Science: Pusztai: the verdict

GM food: safe or unsafe? First we must ask experts in this field how they rate the research work that led to this most recent controversy, says Steve Connor

The case of Arpad Pusztai - the scientist forced to retire over his public comments about genetically modified (GM) potatoes - has become a cause celebre with the environmentalists. He claims to have shown that GM food can stunt the growth of laboratory rats, harm brain development and damage the immune system. If he is right, it represents a hammer-blow to the biotechnology industry, which is keen to exploit advances in genetics. If he is wrong, Dr Pusztai could be accused of whipping up public hysteria.

Last week, a group of 20 scientists signed a memo in support of Dr Pusztai, stating that the Hungarian-born researcher stands fully vindicated. Dr Pusztai's data from experiments he conducted at the Rowett Research Institute in Aberdeen clearly show, the memo says, that when laboratory rats were fed GM potatoes, their internal organs failed to grow fully and their immune systems were suppressed. They concluded that Dr Pusztai's research report would be acceptable for publication in a peer-reviewed scientific journal.

It is this last claim that has been put to the test by The Independent. Dr Pusztai's report became publicly available this week, and we asked one of Britain's leading experts, Professor Tom Sanders of King's College London, to comment on whether it would pass muster with genuine scientific referees. Professor Sanders concludes (see below) that Pusztai's work fails to reach a standard acceptable for a peer-reviewed journal.

Science is only science if it has passed through review by experts before being published in a journal. The 20 "experts", mostly from abroad, who signed the memo say Pusztai's research is of high quality. This is not the view of other scientists who saw it.

The story began in 1995, when the Scottish Office funded a three-year project involving three centres: the Rowett, Durham University and the Scottish Crop Research Institute. The aim was to identify ways of making crops pest-resistant, with minimium side-effects.

Dr Pusztai's role, as an expert on toxic plant proteins called lectins, was to undertake a series of feeding experiments using GM potatoes and laboratory rats. His particular interest was a lectin called GNA, found in the bulbs of snowdrops, which acts as a natural pesticide. According to Dr Pusztai's report no other lectin-producing GM plants were used in the experiments, although he does say he performed some "analytical" work with GM potatoes expressing another, more toxic lectin, Con A, from the jackbean plant.

This is a key point because, according to an audit investigation into Dr Pusztai's work (ordered by the Rowett following his statements on TV's World in Action), he had become confused over whether he was talking about GM potatoes expressing GNA or ordinary potatoes which had Con A added to them in concentrations 5,000 times greater than would occur naturally. The audit, conducted by four scientists, two from the Rowett and two from outside, concluded that the only time the rats in the Pusztai experiment showed any signs of stunted growth was when they were fed Con A in these high concentrations.

"Therefore, the audit committee is of the opinion that the existing data do not support any suggestion that the consumption of rats of transgenic potatoes expressing GNA has an effect on the growth, organ development or immune function," the audit report stated.

In answer to this criticism, Dr Pusztai compiled his own "alternative report" in which he details what he did and the results he obtained. It is this report, which his supporters claim to be of sufficient standard for publication in a peer-reviewed journal, which we passed to Professor Sanders.

Dr Pusztai fed the rats a diet of raw, baked or boiled potatoes. Some of the potatoes, he says, were genetically modified with the GNA lectin and some had GNA added to unmodified potatoes. A diet that solely consists of potatoes is so nutritionally poor that he sometimes added a protein supplement, otherwise the experiment would breach Home Office regulations limiting the suffering of animals. Two types of feeding trial took place: one over a 10-day period, the other over 100 days. His report states there was only one 100-day experiment, where extensive protein supplements had to be used. He found the liver weights of the animals in this trial actually improved when they were fed GM potatoes, and put this down to the protein supplement. Dr Pusztai said the long-term trial was only preliminary.

It was one of the 10 day experiments - code name D242 - where Dr Pusztai claims to have observed significant effects which indicate that GM food is dangerous. According to Dr Pusztai, rats in this experiment were fed GM potatoes, enhanced with the GNA snowdrop lectin, as well as ordinary potatoes and potatoes spiked with "free" GNA. Dr Pusztai claims the growth of rats was significantly reduced on a diet of boiled potatoes, and more so on one of raw potatoes, as expected owing to the low nutritional value of potatoes compared with a high protein diet.

"However, in this instance the difference between the final body weight and empty body weight of rats (accounting for food in the gut lumen) which were fed raw, transgenic potato diets as significantly higher than that of rats given diets containing the raw parent line. This again indicated that digestion and absorption of nutrients of transgenic potato diets was retarded in comparison with ordinary potato diets," Dr Pusztai writes.

