A series of letters to the Editor in response to the recent G-E Seralini / CRIIGEN scientific article about rat feeing tests on GM maize are now available via the following links :
The journal’s In-press page: http://www.sciencedirect.com/
Seralini GE et al. (2012)
Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize.
Food and Chemical Toxicology 50(11): 4221–4231. DOI: 10.1016/j.fct.2012.08.005
Letters to the Editor:
Panchin AY. (2012)
Toxicity of Roundup-tolerant genetically modified maize is not supported by statistical tests.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.039
External URL: http://www.sciencedirect.com/
Seralini et al. (2012) claim to have found evidence for the long term toxicity of roundup-tolerant genetically modified maize (GMM). Using one-tailed Fishers exact test we show that there is no statistically significant increase in mortality rates or the number of tumors in rats fed GMM compared to control groups in the original data.
Seralini et al. (2012) claim to have found evidence for the long term toxicity of roundup-tolerant genetically modified maize (GMM). Using one-tailed Fishers exact test we show that there is no statistically significant increase in mortality rates or the number of tumors in rats fed GMM compared to control groups in the original data.
Cockburn A. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.040
Tribe D. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.042
Schorsch F. (2012)
Serious inadequacies regarding the pathology data presented in the paper by Séralini et al. (2012).
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.043
Hammond B, Goldstein DA, Saltmiras D. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.044
External URL: http://www.sciencedirect.com/
...1. Experimental design
The authors of this study assert that it was conducted in a GLP environment and according to OECD guidelines. They did not follow OECD GLP guidelines nor OECD testing guideline (TG) 453 for conduct of a combined chronic toxicity/carcinogenicity study. OECD GLP’s require “Detailed information on the experimental design, including a description of the chronological procedure [e.g., start date, end date] of the study, all methods, materials and conditions, type and frequency of analysis, measurements, observations and examinations to be performed, and statistical methods to be used (if any)” and … “The study should be conducted in accordance with the study plan”. Apparently, the authors’ original intent was not to conduct a carcinogenicity study “…we had no reason to settle at first for a carcinogenicity protocol using 50 rats per group.” (Seralini et al., 2012), but at some point during the in-life phase, they changed the purpose of the study by extending it for 2 years to assess potential carcinogenicity. Assuming they had a protocol at the start of the study, they did not follow it as they substantially altered the purpose and the design of the study while it was in progress. This should be considered a violation of GLP guidelines as the study was not conducted in accordance with the original study plan. If they wanted to carry out a carcinogenicity study, they should have terminated the existing study, and prepared a new study plan adapted from OECD TG 453. They did recognize, as stated above, that they needed a larger number of animals (a minimum of 50 rats/sex/group) for a carcinogenicity study, instead of the 10 rats/sex/group that they had in their existing study. For reasons which will be discussed later, their study did not have enough animals to draw any meaningful conclusions.
Rodent carcinogenicity studies must be sufficiently powered not only to detect an increased incidence of rare tumor types, but also to discriminate treatment-related effects from spontaneous, or background, incidence of common tumor types. For this reason, US ( [US EPA, 1998] and [FDA, 2006]) and OECD (1995a) regulatory guidelines for the conduct of carcinogenicity studies in rodents specify the use of at least 50 animals per sex per treatment group. In addition, OECD states that “it is unlikely that a regulatory authority would find a study using a lower core number of animals per sex and per group acceptable for regulatory purposes, since a sufficient number of animals should be used so that a thorough biological and statistical evaluation can be carried out.” (OECD, 1995b). OECD further states that “for strains with poor survival such as SD rats, higher numbers of animals per group may be needed in order to maximize the duration of treatment (typically at least 65/sex/group).”(OECD, 1995b). For this reason, the US EPA specifies that survival in any group should not fall below 50% at 18 months or below 25% at 24 months (US EPA, 1998), while the US FDA specifies survival of a minimum of 25 rats per sex per group at study termination (FDA, 2006). The SD rat has been widely used in toxicology research, including numerous chronic studies, but these studies employ many more animals than used by the authors in consideration of their lower survival rate and high background tumor rates, especially mammary tumors in females.
