3. ADULT PROTEIN REQUIREMENTS

Three separate types of studies were conducted to explore questions of adult protein requirement. The first studies followed the standard UNU protocol and were designed to gather data comparable to those previously gathered from a greater variety of environmental situations. The second set of studies examined the nitrogen balance behaviour of individuals over long periods of time (2–3 months), with the question of how variable the individual was in terms of these data. Here the question was how representative of the individual status was the data from the standard protocol. The third set of experiments was directed at the variability among individuals. These experiments examined the balances of large numbers of individuals (50) receiving a constant intake over a short period of time, equivalent to a single level of the standard protocol. The intake levels initially were chosen to test whether the mean requirement plus two standardard deviations, as estimated in previous experiments with the same diet, did indeed exceed the requirement of almost all of the larger sample. (This is the fundamental question of whether the response has a Gaussian distribution.)

3.1 Standard Protocol

Six studies that followed the standard protocol were presented at the meeting; these were conducted in India (paper 6, two studies - one involving men and the other women), China (paper 7), Egypt (paper 8 - women), Turkey (paper 9), and Nigeria (paper 10). An additional report from Brazil (paper 11) is included, although it arrived too late for review by the group. A summary of the status of each project with respect to the important factors listed in the methodology section is shown in Table VI. This table and subsequent discussion include not only these new results, but also those reported to the previous workshop (8). Note that for comparative purposes all results are calculated as indicated in the methodology section (Table IV), and may differ from those figures reported by the original authors. While this method of analysis is not strictly applicable for these data sets, in which protein quality varied, examination of the actual data suggests that the error introduced is well below that error inherent in the experiment.

Studies in males

Predicted intakes to achieve zero nitrogen balance for the studies involving males are shown in Table VII. The predicted N requirements on egg and fish protein diets in three studies from Chile, Taiwan, and Japan range between 94 to 101 mg N/kg/day. The estimate for egg protein from Thailand is higher (124 mg N/kg/day), which could be due to unmeasured sweat losses caused by higher temperature, a habitual rice diet, and a low Kcal intake in some subjects.

On mixed animal and cereal protein diets, the estimated mean values ranged from 97 to 148. Closely similar estimates were observed in studies from Nigeria, Chile, Mexico, Taiwan, Brazil, and Japan, the values being 127, 133, 121 127, 107, and 106 mg N/kg, respectively, and averaging 120 mg N/kg/day. The high estimates of 148 for data from North China is unexplained, but may be due in part to the large amounts of rice consumed, a situation similar to that In Thailand with the egg and rice diet.

Indian subjects showed the surprisingly low estimate of 87 mg N/kg for mean requirement. Factors such as high temperature may have resulted in excessive sweat nitrogen loss, and the poorer quality of the diets supplied at lower levels compared with that at higher levels might contribute to lower zero nitrogen balance intercepts. Initial examination of the laboratory standardization exercise gives no insight into these results.

A = Animal Protein

V = Vegetable Protein

Studies on females

Studies of N requirements of women were presented from India (paper 6) and Egypt (paper 8). Additional data were presented by Calloway from the United States. Table VIII shows how the important factors in the standard protocol effect each of these studies.

In both the Indian and Egyptian studies, protein quality was not constant at the different levels of dietary protein, and the occurrence of excessive integumental losses was possible as the temperatures were high on some days. Integumental losses were not measured in either study. The predicted effect of these two sources of biäs would be a reduction in the predicted N requirement (i.e., intercept).

In the Indian and United States experiments, dietary treatments were presented in a descending order only. This design may lead to a higher estimate of the requirement, but other factors in their studies could have the opposite effect, as discussed in the methodology section.

A summary of the results is presented in Table IX. When egg albumin was fed, the predicted N requirement was 73 mg N/kg. The two values for a mixed diet were 74 and 109 mg N/kg. The Egyptian estimate is lower than expected, and is believed to reflect bias due to excessive integumental losses, since a high of 39°C was reached on some days during the experimental period.

Although there are very few data, it appears that the N requirement for women in mg/kg is not markedly different from that for men. For high-quality protein, the estimate is about 70–75 mg N/kg. The Agarwal study predicted a need of about 110 mg/kg from a mixed diet. If the egg albumin data are adjusted for a dietary protein score of 70, the predicted N need is about 105 mg N/kg, a value quite close to that of Agarwal.

Additional data were presented by Calloway that indicated that, when using short-term balances in women for determining nitrogen (N) requirements, the influence of the menstrual cycle on N utilization should be considered. Recent work Kurzer and Calloway (9) reported a statistically significant biphasic cycle in urinary N excretion during the menstrual cycle in six healthy young women. Urinary N tended to rise during the follicular phase, to be lowest at about the time of ovulation, to increase sharply during the first week of the luteal phase, and to drop again just before or with the onset of menstruation. Urinary urea excretion and a pattern similar to total urinary N. In one subject, the high and low points of the cycle were about ± 15 % from the cycle mean. It is interesting that an infectious disorder in one woman during the study appeared to suppress the cycle.

