Although some GMOs have been approved and marketed for several years, there was no body of scientific research on their impact on the biology of living organisms. This is partly because animal feeding trials are not required in the current safety approval process for GMOs in the EU or USA. Only now is a body of evidence starting to emerge from a small number of animal feeding trials into the health effects and progress in the new science of epigenetics. This indicates that genetic engineering is much more unpredictable and risky than traditional breeding.

Animal feeding trials

Recent studies have found a range of serious, unexplained effects from GM consumption:

* an Australian study of GM peas revealed immunological effects of genetic engineering with the transfer of a ‘safe’ gene to a different plant species producing allergic reactions in mice. A trial by Monsanto also indicated immunological effects with higher white blood cell levels in GM maize fed rats.

* the only long-term feeding trial (24 months, by an Italian team) found GMOs can affect key body organs, changing the cell structure and cell functioning of the liver, pancreas and testes of mice fed Roundup Ready soya. Similarly, a Monsanto trial found rats fed its GM maize Mon863 developed smaller kidneys.

* a Monsanto trial found GM consumption affects the development of the blood with fewer immature red blood cells and changes in blood chemistry in rats fed its GM maize Mon863.

* a Russian rat study found apparent generational effects of GMOs with very high death rates in the young of rats fed GM Roundup Ready soya (56% died) and stunted growth in the surviving progeny.

* a programme of UK studies funded by the Food Standards Agency found that genetic engineering routinely causes a large number of random genetic and chemical changes in GM plants, the health impacts of which are unknown.

* two UK trials, one with humans and one with sheep, found that when GMOs are eaten some of the inserted genes move out and transfer into the gut bacteria.

Additionally, past studies found GM consumption damages the gut wall and is associated with unexplained deaths of test animals:

* studies by three scientific teams of two different GM plants found GMOs have the potential to cause haemorrhage. Feeding trials by two teams found that GM potatoes cause lesions in the gut wall of rats and mice, and two US feeding trials found that GM tomatoes cause lesions in the gut wall of rats.

* at least two trials of different GMOs found unexplained deaths among the test animals, with 7 of 40 rats (17.5%) in a feeding study of GM tomatoes dying within two weeks ; and a 7% mortality rate for chickens fed GM glufosinate-tolerant Chardon LL maize (twice the rate of the non-GM fed chickens).

(It should be noted that these studies were designed to identify health impacts and include toxicological studies involving tissue analysis. These are different to the various non-toxicological feeding studies frequently referred to by the biotechnology industry, which are primarily carried out to test commercial aspects of GM feed).

The study of Epigenetics

The actual causes of these effects are not known, but many possible factors could account for them. It has long been known by scientists that the artificial insertion of the genes physically disrupts other genes through the damage caused by the uncontrolled insertion process ('positional effects'). In addition, the chemical functioning of the new gene interacts with the activity of the plants' existing genes and biochemical pathways, and so disrupts the metabolism in unpredictable ways.

However, research into the new science of "epigenetics" (meaning 'above genetics') is also now showing that genes account for only a part of the control of the biochemistry of organisms, and organisms have a level of control above genes that interact with genes. The exact details of this interaction between the rest of the organism and its genes are still far from known. However, this more complete understanding explains why genetic engineering is so unpredictable, with different results produced by each attempt and why the products are often unstable.

GA, 13.1.2006, GMbriefing19. Updated 13.4.2006