The Working Group on Allergenicity of the Ad Hoc Task Force on Foods Derived from Biotechnology met in Vancouver Sept. 10-12. The industry forces (led by Canada and the US, and actively supported by Australia and UK, all of whom sent high level policy representatives), dominated the meeting and pushed through extremely weak guidelines on assessment of allergenicity which go contrary to expert scientific consensus on the topic. Most of the delegations which normally fight for strong health and consumer protection either did not attend or sent only technical representatives (the EU and Norway were absent, and the representative of Belgium did not feel qualified to speak for the EU Presidency). Under the leadership of the Canadian Chair, the Working Group did not carry out the task given it by the Ad Hoc Task Force-which stated that the Working Group was to develop "detailed procedures for allergenicity assessment" and was to use a decision tree strategy based on the decision tree from the 2001 FAO/WHO Joint Expert Consultation. Instead the group decided to develop guidance using a much less rigorous "weight of the evidence" approach, based on a US position paper. The representative from the UK, Nick Tomlinson, played a key role in producing the weak guidelines, along with Canada and Australia supporting the US. Tomlinson helped to push the notion that the guidelines should be more general in scope (even though the explicit terms of reference called for developing "detailed procedures" for assessing allergenicity) and never objected to abandoning the decision tree. The draft developed at this meeting will be submitted to the full Ad Hoc Codex Task Force on Biotechnology and discussed at its next meeting in March 2002 in Japan.

Background

The Ad Hoc Task Force of Foods Derived from Biotechnology is the Codex body that is developing international Safety Assessment standards for GE/GM crops. At the first meeting of the Task Force, in Chiba in March 2000, the Task Force developed a list of scientific questions on which it needed advice and requested that FAO/WHO convene a joint expert committee to give it answers. An Expert Consultation met in May 2000 and provided a great deal of detailed guidance, most of which was accepted by working groups and became part of the draft Safety Assessment document at the 2001 meeting of the Task Force. But the Consultation noted that further work needed to be done on the allergenicity question, particularly updating of a decision tree that had been developed by International Food Biotechnology Consortium and International Life Sciences Institute (IFBC/ILSI) and initially put forward at a 1996 FAO/WHO Joint Expert Consultation. The 2000 Consultation recommended that a specific technical meeting be held on the allergenicity question. This was held in Rome in January, 2001. The report from that 2001 Consultation clearly spells out this process: "The 2000 Joint FAO/WHO Consultation recommended that 'WHO/FAO should be encouraged to convene an expert consultation on the assessment of the allergenicity of genetically modified foods and novel proteins contained therein. The Consultation should focus on development of an improved decision-tree approach for the assessment of the allergenicity of genetically modified foods and on the standardization/validation of the specific criteria, such as the optimal methods for assessment of digestive stability.' With this background the current Consultation undertook efforts to develop an improved decision-tree approach using as a start, the existing IFBC/ILSI decision tree as adapted by the 2000 FAO/WHO consultation."

The second Expert Consultation did in fact develop a detailed updated decision tree, and issued their report in February 2001. However, at the March 2001 meeting of the Ad Hoc Task Force in March, agreement was not reached on how to incorporate this advice into the draft safety assessment guidelines. Canada suggested setting up a Working Group to develop language and offered to chair it. The specific charge to the Working Group was to develop "detailed procedures for allergenicity assessment" and have it be based on the decision-tree approach (which already is explicitly referred to in the Safety Assessment Guidelines [para 39], which are at Step 5).

The Working Group meeting took place Sept. 10-12, 2001. Canada sent out a rough draft in mid-August, giving delegations relatively little time to look at it. The Canadian working draft made no mention of the decision tree approach, instead referring to a "weight of the evidence approach." Eleven countries submitted written comments on the issue (most such comments were written before receipt of the Canadian draft, but there were also a few comments, submitted as Conference Room Documents on the Canadian draft). Nine of the countries (UK, Belgium, Germany, Netherlands, Finland, Italy, Sweden, Brazil, and Thailand) explicitly supported a decision-tree approach and all said it should be based the 2001 Expert Consultation. The UK's written comments were excellent, supporting and basically consisting of a verbal version of the 2001 decision tree. Here's what the UK's written comments say on the general issue:

"The assessment of allergenic potential of foods derived from modern biotechnology involves a step-wise decision tree based strategy which focuses on the source of the gene,, the sequence homology of the newly introduced protein to known allergens, the immunochemical binding of the newly introduced protein with IgE from the blood serum of individuals with known allergies to the transferred genetic material, and the physiochemical properties of the newly introduced protein (molecular weight, sequence homology, heat and processing stability), effect of pH and/or gastric juices (digestive stability), and prevalence in food."

Only the US and Canada submitted written comments indicating that they did not support the decision tree developed by the 2001 Expert Consultation.

