Contents - Previous - Next


3. The effects of different levels of protein intake on energy metabolism


Only minimal discussion will be given to consideration of the effects of changes in dietary protein intake on energy metabolism, since this is dealt with in considerable detail in other chapters. Much work in this area relates to studies of dietary changes in the treatment of obesity although, at the other extreme, under conditions of dietary deficiency, marasmus and kwashiorkor can occur. While kwashiorkor is normally regarded as being a dysfunction of protein metabolism, it could just as plausibly be viewed as a maladaptation of energy metabolism, induced by a combination of high energy intake in association with low levels of protein. This would be especially true if the views of WHITEHEAD and ALLEYNE (1972) were accepted that the high-carbohydrate diet elevates insulin, which then influences the amino acid pool and causes both low albumin synthesis and hence oedema and also poor fat transport, due to low synthesis of VLDLs, and hence fatty liver.

Protein turnover and thermogenesis in response to high-protein and high-carbohydrate feeding in men were investigated by ROBINSON et al. (1990). These investigators determined the rates of energy expenditure and whole-body protein turnover during a 9-hour period in a group of seven men while they received hourly isocaloric meals of high-protein (HP) or high-carbohydrate (HC) content. Their responses to feeding were compared with those to a short period of fasting (1524 hours). The 9-hour thermic response to the repeated feeding of HP meals was found to be greater than that to the HC meals, and the rate of whole-body nitrogen turnover over 9 hours also increased, especially with HP feeding. By using theoretical estimates (based upon ATP requirements) of the metabolic cost of protein synthesis, some 36% of the thermic response to HC feeding and 68% of the response to HP feeding could be accounted for by the increases in protein synthesis compared with the fasting state.

Stimulation of protein turnover by carbohydrate overfeeding in 11 healthy men was investigated by WELLE et al. (1989). Total urinary nitrogen output during 10 days of carbohydrate overfeeding (1600 extra kcal/d) decreased 27% relative to nitrogen excretion during 10 days of weight maintenance, demonstrating protein accretion during overfeeding. However, postabsorptive nitrogen excretion did not change, indicating that the positive nitrogen balance associated with overfeeding resulted from enhanced postprandial nitrogen retention. Overfeeding reduced postabsorptive glucose concentrations and increased both glucose production rate and glucose clearance. Increases were observed in the plasma concentrations of insulin, glucagon, 3,5,3'-triiodothyronine, alanine and branched-chain amino acid concentrations following overfeeding, but serine, threonine, and asparagine concentrations were reduced. Postabsorptive leucine flux, an index of proteolysis, was also increased following overfeeding when compared with values after 10 days on a weight-maintenance diet. It was suggested that excess food energy intake, without an increase in protein intake, stimulated postabsorptive proteolysis and protein synthesis. Although carbohydrate and energy metabolism were investigated and showed changes, this study could be considered as actually more in the area of the effects of energy on N metabolism than of N on energy metabolism.

A major component of total energy needs is represented by resting energy expenditure. Whether variability among healthy young adults in resting metabolic rate, normalized for the amount of metabolically active tissue (assessed by total body potassium), was related to protein turnover, was investigated by WELLE and NAIR (1990). High correlations between leucine flux (an index of proteolysis) and metabolic rate (R= 0.84) and between the non-oxidized portion of leucine flux (an index of protein synthesis) and metabolic rate (R=0.83) were found. Regression analysis suggested that the contribution of protein turnover to resting metabolic rate was approximately 20% in an average subject, a value similar to that proposed by YOUNG et al. (1992). Metabolic rate and protein turnover were highest in the subjects with the greatest amount of body fat, even after accounting for differences in whole-body potassium.

In investigations concerned with the optimal composition of reducing diets, HENDLER and BONDE (1988) studied the impact of very-low-calorie diets with high and low protein content on triiodothyronine, energy expenditure, and nitrogen balance. Seventeen obese inpatients received 440 kcal/d for 3 weeks. The diets consisted of either 41% protein plus 55% carbohydrate or 95% protein. There were no significant differences produced by diet in weight loss, loss of lean mass, metabolic rate reduction or meal-stimulated thermogenesis. In all, subjects fed the high-carbohydrate diet gained no advantage over subjects fed that containing high protein, and regression analyses revealed no relationship between thyroid hormones, energy deficit, or lean mass with nitrogen losses, suggesting that other or more complex processes govern endogenous protein metabolism during weight loss.

