Contents - Previous - Next

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

PAG/UNU guideline no. 15: Nutritional and safety aspects of protein sources for animal feeding


The need to expand the world supply of protein for human food and animal feed has been well documented. The basic problems are:

a) Increasing worldwide demand for protein.
b) An immediate need to upgrade proteins in many areas of the world.
c) The demand created by both agriculture and industry to modify existing protein supplies and produce new sources and forms of protein.
d) Demands by many governments to evolve objective regulations controlling the quality and safety of modified and novel protein sources, such regulations to be capable of harmonization at the international level to the greatest possible extent.
e) The need to allow unrestricted and unimpeded international export and import of such products, which will require international similarity of national regulations.

These developments have several implications for the future. They imply a possible revolution in animal and human feeding, as people will consume meat, milk and eggs from animals receiving new forms of protein in their feed rations. Before very long, increasing numbers of human beings may receive such protein foods as direct components of their diet. Thus, there is a need to ensure that the new proteins will have a beneficial effect on the nutrition and health of intermediate and final consumers.

After these Guidelines were initially written, a series of single-cell proteins were evaluated using the criteria as originally defined. Their safety, utility and efficiency have been clearly demonstrated. The products have been manufactured and no hazards were detected in the original experimental work or in subsequent commercialisation of these products.

Experience has shown that it is necessary to establish acceptable criteria for determining the nutritive value and safety of protein materials in relation to their use in the rations of the particular animal species con" corned, and to consider also the public health implications resulting from the use of the same animals for human feeding.

The PAG/UNU has weíl-developed Guidelines for the preclinical testing of novel food in laboratory animals (PAG/UNU Guideline No. 6), and for their clinical evaluation in human subjects (PAG/UNU Revised Guideline No. 7).. These Guidelines were stimulated originally by the potential direct use by human beings of microbial biomass grown, in some cases, on unconventional substrates such as hydrocarbons and simple alcohols.

The concepts expressed in PAG/UNU Revised Guideline No. 6 are also relevant to an animal feeding situation, but the need to use laboratory animals as models for testing new food components is removed because of the ability to test directly in the "target species" -the domestic and farm animals for which they are intended. Such experiments are of equal, if not greater, importance than those obtained from conventional experimental animals. However, this does not imply that the latter, which are the biological tools of the toxicologist and nutritionist, cannot provide information relative to the consumption of animal products by humans.


The question arises as to the necessity or usefulness of establishing guidelines for testing protein sources proposed for animal feeds in addition to those already established for precíinical evaluation of novel foods for human consumption. The guidelines have been useful to industrial producers of single-cell -proteins, e.g., selected yeasts, bacteria, etc., as the companies required an internationally and mutually agreed upon standard by which both they and governments could gauge the quality, uniformity, nutritive value, and safety of their products. The need to maintain this is still apparent, as is the need to evaluate more conventional feed protein sources produced with the aid of extensive processing systems.

The material to be evaluated initially cannot be derived from full-scale plant production; testing and product acceptance must therefore involve products from pilot plants. Confirmation that the pilot-scale product is similar to the final product is also required.

Many factors must be considered for prior approval of protein sources, including the quality of the raw materials, the acceptability of the critical processing techniques affecting nutritive value, safety, purity, and uniformity of the ultimate commercial output compared with the pilot plant product previously evaluated and approved.

In the case of SCP products, the genetic stability of the micro-organism subjected to continuous fermentation must also be considered. This requires careful development of physical, chemical, and microbiological specifications as well as nutritional and toxicological measurements appropriate for the target species.

The extent to which it is necessary to repeat toxicological and nutritional studies on micro-organisms that have been genetically engineered will depend on the nature of specificity of the DNA interchange between the vector and the host and an understanding of the relative safety of the original vector donor.


As stated in PAG/UNU Revised Guideline No. 6, it must be understood that a published guideline does not constitute a compendium of mandatory tests that, unless completely and successfully satisfied, would preclude acceptance of any new food component. Indeed, with the accumulation of experience in routine testing of these products and their feeding to livestock under practical conditions, it seems likely that more informative and possibly more routine test procedures will emerge that will permit improved testing efficiency and economy. It is recognized that toxicologists and nutritionists experienced in these matters should have the prerogative to decide on the extent and design of experimental protocols.

