Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Methodology

1. Dermal nitrogen losses in young Japanese men
2. Patterns in urinary nitrogen experiments in long-term studies with constant protein intake
3. Human protein requirements: autocorrelation and adaptation to a lowprotein diet
4. Investigation of a short-term procedure to evaluate protein quality in adult human subjects - a preliminary report
5. Additional studies on very short-term procedures to evaluate protein quality in adult human subjects
6. Nitrogen losses in sweat induced by exercise
7. Nitrogen balances of 15 turkish young adults on a safe level of protein intake for 15 days

1. Dermal nitrogen losses in young Japanese men

Goro Inoue, Takashi Yamamoto, and Kyoichi Kishi

Department of Nutrition, School of Medicine. Tokushima University. Tokushima, Japan

About 15 years ago we measured dermal nitrogen losses of young Japanese men throughout the year. The results show that, when environmental temperature was below 20°C, skin N loss averaged about 7 mg/kg, whereas above 25°C it increased steeply, up to 20 mg/kg at about 30°C, probably as a result of higher sweat losses.

Shortly thereafter, Calloway reported her exact method for measurement and the precise figures she obtained with Caucasians in Berkeley (1). Based on her estimation, a figure of 5 mg/kg has usually been accepted as a reference figure for dermal and miscellaneous losses for adults. However, we felt that it was necessary to estimate the dermal N losses of Japanese living under different climatic conditions. Although these studies are still in progress, some of the data obtained are nonetheless of interest.

Experimental Details

The experiments were carried out with five students during the winter and with three subjects during the summer (two subjects participated in both studies - table 11.

The subjects were given an adaptation diet composed of about 45 kcal/kg of energy and about 1.2 g/kg of protein for four days, followed by three days of dermal N collection. They lived in a metabolic ward and were allowed only one hour of routine work or light sports activity per day. They were clothed in the standard fashion. and Calloway's method of sweat collection was used.

Results

In winter total skin N loss averaged 3.63 mg/kg and in summer it rose to 12.72 mg/kg. The increase of N adhering to clothes was much larger than the increase of N remaining on skin (see tables 2 and 3). As seen in table 4, analysis of N components shows that amino nitrogen contents were 0.61 mg/kg in winter and 0.68 in summer, not significantly different. In contrast, urea and ammonia N increased markedly in summer compared with winter. Presumbly, there is a compensatory reduction in urinary N excretion in summer.

TABLE 1. Characteristics of the Subjects and Environmental Condition

a. Initials indicate the individual subjects

TABLE 2. N Content in Each Fraction (mg N/kg/day)

TABLE 3. N Component in Each Fraction (mg N/kg/day)

TABLE 4 Dermal N Loss in Young Japanese Men (mg N/day)

Finally, dermal N loss clearly increases with higher temperatures in the Japanese summer. The figure of 12 7 mg/kg in summer is comparable to that found by Huang in Taiwanese men whose sweat was collected in a room with artificially controlled temperature

Reference

1. D.H. Calloway, A.C.F. Odell, and S.N. Margen "Sweat and Miscellaneous Nitrogen Losses in Human Balance Studies". J. Nutr, 101: 775-786 (1971).

2. Patterns in urinary nitrogen experiments in long-term studies with constant protein intake

William M. Rand

Department of Nutrition and Food Science, Massachusetts institute of Technology, Cambridge, Massachusetts, USA

1. Urinary nitrogen (UN) excretion was measured daily in 42 subjects who participated in five different experiments in which they received constant protein intakes (0.73-1.8 9 protein/kg body weight) for two to three months (table 1). The individual subjects were healthy, young adults, and most were male (table 2). Throughout the studies all subjects had their meals carefully supervised, otherwise they pursued normal activities, with the exception of those subjects who slept in a metabolic unit in Dr. Uauy's study.