HENDLER and BONDE (1990) have also shown that 800 kcal/d sucrose diets, unlike pure protein diets, maintained resting metabolic rate (RMR) and triiodothyronine (T3) levels. Since thermogenesis from sucrose might reflect protein catabolism, 23 obese women were studied as inpatients for 2 weeks at dietary levels of 50% of their RMRs. These diets contained either 93% sucrose, sucrose plus protein (75%:20%) or fat plus protein (75%:20%). RMR, leucine kinetics and nitrogen balance were measured.

RMR fell with the sucrose-plus-protein and the fat-plus-protein diets, but was maintained on the sucrose diet. Plasma leucine decreased on the two sucrose-containing diets but increased on the diet containing fat. Leucine turnover, oxidation, and non-oxidative disposal all also decreased on the sucrose diets again, in contrast to the fat-containing diet. Cumulative (2 weeks') nitrogen loss was least in the sucrose-plus-protein diet, when compared to the other diets. The authors concluded that the thermogenic effects of sucrose do not depend on protein catabolism, since the inclusion of sucrose in hypocaloric diets maintained RMR, while decreasing leucine turnover, oxidation, and nitrogen loss.

The effect of a 50% reduction in food intake on energy expenditure, protein metabolism, glucose cycling, and body composition was investigated in eight moderately overweight men by STEIN et al. (1991). The pre-study mean energy and protein intake was determined for the eight subjects. They were then maintained on this diet, which averaged 3270 kcal/d and 20 g N/d, for 6 weeks. The diet was reduced uniformly in the major foodstuffs by 50% for the next 4 weeks to an average of 1555 kcal/d and 9.6 g N/d. The subjects lost 4.0 kg fat during the 4 weeks on the reduced energy regimen, and it was determined that both protein turnover and glucose cycling were reduced.

Nutrient oxidation patterns and protein metabolism in lean and obese subjects were examined by BRUCE et al. (1990). The immediate metabolic response to eating was compared in a group of grossly obese subjects (BMI=45) with that in lean controls (BMI=22). The changes in energy expenditure and nutrient disposal with the onset of eating were assessed by a method of combined respiratory gas analysis and intravenous infusion of 13C-labelled leucine. Leucine kinetics were used to quantitate rapid changes in protein oxidation and to assess protein synthesis and degradation. Total energy expenditure was 20-30% greater in obese than lean subjects in fasting and feeding, but energy expenditure expressed per kg fat-free mass was similar in obese and lean subjects in both fasting and feeding.

The onset of eating was associated with increased carbohydrate and protein oxidation with decreased fat oxidation in both lean and obese individuals. In obese subjects, however, both the decrease in fat oxidation and the increase in protein oxidation were significantly smaller than the corresponding increments in lean subjects. The rate of protein synthesis (and probably also protein degradation) was significantly higher in obese subjects, both in the fasting state and in the fed state. The observed differences between obese and lean individuals in protein and energy metabolism in the fasted state and in the immediate response to eating led the authors to observe that their study does not support a hypothesis of greater metabolic efficiency in obesity.

Rates of whole-body amino nitrogen flux were investigated (PENCHARTZ et al., 1988) for adolescents undergoing obesity treatment using a high-protein low-energy diet. The subjects received approximately 2.5 g of animal protein per day and kg ideal body weight, and maintained nitrogen balance throughout the 18 days on the diet. Flux rates were calculated separately from the cumulative excretion of 15N in urinary ammonia and urea following the administration of a single dose of 15N glycine. The pattern of 15N label appearance in urinary ammonia and urea nitrogen was followed for 72 hours after the administration of 15N glycine. Significant amounts of label continued to be excreted in both urinary ammonia and nitrogen for 36-48 hours after label administration.

