This commentary refers to the paper "SCF opinions on the assessment of novel foods.
I. Recommendations concerning the scientific aspects of information necessary to support applications for placing on the market of novel foods and novel food ingredients", which has been formulated by the scientific expert committee of EU as a guideline for the application of the Novel Food regulation.

Below, we have selected and commented quotes from this document with special reference to genetically engineered Novel Foods. The comments will refer to the molecular biologist John Fagans article "Assessing the safety and nutritional quality of genetically engineered foods" (73K). (Because of the length of this article, we have below, for Your convenience, linked to separate files with excerpts from this article).

Abbreviations: SE = "Substantial Equivalence". NF = "Novel Food".

3.3 Substantial Equivalence

"If a food is found to be SE to an existing traditional food or food component, it can be treated in the same manner with respect to safety, keeping in mind that establishment of SE is not a safety or nutritional assessment in itself but an approach to compare a potential new food with its conventional counterpart" (3.3, page 6).... "SE NF are thus comparable, in terms of safety, to their conventional counterpart" (3.3, page 7).

Comment: See 3.4

3.4 Compositional analysis

...."Investigations should focus especially on the determination of the content of critical nutrients (both macro- and micronutrients) and any critical toxicants and anti-nutritional factors which might be either inherently present or process derived".

Comment to 3.3 and 3.4: The concept of substantial equivalence, or "essential equivalence" is based on the erroneous assumption that genetically engineered foods are not more harmful than their natural counterpart, provided their properties in general are the same including appearance, texture, taste, nutritional, toxicological and other biochemical qualities".

In practice, for example in the assessment of the genetically engineered RR-Soy and Ciba maize, this principle has meant that it is sufficient to compare "selected" biochemical including nutritional properties as well as to seek allergens and toxins that are known to appear in the recipient organism or in the organism from which the gene is taken (see . The formulation of the regulation is far too unspecific and seems not to clearly demand any change of the assessment as it has been practised so far.

The insertion of a foreign gene may cause metabolic disturbances resulting in the appearance of unexpected toxins or allergens (see also "How genetic engineering can create hazardous foods"). The molecular biological mechanisms behind this have been explained in detail by John Fagan (1). It is fully possible that the genetically engineered novel food may be not only "substantially equivalent" regarding selected characteristics, but exactly identical except for an unexpected toxin. So establishment of substantial equivalence in the cases of genetically engineered Novel Foods is useless and irrelevant for the assessment of food safety in the case of genetically engineered organisms. This has also been explained in the paper "The failings of the principle of substantial equivalence in regulating transgenic foods" (2) and in a report from an OECD meeting (3).

3.6 Nutritional considerations affecting toxicological testing in animals

A holistic scientific interpretation of the overall wholesomeness assessment data on a case-by-case basis can provide the acceptable justification for the use of safety factors for NF lower than for those traditionally uses in safety assessment".

Comment: Some very vague "scientific interpretation of overall assessment data" is explicitly taken as a justification for accepting a lower safety threshold for NF compared to traditional safety assessment.

So here it is clearly stated that rigorous safety testing is not generally required for Novel Foods.

In practice, for example the "testing" accepted for approval of RR soy only consisted of feeding fish, some bird species, and cattle with the soy in the same amounts as when natural soy is added to the feed as a protein source. This is a very incomplete and unreliable method for assessing the safety of a food. By calling it a "wholesomeness test", the producer seems to have realized that it cannot be called a toxicity test.

3.7 Toxicological requirements

...."three scenarios may be considered: (1) Substantial equivalence can be established to an accepted traditional food or food ingredient, in which case no further testing is needed. (2) SE can be established for a single or a few specific traits of the NF, in which case any further safety assessment should focus specifically on these traits. (3) Neither partial nor total SE can be established. In this case the wholesomeness of the whole novel food or macronutrient has to be assessed using an appropriate combined nutritional-toxicological approach."

Comment: The unreliable and vague principle of substantial equivalence is used as a justification for accepting incomplete safety assessment of foods. As explained in the comment to 3.4 above, this is definitely unjustified from a scientific standpoint in the case of genetically engineered novel foods (see also "Scope of safety testing"). If these recommendations for NF regulations are accepted, there is a definite risk that a large part of the population of Europe will be exposed to potentially serious damage from unexpected harmful substances that may very well appear in "substantially" or "essentially" equivalent genetically engineered food.

3.8 Implications of NF to human nutrition

..."Information will be needed on long term as well as on short term effects of eating the NF. The appropriate information should be derived by combined nutritional and safety post-market surveillance"...