Failure to take account of menstrual cycles in determining N requirements could lead to larger coefficients of variation. For example, if requirements are predicted at the lowest and highest points of urinary excretion in each cycle, the predicted requirements would be 42 ± 15 mg N/kgg and 109 ± 36 mg N/kg, respectively. Selection of 4 points at random in the cycle predicted a requirement of 70 ± 54 mg N/kg. The CV is much greater, however; 77% versus 27%. Thus, the effect of the menstrual cycle on N requirements should be minimal if there is a random distribution of points in the cycle among the population studied, and if the distribution of dietary treatments is random with respect to the cycle.

3.2 Long-term studies

As discussed in the introduction and methodology sections, short-term (5–15 days) nitrogen balance studies are not, by themselves, sufficient to estimate protein needs. The group considered it essential that any method used to arrive at estimates of protein requirement should be validated by long-term studies. The variable time required for metabolic adaptation and present limitations in the sensitivity of indicators requires that the effects of a given protein intake be studied over a period of time of at least 60–90 days before concluding that it is adequate or insufficient to meet protein energy needs.

There are multiple problems in the design and interpretation of long-term metabolic studies; most of these also confound short-term studies and have been discussed in the methodology section. The key issue in the evaluation of long-term data is the selection of indicators that adequately reflect protein nutritional status at intakes close to requirement level. Currently, available methods to determine loss of mean body mass, and functional consequences of insufficient protein intake are not sensitive enough to define requirement precisely. The high variability of nitrogen balance over time within and among individuals further complicates data analysis.

Five studies (papers 7 and 12–15) were presented in which subjects were fed a predominantly mixed vegetable protein. Intake was kept nearly constant for 50 to 120 days at levels derived from short-term N balance responses but above the safe FAO/WHO recommended levels. Other studies in which the safe level or a lower level of protein of high quality were tested are included for comparative purposes. These results are summarized in Table X.

It appears from these data that most young adults on a predominantly vegetable diet fed about 1.0 g/kg/day of protein are close to nitrogen balance and do not show signs of adverse effects in health or nutritional indicators (e.g., serum albumin, creatinine/height, serum transaminase levels, total body potassium, weight) over a 60–90 day period. Several subjects in the various studies showed periods of negative nitrogen balance associated with infections, diarrheal episodes, and psychological stresses. In a study by Garza (10) where .57 g/kg/day of egg protein was given, some subjects showed a decline in TBK and urinary creatinine output and elevation in serum transaminase levels after two months of diet, while their nitrogen balance remained close to zero. The significance of these changes is not fully interpretable, but suggests that this level is insufficient.

* Some subjects had periods of negative N balance.

** Some subjects reached negative N balance in response to documented stress.

The transient negative N balances in several studies, and especially the stress-related negative N balance periods observed in the Korean study, suggest that under real-life conditions protein requirements may change episodically. These added sources of variability interact with the intrinsic variations in N utilization at the individual level.

Based on the variability of measurements of N intake, fecal nitrogen, urinary nitrogen, and the error in using a figure of 5 mg N/kg/day for skin and miscellaneous N losses, a balance of ± 10 mg/kg/day was considered equivalent to equilibrium (N balance of zero). Thus, the data presented here indicate that N equilibrium can be reached at a range of levels of protein intake. On the basis of available indices of protein nutritional status and health indices, it appears that 1.0 kg/day of a predominantly mixed vegetable protein diet, or 0.8 g of a high-quality protein are sufficient to meet long-term protein needs of most normal adults if adequate amounts of energy are provided.

3.3 Large-sample studies

Three investigators (Bourges in Mexico, Ozalp in Turkey, and Uauy in Chile) (papers 16–18) conducted studies using large numbers of subjects to examine the distribution of individuals' short-term balances on constant intakes. All were 10-day balance studies at a single level of nitrogen intake derived from local diets. For two of these studies (Mexico, Chile) the level fed was two standard deviations above the mean requirement, as estimated by previous standard protocol experiments. In the Turkish study the estimated mean requirement was used for the intake level.

The three studies show mean N balances close to zero, and similar intragroup variability (SD 12.0, 12.6 and 15.0 mg N/kg/day, respectively). (See Table XI.)

In the Chilean and Mexican studies, 97.5% of the individuals were expected to be in positive nitrogen balance, but six of 30 and five of 12 were negative, respectively -- significantly different from 2.5%. Since in these studies the protein level fed was below the customary intake, the subjects may well have been adapting, with the results reflecting the variability of adaptation rates. Alternatively, these data suggest that the distribution of individual balances for a constant intake may not be Gaussian, but may have heavier tails than would be expected. The implication of the latter would be that more than two standard deviations would be needed to estimate the safe level of intake for a population.

TABLE XI: LARGE SAMPLE EXPERIMENTS

Objective: To estimate variability of N balance with local diets fed at the level of mean intercept (from multiple level expt) + 36%.

Three Studies: Turkey (paper 17), Chile (paper 18), Mexico (paper 16).