A test of the rats' immune systems during this experiment also indicated that the animals fed transgenic potatoes were almost always more suppressed. Dr Pusztai claims that when free GNA was added to a diet of unmodified potatoes, he did not see this suppression. In other words, there was something about the act of genetic modification itself that has led to the effect he has observed. Environmentalists jumped on this as evidence that all GM food is unsafe.

Dr Pusztai's conclusions were unambiguous: "Four feeding trials were carried out ... In all four experiments, feeding transgenic potatoes to rats induced major and in most instances highly significant changes in the weights of some or most of their vital organs ... The growth rate of rats fed potato diets was slightly but significantly less than that of rats fed a high-quality control diet, but the presence of GNA, whether added to potato-based diets or expressed in the transgenic [plant] had no significant effect on weight gain and weight change compared to parent potato lines."

Dr Pusztai's two outside collaborators at the Scottish Crop Research Institute and the University of Durham have distanced themselves from his conclusions. John Gatehouse, at Durham, is understood to be privately furious at Dr Pusztai's failure "to consider the most elementary tenet of science - that before one reaches a conclusion about cause and effect, it is necessary to demonstrate that causality exists".

Meanwhile, the four members of the original audit committee have looked at Dr Pusztai's alternative report and found that it is not convincing.

Dr Pusztai may have convinced the 20 scientists who signed the memo, and certain sections of the media, but he has failed to win over the experts - including his own colleagues - who are closest to the research.

Tom Sanders, professor of nutrition at King's College London, is one of the most distinguished food toxicologists in Britain. He is a member of the Government's Advisory Committee on Novel Foods and Processes, and has acted as an expert reviewer for some of the leading scientific journals in the field. This is an edited version of his peer review of Dr Arpad Pusztai's "alternative report", which environmentalists have used to support a ban on genetically modified food.

"THIS DOCUMENT reports four feeding studies with transgenic potatoes. The document has not been carefully prepared and is not up to the standard required for publication in a good scientific journal. The tables are not clear, and the captions do not make it clear whether the results are expressed as mean with SEM or SD [two statistical ways of expressing the possible error in calculating an average].

"The food intakes are not adequately described - this data is essential for interpretation of the data. The dietary design of the first three studies is fundamentally flawed, as the diets did not contain adequate amounts of protein and the intake of nutrients and antinutrients differed between the transgenic and control animals.

"An intake of 100g protein/kg diet is regarded as the minimum amount for growing rats, and results in some degree of malnutrition, particularly when under stress such as pregnancy. Most commercial rat diets contain protein at 200g/kg diet to support normal growth. The first three studies use between 55.6 and 72.5g protein/kg. This level of protein was not adequate to sustain normal growth and development in the rats. The transgenic potatoes contained 20 per cent less protein than the parent variety but it appears no attempt was made to ensure that the protein content of the diets was similar. Thus differences between the parent variety and the transgenic animals could be attributed to differences in protein intake.

"It is unclear why a diet of raw potatoes was used, given that they are renowned for containing high levels of natural toxins. The study generalises conclusions made from the use of raw potatoes to the use cooked potatoes.

"It is well documented that protein malnutrition in rats leads to decreased growth rate, changes in gut morphology and hepatic atrophy features that were observed in these studies. The report gives the impression that these dietary aspects of the study were considered post hoc. Why were the potatoes not analysed before the diets were formulated? And why was the lower protein content in the transgenic line not compensated for by additional protein?"

Professor Sanders casts doubt on the statistical analysis that led Dr Pusztai to conclude that the brains of rats fed GM potatoes were affected by their diet. "It is odd that differences in brain weight were reported, as brain weight is generally not influenced by diet in adult rats," he says. "The statistical analyses is muddled. The tables do not indicate the number of animals in each group is not given. ... Many of the reported differences are not significant.

"I would not recommend this paper be accepted for publication in its current form. In my experience as an editor and reviewer it would be rejected by the British Journal of Nutrition, Journal of Nutrition and American Journal of Clinical Nutrition. Although the report is of poor quality, the subject matter is novel so I would not to reject it outright even though the first three studies are fundamentally flawed in their design (deficient in protein). I would invite the author to respond to detailed criticisms and consider a revised version with more detail and suggest further studies, particularly with regard to the lymphocyte proliferation studies [a test of the immune system], which are unconvincing."