Science by press release: Austrians marry in haste, repent at leisure over GM corn dowry

Monsanto scientist’s recalculation of Austrian multi-litter mouse reproductive study.(Full text of Monsanto response below)

News Releases
Independent Scientists Determine Study Conclusions are Flawed

No Evidence of Risk of GM Crops
SAINT LOUIS (November 20, 2008) – On November 11, 2008, the Austrian Ministry of Health, Family and Youth released a report on three studies designed to assess the impact of GM corn on reproduction. One finding in one of the studies was interpreted as a possible impact on reproduction in the test mice. These unpublished studies had not been subjected to peer-review or analysis by independent scientists. At the time the report was released to the press, the author of the study, Dr. Jurgen Zentek, remarked that his team’s three studies showed inconsistent results and should be considered preliminary.

Since then, the unpublished study has been reviewed by several scientists, including two internationally recognized experts in reproductive toxicology; Dr. John DeSesso, Senior Fellow at the non-profit group Noblis and Dr. James Lamb, currently of The Weinberg Group. Dr. Lamb developed one type of study used by the Austrian researchers while at U.S. Department of Health and Human Services, National Toxicology Program. Doctors Lamb and DeSesso have both concluded that there are significant flaws in the study reporting and analysis which bring serious question to the validity of the findings. They agreed with Dr. Zentek that the results were inconsistent but concluded that there was no evidence of any adverse effects of the GM crop.

Within 24 hours of the preliminary findings being released, Greenpeace and the Center for Food Safety (CFS) had issued statements calling for a recall and moratorium on GM crops and foods. “Once again, these organizations have demonstrated that their primary interest is sensational headlines and not scientific substance,” said Jerry Hjelle, Ph.D., Vice President of Monsanto’s Regulatory group. “Every time a preliminary study like this comes out, Greenpeace and the Center for Food Safety cry ‘wolf’. And time and time again, scientific scrutiny finds that GM crops and food are safe.”

More information:

Dr. Lamb’s review of the Austrian mouse study (pdf)
Monsanto’s technical review of the Austrian mouse study (pdf)
Monsanto's position on the Austrian mouse study
About Monsanto Company

Monsanto Company is a leading global provider of technology-based solutions and agricultural products that improve farm productivity and food quality. Monsanto remains focused on enabling both small-holder and large-scale farmers to produce more from their land while conserving more of our world's natural resources such as water and energy. To learn more about our business and our commitments, please visit: www.monsanto.com

CONTACT: Brad Mitchell, +1-314-694-6132, for Monsanto Company

Pundit notes:

The Pundit can confirm that many of the errors picked up by the technical commentaries provided by Monsanto are obviously displayed In the original Austrian report. The failure to analyse for the content of mycotoxins such as fumonisin and aflatoxin (easily done) even in light of evidence that the Canadian samples contain high microbial contamination and mould growth, failure to properly select representative corn samples, and failure to include a reference group in the multi-litter study stood out even on a quick read through the reports. The Pundit also questions why comprehensive statistical information was not included, particularly the analysis of variance (ANOVA).

In a subsequent post, the Pundit notes the curious affair of silence about mistreated mice.

Text of the Monsanto technical response on this issue:

Monsanto Company Response to Austrian Report on Mouse Chronic and Reproduction Studies with NK603 x MON810 Maize
November 20, 2008

Report: Biological effects of transgenic maize NK603xMON810 fed in long term reproduction studies in mice
Authors: Alberta Velimirov, Claudia Binter, and Jurgen Zentek Sponsor: Austrian Ministries for Agriculture and Health
Date: October 2008

The above report, which was released on November 11, 2008, summarizes the results of three studies intended to examine potential long-term and/or reproductive effects in mice fed NK603 x MON 810 maize. According to the authors, the results of the
reproduction studies suggest that the reproductive performance was affected in mice fed GM maize. However, this report lacks sufficient experimental details to fully interpret the results and contains a number of errors which make it unsuitable for risk assessment and/or regulatory purposes. Until sufficient details regarding the methodologies used and individual animal data are made available, it is impossible to conduct a comprehensive assessment of these studies. However, our evaluation of the available information indicates that the authors’ conclusions of adverse reproductive findings are unfounded. The justification for our conclusion and our major comments on other aspects of the studies are summarized below. It should also be noted that no adverse effects on health or survival were noted in a limited but life-time mouse feeding study that was also summarized in this report.

