August 81

Item 3.2.4 of the Provisional Agenda

Joint FAO/WHO/UNU Expert Consultation on
Energy and Protein Requirements

Rome, 5 to 17 October 1981



N.A. Blackburn, D.A.T. Southgate
A.R.C. Food Research Institute, Norwich, UK

1. Introduction

Digestion and absorption should always be considered as inherent parts of protein quality. A protein can be predicted as being of good quality on the basis of its amino acid score, but in practice be of only poor quality because it is poorly digested and/or absorbed. Thus when making recommendations for protein requirements, factors which might affect digestibility or absorption should be considered.

A number of studies have been undertaken to determine what effect ingestion of dietary fibre has on protein digestibility. Experiments have generally been of two types:-

  1. Those where the fibre source is a mixture of polysaccharides, usually in the form of a whole food, and is often associated with one or more of the protein sources.

  2. Those where the fibre source is an isolated polysaccharide, added to a diet containing a separate protein source.

Results from a number of these studies will be reported here.

Measurements are either of ‘apparent protein (Nitrogen) digestibility’, or ‘true protein (Nitrogen) digestibility’.

Apparent Protein (N) Digestibility (%)
True Protein (N) Digestability (%)

I = Nitrogen Intake F = Feacal Nitrogen output on test diet

                               Fk =    "         "         "         on a non protein diet.

The studies with humans reported here determine apparent protein digestibility. The validity of these measurements will be discussed later.

2. Studies in which the Fibre Source is a Mixture of Polysaccharides

McCance and Widdowson (1947), and McCance and Walsham (1948) conducted metabolism experiments on human subjects using wholemeal bread baked from different wheat varieties of differing extraction rates. It was known that raising the extraction rate ( and thus fibre content) of the wheat flour used for making bread increases the amount of N in faeces. The question to be resolved was to what extent the increase was due to undigested and unabsorbed protein in the outer layers of the wheat grain, and to what extent it was due to increased intestinal secretions provoked by the bran layers of the wheat. Results from both studies indicated that the proteins in the various wheatmeals are more or less completely digested and absorbed, and that the N in the faeces is entirely metabolic N. This deduction is supported by the previous results of Heupke (1943), who found that a supplement of bran fed to humans yielded the same increase in fecal N as found when the same amount of bran was ingested whose nutrients had been removed previously by in vitro enzymatic procedures.

Southgate and Durnin (1970) found in a study involving 49 humans of different age and sex that an increase in intake of unavailable carbohydrate in the form of fruit, vegetables and wholemeal bread results in a statistically significant increase in feacal N loss. Apparent digestibility of the protein in the diet was significantly reduced by up to 7%. Urinary N excretion was not significantly changed.

In a study by Walker (1975) in which the diet of South African Negro schoolchildren was supplemented by 350 g oranges/day (approx. 4 g crude fibre, virtually N free), a significant increase in feacal N was found. Urinary N output was not determined. Walker concludes that as the oranges are virtually N free, the extra faecal N must have been non-dietary or endogeuous in origin.

Spiller et al. (1975) compared the effects of various fibre free semisynthetic liquid diets and a control commercial diet on faecal output in pig-tailed monkeys (Macaca nemestrina - these animals have a similar gastrointestinal tract to that of humans). Among the changes in faecal composition noted was an 8-fold increase in faecal N from monkeys fed the commercial diet containing fibre compared to those animals fed the fibre-free diets. The authors conclude that dietary fibre is a major factor affecting faecal output and composition.

Reinhold et al (1976) increased the fibre content of the diet of two human males by substituting bread made partially from wheaten wholemeal flour in place of white bread. There was a statistically significant increase in faecal N excretion, but Nitrogen balance did not change significantly.

Calloway and Kretsch (1978) fed 6 adult human males for 15 day periods on fibre free egg formula diets with and without added oat bran, and a rural Guatemalan diet. The addition of 45g oat bran per day to the egg formula diets resulted in a reduction in apparent protein digestibility of 3 to 4%. There was no significant change in urinary N. The authors suggest a proportion of the extra faecal N could result from increased bacterial N. They also propose that a reduced transit time due to increased fiber ingestion could lead to failure to reabsorb endogenous intestinal N, and N could be trapped in the faecal mass. The largest faecal N loss was when the rural Guatemalan diet was eaten. This diet contained the same level of protein as the egg diet above to which bran was added, but 8-times the level of neutral detergent fibre of the egg and bran diet. Apparent protein digestibility was 69 ± 2% compared to 85 ± 4% on the egg and bran diet, but N balance was maintained in 5 out of 6 subjects. Urinary N was lower on the Guatemalan diet than on the egg and bran diet.