The weight-reducing diet accelerated 15N cumulative excretion in urinary urea, but not in ammonia nitrogen compared with the control diet. Whole-body nitrogen flux rates increased rapidly and significantly on the diet. Using the urea end product, this increase was evident on the 4th diet day, but not by the 7th or subsequent days. On the other hand, using the ammonia end product, flux rate increased markedly and remained elevated throughout the whole study. These results demonstrate adaptive changes in whole-body amino-nitrogen metabolism in response to the reducing diet.

Different patterns of change were seen, depending upon whether an ammonia or a urea end product was used, which adds to the evidence for compartmentation of the body's amino-nitrogen pools. A conclusion from this and other studies discussed above would be that severe energy limitation does not significantly reduce protein turnover when protein intake is maintained at an adequate level to permit N balance.

Resting energy expenditure is reduced in elderly human subjects and, in some studies, even after adjustment for body size and composition, although the age-related differences are small in this case. Extending this observation, the thermic effect of a protein meal was investigated by FUKAGAWA et al. (1991) in young men (20-26 yr; n = 9), old men (70-89 yr; n = 9), and old women (67-75 yr; n = 6). Energy expenditure was measured before and from 1 to 6 hours after presentation of 60 g protein and of a control, non-caloric meal on separate occasions. Despite substantial differences in body size and composition, the protein-induced increment in energy expenditure was similar in all groups.

Although fasting plasma norepinephrine (NE) levels differed among all three groups, their concentrations were not affected by protein ingestion. These and other data, including the blocking of extraneuronal synthesis of dopamine and serotonin which can stimulate energy expenditure, demonstrated that not all mechanisms responsible for EE decline with age, and that protein-induced changes in EE are more a function of the oral load of protein than of the size, age, or antecedent diet of the individual ingesting the protein.

A specific final example from WESTPHAL et al. (1990) involved seven healthy, normal-weight subjects, fed breakfasts of 50 g protein, 50 g glucose, and 10, 30, or 50 g protein plus 50 g glucose in random sequence. A number of biochemical indicators including plasma glucose, insulin, C peptide, glucagon, non-esterified fatty acids, and alpha-amino nitrogen were then determined, and the net area of the response curves were calculated. Ingestion of 50 g protein alone did not change the serum glucose concentration, and the various amounts of protein ingested with 50 g glucose also did not alter the serum glucose response, compared with that observed with 50 g glucose alone. Ingestion of the various amounts of protein did not result in a further increase in insulin concentration when ingested with glucose, except with the 50 g protein dose which produced only a modest increase.

Ingestion of glucose resulted in a decrease in alpha-amino nitrogen and glucagon concentrations, whereas ingestion of protein increased them, as expected. Additions of progressively larger amounts of protein to the glucose meal resulted in a progressive increase in the alpha-amino-nitrogen- and glucagon-area responses. The null point, that is, the protein dose ingested with 50 g glucose at which there would be no change in area response, was estimated to be 9 g protein for alpha-amino nitrogen and 5 g protein for glucagon. As a general comment of the relationships between protein and energy metabolism, it is concluded that while there are real effects of protein on energy metabolism, these would appear to be generally less significant, in the context of establishing macronutrient requirements, than are the effects of dietary energy intake on body protein metabolism and requirement estimates for dietary protein (nitrogen).

4. Protein/energy ratios


Part of the rationale of this workshop on protein-energy interactions was to make recommendations for desirable protein/energy ratios in health and disease, and three of the four panels are constrained to provide such recommendations. It is nevertheless considered that protein/energy ratios (expressed as Pcal%) should also be discussed briefly here, especially as they concern the data presented in this chapter.

When the combined N-balance data were examined, it was noted that the correlation between Pcal% and NB was non significant (R2=0.007) and that therefore Pcal% by itself was of little predictive value. Nevertheless the combination of NI and Pcal% was marginally better in predicting NB than was the regression equation using NI and EI as the independent variables. The regression equation obtained was:

NB = 0.43 NI-4.69 Pcal-22.75

R2 = 0.57

suggesting that 57% of the variation in NB came from NI and Pcal%. The average and SD of the Pcal% (Pcal% = protein g/100 g x 400 divided by kcal/100 g) values in the data set (n = 361) was 6.1 2.9 with a maximum of 30.2 and a minimum of 1.6. When the CALLOWAY and SPECTOR (1954) data were not included, the maximum value was only 11.4, and the mean and SD were reduced to 5.8 2.4. When the data are classified into Pcal% groupings (eight groups from Pcal% 3 to > 9), as shown in Figure 13, a pattern of relationships does emerge in that, for this data set, the increase of Pcal% is mainly a result of increasing NI and there is a parallel increase in NB as Pcal% increases up to about Pcal% = 8. At higher Pcal% values NB is reduced (causing the lack of overall significance) as a result of lower energy intakes.