2. Primary interest was in how an individual's U_N varied over time. The basic model is that the measured values of U_N represent the sum of an equilibrium level (u) that depends on the individual and his situation - i.e. body composition, metabolism, climate, etc. - and a quantity (e) that represents daily variability about this equilibrium - short-term stress, weather, measurement error. etc. This can be written as:

U_n = u + e

The simplest situation is that in which u is constant over time and the daily values of e are independent of one another and follow some nice statistical distribution, preferably Gaussian. We define a pattern as any deviation from this, and our formalization suggests two ways in which patterns can arise: first, the equilibrium may not be a constant but may vary with time. and, second, the daily variations about this equilibrium may not be independent of one another, but correlated. Practically, these two terms represent different aspects of nitrogen metabolism, slow and fast response, and can be considered as different points on a single continuum of response time.

3. These two effects would manifest themselves in the data in different ways.

TABLE 1. Experiments Included

3.1 The first, a varying equilibrium, would be seen by a significant fitting of the data to some function of time. While we could postulate an almost endless number of such functions (exponential equations based on compartmental models or trigonometric equations based on postulation of periodic effects), we choose to work with low-order polynomials (considering just linear, quadratic and cubic) for purely practical reasons. We are interested in describing. and eventually removing, the long-term trend, and polynomials are an easy way to start. We assume no model of "behaviour," merely using polynomials because they are "smooth" functions and easy to work with.

3.2. The second effect. lack of independence between the daily error terms, would appear as a serial or autocorrelation of the data after the trend has been removed. It is important to point out that investigation of the data must be carried out by first removing the long-term trend before autocorrelations of the data are calculated, because the trends themselves will show up as autocorrelations.

4. The first question of interest was whether our individuals were at some sort of equilibrium. We note two disquieting features in the data (table 3):

4.1. In 33 of the subjects (79 per cent) weight was significantly correlated with time-these subjects either gained or lost weight consistently throughout the entire experimental period (figures 1, 2, and 31. To see how this might affect nitrogen excretion we examined the correlations between weight and UN and found this to be significant in 13 (31 per cent) of the subjects. While we might expect this as a result of changes in "pool" sizes, for six individuals these correlations were positive and for seven they were negative, suggesting that changes in weight are not indicative of "pool" size change Further evidence of this was that 23 (55 per cent) of the individuals had significant weight changes with no corresponding changes in their UN

TABLE 2. Subject Characteristics

TABLE 3. Correlations

U_N = urinary nitrogen (9 N) on day t. t = day of study Wt = weight Cr = creatinine U_N(-1) = urinary nitrogen (g N) on day t-1

FIG. 1 Urinary N (UN), Creatinine (Cr), and Weight (Wt) of Subject 31.

FIG. 2 Urinary N (UN), Creatinine (Cr) and Weight (Wt) of Subject 18.

FIG. 3. Urinary N (UN), Creatinine (Cr). and Weight (Wt).

FIG. 4 Urinary N (UN) NO Pattern Subjects 17, 24 and 42.

FIG. 5. Urinary N (UN) Linear Trend, Subjects 18, 13, and 15.

FIG. 6. Urinary N (UN) Quadratic Trend, Subjects 7, 35, and 38.

FIG. 7. Urinary N (UN) Cubic Trend, Subjects 19, 21, and 31.

FIG. 8 Urinary N (UN) Autocorrelated, Subjects 28, 6, and 33.

4.2. Nine subjects (21 per cent) had significant correlations between their urinary creatinine excretion and time. Of these, five had significant correlations with between Cr and UN in addition to the eight others with significant correlations between Cr and UN (all positive). This suggests that creatinine excretion is related to "pool" size (although there are obvious examples where it is related to collection error). We can correct for this before examining for trend of UN with time by first regressing UN on creatinine and working with the residuals. When this was done it was found not to change the results for any individual.

5. To explore the primary question of patterns over time, the UN data for each individual was regressed on t, t?, and tl (where t is day of experiment), and the significance of the coefficients was noted (table 4).