2. Statistical analysis and presentation of data
The authors have a history of inappropriate application of statistical methods to analyze toxicology data ([Seralini et al., 2007] and [Spiroux de Vendomois et al., 2009]) which has been criticized by regulatory agencies and other experts ( [EFSA, 2007], [EFSA, 2010], [FSANZ, 2009a], [FSANZ, 2009b], [HCB, 2009] and [Doull et al., 2007]). There are numerous problems in the way the data were statistically analyzed in this study.
For example, in Table 3, mean values are not presented for each group and sex to allow comparison of measured parameters. Control data are not presented. Instead, the authors used a statistical method that is not traditionally used to present toxicology data, a multivariate technique called Partial Least Squares Discriminant Analysis (PLS-DA). Mean differences (%) of variables (discriminant at 99% confidence intervals) were presented to investigate the relationship among 48 blood and urine measurements relative to the different treatment groups. PLS-DA can be used to identify patterns in the data and to develop a function which can be used to discriminate between the groups. However, any differences between groups must be further evaluated for toxicological relevance. Presentation of the data in this manner does not lend itself to straightforward interpretation of the study findings.
In Fig. 5, the same PLS-DA procedures were followed with jack-knifed confidence intervals at 99% confidence level. This procedure may be familiar to statisticians, but it is not commonly used to present toxicology data and is difficult to interpret, particularly when the data used to construct these graphs are not presented. Examination of Fig. 5a would suggest that the majority of measured parameters fall within 99% confidence intervals with the exception of serum and urine electrolytes. Unfortunately, no data were provided from other intervals when these data were collected to determine if the same patterns were evident. No lab historical data were provided to put these data in perspective. As stated earlier, just because one can discriminate between the groups, it does not make the result toxicologically relevant. There was no presentation of actual statistical analysis to compare the means for each measured parameter.
To determine if there are patterns of differences in toxicologically related findings, the toxicologist expects to see the actual mean data for each parameter/group and the standard deviation and the control data should also be provided for comparison. The test and control values for measured parameters should also be compared to the historical control data from the testing laboratory and/or the literature to determine if differences were within or outside of the normal range. As presented, the reader has no way of determining whether the conclusions drawn by the authors are supported by the actual data, or are merely statistical anomalies resulting from non traditional analysis. The manuscript contained figures with graphs that were difficult to read because lines overlapped, and percent variations were presented rather than the mean test and control data which is the more standard practice in presenting toxicology data. For instance, incidences of 1 vs. 2 or 5 vs. 10 both represent a change of 100%, however, these absolute values would likely result in different conclusions...
Ollivier L. (2012)
A Comment on “Séralini, G.-E., et al. Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize. Food Chem. Toxicol. (2012),”http://dx.doi.org/10.1016/j.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.045
Tester M.. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.046
Wager R, Lerayer A, Fedoroff N, Giddings V, Strauss SH, Leaver C, Shantharam S, Potrykus I, Fellous M, Burachik M, Jany KD, Trewavas A, Kameswara C, Prakash CS, Miller HI, Bradford K, Cetiner S, McHughen A, De Stefano-Beltrán L, Chassy BM, AlMomin S, Newell-McGloughlin M, Ammnann K, Herring RJ, de Souza L. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.047
Pilu R. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.048
External URL: http://www.sciencedirect.com/
...It is well known that maize kernels can be affected by fungal infections (mainly fungi belonging to the genus of Fusarium and Aspergillus) leading to the contamination of grains with mycotoxins such as aflatoxins and in particular fumonisins ( [Shephard et al., 1996] and [Pilu et al., 2011]). These toxins are associated with a number of serious animal and human diseases, in particular cancer (Ross et al., 1992). For this reason the FAO/WHO Expert Committee on Food Additives, the U.S. Food and Drug Administration and the European Union have established for direct human consumption the threshold of fumonisin content in maize at the level of 1 ppm (Rule EU 1126/2007). It is also well known that incorrect conditions for seed and flour storage (e.g. high temperature and moisture) cause increases in the levels of contamination.
The authors wrote in Materials and Methods paragraph (2.2. Plants, diets and chemicals): “The varieties of maize used in this study were the R-tolerant NK603 (MonsantoCorp., USA), and its nearest isogenic non-transgenic control. These two types of maize were grown under similar normal conditions, in the same location, spaced at a sufficient distance to avoid cross-contamination.” This distance is of about two hundred meters, which may determine in some cases dramatic differences in the kernel quality due to slight environmental differences in the field. This work does not report or mention the analysis of the mycotoxins content of the material obtained...
