The simple answer is: The criticism of Séralini’s use of rats is not about the fact that they were Sprague Dawley (SD) rats. It is that the number of SD rats he used was not appropriate to draw the conclusions he did. SD rats are acceptable to use in carcinogenicity tests as long as the experiment is designed to account for the fact that SD rats are known to have a high rate of certain spontaneous diseases (e.g., mammary tumors) (Brix et al., 2005).
The more technical explanation is: Spontaneous disease is not unique to SD rats; other rat strains have high background rates of different diseases as well. Thus, rodent carcinogenicity studies must be appropriately populated (i.e., have sufficient statistical power), 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.
Perhaps a good way to illustrate this is to consider an extreme example:
A group of researchers want to use noninvasive techniques (e.g., ultrasound) to determine if a certain diet increases heart damage in dogs, and need to pick a breed to study. One breed, the Cavalier King Charles spaniel, is known to get heart damage—usually involving the mitral valve—at a rate of approximately 90 percent. Given this high rate of background incidence, it would be difficult to pick up an increase in mitral valve damage in the animals fed the test diet. However, the researchers could address this issue through two remedies: 1) use another breed, one that does not have a high background incidence of heart damage or 2) use the Cavalier King Charles spaniel, but use more dogs. More dogs lead to greater statistical power to detect differences, even with the background incidence.
This is especially critical in the field of toxicology, where it has long been established that the dose is important (i.e., more of a toxic substance is more toxic than less of that substance). This problem can be solved 1) by using a strain with a lower expected incidence of a certain finding (this can be difficult to determine ahead of time) or 2) by using more rats (referring back to the Séralini study using the SD rats) to improve the ability to detect the difference between treatment groups.
For the reasons provided above, “US (US EPA, 1998; 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 ...’” (Hammond et al., 2013). The Séralini study (2012) included only 10 rats/sex/group, and this low number is inadequate to make meaningful comparisons in tumor incidence between groups at the end of a chronic study—especially when one considers that older female SD rats are known to have high spontaneous rates of mammary and pituitary tumors (EFSA, 2012; Hammond et al., 2013). EFSA (2012) pointed out that these facts were ignored in the discussion of findings in the Séralini publication (2012), in which the authors claimed that the mammary and pituitary tumors observed in test animals were treatment related.
These OECD guidelines are very familiar to toxicologists. Considering that all of the information above is publicly available, one can only wonder why Séralini et al. chose not to discuss their findings in the context of existing knowledge of spontaneous disease rates in SD rats.
Brix, A.E., Nyska, A., Haseman, J.K., Sells, D.M., Jokinen, M.P., Walker, N.J., 2005. Incidences of selected lesions in control female Harlan Sprague-Dawley rats from two-year studies performed by the National Toxicology Program. Toxicol. Pathol. 33, 477–483.
US EPA, 1998. Carcinogenicity. Health Effects Test Guidelines OPPTS 870.4200. August 1998.
EFSA, 2012. Statement of EFSA: Review of the Séralini et al. (2012) publication on a 2-year rodent feeding study with glyphosate formulations and GM maize NK603 as published online on 19 September 2012 in Food and Chemical Toxicology. EFSA J. 10, 2910.
FDA, 2006. Carcinogenicity Studies with Rodents. Redbook 2000: C.6. January 2006 (Chapter 4).
OECD, 1995a. Carcinogenicity Studies. OECD Guideline for the Testing of Chemicals, No. 451, adopted 07 September 2009.
OECD, 1995b. Guidance Document 116 on the Conduct and Design of Chronic Toxicity and Carcinogenicity Studies, Supporting Test Guidelines 451, 452 and 453, second ed. OECD Series on Testing and Assessment, No. 116, adopted 13 April 2012.
Hammond, B.G., Goldstein, D.A., Saltmiras, D.A., 2013. Response to original research article, in press, corrected proof, ‘‘Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize’’. Food Chem. Toxicol. 53, 459-464.
Séralini, G.-E., Clair, E., Mesnage, R., Gress, S., Defarge, N., Malatesta, M., Hennequin, D., de Vendômois, J-S., 2012. Long term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize. Food Chem. Toxicol. 50, 4221-4231.