Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Revised PAG guidelines

PAG/UNU guideline no. 6: Preclinical testing of novel sources of food
PAG/UNU guideline no. 12: The production of single-cell protein for human consumption
PAG/UNU guideline no. 15: Nutritional and safety aspects of protein sources for animal feeding


Probably no activity of the former Protein Advisory Group of the United Nations System has been more useful or influential than its issuance of Guidelines for the development of new protein sources for animal and human feeding. Guideline No. 6 on the Preclinical Testing of Novel Sources of Food has become the international standard for judging when new protein sources can be safely subjected to trials in human subjects. Guideline No. 12 was developed subsequently to find additional criteria for single-cell protein produced on petroleum hydrocarbons and intended for human consumption. Guideline No. 15 was designed to insure that new protein sources for animal feeding would not cause problems when the animals consuming them or their products were incorporated into human diets. In view of the lapse of over 10 years since these guidelines were originally issued, it was deemed appropriate to re-examine them in the light of subsequent experience and accept revisions as might be indicated.

To accomplish this, the United Nations University invited each of the three scientific unions principally concerned, the International Union of Nutritional Sciences (IUNS), the International Union of Food Scientists and Technologists (IUFoST), and the International Union of Pure and Applied Chemistry (IUPAC), to supply delegates to a small task force for the specific purpose, and several additional consultants were added. Three of these revised Guidelines appear on the following pages. The changes are minor, and it is anticipated that they will continue to be of value in guiding the development of new protein sources for animal and human feeding.

Guideline No. 7, on Clinical Evaluation of New Protein Sources, was also reviewed at the same time, but is being somewhat more extensively revised. A new version will appear in the next issue of the Bulletin. The Bulletin will continue to publish up-dated versions of the PAG Guidelines from time to time, and such new ones as may be developed under the auspices of SCN.


PAG/UNU guideline no. 6: Preclinical testing of novel sources of food

Novel foods are defined as those not previously eaten by a human population, and that cannot be considered to be minor processing variants of conventional food. This guideline is directed towards commonly used foods processed by new techniques and to raw materials not so far used as human food either directly or by inclusion in food products. Similar attention needs to be applied to new varieties of conventional foods or those that have been genetically changed. Clear examples for application of the guidelines are new foods developed by isolation from conventional sources by unusual techniques, and yeasts, bacteria, molds, or algae, i.e., the so-called single-cell proteins. It is recommended that all new foods, including proteins, be evaluated with respect to their safety for use and nutritional value before application as a human food source.

It is intended that this guideline serve as a general recommendation rather hen as a series of mandatory procedures. The guidelines have been developed in general terms to describe the categories of data that need to be provided in some cases, but not necessarily in all. Thus, processes involving the use of solvent extraction or unusual heating conditions, or the utilization of food additives in a variety of combinations, would not require a full preclinical evaluation despite possible changes in digestion, absorption, or metabolism.

The development of a protocol for a specific food material will depend upon its similarity to a conventional food, the kind of technological process applied in its preparation, and the conditions of its intended use as prepared for consumption. Prior history of safe use may be taken into account in the evaluation of a novel food proposed for general consumption, but this alone is not necessarily sufficient to preclude adequate preclinical testing by currently available, more objective, laboratory animal feeding studies.

The extent of animal testing considered necessary before undertaking trials in human subjects will depend on the degree of novelty of the food. in the event that the observations and results of a preclinical appraisal of a novel food are to be submitted to a regulatory or institutional agency as a basis for clinical trials, it Is advisable to review the proposed protocol in advance with such an agency in the interest of saving time and effort. In principle, products intended for incorporation into animal feeds may not require such extensive testing as is suggested for human foods, but Guideline 15 should be consulted for tests considered essential.

The physical and chemical identity of the industrial product should be established to be essentially the same as that of the material tested experimentally. To be truly significant, the studies should be conducted on the product as made on a production scale rather than by laboratory batch. Where this is not possible, the process used to produce the test material and its specifications should not be significantly different from the process and product to be marketed. If the organism or source material has been genetically modified, some of the preclinical and clinical studies may require repetition. However, minor variations in processing conditions need not necessitate repetition of the entire series of preclinical or clinical studies. Attention must be given to contaminants arising from extraction or refining, as well as reaction products resulting from heat processing. Substances used as lubricants or binders (e.g., in texturization) should also be considered in this connection.