Kelsay et al. (1978) fed a group of 12 men 2 different diets equivalent as far as possible in all respects except fibre content. The low fibre diet contained fruit and vegetable juices, whereas the high fibre diet contained whole fruit and vegetables. Neither diet contained cereal products. On the high fibre diet there was a statistically significant reduction in apparent N digestibility of approx. 9% compared to the low fiber diet. There was no significant reduction in N balance or urinary N excretion.

Farrel et al (1978) increased the fibre content of the diet of 14 adult human males (additional fibre in form of wheat bran) and found a statistically significant reduction in apparent protein digestibility of approx. 2%.Urinary N was not determined.

Stephen and Cummings (1979) fed a British type diet to 2 groups of 6 human subjects, with or without the addition of 18g dietary fibre/day in the form of cabbage or bran. Both bran and cabbage increased faecal N excretion. An increase in faecal excretion of bacteria was also noted, and the authors suggest that this increased output of bacteria accounts for the increased N excretion.

Urinary N excretion was not determined.

3. Studies with isolated polysaccharides

3.1 Cellulose

Keim and Kies (1979) fed cellulose, hemicellulose and lignin to weanling mice at levels of 5, 10 or 15% w/w in a semipurified casein basal diet. In general, as the level of fiber incorporation increased, protein efficiency, N balance, and apparent protein digestibility decreased, and faecal N excretion increased. Urinary N also tended to increased with increased intake of hemicellulose and lignin, but not with cellulose.

Hove and King (1979) found that increasing the level of cellulose in the diets of weanling rats from 0 to 20% resulted in a decreased apparent protein digestibility, when protein was fed at both 7.6 and 20% levels in the diet. Biological value of the protein was not however changed.

Slavin and Marlett (1980) discovered that the addition of 10g cellulose/day to the diets of 7 young women had no effect on apparent digestibility of protein. Mickelson et al (1979) previously found that the addition of 30.2g cellulose/day to the diets of college males also had no effect on faecal N output.

3.2 Hemicellulose

The results of Keim and Kies (1979) are reported above. Hemicellulose had the greatest effect on N balance of the three forms of fibre studied.

Kies and Fox (1977) conducted experiments to determine the effect of graded levels of hemicellulose (4.2, 14.2, 24.2 g/day) added to a constant plant based diet on the protein nutritional status of adult men. Graded increases in hemicellulose had no effect on N balance of subjects in strong positive N balance. However, subjects in marginal or negative N balance tended to show poorer N balance as the level of dietary hemicellulose was increased. Both urinary and faecal N excretion increased with increasing hemicellulose supplements, but these differences were not statistically significant at all levels. The authors propose that these results suggest but do not prove that the poorer N balances achieved by some subjects with higher hemicellulose intakes were due to interference at some stage with protein absorption.

3.3 Pectin

Hove and King (1979) also studied the effect of incorporating pectin (0 to 10%) in the diets of weanling rats. Apparent digestiblity of protein was again depressed as pectin levels were increased, at both levels of protein in the diet. However, as with cellulose, biological value of the protein was not significantly reduced.

In a study by Cummings et al (1979), the addition of 36g pectin/ day to the diets of five male students resulted in a 47% increase in faecal N excretion. Urinary N was not measured. The authors suggest that the increase in faecal N, along with other changes in faecal output observed, could be explained by an increase in faecal bacterial mass.

3.4 Others

Sauer et al (1979) conducted N balance experiments on growing rats in order to assess the effects of level and source of fibre on casein utilisation. Fiber sources used were ‘Alphafloc’, ‘Methocel’, ‘Celufil’, barley straw, and oat hulls, all fed at 5, 10 or 15% levels in the diet. With the exception of ‘Alphafloc’, they found that with increasing levels of the same fibre source, there was no decrease in true protein digestibility, biological value, or net protein utilisation.

Harmuth-Hoene et al (1978) added 5% guar gum to a diet containing 10% casein, and found no change in overall N retention of young rats, but there was a significant shift in N excretion from urine to faeces, with the result that there was a decreased apparent protein digestibility.