Opening session

The US, whose draft text ultimately served as the basis for discussion, and Canada were able to get what they wanted through a number of maneuvers. Initially, Dean Metcalfe, head of the US Nat'l Institute of Allergy and Infectious Disease (highest ranking person with allergy specialty in the US), who would have given a good strong introduction to the report of the FAO/WHO joint expert consultation and would have been strong on public health, was set to give the introductory talk, This was switched, for reasons still unknown, and Dr. Steve Taylor, who gets significant funding from industry, gave the introductory report from the expert consultation. Taylor gave a very biased talk where he tried to say that all the scientists at the expert consultation agreed that the decision tree approach wasn't really that useful and that a "weight of the evidence" approach should be taken. (A decision tree is a step by step assessment where a GMO is subject to a series of tests. If it passes A, you go to test B , etc, and only when the GMO passes all the hurdles is it deemed to have a low chance of being an allergen. In a weight of the evidence approach, regulators look at all the relevant data and testing at once, and form a judgment. This approach gives regulators much more latitude. The US and industry may want this latitude in dealing with Starlink, which would no longer be a problem if they could declare that it is not a human allergen.) Taylor's talk then proceeded to downplay any risks of allergenicity, saying that we should all agree that the first wave of GE products show no harm. When discussing the various decision trees, he stated that all were more or less based on a "weight of the evidence" approach. He stated that all the experts at the 2001 FAO/WHO Expert Consultation would agree the decision tree they developed shouldn't necessarily be used in practice! From conversations with other participants (as well as from the reports of the consultation), this is a gross distortion of what actually happened at the Expert Consultation.

During the discussion following Taylor's talk, I challenged him, reading from the 2000 and 2001 consultations, showing how they were set up to look at detailed procedures for developing allergenicity testing based on a decision tree and saying that I could see no reference to the term "weight of the evidence approach" in either of the reports. Indeed, I couldn't even find the exact terminology "weight of the evidence"in either of them. Taylor did admit that such language appeared nowhere in the report of the 2001 Consultation, but none the less said the sense of the meeting was that their decision tree wasn't really appropriate to use.

In sum, Taylor was allowed to get away with giving a highly distorted view of the expert consultation. What he presented were clearly his own thoughts that were not representative of what actually took place at the 2001 Expert Consultation. But after the "official representative" from the expert consultation, Dr. Taylor, had said that all the scientists at the Expert Consultation had agreed that a "weight of the evidence" approach was the way to go, there was much less resistance when they introduced the specific language in the guidelines themselves. This set the tone for the fundamental hijacking of the rest of the meeting.

Key Issue: Weight of evidence approach vs stepwise decision tree approach

The discussion of the Canadian draft followed. The US immediately said they were very pleased with the Canadian document but that their "thinking has changed over the last few months" and then the US put forward its modification of the Canadian draft. It was significantly weaker than the Canadian proposal, going even further towards weight of the evidence. Australia (Marion Healy) spoke up and said that we needed to discuss what level of detail should be in the document. She felt that it should be very general. [Australia had no written comments before the meeting.]

Nick Tomlinson of the UK spoke next, saying that we should look at the present US draft document "in conjunction with the overall Guidelines." He also said further studies need to be done, such as what are the practical constraints of getting the data suggested by the 2001 Decision tree. As an example he mentioned the supposed down-side to reduce the number of contiguous amino acids in a amino acid homology search from 8 to 6 is an increase in rate of false positives. He then asked rhetorically if we really could incorporate the decision tree itself into the document we develop. He then went on to say he liked the level of detail he found in the present e.g. US document. So he did not speak out explicitly for the 2001 Decision tree; rather he spoke against it. This in spite of the fact that the official comments of the UK explicitly supported the decision-tree approach as the way to go.

Japan then proceeded to support of the approach of Australia and the UK. The Netherlands echoed Australia, UK and Japan, saying there was a need for general discussion of the focus of the document and that it should not be overly prescriptive.

Denmark then spoke up to say that vis-à-vis how much guidance the document should give, they though the document should be very precise. Belgium agreed with Denmark, as did CI.

Canada then supported the US proposal and then the discussion then turned to a paragraph by paragraph discussion of the US draft document.

When we got to paragraph 2, where there was the first mention of "weight of the evidence approach," the first real fight of the meeting started. I (CI) immediately proposed alternative language saying to use a "step-wise decision tree approach" that's based on the decision tree produced by the 2001 FAO/WHO consultation. Sweden proposed "use an integrated step-by-step approach." UK suggested much weaker language: "a structured approach as followed below." Canada suggested modifying the UK language by adding "from several types of information described below." Australia wanted to explicitly include "weight of the evidence" and proposed: "The weight of the evidence should be considered including its scope and veracity." Netherlands supported Australia and US saying we're using a weight of the evidence approach. CI, France and Germany supported the Swedish proposal. The final resolution of the issue (which we made some progress on) was the following language: "it is recommended that an integrated, stepwise, case by case approach, as described below, be used in the assessment of possible allergenicity of newly expressed proteins. This approach takes into account the preponderance of evidence derived from several types of information and data."