Comment: There is no ethically acceptable justification to leave the study of short and long term effects of eating the NF to "post-market" surveillance. This is nothing less than experimenting with the whole population of consumers. In stead long term pre-market studies on large groups of volunteers for at least 1,5-2 years should be done before any such food is marketed as suggested by John Fagan (1).

3.10. Allergenic potential

..."As a general principle of assessment, the immunological reactivity of individuals who react to the traditional food counterpart should be tested in vitro and in vivo against the NF. ..."If the novel protein is expressed by genes derived from a source known to be associated with food allergy, sera of people with confirmed allergies to that source can be subjected to specific immunological tests, e g Western Blotting or RAST. If in vitro tests are negative, in vivo skin prick tests or clinically supervised double blind placebo controlled challenges in these people may be performed."...."In the present state of knowledge, the allergenicity of a novel food from a GM source should include consideration of the allergenic potential of the donor and of the recipient organism".

Comment: Note that, in genetically engineered NF, only known allergens in the recipient organism or the gene donor organism are targeted at. This neglects the definite possibility that unexpected allergens may appear because of the disturbance induced by genetic engineering (1). To arrive at a more reliable result it is necessary to test humans for allergy in a far greater scale than only some selected persons who are allergic to known allergens in the food.

So also here, the protection of the consumers will be incomplete if this recommendation is applied.

5. XIII Toxicological information on the NF

"If substantial equivalence to a traditional counterpart cannot be established, the safety assessment based on a case-by-case evaluation must consider the following elements: consideration of the possible toxicity of the analytically identified chemical components. .....toxicity studies in vitro and in vivo including mutagenicity studies, reproduction and teratogenicity studies as well as long term feeding studies, following a tiered approach on a case-by-case basis......studies on potential allergenicity"

"....toxicological...programme should include at least 90 day feeding in a rodent species, whereby special attention is paid to the choice of doses and the avoidance of problems of nutritional imbalance..." "....in vitro mutagenicity studies will need to cover the usual major endpoints". "...There may be cases where feeding studies in a second species and an effect on the composition of the intestinal flora may be needed. Also chronic toxicity/carcinogenecity studies may be necessary. The allergenic potential also needs to be investigated"

Comment: Even the above described much higher demands for safety assessment in the case of absence of "substantial equivalence" are unsatisfactory. They are vague and imprecise and insufficient for providing satisfactory safety.

It is unacceptable that only short term animal studies are required for toxicological assessment. Low level toxins may not cause measurable damage within only 90 days as recommended above. Therefore the wording:
"chronic toxicity/carcinogenecity studies may be necessary" must be changed to:
"chronic toxicity/carcinogenecity studies must be done in every case without exception".

Most importantly, no genetically engineered Novel Food can be exempted from this requirement, considering the serious inadequacy of the principle of substantial equivalence.

It is a most important fact that animal studies are insufficient for the safety assessment of toxicity/carcinogenecity, and allergenicity of a NF. There is ample experience from toxicology that animals are not reliable for safety assessment, because of metabolic and physiological differences. So it is necessary to add an additional requirement:

Safety assessment of all genetically engineered NF must always include short and long term toxicity/ carcinogenecity, and allergenicity studies on humans. The long term studies should last for at least 1,5-2 years to yield satisfactory levels of safety according to John Fagan (1). But, as pointed out by Fagan, even so there is a residual risk that some slow-acting harmful substance will remain undetected (1).  Therefore, post-market surveillance of a large test population (100.000-300.000 people)during an additional 2-3 years is necessary even after rigorous testing. For this end, mandatory labelling of all products that contain even a small proportion of GE-products, is necessary also after the post-marketing test period.

To secure satisfactory safety assessment of Novel Foods, the regulation text should consequently be as follows:

All genetically engineered Novel Foods, withou any exception, must be subjected to most rigorous safety assessment including long term testing on humans, even if they appear to be completely identical with their natural counterpart. Mandatory labelling of all foods containing GE-products is necessary to secure early discovery of hazardous substances that have remained undetected.

Anything less than the requirement above will expose the consumers to unpredictable health hazards from eating genetically engineered food.

References:

1. John B. Fagan 1996. "Assessing the safety and nutritional quality of genetically engineered foods".

2. John B. Fagan 1997. The failings of the principle of substantial equivalence in regulating transgenic foods.

3. OECD, DSTI/STP (95)18, Paris, 1995, pages 79-87.

Till den svenska hemsidan