The mean Pcal% for the US diet, as calculated from USDA (1983) food availability data, is much higher than these values and averages about 16% (PELLETT and YOUNG, 1990). The range for requirement data (Table 5) is, however, only from 3.6 to 8.2% at moderate activity levels. Since energy allowances are averages and protein allowances are safe levels (means + 2SD), the direct ratios of allowances (Column A in Table 5) are misleading, and the ratios (Column B) calculated from average protein needs are more appropriate. It can thus be seen that the experimental data represent the low end of the range for Pcal% values. This is not surprising, when the rationale behind the experiments is considered, which was often to examine the effect of excess energy at low to normal protein intakes.

For a number of years, Pcal% has been suggested as a useful dietary indicator for protein sufficiency (MILLER and PAYNE, 1961; BEATON and SWISS, 1974), but its limitations have also been recognized (BEATON and SWISS 1974; PAYNE, 1975; FAO/WHO/UNU, 1985). The relevance of the simple ratio, even when correctly calculated (average protein: average energy), as a basis for assessing diets has been questioned since it does not take into account individual variability in the needs for energy and of the extent to which these are independent of variability in protein requirements. For the US diet, the Pcal% stays at about 16% (15.4 to 17.5) across a wide range of age, sex and income groups, despite a more than two-fold range in protein availability (PELLETT and YOUNG, 1990), and even for poor developing countries, average Pcal% values from food balance sheet data (YOUNG and PELLETT, 1991) are usually in excess of Pcal% = 11%. These ratios can be seen to be well in excess of the ratios calculated from requirement data (Table 5).

Figure 13. Relationship between N balance, energy and nitrogen intakes when classified by protein/energy ratios (Pcal%).

The major problem in assessing the relevance of Pcal% (or any other indicator of protein/energy ratios), is to evaluate the range over which individuals can adapt either energy intake to suit expenditure, or expenditure to suit intake, without detriment to health or growth (PAYNE, 1975). Two solutions to the problem have been suggested. The first is to use observed variability data for energy intakes in populations (BEATON and SWISS, 1974), while PAYNE (1975) has suggested use of experimental evidence for the minimum energy intake for the maintenance of body energy content. PAYNE (1975) suggests that both give quantitatively similar results.

Evaluation of the data set in terms of Pcal% is unable to address these considerations of adaptation. What, however, does emerge, is that both EI and NI seem to be individually effective in improving N balance, and no stepwise progression could be observed. In other words, the accepted concepts from both CALLOWAY and SPECTOR (1954) and MUNRO (1951; 1978) appear true only under much more extreme conditions of protein or energy limitation than is met within the analyzed studies.

Within the range examined, it appears that increasing either EI or NI or both can individually improve nitrogen balance even at the lowest levels of the other. This can make the use of Pcal%, or any other form of P/E ratio for assessing the value of diets, somewhat confusing. As can be seen, if it be true that increases in either protein or in energy intakes lead to improved N balance, then the former will increase the Pcal% while the latter will reduce it, yet both actions have a positive impact on NB. This is, of course, borne out by the lack of correlation between NB and Pcal% using the full data set.

A similarly confusing situation can be seen by examining Table 5, where at first sight it is surprising that the 'requirement' values for Pcal%, increase with age in apparent contradiction to the widely held view that the protein value of diet is of greater importance for the younger age groups than for the older. This occurs because the decrease in average energy needs per unit body weight with age is far steeper than the decrease in protein requirements. As is emphasized elsewhere in this volume, the use of a protein/energy ratio for assessment of dietary protein value requires a considerably greater degree of sophistication in its interpretation than the simple nature of the ratio would seem to imply.