Overall, we were able to mitigate the damage by supporting "integrated, stepwise, case by case approach," as proposed by Sweden, after it became apparent that the CI proposal, which explicitly said "decision tree approach," was going nowhere. However, we weren't successful in getting explicit reference to a decision tree approach based on the 2001 FAO/WHO consultation, even though 9 of the 11 countries that submitted comments before the meeting supported a decision tree approach. Only the US, Canada and Australia voiced support for a "weight of the evidence approach, " although the UK played a key role by supporting a more general approach to the document under development while at the same time talking of some of the difficulties of trying to follow the decision tree as laid out by FAO/WHO.

After agreeing explicitly not to endorse the 2001 decision tree, the rest of the meeting became a fight over smaller issues. Although, para 2 doesn't explicitly refer to the 2001 FAO/WHO decision tree, there were a number of other attempts to get language consistent with the 2001 decision tree into the document, but it was resisted or watered down. On a few occasions during Monday and Tuesday, both Sweden and Belgium tried to get reference to the FAO/WHO 2001 decision tree in the document, but failed. The most dramatic attempt came on Wednesday, when Denmark suggested adding an explicit reference to the right hand side of the FAO/WHO 2001 decision tree. Incredibly, the chair (Paul Mayer) replied that you couldn't refer to the FAO/WHO scheme in the text of this Codex guideline as it's "outside the scope of Codex"!! This assertion is completely false and was a clearly blatant attempt to block any mention of the decision tree from the 2001 meeting.

Amino acid sequence homology

Para 7, the first para in the Section 3.2 Amino Acid Sequence Homology. There was a fight on this issue and Tomlinson was on industry's side. First, he said he wondered if there was too much "detail" in this para. Then he chimed in on the question of whether to recommend using 8 or 6 sequential amino acid segments in the searches for homology. The US draft text in para 7 explicitly used the industry figure-"8"-rather than the figure suggested by the FAO/WHO 2001 consultation-"6." The UK didn't protest this in principle (e.g. having the industry figure in the guidance rather than the figure suggested by their own expert committee), although the delegate from France did protest this point. In the discussion of the language, UK explicitly said that the first sentence-"The smaller the peptide sequence used in the stepwise comparison, the greater the likelihood of identifying false positives, thereby reducing the utility of the comparison"-should be kept. However he didn't speak of the need for a counterbalancing sentence-"The longer the peptide sequence, the greater the likelihood of identifying false negative"-to be created. The latter was suggested by others and ultimately included.

Serum Screening

In the discussion of para 13 , the first para in Section 4, which talks about specific serum screening for those protein which originate from a source known to be allergenic, the UK made the first comment, suggesting that the whole paragraph be deleted and it goes into too much detail, and that the guidelines are supposed to be general.

"13. For those proteins which originate from a source known to be allergenic, or have sequence homology with a known allergen, testing of protein extracts in immunological assays is recommended. Sera from individuals with clinically validated allergy to the source of the protein can be used to test IgE-binding of the protein in in vitro assays. A critical issue for testing will be the availability of human sera from sufficient numbers of individuals. In addition, the quality of the sera and the assay procedure need to be standardized to produce a valid test result."

The UK's suggestion was not followed up on, and although slight changes were made, the paragraph as a whole remained in the final text.

Action on Known Allergen

Para 15 makes an extremely important recommendation about discontinuing development of GE foods that have been shown to contain introduced allergens. The Canadian draft text said that such foods should be "discontinued." The UK suggested deleting paragraph 15 entirely, in line with their thinking that the guidelines shouldn't contain too many "details" . The US draft that we were working from had weakened this further by putting a disclaimer that said perhaps you could deal with this problem through labelling. I was successful in getting all mention of "labelling" removed from the language of the text so that the para now reads: "15. The identification of a newly expressed protein as an allergen through immunological assays suggests that further development of the product be discontinued, unless adequate risk management and risk communication measures could be assured throughout the marketing and distribution of the product, since segregation and identity preservation of the new source of this allergen may be difficult or impossible to enforce."

I don't like the final text of para 15 (the middle clause should be deleted), but there was nothing I could do about it. But Tomlinson's intervention would have weakened the guidance even further than what the US was proposing.

The few attempts to strengthen the draft report (for ex., we won on the "amino acid sequence homology" front with getting mention of "8" in the main para removed and we got the inclusion of language that using longer sequence (such as 8 amino acids, which is industry's choice) can increase the risk of false negatives) were done with out Tomlinson's help.

Actions of Australia

Even more so than the UK, Australia (represented by Marion Healy) played a key role in supporting the industry-desired move away from the 2001 FAO/WHO decision tree approach. Australia was the third to speak up, after Canada and the US had opened discussions on the first day with their versions of the draft working document, and immediately framed the discussion by saying that two larger issues needed to be discussed: i) whether we're creating a stand-alone document or one to be read in conjunction with the larger risk assessment, and ii) what level of detail to go into in the document. So, Australia, proposed keeping the document "general," to keep prescriptive details out of the document. As previously stated, Australia even suggested language for para 2 that explicitly referred to a "weight of the evidence approach."