When evaluating a new product, the following categories of information are likely to be needed:

1.1 Specification of the material. This is a basic requirement; it includes a description of how the product is produced, the results of chemical and microbiological analyses, and details of manufacturing specifications.

1.2 Nutritional value, as predicted from chemical composition, with particular emphasis on nutritional value, bioavailability, and digestibility obtained in viva.

1.3 Sanitation, with respect to the source of the raw material and the conditions under which it is processed. Hygiene aspects as well as potential pathogenicity, if appropriate, should be taken into account.

1.4 Toxicological safety, as predicted from information concerning methods of production, chemical and physical properties, content of micro-organisms and their metabolites. Toxicological studies in laboratory animals may be required. The extent of the preclinical testing programme should be decided for each new product, based on a consideration of its source, composition, and the nature of the process employed in its production. As examples, the species of fish used for the production of fish protein concentrate, the micro-organisms used as single-cell proteins, and the extraction systems employed in processing, may determine how much preclinical evaluation is required. Choice of procedures should be exercised with judgment based on experience. A product intended for use as a milk substitute or a weaning food will demand additional preclinical evaluation compared to products intended for use by older children or adults.

1.5 Technological and physical properties from the point of view of incorporation of the product into current fly acceptable foods, or the fabrication into new food items and procedures to be used are.

2.1 Chemical analyses to define the approximate composition of the material, including the presence of contaminants, pesticide residues, solvents, naturally occurring or adventitious toxins, specific additives, and natural components with unusual structure.

2.1.1. Proximate composition, i.e., moisture (and total solids), total nitrogen, "fat" (ether extract), ash, crude fibre, and "available carbohydrate".

2.1 2. Protein

a) The nitrogenous components should be hydrolyzed and the amino acid spectrum determined. The amino acid composition should be expressed per 16 9 N. Since lysine is the principal (although not the only) essential amino acid likely to become bound and thus unavailable as a result of heat processing, the slightly modified Carpenter method is especially useful as a quality control procedure (1).

b) The presence and amount of non-protein nitrogenous components such as glucosamines, amides and amines should be determined, particularly in the case of products derived from animal sources.

c) The content of nucleic acid should be determined in single-cell proteins.

2.1 .3. Fat
The solvent extract should be analysed for the presence and content of triglycerides, unsaponifiable lipids (steroids) and phospholipids. If the ether extract is greater than 1 per cent, the fatty acid profile should be determined; the ratio of polyunsaturated to saturated fatty acids should also be calculated. Single-cell proteins derived from petroleum hydrocarbons should be analysed for total and carcinogenic polycyclic aromatic hydrocarbons by a suitable quantitative method.

2.1.4. Ash
Ash should be analysed for its content of calcium, phosphorus, iron, iodine, alkali and alkaline earth elements, and heavy metals. Products of marine origin should also be analysed for mercury, arsenic, and fluorine. In the light of concern oyer mercury contamination of fish from lakes, streams, and marine waters, attention should be directed to the possible presence of inorganic and particularly alkyl mercury and cadmium in protein concentrates derived from fish or algae.

2.1.5. Undigestible Material
True digestibility should be defined, and the nature of the undigested material ascertained. Examples are cellulose, hemicelluloses, pectic substances, gums, lignin, and some other cell wall-associated substances. While such indigestible material may interfere with the absorption of certain nutrients, it may, on the other hand, be beneficial in a way similar to dietary fibre.

2.1.6. Vitamins
Analyses should be conducted for all of the major vitamins except those for which a low lipid content or instability under processing conditions indicates little likelihood of their presence in significant amounts.

2.1. 7. Food Additives
Food additives used should be declared and levels specified, based on quantitative analysis.

2.1.8. Processing Damage
Useful information concerning the effect of heat on the product may be obtained not only by determinations of available lysine, but also be products of the Maillard (browning) reaction. Understanding of the effect of alkali treatment of the product may be obtained by the determination of Iysinoalanine.

2. 1.9, Miscellaneous
Depending upon the nature of the raw material and the conditions employed in its production, special analyses of the product should be conducted for:

a) Solvent residues, such as polycyclic or chlorinated hydrocarbons,
b) Pesticide residues,
c) Naturally occurring toxic substances, e.g., gossypol, haemagglutinins, and marine toxins (it should be noted that there are no satisfactory non-biological tests for the latter category of substances).