Harmuth-Hoene and Schwerdtfeger studied the effects of five indigestible polysaccharides (guar gum, carob bean gum, sodium alginate, agar-agar and carrageenan, all fed at a 10% level in the diet) on N balance in growing rats. The polysaccharides all reduced apparent protein digestibility, but N retention was only significantly lower after ingestion of agar-agar and carrageenan, Urinary N excretion was reduced to such a level in rats receiving guar gum, carob bean gum, and sodium alginate, so as to compensate for increased faecal loss. The authors postulate two mechanisms to explain the results:- (i) The presence of undigested guar gum and carob bean gum in the large intestine results in enhanced bacterial activity, and therefore increased microbial N requirement. The reduction in urinary N seen in the animals receiving guar gum and carob bean gum indicates the additional N is being supplied by the host organism from endogenous sources. The increase in faecal N loss is therefore most likely of bacterial origin. Thus true digestion of dietary protein is not being reduced. (ii) In the case of carrageenan and agar-agar, the authors suggest the increased faecal N excretion is due to interference in protein digestion by the polysaccharides. Carrageenan was observed to decrease trypsin activity in vitro. In this situation the protein will be less well digested.

Prynne and Southgate (1979) supplemented the diets of 4 humans with 25g/day Isphagula husk (Isogel) for 3 weeks. Only one subject showed a decrease in apparent protein digestibility.

4 Individual Variation in Protein Digestibility

Differences in diets, analytical procedures and experimental techniques between the studies are such that the extent of individual variation in protein digestibility should only be assessed within individual studies.

McCance and Glaser (1948) found that the apparent digestibility of oatmeal protein varied from 64.9% to 81% in the 6 human male subjects they studied.

Southgate and Durnin (1970) found a statistically significant difference in the apparent digestibility of protein between the groups of young men and young women they studied.

In the study by Kelsay et al (1978) individual variation in apparent protein digestibility between 12 men on a high fibre diet was small (81.1 + 1.1% apparent protein digestibility).

Variations such as those mentioned above are found in all the studies reported. However, experimental error, even within studies, will account for a proportion of this variation.

5. Conclusions

Interpretation of available evidence is made difficult by two factors. Firstly, there is considerable variation in experimental design, making it difficult to quantitatively compare results from even the same fibre source. Secondly, studies with humans have all involved determinations of apparent protein digestibility, and a reduction in apparent digestibility indicates only an increase in faecal N output. There are 3 possible sources of the extra faecal N detected:-

  1. Bacterial

    An increased fibre intake stimulates bacterial growth in the large intestine resulting in increased faecal bacterial output, and therefore increased N output. [Stephen and Cummings (1979); Slavin and Marlett (1980); Cummings et al. (1979); Harmuth-Hoene and Schwerdtfeger (1979); Calloway and Kretsch (1978); Kelsay et al. (1978); Sandberg et al. (1981)].

  2. Endogenous

    An increased fibre intake increases intestinal secretions and sloughing of intestinal epithelial cells. [McCance and Walsham (1948); Walker (1975)].

  3. Undigested Food Protein

    This can result from:-

    1. Inhibition of proteolytic enzymes by dietary fibre [Harmuth-Hoene and Schwerdtfeger (1979); Sauer et al. (1979)].

    2. A reduction in transit time induced by increased intake of dietary fibre allows less time for nutrient absorption [Southgate and Durnin (1970); Calloway and Kretsch (1978)].

    3. Protein in intact plant cells could be less accessible to digestive enzymes and therefore be less well digested and absorbed. [Southgate and Durnin (1970); Calloway and Kretsch (1978)].

When the effects of dietary fibre on protein digestibility in relation to protein quality is the main consideration, an increase in faecal N is of significance only if one of the sources of faecal N is undigested and unabsorbed food protein. Measurements of apparent protein digestibility do not identify the source of faecal N, and therefore a reduction in apparent digestibility is not necessarily significant in this respect. A reduction in true protein digestibility (as defined by the equation given earlier) is also not necessarily significant. The Fk Nitrogen component in the definition is intended to be a measure of endogenous (and bacterial) faecal N output such that (F - Fk) represents N from undigested food protein. However, unless Fk is determined on a diet identical in all respects to the test diet, but minus the protein component (this is difficult as fibre and protein sources are often associated in the same food), an incorrect result will be obtained, as some fibre sources appear to increase endogenous and bacterial N output [McCance and Walsham (1948); Walker, (1975); Stephen and Cummings (1979); Sandberg et al (1981)]

When protein quantity, not quality, is the main consideration, then the source of any increase in faecal N content is immaterial. Thus a reduction in true digestibility or apparent digestibility is likely to be relevant, though there are 2 possible problems. Firstly, in view of the results of Harmuth-Hoene and Schwerdtfeger (1979), urinary N output should be determined. These results suggest that urinary N loss can be reduced in the rat in order to compensate for increased faecal N loss. This does not seem to apply in the results from human studies at present, but the possibility should be given consideration. Secondly, Oomen and Corden (1970) proposed the existance of N fixing bacteria in the large intestine of New Guinean sweet potato eaters in order to explain the consistant negative N balances found in these healthy people. This phenomenon is unlikely to be widespread, but it should not be overlooked in populations with similar lifestyles and diets to the New Guineans studied.