For the guidance of those who may lack the necessary resources in these highly specialised disciplines, it is considered useful to suggest procedures to support acceptable conclusions as to both functional utility and safety of protein sources. Many aspects of the criteria described in PAG/UNU Revised Guideline No, 6 are applicable to proteins intended for animal feeding; particularly relevant are the sections dealing with chemical, physical, and biological evaluations.

However, the criteria for the prediction of protein quality of foods used in human nutrition, e.g., bioassays for protein efficiency ratio, net protein utilisation and biological value, while of interest, are not directly applicable to animal nutrition. This explains the importance of establishing the quality and safety of feed ingredients for nourishment and productivity directly in the target species. Products intended for incorporation into animal feeds may not require as extensive testing as that suggested for human foods, but foods derived from animal sources must be considered from the viewpoint of the possible presence of residues in meat, milk or eggs transmitted from and" mat feeds.


When reviewing the elements of this Guideline, reference should first be made to the introductory pages of PAG/UNU Revised Guideline No. 6, particularly those dealing with the categories of information needed (1.1-1.5) and evaluation procedures (2.1- 2.4). With respect to ingredients intended for animal feed purposes, the Guideline covers aspects specifically applicable to the feed requirements of the species concerned as well as aspects applicable to foods derived from these animals and intended for human use. With respect to the species involved, the Guideline provides for the evaluation of the quality and efficiency of the protein for maintenance, growth and reproduction, its safety for the target species, and its effect on productivity from the economic standpoint. With respect to foods derived from these animals, consideration of acceptability in terms of flavour, colour, texture, etc., would have been taken into account before marketing any such product. Safety for humans would involve the possibility of contaminants or residues arising from processing or source materials, e.g., substrate or media employed in the case of SCP. In general, protein molecules themselves would be unlikely to present a hazard. The quality and safety of food from animals fed unconventional protein should be predictable from the nutritional and toxicological studies of the protein in the animals fed.

Because of the impracticality of identifying, by means of animal feeding studies, the minute traces of residues or contaminants that might be transmitted from the feed into meat, milk or eggs, detection of any such substances that might be suspected to be present must depend on highly sensitive chemical analytical procedures. The feeding of meat/fat from feed animals to laboratory animals to assess toxicity has been suggested as a means of gaining an insight into the residues of animal products. Experience has shown that, although such protocols might identify very potent toxic or pharmacologically active substances, in all other cases the dilution factors involved in the passage of these supplements through the food animals are so vast as to make these kinds of study relatively meaningless.

In view of the fact that some types of protein sources may be produced in the form of microbial biomass under industrially controlled conditions using agricultural or industrial waste, including both human and animal sewage, special attention must be directed toward ensuring the absence of pathogenic organisms and metabolites (PAG/UNU Guideline No. 12), as well as toxic residues of pesticides or drugs. In the case of inocula for SCP production from any sub" strafe, the micro-organisms selected should be nonpathogenic, notwithstanding the fact that the fermentation process need not involve the use of pure cultures grown aseptically.

After the processing conditions have been established to ensure uniform quality, safety, and nutritive value of the final product, chemical criteria should be determined, as suggested in Table 1.

Sensitive, quantitative analytical procedures must be used to determine the levels of any undesirable residues that could arise from the source or media from which the proteins are derived, or from solvents and other agents employed in any stage of processing. For in-plant control, the methods of examination and analysis should be capable of giving valid, reproducible results.


The nutritional value of novel proteins for animal feeding should be determined in the species for which its use is intended and as far as possible under simulated practical conditions. In experimental diets conventionally employed for evaluation in rats, the test protein is incorporated as the sole source of nitrogen. Since new protein sources would be used in animal feed as a supplement or partial replacement of soybean, fish meal, or other protein component of the total ration, test diets should reflect likely local, practical rations. Occasionally it may be necessary to effect a total protein replacement, but supra-high protein levels resulting from high test product incorporation should be avoided, as these could introduce irrelevant effects of no significance in the evaluation of safety.

To achieve the best utilization of protein sources used in animal feeds, consideration may be given to supplementation with synthetic amino acids essential for the species in question; e.g., Iysine for pigs and methionine for poultry.

In the design of experimental rations for nutritional or toxicological evaluation, care must be exercised to ensure that normal dietary requirements for the species in question are satisfied in order not to con" fuse nutritional effects with toxic effects. Dietary deficiencies or imbalance may yield apparent toxicological symptoms that are not due to the test material as such, e.g., mineral imbalance causing nephrocalcinosis in rats, or impaired growth response.