2.2 Microbiological examination is necessary for viable micro-organisms, both pathogenic and non-pathogenic, aerobic and non-aerobic, vegetative, and spore-forming. In the case of single-cell protein, the microbiological tests should also indicate the taxonomic and potential pathogenic status of the organisms grown, and the sanitary nature of the fermentation or processing conditions. These proteins should contain no viable cells of the producing organisms, thus eliminating any problems of pathogenicity.

2.3 Nutritional Evaluation

Gross and available energy should be defined.

The value of a protein product in promoting growth and nitrogen retention may be obtained with young rats or other laboratory animals when fed as a sole source of protein or as a supplement to other foods. The digestibility of a protein product should be determined in viva. For a description of the nutritional evaluation of protein foods, including their energy assessment, see Reference 1, Chapter 5

2.3.1. Protein quality studies may be conducted with young rats or other laboratory mammals, to indicate the value of a protein product for promoting growth and nitrogen retention when fed as the sole source of protein and as a supplement to others foods.

2.3.2. Digestibility of the protein product is best determined in viva. Investigations to determine the rate and degree of hydrolysis by pepsin and pepsin plus trypsin in vitro may be of value to simulate conditions within the human gastrointestinal tract. Calculations based on the essential amino acid content of enzymic hydrolysates have been adapted for estimating the biological value (utilisation) of proteins.

2.4 Safety Evaluation

The identity and reproducibility of the test material must be established with that produced commercially by chemical and other relevant procedures. Safety evaluations using such test materials are based on feeding studies in rodents and other experimental mammals. It is assumed that the data will be required for regulatory approval purposes, and therefore any reports of investigations submitted must include full details and data for control as well as test groups and appropriate statistical analysis of the findings.

Brief descriptions of the observations and conclusions are not acceptable. The initial short-term studies should be followed by long-term tests, as appropriate, to define the presence of toxic substances.

Attention should be given to the following:

2.4.1. Naturally-Occurring Toxic Substances
Naturally-occurring toxic substances found in plants include carcinogens (e.g., cycad nuts, oil of sassafras), goitrogens (Brassica species), haemagglutinins now called lectins (e.g., phaseolotoxin in legumes), lathyrogens (e.g., vetch, sweet peas), cyanogenetic glycosides (cassava and certain beans and nuts), and estrogens (in seeds and leafy vegetables). Marine sources of protein, such as fish or shellfish and the algae or plankton on which they feed, may contain highly toxic substances. Naturally occurring toxic agents can be avoided either by care in the selection of the raw materials or by appropriate methods of storage, heat processing, or extraction.

2.4.2 Microbial Toxins
Raw materials subject to microbial contamination and spoilage must be examined for the presence of staphylococci and clostridial toxins. Raw materials exposed to warm, humid conditions that induce fungal growth must be examined for the possible presence of mycotoxins, such as the aflatoxins.

2.4.3. Residues
Protein concentrates that have been isolated or refined by means of solvent extraction should be analysed for the possible presence of solvent residues and any products that may be formed, particularly by the use of reactive chlorinated hydrocarbons. In the event that any such rest. dues are present, toxicological data should be available to establish safe limits. Depending upon the nature of the raw materials, the media. and the conditions of processing, analysis for the possible presence of impurities or contaminants such as solvent residues, heavy metals, fluoride, etc., should precede any toxicological feeding studies.

2.4.4. Nutritional Adequacy of Test Diet
It is essential to differentiate between nutritional inadequacy and appetite depressant effects, as well as toxicity, all of which can suppress growth. The basic diet to which test food is added should itself be nutritionally adequate for normal growth and development of the animal species employed. The extent to which the test food may contribute high levels of lipid, carbohydrate, protein, minerals, or indigestible material may create the need for adjustments to balance out these factors between the test and control diets. This applies to most proteins, especially where the amino acid content provided by the test material should be added to basal diet components in order to satisfy nutritional requirements.

Physical form of the diets is important. Experience with laboratory rodents has shown that for a short-term study, i.e., a 3-month test, either a "synthetic" type diet, based on casein and starch, or "natural" diets, based on normal food ingredientes are satisfactory. However, for long-term chronic toxicity or carcinogenicity studies, natural diets are essential.

