Contents - Previous

Appendix: Criteria for valid nitrogen balance measurement of protein requirements

Long-term studies
Analysis of multilevel short-term balance studies

NS Scrimshaw

UNU Food and Nutrition Program, Charles Street Station, P.O. Box 500, Boston, MA 02114-
0500, USA

Descriptors: nitrogen balance, protein requirements, methodology

Estimates of human protein requirements are almost wholly dependent on nitrogen balance measurements. This was also true for the determination of amino acid requirements until the recent development of stable isotope methodology. Because there are a number of inherent sources of error in nitrogen balance measurements and a number of requirements that are not always observed, the methodology has been much criticized. Nitrogen balance data can be reliable and consistent or of doubtful validity and consistency depending on how they are obtained. The following precautions are essential:

1. Caloric intake must match caloric requirement

Nitrogen balance is very sensitive to changes in dietary energy intake above and below that required for zero energy balance. Therefore, estimates of normal protein needs from nitrogen balance are valid only when energy intakes closely approximate those when the individual is consuming his usual diet while maintaining his usual level of activity. Many past studies are invalid because energy levels were too high or too low.

2. An appropriate stabilization period must precede each measurement

Adaptation to a new protein level is relatively rapid and depends on the magnitude of the change. For subjects who are protein replete, a five-day adaptation period has been suggested as sufficient for determining incremental changes in N-balance. For randomized, multilevel studies in adults an initial one-day nitrogen-free period helps to ensure the adequacy of the subsequent adaptation period.

3. Periods on the same diets must be long enough to establish a stable response to the dietary change

Experience has shown that consistent results are obtained using two five-day or three three-day periods per level of intake.

4. Degree of depletion of the subject must be taken into account

Chronically depleted subjects utilize protein more efficiently than those who are normally replete and hence give an overestimate of normal nitrogen retention. They are inappropriate for determining normal protein requirements.

5. No infections, even of seemingly mild degree, can be present

Infections, no matter how mild, increase catabolic nitrogen losses and divert protein for the synthesis of immune proteins (Scrimshaw et al, 1991). It is also virtually impossible to maintain the specific protein intake because anorexia is an early characteristic of acute infections, even when they are subclinical.

6. Variations in daily physical activity must be minimal

Because physical activity is such a large component of energy requirements, it is impossible to maintain an appropriate energy intake if there are large daily variations in activity.

7. Fluid intake must be controlled

A large fluid intake increases urinary nitrogen retention so that variability in fluid intake causes variability in nitrogen balance results.

8. Correction must be made for integumental and miscellaneous losses

Nitrogen balance studies should be conducted under conditions of minimal sweat loss or, if this is not possible, sweat losses should be measured and taken into consideration. Studies of the same individuals showed higher integumental N losses and correspondingly lower urinary N losses in summer than in winter (Huang et al, 1972).

9. Subjects should not be anxious or otherwise disturbed

Psychological stressors cause metabolic responses that are qualitatively similar to those observed with infections.

10. Protein intake be precisely monitored

Misleading results have been obtained in studies where, unknown to the investigators, the subjects consumed additional food or failed to consume some of the food prescribed.

11. Urine collections must be precisely timed and complete

Incomplete urine collections or varying duration of daily collections can introduce large errors into N balance studies. Twenty-four hour urinary creatinine excretion should be monitored as one indicator of the quality of the urine collections.

12. Faecal collections must be complete and very well homogenized before a sample is taken for analysis

A common and serious error is taking a faecal aliquot for analysis without adequate homogenization of the total sample.

13. Determinations of nitrogen in food, urine and faecal collections must be accurate

Nitrogen should be directly determined by analysis and not estimated from food composition data. Inaccurate analyses of the N content of samples is a common source of error in balance studies.

14. The experimental design must be appropriate for the intended purpose

Correspondence to: NS Scrimshaw.

To determine the amount of protein in a diet required for nitrogen balance in adults, a standardized protocol has been successfully applied in multicenter studies sponsored by the United Nations University, the results of which were the basis for the 1985 revised recommendations for adult protein requiremens (Rand et al, 1984). This protocol was devised by an expert working group (UNU/WHP, 1979) and can be summarized as follows.

Taking into consideration all of the above requirements, protein should be given at a minimum of four levels, the highest near the anticipated mean requirement. The 1979 working group recommended levels of 0.03, 0.04, ().05 and 0.06 g/kg for adults and appropriately higher levels for children, offered in a Latin-square protocol. They proposed:

For adults, the difficulties in obtaining accurate estimates of faecal output, and in consideration of adaptation to reduced nitrogen intakes, make desirable, for each test protein, one day on a nitrogen-free diet, five days for adaptation to the test level, and then five days for the urine collection. The collection period should then be followed by a break period of three or more days. Faecal collections are best pooled for the last eight days at each test protein level.'

In measuring the nitrogen-balance response of children, each protein level should be fed for seven days, and estimated nitrogen balance should be calculated on the basis of the last three days, the protein-free day should be omitted, and a three-day break period interposed between the randomized test diets. Children should receive, during these three days, the same diet that was given for the three days prior to the beginning of the study.'