Comments on Reproduction Data
The report summarizes the results of two reproduction studies, a multi-generation (MG) and a continuous breeding (RACB) study, both conducted in mice. With a few exceptions, these studies appeared to have been relatively well-designed but, as described below, not well analyzed or reported. In addition, no information was provided
on whether the testing laboratory had ever conducted similar reproduction studies before. The lack of historical control and insufficient use of reference control groups makes it very difficult to properly interpret data that the authors acknowledge are highly
variable. Only one reference group was included in the MG study and no reference groups were included in the RACB study. The lack of historical data and/or reference groups in the RACB study is especially concerning since the authors acknowledged that they utilized an atypical design that puts the maternal animals under substantial stress.
The reporting of the data for both reproductive studies appears to be seriously flawed.
The method of calculation for several key reproductive parameters was inconsistent between the two studies and did not follow standard scientific and regulatory guidance that indicates the litter is generally the most appropriate experimental unit. There also appear to be a number of calculation errors in reproductive parameters, especially in the RACB study. Insufficient data were provided in the report to verify all values and confirm the statistical analysis that was done. However, these reporting errors magnified the apparent differences between the GM and control groups and likely were responsible for several of the findings in the RACB study to be reported (possibly
erroneously) as statistically significant.

Specific comments on the MG and RACB studies are summarized below:

MG Study
1. Table 36: Although no statistically significant differences in fertility or other reproductive parameters were noted in the GM group, the authors concluded that the number of pups at birth and at weaning were lower in the GM group than in the controls. However, with one apparent exception, the numbers of pups at birth and
weaning were inappropriately expressed as numbers of pups per co-housed pair rather than numbers of pups per litter. The one exception is the value for the number of pups at birth/pair for the GM F2 litter. This was reported as 8.92, which do does not appear to represent either the number of pups per co-housed pair (208/24=8.67) or the number of pups per litter (208/21=9.90). In addition, it is not
clear how the numbers of pup losses/group were calculated since none of the values could be derived from the data presented. The largest discrepancy appears to be for the GM F2 pups where the reported value (2.95 pups lost per group) appears to be inconsistent with a total of 25 pups lost from 21 litters. Finally, fertility (number of
deliveries/group) was apparently decreased in both the control (ISO) and GM groups in the 3rd (F2) and 4th (F3) generations. This is unusual and could indicate the presence of some confounding factor (disease, environment, etc) yet no explanation was provided. In addition, this decline resulted in an apparent but misleading reporting of a decrease in litter size when the authors inappropriately expressed the data as pups per co-housed pair rather than per litter.
2. Table 40: Despite the incorrect heading (ISO vs. GM), this table actually compares data from the ISO and A REF groups. However, it is not clear why, if they were conducted at the same time and shared the same control group, the A REF data were not presented (and statistically analyzed) along with the GM data. In addition, this table appears to share the same erroneous calculations as those discussed
above for Table 36.
3. Litter size can impact pup weight, maturation rate and survival. Therefore, to avoid complications in interpreting data, pups are generally culled to 8 (or sometimes 10) pups per litter. OECD recommends that, if culling is not conducted, litter size be included in the statistical analyses, e.g., by covariate analyses (OECD, 2008). This was not done in these studies. Instead, the authors took the highly unusual approach of separately reporting some offspring data from "litters with more than 8 pups" and "litters with 8 or less pups". No justification for this apparently arbitrary distinction was provided.
4. The report did not indicate whether or not sibling matings were avoided when selecting pups for subsequent generations. Failure to avoid sibling matings can lead to confounding results, especially in later generations.