The available evidence, in view of the problems discussed, is such that only the following points can be made:-

  1. It is incorrect to generalise that dietary fibre reduces protein digestibility, either true or apparent. Effects of different fibre sources vary.

  2. Reductions in apparent protein digestibility have been found in humans after ingestion of some fibre sources but as such do not necessarily mean the protein in the diet is less well digested and absorbed. The reductions in apparent digestibility are usually less than 10%.

  3. There is variation in apparent protein digestibility between human subjects. The 30% increase in recommended protein intakes to allow for differences in individual requirements should cover these variations.

Obviously, more work needs to be done, especially in relation to the sources of increased faecal N output. Work with ileostomy patients (Sandberg et al. (1981)) should yield valuable results on how well particular proteins are digested. Work by Rich et al. (1980) suggests that the rat might be a good model for use in protein digestibility studies, but in view of the conflicting results on the effects of cellulose on protein digestibility [Keim and Kies (1979); Hove and King (1979); Slavin and Marlett (1980); Mickelson et al. (1979)] results from rat studies should be treated with caution. Digestibility determinations using in vitro enzyme techniques [Hsu et al (1977)] could also be used.


McCance, R.A. and Widdowson, E.M. (1947). J. Hyg. Camb., 45, 59.

McCance, R.A. and Walsham, C.M. (1948) Br. J. Nutr., 2, 26.

Heupke, W. (1943) Deut. Z. Verdauungs-Stoffwechselkrankh, 7, 49.

Southgate, D.A.T. and Durnin, J.V.G.A. (1970) Br. J. Nutr., 24, 517.

Walker, A.R.P. (1975) Am. J. Clin. Nutr., 28, 1161.

Spiller, G.A., Saperstein, S., Beigler, M.A., Amen, R.J. (1975) Am. J. Clin. Nutr., 28, 502.

Reinhold, J.G., Faradji, B., Abadi, P., Ismail-Beigi, F. (1976) J. Nutr., 106, 493.

Calloway, D.H., Kretsch, M.J., (1978) Am. J. Clin. Nutr., 31, 1118.

Kelsay, J.L., Behall, K.M., Prather, E.S. (1978) Am. J. Clin. Nutr., 31, 1149.

Farrel, D.J., Girle, L., Arthur, J. (1978) Aust. J. Exp. Biol. Med. Sci., 56, 469.

Stephen, A.M., Cummings, J.H. (1979) Proc. Nutr. Soc., 38, 141A.

Keim, K., Kies, C. (1979) Cereal Chem., 56, 73.

Hove, E.L., King, S., (1979) J. Nutr., 109, 1274.

Slavin, J.L., Marlett, J.A. (1980) J. Nutr., 110, 2020.

Mickelson, O., Makdani, D.D., Cotton, R.M., Titcomb, S.T., Colney, J.C., Gatty, R. (1979) Am. J. Clin. Nutr., 32, 1703.

Kies, C., Fox, H.M. (1977) J. Food Sci., 42, 440.

Hsu, H.W., Vavak, D.L., Satterlee, L.D., Miller, G.A. (1977) J. Food Sci., 42, 1269.

Cummings, J.H., Southgate, D.A.T., Branch, W.J., Wiggins, H.S., Houston, H., Jenkins, D.J.A., Jivraj, T., Hill, M.J. (1979) Br. J. Nutr., 41, 477.

Sauer, W.C., Eggum, B.O., Jacobsen, I. (1979) Arch. Tierernährung 29, 533.

Harmuth-Hoene, A.E., Jakubick, V., Schelenz, R. (1978) Nutr. Metab. 22, 32.

Harmuth-Hoene, A.E., Schwerdtfeger, E. (1979) Nutr. Metab. 23, 399.

Prynne, C.J., Southgate, D.A.T. (1979) Br. J. Nutr. 41, 495.

McCance, R.A., Glaser, E.M. (1948) Br. J. Nutr. 2, 221.

Sandberg, A.S., Andersson, H., Hallgren, B., Hasselblad, K., Isaksson, B., Hultén L. (1981) Br. J. Nutr. 45, 283.

Oômen, H.A.P.C., Corden, M.W. (1970) Metabolic Studies in New Guineans ; Nitrogen Metabolism in sweet potato eaters.Technical Paper No. 163.South Pacific Commission, Noumea, New Caledonia.

Rich, N., Satterlee, L.D., Smith, J.L. (1980) Nutr. Rep. Int., 21, (2), 285.

Top of Page