (20), but it imposes a considerable burden on subjects and the risk
of an adverse
effect
on cooperation
rates.
The finding that those who weighed their food did not differ from an expected East Anglian population in terms of age, sex, or social class however indicates that no major bias
was
introduced
into
the
sample.
The act of weighing food may disturb normal dietary habits, and the average energy intake of 9.1 MJ recorded in this survey was only 91% of the recommended level (21). However, according to the National Food Survey, where food is not age energy intake of the
has declined steadily MJ in the first three
weighed, British
26
of individual
consuming taking
24
the averpopulation
since 1970, and was 9.3 quarters of 1977, or 93%
28
dietary
30
fiber
32
g
Total
dietary
fibre/day
intakes.
of the recommended tively low energy individual survey,
intake (19). intake recorded in which all food
The relaby this
and drink consumed was measured, is thus in agreement with National Food Survey findings, and supports the downward trend in energy intake in the British population over the past 7 years. The average intake of dietary fiber was 19.9 g/day. This may be taken as a fair estimate of national fiber intakes in Britain, since there was close agreement, Table 2, with the calculations undertaken from an independent source, the 1976 National Food Survey. Previous work (18) has indicated that total fiber intakes in Britain are now lower than 20 years ago. Calculations from the National Food Survey show a fall in fiber intakes of about 1 g every 10 years, and that most of this fall is due to a decline in cereal consumption. Whereas cereals used to supply the majonty of dietary fiber, vegetables are now the main
source,
as has been
found
in this
survey.
If dietary fiber from cereals, rather than fruits and vegetables, is protective against ischemic heart disease (22) this trend is undesirable. Current hypotheses attribute the protective effect of fiber in large bowel disorders, including cancer, to its capacity to dilute large bowel contents and increase fecal weight. The
BINGHAM
1318
shortened transit time associated with increased fecal weight, and diluted contents, would reduce carcinogen contact with the large bowel mucosa. Populations assumed to have high intakes of dietary fiber have large bulky stools and low incidence oflarge bowel disorders. However, there are only two populations where fiber intakes have been measured by the same method of analysis used
here this
(9). Total group
was
dietary similar
fiber
for the
to fiber
intakes
men
in
in men
in urban Denmark (17.2 ± 3. 1 g), and lower than those of rural Finnish men (30.9 ± 11.3 g). The British intake of about 20 g/day may be minimal compared with that of the rural African. A typical rural Ugandan diet for instance might contain 100 to 170 g of dietary fiber per day (S. Bingham and D. A. T. Southgate, unpublished data). With its low intake of dietary fiber the average fecal weight of the British population is correspondingly low. intakes containing
Subjects on controlled 17 to 22 g of dietary
diet
fiber
per day produce stool weights of around 100 g/day (1 1, 23) which compares with fecal weight of over 400 g/day in rural Africa (24). The different fractions ofdietary fiber may be more important than total intakes, particularly in large bowel disorders. In studies in which dietary fiber from apple, cabbage, carrot, bran, and guar gum was fed in equal amounts to subjects, changes in colomc function
were
closely
related
to changes
in intake
of the pentose fraction of the noncellulosic polysaccharides (11). To achieve a 50% increase in fecal weight pentose intakes would have to be increased by about 3 to 4 g/day. Good sources of pentose are bran, wholewheat cereals, peas, and green vegetables. About 330 g of cabbage would have to be eaten to supply an extra 3 g of pentose. All Bran, 25 g, or 13 g of bran would be a more feasible supplement. However, white bread supplies less than 0.2 g of pentose in the average British diet and contains only a tenth of that in wholemeal bread. Replacement of the average intake of 110 g of white bread (17) with 150 g (four slices) of wholemeal would supply 3.0 g of pentose and raise total dietary fiber intakes by 10 g. There have been no other detailed dietary surveys for epidemiological purposes in Britain. When seeking disease incidence,
national reliance
correlations for food
with intake
ET AL.
has been placed on the National Food Survey (25, 26). Average food and nutrient intakes by the standard regions, income group, size of family, and rural or urban area are published each year. The validity of using National Food Survey data in this way is supported by its close agreement with the results of this individual survey. However, the National Food Survey is unique and cannot be used for international comparisons.
data
National sumption
statistics of food have been used
available for confor international
correlations (26) between food intake and disease incidence but the large discrepancy that usually exists between available food calculated from these sources and actual consumption will contribute to the inconclusive results obtained from such exercises. Moreover, this individual survey has shown that there is a large individual variation in fiber intake within the average. The 4-fold range from 8 to 32 g/day is not accounted for by differences between the sexes or declining intakes with age. This variation may be of great importance since differences in disease incidence may be accounted for by individual, rather than average, differences in food intake (27). Careful individual studies, at both national and international levels, might yield more useful information. In conclusion, dietary fiber intakes in Britam are around 20 g/day and are low compared with estimated fiber intakes in less industrialized communities. Such intakes are associated with low fecal weights and with a relatively high incidence of bowel disorders. In Britain there is scope for increasing the average fiber intake by reversing the trend of declining bread and cereal consumption. A significant alteration would be brought about by a greater consumption of whole wheat cereals. But there is considerable variation between individuals in fiber intakes and practically nothing is known of the way in which the intake of fiber and its different components varies nationally and internationally in relation to the incidence of these diseases. More information, from comparable individual studies using new methods of analysis for dietary fiber and its fractions, is needed. U The authors are greatly indebted to those of the village, and their relatives, who made the study possible by agreeing to undertake the weighing of their food. The authors also thank Mr. R. Tibbs of the Cambridge
DIETARY
FIBER
IN BRITISH
Evening News and the village organisations for publicizing the survey, and Dr. D. Lewis for his help. Miss T. Breacker gave valuable help with the dietary analysis and Dr. David Southgate kindly made available much unpublicized data of the fiber content of food.