RACB Study
1. Table 59: Despite the fact that the same terminology (pups per "pair') was used as in Tables 36 and 40, most of the values in this table were appropriately expressed as numbers of pups per litter rather than per co-housed pair. Unfortunately, there appeared to be three major exceptions, all in the GM groups: pups at weaning for
the 3rd litter, pups at birth for the 4th litter, and pups at weaning for the 4th litter. The results of calculating these values on a per litter born rather than per co-housed pair basis, as was done for the other data in this table, are shown in Table 1. In addition, it appears that there were five instances of total litter loss after birth in the 3rd litter and one instance in the 4th litter. The reason for the total litter loss was not discussed but it would probably be most appropriate to use the total number of surviving litters as the denominator when expressing number of pups at weaning, not the numbers of litters at birth. The results of this calculation are also shown in Table 1.
Furthermore, as in Tables 36 and 40, it is not clear how the sum of pup losses per group were calculated . Finally, although it had no impact on the conclusions, the value for the sum of pup losses per group in the 3rd GM litter appears to also be incorrect (213-207=6, not 2).
2. According to Table 59, one of the GM pairs in the RACB study apparently were never successful in delivering a litter (23 first litters delivered/24 co-housed pairs).
Since no effect on fertility was noted in the MG study, it is likely that one of the animals from this pair (either male or female) was infertile prior to exposure to the GM corn. If so, the results from this pair should be excluded from the table.
3. What was the cause for the loss of entire litters in the RACB study? Was the cause of death for any of the pups determined? Were males left in the cages with the litters? If so, what steps were taken to minimize or document cannibalization which commonly occurs in such studies?
4. Despite the fact that the dams were much older and had been subjected to the stress of continuous breeding and raising of offspring , the numbers of pups/litter at birth and weaning of the 3rd and 4th litters of the GM-fed animals were greater than in either the control or GM animals from the 1st litter. This suggests that the minor differences between the GM and control group were just a result of normal biological variation and were unrelated to feeding of GM maize.

Comments on Microarray Data
1. The analysis of the microarray data appears to be biased to find differences in the GM-fed samples. The authors pooled together the two non-GM samples, thereby eliminating the opportunity to understand the natural variability in gene expression between two groups of animals fed conventional diets and, instead, compared this
pooled set against the GM-fed sample, reporting 2374 significantly deregulated genes (most of which h had differences < 2-fold, which are generally ignored in microarray studies). The way the study was analyzed and presented makes it appear that 10520 genes were expressed differently and uniquely in the GM sample (the red sector) compared to the A REF and ISO samples, when a typical Venn diagram of transcriptome analysis would have the largest fraction to be in the
intersection of all three samples, where most genes would be expressed similarly in all three samples, thereby focusing on the unique differences that were specific only to the GM sample.
2. The authors highlight the fact that there were 440 differences in gene expression between the ISO and GM groups but not that 1016 differences were noted between the ISO and REF group.
3. With respect to samples of the intestine taken for microarray analysis, did the authors ascertain that the intestines were empty? If not, how did they ensure that they were not isolating DNA and RNA from intestinal contents?
4. According to the introductory statements, the authors postulate that changes might be seen in the absorptive epithelia of the intestines. If so, it seems appropriate to test that tissue. How did the authors ensure that their 50 mg samples of frozen tissue came from absorptive epithelium and not from the intestinal serosa or
muscular layers? How do they know that all of the samples taken represent the same regions of tissue? If all samples for one group are serosa while the others are from the muscular layer, the results are likely to differ and neither would provide information about the status of the absorptive epithelium.