Williams B. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.049
External URL: http://www.sciencedirect.com/
...
...
1.2. ARRIVE guidelines
We have evaluated the study against guidelines produced by NC3Rs, the National Centre for the Replacement, Refinement and Reduction of Animals in Research. The ARRIVE guidelines have been produced to improve standards of reporting in scientific papers. They help ensure that a paper provides “key information on how the study was designed, conducted and analysed”, without which its “value in informing future scientific studies and policy” is limited.
It is recommended that the introduction of a paper should “include sufficient scientific background (including relevant references to previous work) to understand the motivation and context for the study, and explain the experimental approach and rationale.”
The authors of the current study state in their introduction that “there is an ongoing international debate as to the necessary length of mammalian toxicity studies in relation to the consumption of genetically (GM) plants”. However, the only evidence used to substantiate this claim is a paper by the leading author of the current study.
Further claims made about GMOs reference only papers that include authors of the current study. This does not constitute “sufficient scientific background”.
Where the current study falls short of the ARRIVE guidelines in other respects, we have highlighted the shortcomings...
Trewavas A. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.050
External URL: http://www.sciencedirect.com/
...If inbred rodents like Sprague Dawley are kept in a cage for 2–3 years most die of cancer. The company Harlan (www.harlaneurope.com) who market Sprague Dawley rodents describe their life characteristics under Life Span and Spontaneous Disease. I quote- pituitary gland tumours were found in 20% of the males and 39% of the females. This relatively low incidence had little effect on the survival of the females (50%) due to the high incidence (76%) of mammary gland tumours (predominantly fibroadenomas) resulted in unscheduled sacrifices of many females. Other common neoplasms were benign medullary tumours (27% in males, 11% in females) and endometrial stromal polyps (22%) in females. These are spontaneous tumours of course with no known inducing agent and so far as can be observed develop at random with increasing age. These figures for spontaneous tumours are identical with those reported by Seralini et al. but had no special diet or treatment.
Spontaneous tumours have also been described in detail in Sprague Dawley rats in Schardein et al. (1968) Pathol. Vet. 5, 238–252. These detailed studies on large number of rats report that 6 rodents out of about 3000 develop spontaneous tumours before 90 days, the obvious reason that most tests stop at that time to avoid spontaneous tumours causing conflicting results. 10 out of 700 develop tumours between 3 and 6 months; 6 and 9 months, 20 out of 400 rodents with tumours and so on. Their cumulative% of rats developing tumours over 18 months were 49% for males and 70% for females; the latter being due to high numbers in the mammary gland. The numbers involved are more accurate than any others I have seen including the two references produced by Seralini et al. The paper by Seralini et al. made great claim of early developing tumours at four months but these are obviously normal spontaneous characteristics which again require no known inducing agent. In all cases in cancer the accumulation of mutations is a time dependent phenomenon random in its characteristics in the absence of an inducing agent as to when the mutations bunch up sufficiently in character and type (such as p53 changes)to induce carcinomas. McComb et al., 1984. Journal of the National Cancer Institute, 73, 1143–1186) summarised over 40 reports of spontaneous pituitary adenomas in rats of a number of lines. Again I quote. Depending on the age, sex, or strain the prevalence of spontaneous pituitary adenomas varies from 0% to100%. The frequency increases with age in all strains but exact figures may vary considerably from strain to strain from study to study. Note the figure 0–100%. How many rodents are required to enable a significant difference, with that degree of variation, a lot more than 10 as controls. The numbers of spontaneous pituitary and mammary tumours and their appearance developed at random match those described by Seralini et al. and in the case of Schardein et al. with 15-fold greater number of rodents and thus greater accuracy...
Grunewald B, Bury J. (2012)
Comment on ‘‘Long term toxicity of a Roundup herbicide and a 4 Roundup-tolerant genetically modified maize’’ by Séralini et al.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.051
External URL: http://www.sciencedirect.com/
...Seralini et al. fed rats during a period of two years with either NK603 maize (three different doses: 11%; 22% or 33%), with NK603 maize (the same three different doses) that was sprayed with Roundup, or provided drinking water in which Roundup was present (three different concentrations: 50 ng/l, 400 mg/l or 2.25 g/l). In this way nine different treatments were tested and were compared with only one negative control, being a diet with one dose (i.e. 33%) of near-isogenic non-GM maize. Each diet was fed to 10 male and 10 female rats, which means that in total 200 rats were tested. Because only one negative control was used, this control group served as a reference for all nine treatments. As rats, or any other animal, have a serious chance of spontaneously developing different pathologies with age, this one negative control is far too little. On the basis of simple probability calculations, it is already obvious that the chances of finding effects in the group of treated animals are much higher than finding effects in the control group. There should at least have been one control per treatment group. This illustrates the first fundamental flaw of the study: there are too little negative controls....
