The approval of Monsanto's genetically engineered Roundup Ready Soy as a food

An example of approval on the basis of the principle of Substantial Equivalence

In the table below, the testing accepted for approval on the basis of substantial equivalence is compared with the testing required from a scientific standpoint.

The left column presents the test results accepted as sufficient for approval by FDA in the US and by the European Union. - The test results were provided by Monsanto (not by any independent body).

The right column presents the testing procedures that are required from a scientific standpoint to minimize the risk of hazard from unexpected harmful substances.

In the end of the table you find our evaluation and conclusion


The data accepted as sufficient for approval

Requirements from a scientific standpoint and comments

Chemical evaluation of equivalence was based on selected substances in the engineered product was compared with those in the natural counterpart. (The selection was done by the manufacturer). No essential difference was found.

Comment: Biochemical comparison of "selected" substances does not exclude the possibility of important differences among substances that were not selected.

Nutritional evaluation of equivalence was based on biochemical comparison of selected substances. (The selection was done by the manufacturer). No essential difference was found.

Comment: There might have been important differences in the substances not selected. There might be unexpected substances in the GE soy that have an anti-nutritional effect. Therefore biochemical evaluation is insufficient. Long term nutritional studies on animals that monitor organ development and metabolic parameters are required.

Animal testing: Chicken, bobwhites, cattle, and fish were given GE soy as part of their food in the same proportion as natural soy has been used in the food as a protein supplement. Their growth during a 6 week period was checked and found normal. Monsanto did not call this toxicity studies, but used the term "wholesomeness studies", indicating that they themselves realized that this was not sufficient to reliably detect toxicity.  

Comment: This is a very superficial and gross test that is very different from what is required for reliable detection of harmful effects.

Required testing: Biochemical indicators such as those for liver, kidney and bone marrow function should have been measured. In the absence of such tests as well as immuno- logical, neurological, metabolic and endocrinological tests, monitoring of organ development and histological exami- nation (tissue microscopy), it is impossible to judge the safety of the tested food.

No toxicological testing was done on animals

Comment: Without toxicological testing it is impossible to detect unexpected hazardous substances. For safety, the first testing is commonly made on animals.

Required testing: Short and long term toxicity tests on mice with monitoring of all the above mentioned parameters. If this shows no indications of harmfulness, it is necessary to proceed to testing on humans.

No toxicological testing made on humans

Comment: As animals are unreliable predictors for human toxicity, it is necessary to test on humans.

Required testing: Toxicological testing on humans in four phases (based on a diagram by Dr. J. Fagan):

Phase 1: For 90 days at maximal feasible intake with close monitoring of biochemical and (ctd.) physiological parameters including a.o. tests on bloods cells, liver, kidney, blood proteins, lipids, immune system, blood pressure.
Phase 2: For 6 months at maximum dietary level with tests as in #1.
Phase 3: For 1,5-2 years at maximum dietary level. Tests as in #1. In addition, early detection of cancer and mutagenic effects, as these may not appear until after a longer time.
Phase 4: Monitored premarketing test on a population of about 2.000.000-3.000.000 people during 2-3 years with close surveillance of its health status.

Biochemical allergen markers were checked. Known allergenic proteins were analyzed and found to be equivalent in concentration. Monsanto said that they know that you can't predict whether a protein is allergenic based on it's biochemical characters but "we're going to use those criteria anyway". The test results were accepted as sufficient by the authorities for assessing the allergenicity.

Comment: Biochemical allergen tests are known to be unreliable as indicators of allergenicity. Even animal tests are not sufficient to assess human allergenicity.

Required testing: Short and long term testing on animals (first step) as well as humans (second step) is required to investigate allergenicity as genetic manipulation may give rise to unexpected new allergens.



The testing is highly insufficient for detecting unpredictable potentially hazardous substances that may appear due to the insertion of a foreign.


With the test system presented above in this column, the risk for health hazard from unexpected substances is minimized but not eliminated (for more details, see


With this superficial assessment, there is no scientific basis for declaring that the product is as safe to eat as its natural counterpart. Unless it is withdrawn for the market, labeling is necessary to facilitate early detection of unexpected harmful effects of the product or it's derivatives.


Even with this rigorous testing system, it cannot be ascertained that the product is as safe to eat as its natural counterpart. Therefore, labeling of all products containing the tested food is necessary in order to facilitate early detection of unexpected effects of the product or it's derivatives.

In June 1999, it was revealed that FDA documents show that its own scientists were against such superficial testing as that based on substantial equivalence, but the FDA officials ignored their warnings in order to "foster Biotechnology". See FDA records support the lawsuit challenging its policy  [EL]   

In 2002, research was published that reported changes in the cell nuclei of liver, pancreas and testicles of mice who had been fed with RR soy, see Liver changes.... The procedure of Substantial Equivalence does not require such analyses.

Physicians and Scientists for Responsible Application of Science and Technology (PSRAST)

This document was created at Nov 2, 1998. Last update Sept 7, 2007.

For a more details about the methods used by Monsanto in their assessment of Roundup Ready soybeans, see "The Problem with the Safety of Roundup Ready Soybeans"  [ML]  by Judy Carman.

For further details on safety assessment, see "Testing the safety of genetically engineered foods" by John Fagan at: [A] (Scientific text).

"Genetically Engineered Food - Safety Problems"
Published by PSRAST

Siteguide   Starting points   Website search   Site Map   Start page   

News   Introductory articles   Health hazards   Environmental hazards   

Global issues   Safety issues   Alternatives to GE   FAQ   

About us   What You can do   Membership   E-mail   How to sponsor us