5.1. We found that 25 (60 per cent) of the individuals had no long-term patterns (see figure 4 for three examples). Of the remaining individuals, seven, five, and five showed linear (see figure 5 for three examples), quadratic (see figure 6 for three examples), and cubic (see figure 7 for three examples) patterns, respectively. Thus, in 17 (40 per cent) we did find significant long-term trends.

5.2. In one of the 25 individuals without a long-term trend, and in three of the 17 individuals with a longterm trend, we found significant autocorrelation (the 17 were examined after removal of the trend). Thus, in four (10 per cent) we do have evidence of correlated residuals (see figure 8 for three examples of this).

Interpretation

6.1. In many individuals there is no constant equilibrium - UN slowly varies as they adapt to whatever are their long-term stresses (balance the various perturbances in their environment). There appears to be no pattern in these responses, because an individual cannot be identified as varying from either a short section of his record or his individual characteristics.

6.2. The autocorrelations arise from well-defined episodes (see fig. 8). An explanation for these episodes is that the individual is here responding to stress of some sort, and that this reponse takes several days to bring the system back into equilibrium. It is possible that this phenomenon might have been observed in more individuals if they had been studied under less controlled circumstances.

Implications

7.1. Most of these individuals either gained or lost weight consistently throughout the study period. All were on a constant diet and activity patterns were reasonably controlled. This suggests that weight control is, to a large extent, maintained by variations in intake (and somewhat by exercise) rather than modification of efficiency of utilization.

TABLE 4 Patterns

a. If first 29 days (ACF [U_N] = 0.11 ) are deleted.
SU_N = Standard deviation of urinary nitrogen
U_N Pattern-t = Type of pattern in urinary nitrogen over time: no = no pattern. yes = pattern. Iin = linear trend. quad = quadratic trend. cub = cubic trend.
R² = multiple R² faction of variability explained by the indicated pattern.
SU_N/Cr =Standard deviation of urinary nitrogen, corrected for creatinine variability
U_N/Cr ACF (1 ) = Autocorrelation of urinary nitrogen corrected for creatinine variability
U_N/Cr Pattern = Type of pattern in urinary nitrogen. corrected for creatinine
SU_N/Cr, t = Standard deviation of urinary nitrogen. corrected for creatinine and time patterns U_N/CT. t ACF (1) = Autocorrelation of urinary nitrogen corrected for creatinine and time patterns
U_N ACF (1) = Autocorrelation of urinary nitrogen without correction

7.2. It is probably not necessary to normalize nitrogen measurements by dividing by weight for adults within the weight range studied (45 to 103 kg), it is definitely not essential to divide by daily weights.

3. Human protein requirements: autocorrelation and adaptation to a lowprotein diet

Ned Durkin, Dale A. Ogar, Shobha G. Tilve, and Sheldon Margen

School of Public Health, University of California, Berkeley, California, USA

Objectives

Experimental Details

The investigators wished to test whether the model of protein deficiency described by Sukhatme and Margen (1) would predict the experimental outcome. The results are incorporated in the attached tables.

In order to test the hypothesis that individuals can make significant physiological adaptations to varying (particularly low) protein intake, a 77-day metabolic experiment was carried out on six healthy young men, ages 20 to 31. All subjects received a diet containing 57 gm N/kg (0.356 9 protein/kg) derived from egg albumin. Caloric content was adjusted during the first 17 days to maintain a constant weight. After that time the individual diets remained unchanged except for the addition of fibre where needed. Activity was programmed, and biochemical and physical performance was monitored.

Results

The significant findings are briefly summarized below.

1. By day 77, all six subjects had "adapted" or achieved balance (in a statistical stochastic sense). Some were in "slight positive balance;" others were in "slight negative balance;" but all were within the balance range. (Balance was calculated as intake minus urinary and faecal losses.)

2. All but one of the six subjects showed autocorrelations of nitrogen balance with an exponential decay over time after the steady state had been achieved. This indicates to us that, at this level of N intake, the subjects were in a steady state and within the range of adequate nitrogen intake, thus demonstrating regulation and homeostasis.

Continues

Contents - Previous - Next