5. Summary and conclusions


Changes in dietary intakes of protein and energy will significantly influence human nitrogen metabolism. This arises both from changes in the supply of amino acids that serve as substrates for the formation of polypeptides, and from changes in the amounts and sources of chemical energy for the elaboration of high-energy intermediaries (ATP and GTP) required for the formation of the initiation complex, amino acyl-tRNAs, peptide bond formation and for the release of amino acids from dietary and endogenous proteins into the tissue free amino acid pools. In addition, the status of both body nitrogen and energy metabolism is determined by endocrine function and balance which are themselves affected by protein and energy intakes.

While the relationships between protein and energy were well known in the early era of nutritional investigation, it was not until the extensive reviews of MUNRO (1951) and CALLOWAY and SPECTOR (1954) appeared that the relationships were again considered seriously in nutrition research, and it had become clear that many older data on human protein and amino acid needs were conditioned by the levels of food energy given during the nitrogen balance experiments from which conclusions about requirements had been drawn.

Although we now appreciate that there are a vast number of factors involved in determining the relationships between protein and energy intakes and protein metabolism, the major focus of this review has been on the effects of changes in dietary energy intake on protein metabolism, as reflected by alterations in nitrogen balance in adult subjects. Brief consideration has also been given to the needs of children, including those under the special conditions of prematurity and those recovering from malnutrition.

Nitrogen balance has been demonstrated as being influenced by changes in food energy intake, below or above energy needs. Thus, nitrogen balance must be seen as a result of levels of both energy and protein intake, and both need to be considered before any reliable interpretation of nitrogen balance can be made in relation to protein requirements. By current theory, the improvement in nitrogen balance caused by an increase in energy intake can be frustrated, however, if intake of protein is inadequate; conversely, the beneficial effects of an increase in protein intake can be inhibited by an inadequate energy intake. The results of our analysis, however, question this interpretation in that it may only be true under highly restrictive conditions.

A number of studies from the literature over the last three decades have been evaluated, and, by expressing the results in the same units, a single data set was produced which allowed a more detailed examination of the possible statistical correlations between nitrogen and energy intakes and nitrogen balance. By using simple linear regression, 33 to 36% of the variation in NB could be explained separately by NI and EI, while by using multiple linear regression some 53% of the variation in N balance could be explained by NI and EI in combination. Despite the perceived importance of dietary protein/energy ratios, the relationship between NB and Pcal% was non significant and confirmed that Pcal% by itself is of little predictive value. Nevertheless, Pcal% in combination with protein intake (NJ) was a better predictor of nitrogen balance (R2=0.57). Further analysis, using the full data set grouped for EI and NI, demonstrated that over the range for EI and NI, from ca. 30 to 60 kcal/kg and 40 to 200 mg N/kg, increases in either El or NI improve NB, even at the lowest levels of El (< 30 kcal/kg) or NI (< 50 mg N/kg).

It would appear, therefore, that the statements of both CALLOWAY and SPECTOR (1954) and MUNRO (1951; 1978), which claimed that the improvement in nitrogen balance caused by an increase in energy intake could be frustrated if intake of protein were inadequate, and that, conversely, the beneficial effects of an increase in protein intake could be inhibited by an inadequate energy intake, were applicable only under much more extreme conditions of protein or energy limitations than was generally the case for the analyzed studies. Within this range, that is from about 2100 kcal and 18 g protein to about 4200 kcal and 90 g protein/d for a 70 kg male, increases in both EI and NI appeared to be separately and individually effective in improving NB, and no stepwise progression, i.e., expected improvement in NB by increases in either NI or EI being frustrated by a lack of increase of the other, could be observed.

It is also concluded that, while there are very real effects of protein on energy metabolism, they are generally less significant, in the context of requirement estimates, than are the effects of energy on protein (nitrogen) metabolism. Finally, although Pcal% has been suggested as a useful dietary indicator for protein sufficiency, there are considerable limitations in its practicality. This emerges, in part, from the fact that it is not generally directly correlated with nitrogen-balance response; the use, therefore, of protein/energy ratios for assessment of dietary protein value requires careful and detailed evaluation of the specific circumstances involved. Nevertheless, when energy needs are met, a minimum percentage of protein can be specified.


Contents - Previous - Next