Long-term studies

Conclusions regarding protein needs derived from short-term balance studies should be confirmed by feeding at the proposed safe level for three to six months. It is not practical to conduct the number of long-term studies that would be required to include the many variations in age, sex, physiological state, and genetic and environmental factors. However, several excellent studies established the adequacy of the level of 0.8 g/kg of protein predicted from the mean and confidence limits of the short-term studies (Rand et al, 1984).

Analysis of multilevel short-term balance studies

Linearity and significance of the nitrogen balance response to increasing levels of nitrogen intake should be verified for each individual's response. For subjects not meeting at least one of these criteria their values should be eliminated from the pooled data and reported separately, with every effort made to repeat the study with such subjects to determine whether the anomalous results were due to experimental error or to some metabolic difference. A regression line, the mean, and 97.5% confidence intervals for the zero-balance intercept should be calculated using the pooled data.


Huang PC, Chong HE & Rand W (1972): Obligatory urinary and faecal nitrogen losses in young Chinese men. J. Nutr. 102, 16051613.

Rand WM, Uauy R & Scrimshaw NS (1984): Protein energy requirement studies in developing countries: results of international research. Fd Nutr. Bull. Suppl. 10.

Scrimshaw NS, Bistrian BR, Brunser O. Elia M, Jackson AA, Jian Z-M, Kinney JM, Rosenberg IH & Wolfe RR (1991): Effect of disease on desirable protein/energy ratios. In Protein-energy interaction, eds NS Scrimshaw & B Schürch pp 385-398. Lausanne: IDECG.

UNU/WHP (1979): Protein energy requirements under conditions prevailing in developing countries: current knowledge and research needs. Fd Nutr. Bull. Suppl. 1.

European Journal of Clinical Nutrition
Directions to Contributors

Manuscripts should be sent to either:

Professor JS Garrow
Dr JC Seidell
The Dial House
Head Department of Chronic Disease and Environmental Epidemiology
93 Uxbridge Road
Antonie van Leeuwenhoeklaan 9 PO Box 1
3720 BA Bilthoven The Netherlands

Send original plus two copies of paper and illustrations: type preferably on A4 paper (210 × 298 mm) single-sided and double-spaced with 25 mm margins. A covering letter, which should be signed by all authors, must state that the material has not been submitted for publication elsewhere while under consideration for EJCN. Identify the name, full postal address, ad fax number, of corresponding author. Typescripts will not normally be returned to authors.

The layout of the paper should be as follows:

1. Title page, giving a concise but informative title, and the names and affiliations (but not degrees) of all authors.
2. Structured abstract of not more than 250 words, under the headings:

Objective - what was the main question or hypothesis tested ?
Design - was it randomized, case-control, retrospective, etc.?
Setting - where was the study done? primary or tertiary care?
Subjects - how recruited, how many started, how many dropped out?
Interventions - methods used, and duration of administration.

Main outcome measures:
Results - indicate 95% condifence intervals and exact P value for effects.
Conclusions - answer (significant or not) to main question.
Sponsorship - University? charity? commercial organization?
Descriptors - preferably use up to 6 words, chosen from MEDLINE MeSH, which best describe your paper. Theses will be used for indexing your paper in the EJCN, and also help retrieval from computer databases.

NOTE: Further advice on preparation of structured abstracts can be obtained from Haynes RB et al. Annals of Internal Medicine 1990; 113: 69-76. Failure to provide a structured abstract will delay editorial processing of the paper.
3. Text: organized as Introduction, Methods, Results, Discussion, References. Number pages consecutively. Indicate where tables and figures should appear. References in text are indicated in the text by name and date, ad listed at the end of the paper in alphabetical order of first author. See recent issues of the EJCN for citation style.
4. Tables and Figures with legends which enable them to be understood withour reference to the text should be on separate sheets at the end of the text. Artwork should be suitable for direct photographic reproduction.

Units and abbreviations. Use Sl units where possible. Energy should be given as kJ or MJ (with equivalent kcal or Mcal in parentheses).

Statistical methods. For normally distributed data, mean (SD) is the preferred summary statistic. Relative risks should be expressed as odds ratios with 95% confidence interval. To compare two methods for measuring a variable the method of Bland & Altman (1986) Lancet i, 307 should be used: for this, calculation of P only is not appropriate.

Ethical approval. Research studies involving human subjecsts should e approved by the relevant review committee, and such approval must be stated in the Methods section of the paper.

Dietary survey methods. See guidelines published in EJCN 1993: 47: 379-380 or apply to Editorial office for an offprint of these guidelines.

Publication. Proofs will be sent to the corresponding author for marking of printers' errors, together with an order for offprints and reprints.

Funding for this l/D/E/C/G workshop was provided by the United Nations University and the Nestle Foundation.
This publication is available free of charge from
The Secretariat of IDECG c/o Nestlé Foundation P. O. Box 581 1001 Lausanne Switzerland

thanks to a grant from the Nestlé Foundation.

Contents - Previous