TABLE 1. Suggested Criteria for Defining Protein Sources to be Used for Animal Feeding

For All Novel Protein (Dry Weight Basis):

Total nitrogen, full amino acid profile, nucleic acid, and other non-amino components

Total lipid, fatty acid profile, glycerols, phospholipids Carbohydrates

Fibre content
Ash, inorganic components

Trace mineral components (beneficial: Co, Fe, Zn, On, Cu) (harmful: As, Hg, Pb, Cd)


Gross and available energy values

Residual substrates, processing aids

Residual harmful components (e.g., benzo(a)pyrene,* nitrosamines)

Availability of nutrients should be defined where possible

*If specific carcinogenic polynuclear aromatics are found, this might need additional teeing


In considering the test species for toxicological testing, there are two major points of interest: first, human beings will consume the animals fed the protein, and second, the health and well-being of the recipient animal. Regarding the former, there is the ideal situation of identifying all the constituents of animal products so as to be assured of their safety.

However, the technical difficulties manifest for the products under consideration make it necessary to adopt indirect means. Therefore, the types of studies that are relevant to establish preclinical safety are also suited to this context (PAG/UNU Revised Guideline No. 6). Therefore, as there is an intermediate host whose well-being can also be examined in this type of study, the classical laboratory animal studies can be less extensive unless the data obtained suggest that there is such a need. When considering the choice of animals, a rodent species (preferably the rat), and a non-rodent mammal (which may be one of the target species) are generally employed. As there is emphasis on biological quality as wall as toxicological safety, the feeding studies with the target species (e.g., chickens, pigs, fish, calves, etc.) are of considerable value.


7.1 Laboratory Animals

Reference should be made to PAG/UNU Revised Guideline No. 6, Section 2.4 for an outline of the basic method for safety evaluation of novel foods in rats. The studies should be conducted with groups of experimental animals of sufficient size to yield statistically significant results. Physical inspection of the animals, including observations of weight changes, feed conversion, gross and histopathology, haemocytology, and blood and urine chemistry should be conducted at appropriate intervals.

Reproduction studies in a suitable species are considered essential, extending for ar least two generations.

When there is evidence to suggest the need to do long-term studies to establish the absence of potential carcinogens or chronic toxicity due to cumulative effects, these should be done in at least two laboratory species

7.2 Target Species

The ultimate tests of the nutritional value for animal feeds are the practical trials in which grovvth, feed efficiency, reproduction, health, survival, and productivity are the bases for eva-luation. Short-term feeding studies are only minimal pre requisites for the introduction of protein sources. Because of the short duration of such tests relative to the total life span, it is not possible in these tests to observe evidence of long term (chronic) or cumulative effects. Thus, studies for reproductive performance and for carcinogenicity are required where substrates or processing adjuncts suggest the need. If the proposed feed ingredient is to be incorporated in the ration of breeding animals, the appropriate reproductive data should be collected. Where the target species serves for the non rodent mammal in toxicological studies, biological data should be similar to those gathered from the rat.


Information concerning certain proprietary process details may be required to provide assurance of safety and to ensure adequate quality control as well as uniformity of the final product. Examples of process details that should be reported in the case of single cell protein production are the nature and properties of the microorganism, the qualitative composition of the substrate raw materials employed, including major nutrient supplements, and the agents used for special purposes, such as defoamers, emulsion breakers, etc. General processing conditions, such as fermentation method, extraction processes, concentration procedures, drying methods, etc., should be described to the extent needed. Similar criteria for other protein sources will need to be described for each product.

Conformity of the commercial product with the registered product should be based on appropriate sampling of products in the manufacturing plant or in commercial channels. The chemical and biological testing procedures, therefore, will include general methods applicable to all protein sources and feed in general, and to methods selected, especially to relate to the registered product. Certain data to establish criteria of identity should be included in label information.

Approval of a protein source would require statements within the submission detailing nutritional quality, and all safety criteria as outlined within this document. Changes in processing that might affect these features must be reported before the changes are effected.

Chemical, nutritional, and microbiological quality characteristics referred to in this document may be found in PAG/UNU Revised Guideline Nos. 6, 7, and 12. Requirements for additional or improved analytical methods for evaluating novel proteins will be referred to appropriate committees of the International Union of Pure and Applied Chemistry (IUPAC).

Contents - Previous - Next