2.4 5 Highest feasible Feeding Level of Jest food
The nutritional and physical constraints placed upon a food preclude the testing of large multiples of potential-use levels, wich is accepted practice for the safety evaluation of food additives. Nevertheless, the highest dose level practicable should be included, keeping in mind that excessive amounts of high-quality foods may depress growth and feed efficiency. If feasible, graded levels of the test material should be reflected in the experimental design, but it is not realistic to establish a doseresponse curve.

2.4.6. Choice of Animal Species
The rat is the preferred species. Mice are also used, but less is known of their nutritional requirements, and their size precludes obtaining sufficient blood or urine for examination. Among the non rodent species, beagle dogs, Rhesus monkeys, and miniature pigs have been used for short-term but not for chronic life-span studies.

2.4.7. Selection of Animals, Animal Husbandry,
Nature and Frequency of Observations

These procedures should be in compliance with Good Laboratory Practice recommendations. Technical details for conducting individual tests may be found in published international and national guidelines.

2.4.8. Nature of Studies to be Performed For screening purposes, novel food products should be subjected at least to short-term toxicological tests in one rodent and one non-rodent species. They should also be examined for evidence of components with mutagenic potential by the use of a battery of mutagenicity tests in both prokaryotic and eukaryotic systems, looking particularly for evidence of point mutations, chromosomal changes, and interference with DNA metabolism. The degree of novelty of a potentially important food item (both source and method of production) should determine the need for definitive studies. Among these, reproduction and lactation studies are important. While they may not be indicated in the case of a fish or cereal protein concentrate, such tests should be included in a protocol for safety evaluation of single-cell protein. These tests should extend to at least two filial generations, and could be continued with teratology and dominant lethality studies. If necessary, the F1 generation could be used for chronic toxicity and carcinogenecity studies.

2.4.9 Statistical Analyses and Interpretation of Findings
In the interpretation of the responses to toxicological tests, the statistical significance of differences of responses between test and control groups plays an important role. Hence, the size of experimental groups as well as the quantitative rating of both objective and subjective observations are particularly relevant.

However, whatever statistical probability is adopted as the basis for defining significance, the chance that a single group may deviate from the norm without actually indicating a biological aberration cannot be ignored. Judgment founded on experience of the investigator and past performance of the particular strain and colony of animals must be given due weight. Interpretation of experimental findings should take into account the quantitative relationship of the experimental test levels versus use conditions of the product under investigation, interspecies variations, the limited number and variety of observations incorporated into the safety evaluation programme, and the relative size of the test and human populations.


1. Peter L. Pellett and Vernon R. Young (Eds.), Nutritional Evaluation of Protein Foods, The United Nations University World Hunger Programme Food and Nutrition Bulletin Supplement 4, Tokyo, 1980, p. 95.


PAG/UNU guideline no. 12: The production of single-cell protein for human consumption

The following is presented to supplement PAG/UNU Guidelines - and 7 as a general guide for producers on processes for, and products from, single-cell protein (SCP). Because of the large number of potential organisms and processes, a complete description of any individual process is beyond the scope of this Guideline.


SCP is defined here as a biomass of yeasts or other fungi, bacteria, and algae to be produced and/or processed as food for human consumption. Such food in addition to protein has a significant caloric and other nutrient content, e.g., vitamins. The nutritive value of the protein component is of particular importance if the product is to be used as a protein supplement

The traditional concept of "good manufacturing practice" should apply in the send that it is currently used in the food industry. Plants and equipment should be of sanitary design, and special care should be exercised in all aspects of processing, including raw material selection, quality control, sanitation, handling, and packaging.


Various types of organisms have been considered as potential sources of SCP and appear to be capable of meeting the criteria established in this Guideline. It is preferable to select an organism of a species known not to produce pathogenic or toxic variants.

To comply with non-pathogenicity requirements, the final product should contain no living cells derived from the fermentation process. Each individual strain eventually should be evaluated under the exact condition proposed for industrial production.


3.1 Raw Materials

Potential carbon sources include carbohydratecontaining materials Molasses, various sources of starch or other polysaccharides, cellulose, whey, sulfite liquor, etc.), various classes of hydrocarbons (methane and longer-chain normal alkanes), alcohols (methanol and ethanol), organic acids, and carbon dioxide. In view of the variety of carbon substrates, attention must be directed to the composition of the media from the standpoint of the possible presence of chemical components (e.g., polycyclic aromatic hydrocarbons from petroleum) that are regarded as health hazards. In addition to the carbon substrate, other materials involved in the process, include sources of nitrogen, buffer salts, minerals, vitamins, air or oxygen, and many contain chemicals such as anti-foam agents, detergents, and flocculants. No material should be introduced into the process that cannot later be removed from the SCP product if necessary to meet safety requirements.

