GM-free Scotland

A self-selected panel of reviewers who examined a very disturbing report of a study on rats whose mothers had been fed GM soya made some very valid comments (see IRINA ERMAKOVA AND A CASE OF TABLOID SCIENCE?– News, April 2008). However, their criticisms about the adequacy of the experimental design are, unfortunately not unusual in the infant science of GM safety.

Russian scientist, Irina Ermakova, carried out several experiments comparing the effects of a maternal diet containing GM herbicide-tolerant soya on rat offspring. She found that the pups had low survival rates and stunted growth. The GM soya tested has been in commercial production since 1996 and has been in our food chain ever since.

The scientists who volunteered themselves to criticise Ermakova's study pointed out extensive problems with the experimental soy feed used. The GM soya came from Archer Daniel Midland, a major bulk grain trader in Europe. It could not have been identity-preserved, and would most likely have been a mixture of GM and non-GM crops in unknown proportions. The control, non-GM, soya would be also have had a mixed composition. In order to see clear evidence of differences produced by genetic transformation, the GM crop must be compared with a control which is as near-identical as possible except for the GM event. This necessitates, for example, growing the test and control crops side-by-side to reduce differences due to environmental effects, ensuring the two strains are genetically as similar as possible, and that they are providing as near-identical food quality, energy and anti-nutrients as possible.

Much was made by the commentators of Ermakova's departures from 'standard practice' and her failure to use standard protocols to make studies acceptable to the scientific community and regulatory agencies.

The points made are important, but 'standard practice', 'standard protocols' and correctly sourced control material are not things which have ever featured largely in GM safety testing.

Access to GM and related non-GM seed is controlled by the biotech industry: “it is literally illegal to undertake a feeding study without industry consent “ (Dr. Ann Clark literature review). A 2008 analysis of 22 papers purporting to demonstrate substantial equivalence in commercial livestock using GM feeding studies, found no consistencies in the strain or source of the comparator non--GM seed used, nor in the validation of the genetic purity, nor in the nutritional balance. The test GM feed ranged from a parental strain to just something non-transgenic, and might be grown side-by--side with the GM crop, or, in another state, or even in another year. The two might have the same protein content, or the same energy content or, in one case, have had sand and cardboard added to bring the quality of the comparator down to the level of the GM feed.

Regulators have been accepting the above mishmash of experimental design for years, and the scientific community, beyond Ermakova's reviewers, hasn't been doing much complaining.

Whether Ermakova's control chow provided a good comparison or not, something in the GM soya-supplemented diet was causing an unmistakable stunting of the pups' growth and facilitating their early death. It may not have been anything to do with the artificial DNA in the chow, but the soya used wasn't a pure material bred for the laboratory, it was from Archer Daniel Midland, and was the same one you may be eating.

Ermakova's reviewers felt that the rats showed signs of poor husbandry, evidenced by a higher than expected mortality and poor performance in the control-fed animals. An experiment on animals with ill-health throws in an extra complication and is difficult to interpret, because clear evidence of any ill-effects from the GM food would be muddied by the underlying sickness. However, where the food has toxic qualities, a sick or stressed animal will tend to succumb where a healthy animal can shrug it off. When you are actively searching for adverse effects, a sick animal is a more sensitive subject: when you want to make sure you only find adverse effects so bad they can't be ignored, you test a healthy animal. How good is an optimally-fed rat kept in a quiet, safe cage as a model for a stressed-out, chronically sick human being?

Much is made in the criticism levied at Ermakova's experimental design that it contradicts other studies. A closer look at the few studies quoted isn't quite so convincing: none examines the effects on rats of a GM diet during conception, pregnancy and lactation; results found in an experiment on mice (a different animal) during pregnancy and lactation (a different time span) gives supportive mammalian/rodent evidence, but is not contradictory; experiments on growth rates in young rats are measuring something much more limited; tests on cows, chicken and catfish are irrelevant to both rats and humans.

At the end of the day, the lack of adverse effects noted previously from feeding GM soya may be just another symptom of the “exceptionally consistent” finding that GM and non-GM feed appear not to differ in any measured parameter (Ann Clark literature review). More sinisterly, the GM soya used by Ermakova (and eaten by you) may be quite different from anything used in previous tests: unexpected new variants of the GM crop may be popping up in fields whenever the transgenes decide it's time for another move (see DANGERS OF DNA TRANSFER ARE UNKNOWN – News March, 2008).

SOURCES:

• Andrew Marshall, GM soybeans and health safety – a controversy reexamined, Nature Biotechnology, 25:9 September 2007
• Dr. Ann Clark, GM Feeding Trials – systematic scientific fraud?, February 2008, forwarded by GM Free Cymru