References 1.
2.
CLEAVE,
T. L., G. D.
Diabetes, Disease.
Coronary Thrombosis Bristol: John Wright,
D. P.,
BuRKIrr,
Press, 3.
N.
PAINTER,
rectum.
7.
London:
D. P.
AND
Cancer
disease
1289,
fibre
and
AND
J. B.
ROBERTSON.
What
is
in food? Nutr. Rev. 35: 12, 1976. 9. IARC Intestinal Microecology Group. Dietary fibre, transit time, faecal bacteria, steroids, and colon cancer in two Scandinavian populations. Lancet 2: 207,
The
response
to dietary
fibre
from
HMSO, 1969. Moanis, J. N., J. W. Mita Diet and Heart: a postscript. 1977. 23. CUMMINGS, J. H., M. J. Hiu, PEARSON
24.
25. 26.
Food
Brit. J. Nutr.
Security.
Recom-
AND
H. S.
D. 0. CLAYTON. Brit. Med. J. 2: 1307,
AND
D. J. A.
WIGGINS.
JENKINS,
Changes
J. R. in fecal
composition and colonic function due to cereal fiber. Am. J. Clin. Nutr. 29: 1468, 1976. Buaiurr, D. P., A. R. P. WALKER AND N. S. PAINTER. Effect of dietary fibre on stools and transit times, and its role in the causation of disease. Lancet 2: 1408, 1972. KNox, E. G. Ischaemic heart disease mortality and dietary intake of calcium. Lancet 2: 1465, 1973.
B. K., J. I.
ARMSTRONG, AND
F.
455,
1975.
MANN,
A. M.
ADELSTEIN
Commodity consumption and ischaemic heart disease mortality with special reference to dietary practices. J. Chronic Diseases 28:
carrot,
apple, bran and guar gum. Lancet 1: 5, 1978. 12. Office of Population Censuses and Surveys. Classification of Occupations. London: HMSO, 1970. 13. Office of Population Censuses and Surveys. Census 1971. Economic Activity Part IV. (10% Sample) London: HMSO.
Foods.
274: 51,
don:
22.
WALKER.
colonic
Social
ROBERT-
Nature
A. M. Diet in pregnancy. and
J.
AND
diet.
mended Intakes of Nutrients for the United Kingdom, Rept. Public Health Med. Subj. no. 120. Lon-
fibre
D. A. T., B. BAILEY, E. COWNSON AND A guide to calculating intakes of dietary fibre. J. Human Nutr. 30: 303, 1976. 11. CUMMINGS, J. H., D. A. T. SOUTHGATE, W. BRANCH, H. HOuSTON, D. J. A. JENKINS AND W. P. T. JAJtas.
The
WIDDOWSON.
BINGHAM
fibre in the British
21.
SOUTHGATE,
A. F.
D. A. T., S.
Dietary
12: 446, 1958. Department of Health
1977.
10.
E. M.
AND
1978.
SOUTHGATE,
THOMSON,
1971.
P. J.,
A.,
20.
Afr. Med.
Rev. 35: 60, 1976. SOEST,
18.
of the colon
J. Changes in the fibre content of the British diet. Nature 238: 290, 1972. SOUTHGATE, D. A. T. “Food and Fibre” Discussion. VAN
R.
composition offoods. Med. Res. Council Spec. Rept. Series 297. London: HMSO, 1960. 16. Accessory Food Factors Committee. The nutritive value of war time foods. Med. Res. Council (War) Memo no. 14. London: HMSO, 1945. 17. National Food Survey Committee. Household Food Consumption and Expenditure 1976. London:
1978. 19. Ministry of Agriculture, Fisheries and Facts no. 9. London: MAFF, 1977.
of Western
South
and Surveys. MortalDH5 no. I. London:
1974.
MCCANCE,
SON.
28: 3, 1971.
evaluation.
HMSO, 15.
Academic
Diverticular
BURKITr.
a deficiency
1319
14. Office of Population Censuses ity Statistics by Area, Series
HMSO,
Car-
ROBERTSON,
Nutr. 8.
Disease.
A. M. Dietary
LUBBE,
J.45:
6.
Refined
TROWELL.
Brit. Med J. 2: 450, 1971. D. P. Epidemiology ofcancer
BURKII-r,
and
S.,
of the colon:
civilisation.
5.
and
AND
1975.
disease 4.
H. C.
AND
Foods
bohydrate
N. S. PAINTER. and the Saccharine 1969.
CAMPBELL
POPULATION
cabbage,
27.
WHITEHEAD,
ESKIN.
R. G. Some quantitative considerations to the improvement of the nutritional rural children. Proc. Roy. Soc. Lond. 199:
of importance status of 49, 1977.