Comments on Electron Microscopy
The description of the methods used regarding the electron microscopic analyses were inadequate to allow proper evaluation. For example:
1. Why are there no example micrographs?
2. How many blocks of tissue were processed per organ and how many blocks contributed to the analysis? How did the authors ensure that the areas of the organs sampled were the same for all groups? If similar areas were not analyzed, how do the authors know that their descriptions are not due to regional differences in the organs?
3. How were the nuclear pore data collected? Identification of nuclear pores requires cuts that are perpendicular to the nuclear envelope and it is not clear whether or not this was done in this study.
4. The authors pointed out differences in fibrillar centers and dense fibrillar components in the liver and pancreas of the GM group compared to the ISO controls, but when the ISO controls were compared to the reference group, similar variability was
observed. Given the variability apparent between two non GM crops, it is difficult to draw any conclusions about the relevance of differences between the ISO and GM groups.
Comments on Test Material and Diet Formulation
1. The production of the test materials used in these studies did not appear to comply with EFSA guidelines. According to EFSA guidelines (EFSA, 2006), the data should be obtained from a comparison of the GM plant and the most appropriate control line grown in the same field under comparable conditions. However, according to the
report, the production locations for the 2005 GM and ISO grain batches used in the MG study were separated by up to 10 to 20 km. No distances were provided for the 2007 test materials used in the RACB study. In addition, EFSA states that the grain should come from "geographical locations representative of the various environments in which the GM plants will be cultivated". However, Nova Scotia is
not a "typical" agronomic environment for corn and is not representative of the locations where the vast majority of GM maize will be grown.
2. Both lots of GM grain had detectable levels of mycotoxin (DON in the 2005 grain and zearalenone in the 2007 grain) and total microbial count for the 2005 GM grain was 7-fold higher than ISO and >80-fold higher than the "A-REF" grain. Therefore in addition to the presence or absence of the GM trait being a difference between GM and control diets, these microbial contamination differences were also present,
thereby complicating interpretation of the results from these studies. Also, given detectable levels of DON and zearalenone in the GM grain, it is unclear why the authors didn't also assess fumonisin, another important mycotoxin commonly found in maize grain.
3. Was there analysis of the GM, ISO, and A-REF diets to confirm that the diets were formulated correctly, e.g., the GM diet had GM grain and the ISO and A-REF diets did not?

Monsanto Conclusion
No adverse effects on health or survival were reported in mice following life-time feeding of NK603 x MON 810 maize. Lack of sufficient experimental detail, high variability and concerns about some of the approaches used to evaluate the data make it difficult to
draw any conclusions from the microarray and ultrastructural data, which the authors themselves described as weak and/or inconclusive.
The authors of the report concluded that the GM maize affected reproduction. However, the reported effects on reproduction parameters were not consistent within or between the studies and are not apparent when the data are calculated properly. No significant
differences were noted in any of the four generations in the MG study. A few statistically significant differences in reproductive parameters were reported in the RACB study but most appear to be a result of calculation errors. Therefore, in our opinion, the data from these studies do not indicate any adverse effects on fertility or
other reproductive parameters from long-term feeding of GM maize.
Two internationally-recognized experts on reproduction have also reviewed the online report. Dr. John DeSesso, Senior Fellow at the non-profit group, Noblis, and member of the Editorial Board of the journal, Reproductive Toxicology, observed that “The continuous breeding study used a non-standard study design, collected data that are not typically measured in these types of studies, and presented data in displays that are both difficult to understand and have mathematical errors.” Dr. James Lamb, currently of The Weinberg Group, and Associate Editor of the journal, Environmental Health
Perspectives, has also reviewed and provided comments on this report. Dr. Lamb had developed the reproductive assessment by continuous breeding (RACB) protocol in the 1980s as Head of the Fertility and Reproduction Group in the U.S. National Toxicology
Program. A modified version of the RACB protocol was used by the Austrian laboratory. Dr. Lamb concluded that while the RACB and multi-generation (MG) study design utilized in this study were appropriate for the hypothesis being tested, the report was not sufficiently detailed to interpret fully the study. Dr. Lamb noted that there were a number of errors in the report and concluded that, “the computational errors in critical tables raise serious questions about the other data in the report and the quality assurance methods that were or should have been applied before the conclusions were
drafted and the report was released. These errors are relatively simple mistakes, but they inadvertently have a profound effect on the study interpretation. When properly analyzed, these data do not appear to support an effect on fertility or reproduction from consumption of GM corn.”

References
EFSA (2006). Guidance document of the scientific panel on genetically modified organisms for the risk assessment of genetically modified plants and derived food and feed. May 2006. Available at: http://www.efsa.europa.eu/EFSAIefsa (accessed 19 November, 2008)
OECD (2008). Series on Testing and Assessment. Number 43. Guidance Document on Mammalian Reproductive Toxicity Testing and Assessment. 24 July 2008. ENV/JM/MONO(2008)16. Available at:
http://www.olis.oecd.org/0Iis/2008doc.nsf/linkto/env-jm-mon0(2008)16
(accessed 19 November, 2008).