...Seralini et al. fed rats during a period of two years with either NK603 maize (three different doses: 11%; 22% or 33%), with NK603 maize (the same three different doses) that was sprayed with Roundup, or provided drinking water in which Roundup was present (three different concentrations: 50 ng/l, 400 mg/l or 2.25 g/l). In this way nine different treatments were tested and were compared with only one negative control, being a diet with one dose (i.e. 33%) of near-isogenic non-GM maize. Each diet was fed to 10 male and 10 female rats, which means that in total 200 rats were tested. Because only one negative control was used, this control group served as a reference for all nine treatments. As rats, or any other animal, have a serious chance of spontaneously developing different pathologies with age, this one negative control is far too little. On the basis of simple probability calculations, it is already obvious that the chances of finding effects in the group of treated animals are much higher than finding effects in the control group. There should at least have been one control per treatment group. This illustrates the first fundamental flaw of the study: there are too little negative controls....
Berry C. (2012)
Letter to the Editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.053
Dung LT, Ham LY. (2012)
Comments on “Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize”.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.054
Heinemann J. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.055
External URL: http://www.sciencedirect.com/
...A number of criticisms of this paper have appeared in the media. They appeared shockingly quickly which caused me concern because I find it takes more time to properly and thoroughly read a scientific paper of this complexity. Most criticisms were of a general nature and without substance worthy of entering the scientific debate. Some criticisms were specific, referring to the type of rat used, the kind of statistical analysis, and the interpretation of the response to increasing concentrations of the agrichemicals, Roundup, or genetically modified plant ingredient...
...The proper pathway forward is for any uncertainty in the findings to be put to rest through: the establishment of a consensus protocol developed through a transparent and openly peer-reviewed methodology; definitive study using this protocol to be conducted by industry-independent scientists of appropriate qualifications, such as Seralini et al., with reasonable access for observation by those nominated by the industry and regulatory communities...
...A number of criticisms of this paper have appeared in the media. They appeared shockingly quickly which caused me concern because I find it takes more time to properly and thoroughly read a scientific paper of this complexity. Most criticisms were of a general nature and without substance worthy of entering the scientific debate. Some criticisms were specific, referring to the type of rat used, the kind of statistical analysis, and the interpretation of the response to increasing concentrations of the agrichemicals, Roundup, or genetically modified plant ingredient...
...The proper pathway forward is for any uncertainty in the findings to be put to rest through: the establishment of a consensus protocol developed through a transparent and openly peer-reviewed methodology; definitive study using this protocol to be conducted by industry-independent scientists of appropriate qualifications, such as Seralini et al., with reasonable access for observation by those nominated by the industry and regulatory communities...
Langridge P. (2012)
Letter to the editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.056
External URL: http://www.sciencedirect.com/
...Further, the comment “All data cannot be shown in one report, and the most relevant are described here” is clearly unacceptable. I find it hard to believe that your journal is unable to support supplementary data files, particularly since these are unlikely to be substantial relative to other studies. I have joined colleagues in signing a petition to urge Dr. Séralini to make all of the data publically available. It is unfortunate that the scientific community should need to initiate such a petition. Full release of data should have been a requirement of your journal...
...Further, the comment “All data cannot be shown in one report, and the most relevant are described here” is clearly unacceptable. I find it hard to believe that your journal is unable to support supplementary data files, particularly since these are unlikely to be substantial relative to other studies. I have joined colleagues in signing a petition to urge Dr. Séralini to make all of the data publically available. It is unfortunate that the scientific community should need to initiate such a petition. Full release of data should have been a requirement of your journal...
De Souza L.. (2012)
Letter to the Editor.
Food and Chemical Toxicology DOI: 10.1016/j.fct.2012.10.057
No comments:
Post a Comment