3.2 Process Variables

Operating variables of fermentation, such as temperature, air or oxygen supply, pH, cell growth rate, and cell concentration must be carefully controlled to ensure the required product quality and uniformity. The operations following cell production should also be monitored to ensure the required product quality and uniformity. In general, this will necessitate careful and continuous monitoring throughout the course of production and an understanding of the influence of these variables on the composition of the product.

3.3 Quality Control

In addition to the quality control procedures associated with factors listed under point 5 (nutritional evaluation), careful attention should be paid to the maintenance of the integrity of the original strain of the organism. An approvers ate series of microbiological and biochemical tests should be worked out both to demonstrate the stability of the organism and the absence of undesirable contaminants.


The procedure for safety evaluation is outlined in PAG/UNU Revised Guideline No. 6, "Preclinical Testing of Novel Sources of Food." Following this testing, thorough clinical evaluation is particularly necessary because of effects such as allergenicity, uricogenesis, subjective responses, and other reactions that can be determinend only in humans. These tests area outlined in PAG/UNU Revised Guideline No. 7, "Human Testing of Novel Foods.".


The nutrition quality of the SCP is to be determined for protein and for other nutrients, as recommended in the above-mentioned PAG/UNU Guidelines.


6.1 Protein

Crude protein is total N x 6.25. Total N includes non-protein N and non-amino acid N. such as nucleic acid, urea, amines, and ammonia. The N conversion factor for true protein depends upon amino acid composition. Because of these variables, true protein should be determined by amino acid analysis. As an approximation, corrected protein N can be calculated by subtracting 1.4 x purine N from crude protein N.

6.2 Nucleic Acids

There is a limitation to the amount of nucleic acids that should be introduced into human diets, because nucleic acid purines are excreted by man principally as uric acid. In susceptible individuals elevated levels of serum uric acid increase the risk of gout, and increased urinary concentration of uric acid may result in the for motion of uric acid calculi (1).

The currently available information (2,3) suggests that there should be not more than two grams of nucleic acid per day introduced into the diet by SCP for adults, and correspondingly less for children, depending on their weight. Because of variations in the nucleic acid content of the biomass of different species, the nucleic acid content can be most usefully expressed as per cent of product

6.3 Lipids

Total lipids should be analysed. If lipid content is greater than one per cent, the fatty acid profile should be determined. Triglycerides, phospholipids, and steroid values need to be expressed.

6.4 Ash

This should be analysed for minerals of biological importance such as iron, iodine, alkali, alkaline, earth, and heavy metals. The bioavailability of the minerals of nutritional importance can be determined only by measurements in human beings.

6.5 Dietary Fibre

Non digestible materials should be analysed and taken into account in the determination of energy value of SCP. The effect of any such materials on the availability of essential nutrients from the remainder of the diet should be determined.

TABLE 1. Suggested Limits for Viable and Contaminating Micro organisms

Micro-organism Number per Gram
Viable bacteria 100,000
Viable yeasts and moulds 100
Enterobacteriaceae 10
Salmonella 1 per 50g
Staphylococcus aureus 1
Clostridia, total 1,000
Clostridia perfringens 100
Lancefield Group D Streptococci 10,000

6.6 Solvent Residues

These should not exceed those compatible with good manufacturing practice. Methods of analysis are given in the 14th Report of Joint FAO/ WHO Expert Committee on Food Additives,

FAO Nutrition Meetings Report Series No. 48, and accompanying monographs, including those for determining residues in foods.


The microbial standards should comply with Table 1.


1. PAG Ad Hoc Working Group Meeting on Clinical Evaluation and Acceptable Nucleic Acid -levels of SOP for Human Consumption, Geneva, February 1975.

2. J.C. Edozien, "Yeast for Human Feeding; New Data on Safety." FAO/WHO/UNICEF Protein Advisory Group, United Nations, PAG Document 2.23/1 (United Nations, New York, 1969)

3. D. H. Calloway, "Safety of Single-Cell Protein," FAO/WHO/ UNICEF Protein Advisory Group, United Nations, PAG Document 2.23/2 (United Nations, New York, 1969).


Contents - Previous - Next