Update from GMO Pundit

EFSA comments:

Request from the European Commission related to the safeguard clause invoked by Austria on maize MON810 and T25 according to Article 23 of Directive 2001/18/EC

Question number: EFSA-Q-2008-314
Adopted date: 4 December 2008
Summary (0.1Mb)
Opinion (0.3Mb)

Summary
On 10 June 1999 and on 8 May 2000, Austria invoked Article 16 of Directive 90/220/EEC (safeguard clause) to provisionally prohibit the placing on the market of the authorised genetically modified (GM) maize events MON810 and T25 on its territory. In February 2004 and November 2007, Austria provided additional information to support the national safeguard measure to be considered under Article 23 of Directive 2001/18/EC. To define whether the information submitted by Austria comprises new information that would affect the environmental risk assessment for the uses laid down in the corresponding consent, the European Commission requested in a letter, dated 18 April 2008, a scientific opinion from the European Food Safety Authority (EFSA).

Following investigation of the evidence presented in the Austrian submission, the Scientific Panel on Genetically Modified Organisms (GMO Panel) of EFSA concludes that there is no new scientific evidence that would invalidate the previous risk assessments of maize MON810 and T25. Therefore, no specific scientific evidence, in terms of risk to human and animal health and the environment, was provided that would justify the invocation of a safeguard clause under Article 23 of Directive 2001/18/EC for the marketing of maize MON810 and T25, for its intended uses, in Austria.

Details relating to animal tests:

Adopted part of the minutes1 of the 46th plenary meeting of the Scientific Panel on Genetically Modified Organisms held on 3-4 December 2008

GMO Panel deliberations on the Austrian report "Biological effects of transgenic maize NK603x MON 810 fed in long term reproduction studies in mice" as adopted at the plenary meeting of3-4 December 2008.

On 11 November 2008 the Austrian Federal Ministry of Health, Family and Youth released a research report on studies in mice, which were conducted to assess the impact of genetically modified (GM) maize NK603 x MON 810 on reproduction (Biological effects of transgenic maize NK603 x MON 810 fed in long term reproduction studies in mice, Dr. Alberta Velimirov, Dr. Claudia Binter, Univ. Prof. Dr. Jürgen Zentek).

The report includes three studies, a life-time study, a multigeneration study (MGS), and a reproductive assessment by continuous breeding study (RACB). According to the authors the life-time study showed no statistically significant differences in survival between mice fed with kernels of maize NK603 x MON 810 and the controls. They also reported that, in the MGS study, no significant differences in reproductive traits were found between mice fed with kernels of maize NK603 x MON 810 and the controls. In the RACB study, the authors used a modified protocol of the original RACB study developed at the U.S. National Toxicology Program (NTP) for the testing of chemicals. Male and female mice were housed as breeding pairs for approximately 20 weeks and allowed to produce litters continuously throughout the cohabitation period.

The authors identified differences in reproductive parameters between mice fed with the GM maize and the controls. They reported that there were statistically significantly fewer pups born in the GM group in the 3rd and 4th delivery and fewer pups weaned in the 4th litter compared with the control group.

The GMO Panel considered this report and came to the following conclusions.

Regarding the RACB study, the summary Table 59 contains calculation errors and inconsistencies in the treatment of the data regarding the 3rd and 4th litters. In addition, it seems that the authors have calculated the number of pups at birth per pair and not per delivering pair, which is standard practice.Also, there appears to be methodological deficiencies in the statistical analysis that seriouslycompromise the interpretation of the data. For the reasons stated above, individual data are required for a proper assessment.

In addition, more detailed information regarding the breeding scheme is needed. In particular, it should be clarified whether in the 3rd and 4th pairing the same or different pairs failed to reproduce.Information regarding the normal variation of the parameters examined in this study for the mouse strain used (historical control data) is required before any conclusion may be drawn on possible alterations in reproductive performance.

In addition, further information on the estrous cycle and histopathological parameters including spermatogenesis, follicle and oocyte counts is essential for assessing the claims of reduced fertility.The GMO Panel also notes that information on the genetic identity and characteristics of the tested materials is not sufficient.On the basis of the data presented the GMO Panel is of the opinion that no conclusions can be drawn from the report.

1 Published at http://www.efsa.europa.eu/EFSA/efsa_locale-1178620753812_1211902199319.htm. The complete minuteswill be adopted at the 47th plenary meeting (28-29 January 2009) and will be published shortly afterwards.