Expanded Statistical Decision Rules for Interpretations of Results of Rodent Carcinogenicity Studies of Pharmaceuticals

Abstract

The FDA draft Guidance for Industry, Statistical Aspects of the Design, Analysis, and Interpretation of Chronic Rodent Carcinogenicity Studies of Pharmaceuticals (US Department of Health and Human Services 2001) recommends that a positive dose–response alone in tumor incidence of an individual tumor type be tested at 0.005 or 0.025 and that a pairwise comparison alone of an increase in tumor incidence in a treated group over the control group of the tumor type be tested at 0.01 or 0.05 level of significance for a common or a rare tumor, respectively, in a standard NDA or IND submission with two chronic studies in both sexes of rats and mice. The use of these decision rules (levels of significance) in statistical tests of the drug effect on individual tumor types results in an overall false-positive rate of about 10% in such a submission. However, the decision rules for two types of submissions other than those with two chronic studies in rats and mice were either not discussed or barely mentioned in the 2001 draft guidance document. They are submissions including one chronic study in rats and one six-month study in transgenic mice and submissions including only one chronic study in either rats or mice.

Also after the issuance of the 2001 draft guidance document for a 90-day public comment, some practicing pharmacologists/toxicologists have used a so-called joint test in their final interpretation of carcinogenic effects of new drugs in order to further reduce the false-positive rate. In the joint test used by those pharmacologists/toxicologists, the results of both a trend test and a pairwise comparison test between the control and the high groups have to be statistically significant simultaneously at the above levels of significance for the separate tests recommended in the 2001 guidance document in order for the drug effect on the development of an individual tumor type to be considered as statistically significant. Results of simulation studies we have conducted show that there is a serious consequence of huge inflations of the false-negative rate in the use of the joint test with the levels of significance for separate tests in the final interpretation of the carcinogenicity potential of a new drug. The inflations are most serious in the situations in which the dose has the large effect on tumor prevalence. The inflation can be as high as 204.5% (or more than three times) of the false-negative rate when the trend test alone is performed in determining if the effect on the development of the individual tumor type is statistically significant.

This book chapter includes two major parts in our efforts to update and to expand the original decision rules in the 2001 draft guidance for industry document mentioned above. The first part, serving as the important statistical basis for the second part, includes presentations of results of our simulation studies investigating the impacts of using the levels of significance for the separate tests recommended in the 2001 draft guidance document in the joint test on the determination of carcinogenic potential of a new drug. The second part includes our recommended sets of expanded decision rules for the separate tests and the joint test to be used in the three types of new drug application submissions.

The article reflects the views of the authors and should not be construed to represent FDA’s views or policies. Parts of this book chapter were also included in a manuscript that have been published online in the Journal of Biopharmaceutical Statistics.