Letter to the Editor

An extensive review article titled “Issues in the Design and Interpretation of Chronic Toxicity and Carcinogenicity Studies in Rodents: Approaches to Dose Selection” was published recently in Critical Reviews in Toxicology (37:729–836, 2007). The authors include prominent representatives of the U.S. FDA, EPA, the chemical and pharmaceutical industries, and academia. This article goes far beyond issues of dose selection alone and considers different goals of carcinogenicity studies in different settings, additional nonstandard endpoints that may be added to achieve specific objectives, and different study designs that may be employed. The article includes an extensive discussion of evaluating and estimating the maximum tolerated dose in range finding and pivotal studies. Those who have an interest in designing and interpreting rodent carcinogenicity studies are encouraged to seek out and review this paper. Great volumes of information, interpretative perspective, and practical examples are available within this single source. In an attempt to be inclusive of all possible objectives of a carcinogenicity study, the authors suggest numerous options that do not represent standard practice for studies intended primarily to identify carcinogenic potential of industrial chemicals, consumer products, or pharmaceuticals. An uninitiated reader may conclude that some endpoints better evaluated in shorter studies, such as clinical pathology parameters, organ weights, and ophthalmic findings, should be standard components of rodent carcinogenicity studies. The Scientific and Regulatory Policy Committee of the Society of Toxicologic Pathology has carefully reviewed the article and offered its opinions as a letter to the editor of Critical Reviews in Toxicology with the request that the letter be published in an issue of that journal.

This letter, including references, is inserted below:

In their manuscript entitled “Issues in the Design and Interpretation of Chronic Toxicity and Carcinogenicity Studies in Rodents: Approaches to Dose Selection” (Critical Reviews in Toxicology, 37:729–837, 2007), LR Rhomberg et al. presented recommendations in the design and interpretation of chronic rodent studies. This manuscript, whose authors include employees of regulatory agencies, is extensive (110 pp) and comprehensive. We commend the efforts of these authors to give assistance in the complicated design and evaluation of carcinogenicity studies.

The Scientific and Regulatory Policy Committee (SRPC) under the guidance of the Society of Toxicologic Pathology (STP) and the Guideline Committee (GC) of the European Society of Toxicologic Pathology (ESTP) are concerned that readers will be left with the impression that this document is officially sanctioned by regulatory authorities, and that its contents are recommendations from regulatory authorities. In addition, some readers may misinterpret the authors’ recommendations regarding the addition of non-standard endpoints to these studies and interpretation of the MTD.

Our concerns are three-fold. First, while the authors clearly state that the endpoints must be carefully selected to achieve the specific aims of the study, the document appears to promote the routine collection of organ weights, clinical pathology data, and other ancillary endpoints in carcinogenicity studies. The effects of aging and intercurrent disease create such variability in clinical pathology and organ weight parameters beyond 12 months of dosing that the data would be uninterpretable, uninformative, or misleading (Weingand et al., Long and Symanowski). In Appendix 2 it is mentioned that clinical pathology assessments are not appropriate for two-year rodent bioassays. This point is not highlighted sufficiently nor brought into the body of the text. The SRPC and GC believe that organ weights and clinical pathology assessments should generally not be recommended nor required for two-year bioassays. Recently, the STP has published its recommendation on organ weights in toxicology studies, which includes the recommendation that they not be performed in two-year rodent bioassays (Sellers et al.).

Second, some definitions for identification of the maximum tolerated dose (MTD) based on pathology endpoints may be too conservative. The FDA/CFSAN Redbook recommends that the high dose (MTD) “should be sufficiently high to induce toxic responses in test animals, and should not cause fatalities high enough to prevent meaningful evaluation of the data from the study” (Redbook 2000). Similarly, the National Toxicology Program defines the MTD, for the purposes of dose selection for a two-year rodent bioassay as “that dose which, when given for the duration of the chronic study as the highest dose, will not shorten the treated animals’ longevity from any toxic effects other than the induction of neoplasms.” These are also consistent with the ICH guidances. A weight-of-evidence approach should be used to determine the MTD. In some cases, setting the MTD based on a single microscopic or hematologic finding may result in selection of a high dose that is insufficient to maximize detection of carcinogenic effects.

Third, the SRPC and GC recognize that, whenever possible, information derived from studies should be maximized so that we may minimize animal use. However, some of the study designs suggested may inadvertently cause increased animal use by attempting to do too much in a single study and thus potentially compromising the primary intent. The SRPC and GC propose that combining multiple purposes in a single study should be done cautiously. For example, if a study was designed to identify inflection points in pharmacokinetics/pharmacodynamics and simultaneously be used for determination of carcinogenesis, there is potential that the study could fail on both counts if sufficiently detailed information from shorter-term studies was not available. Even if such data were available, toxicologists and pathologists are occasionally surprised by the lack of repeatability of observations as studies become longer in duration.

We hope that the editors will publish this letter as a reminder to readers that endpoints unrelated to carcinogenicity assessment should be added cautiously to carcinogenicity studies, that interpretation of the MTD should be flexibly applied to meet the study objectives, and that specific targeted studies are often a better approach to studying mechanistic, pharmacokinetic, and pharmacodynamic responses than incorporating these endpoints into a rodent carcinogenicity study.