ANTICARCINOGENIC RESPONSES IN RODENT CANCER BIOASSAYS ARE NOT EXPLAINED BY RANDOM EFFECTS

Anticarcinogenicity in a long-term rodent bioassay is defined as a sta tistically significant decrease of a specific tumor type in a dosed gr oup following chemical exposure. About 92% of chemicals tested by the National Toxicology Program prior to 1983 reveal at least one site wit h a significant (p less than or equal to 0.05) tumor rate decrease in one or more tested groups, a result consistent with those of J, K. Has eman and F, M. Johnson (1996, Mutat. Res. 350, 131-141) for a database of recently tested chemicals, Detection of tumor decreases in a speci fic site can be explained not only by biological effects, but also as a result of random variability in the background tumor rates, decrease s in body weight, or decreases in survival of treated animals. This pa per evaluates the rate of false-positive anticarcinogenic findings due to random effects (variations in tumor rates and the multiple compari sons undertaken in evaluating a bioassay), while a companion paper add resses the influence of weight and survival depression. Monte-Carlo si mulation was conducted to assess the contribution of random effects. T his contribution was found to be important even when a statistical sig nificance cutoff of p(o) less than or equal to 0.05 was chosen. If a m ore stringent statistical criterion was used (p(o) less than or equal to 0.01 or p less than or equal to 0.005), the proportion of false pos itive determinations diminishes. The number of anticarcinogens in the database remains substantially higher than predicted by the simulation s. An examination of the distribution of all p values (T. Schweder and E. Spjotvoll, 1982, Biometrika 69, 493-502) also indicates that stati stically significant anticarcinogenic responses are found in the datab ase at a higher rate than would result from purely random responses. F inally, the cross-species prediction of anticarcinogenic responses was examined in a manner similar to a study of cross-species prediction o f carcinogenic responses (G. M. Gray et al., 1995, Reg. Toxicol. Pharm acol. 20, 281-301). The analyses show that anticarcinogenic effects in one rodent species predict well anticarcinogenic effect in another ro dent species. It seems likely that biological factors are involved in anticarcinogenic responses observed in rodent cancer bioassays. (C) 19 98 Society of Toxicology.