Teratogenic evaluation of 2,4,5-T.

The herbicide 2,4,5-trichlorophenoxyacetic acid is teratogenic and fetocidal in two strains of mice when administered either subcutaneously or orally and in one strain of rats when administered orally. The incidences of both cystic kidney and cleft palate were increased in the C57BL/6 mice as well as the incidence of cleft palate in the AKR mice. The incidence of cystic kidney was also increased in the rats. In addition, an increase in the ratio of liver weight to body weight in the mouse fetus and the occurrence of hemorrhagic gastrointestinal tract in the rat fetus suggest that this compound also has fetotoxic properties.

Report of an Expert Panel on the reanalysis by of a 90-day study conducted by Monsanto in support of the safety of a genetically modified corn variety (MON 863).

MON 863, a genetically engineered corn variety that contains the gene for modified Bacillus thuringiensis Cry3Bb1 protein to protect against corn rootworm, was tested in a 90-day toxicity study as part of the process to gain regulatory approval. This study was reanalyzed by Séralini et al. who contended that the study showed possible hepatorenal effects of MON 863. An Expert Panel was convened to assess the original study results as analyzed by the Monsanto Company and the reanalysis conducted by Séralini et al. The Expert Panel concludes that the Séralini et al. reanalysis provided no evidence to indicate that MON 863 was associated with adverse effects in the 90-day rat study. In each case, statistical findings reported by both Monsanto and Séralini et al. were considered to be unrelated to treatment or of no biological or clinical importance because they failed to demonstrate a dose-response relationship, reproducibility over time, association with other relevant changes (e.g., histopathology), occurrence in both sexes, difference outside the normal range of variation, or biological plausibility with respect to cause-and-effect. The Séralini et al. reanalysis does not advance any new scientific data to indicate that MON 863 caused adverse effects in the 90-day rat study.

Occurrence of tumors among litters of BALB/c female mice.

In the application of statistical techniques to tumor incidence data it is generally assumed that animals respond independently with regard to tumor occurrence, with littermates being no more alike than are animals from different litters. Data for nine different types of tumors from the large ED01 study conducted at the National Center for Toxicological Research were used to compare tumor prevalence rates among litters with the tumor rates expected under the assumption of homogeneity of tumor rates among litters. These data did not provide sufficient evidence to reject the assumption of homogeneity of tumor prevalence rates among litters of inbred female BALB/c mice for either spontaneously occurring tumors or bladder tumors produced by exposure to 2-acetylaminofluorene (CAS: 53-96-3). However, there does appear to be a difference in liver tumor prevalence rates among litters at 24 months of age. Thus litter effects are a factor that should be considered in the assignment of animals to treatment groups in carcinogenesis studies.

Influence of cause-of-death assignment on time-to-tumor analyses in animal carcinogenesis studies.

The importance of cause-of-death determination in an animal carcinogenesis study with respect to estimation of time-to-tumor distributions of internally occurring (occult) tumors is discussed. A nontechnical description of time-to-tumor estimation is presented. The information obtained from time-to-tumor estimation when cause-of-death designation was used is illustrated for liver tumors in female mice of the inbred strain BALB/cStCrlfC3Hf/Nctr from the ED01 study with N-2-fluorenylacetamide done at the National Center for Toxicological Research. A time-to-tumor analysis of reticulum cell sarcoma data from the same study has provided insight into some difficulties involved in routine case-by-case determination of cause of death. A more flexible system for assigning of cause of death to dead animals and cause of morbidity to moribund animals is described as a way to improve cause-of-death assignment.

Nonlinearity and thresholds in dose-response relationships for carcinogenicity due to sampling variation, logarithmic dose scaling, or small differences in individual susceptibility.

Nonlinear and threshold-like shapes of dose-response curves are often observed in tests for carcinogenicity. Here, we present three examples where an apparent threshold is spurious and can be misleading for low dose extrapolation and human cancer risk assessment. Case #1: For experiments that are not replicated, such as rodent bioassays for carcinogenicity, random variation can lead to misinterpretation of the result. This situation was simulated by 20 random binomial samplings of 50 animals per group, assuming a true linear dose response from 5% to 25% tumor incidence at arbitrary dose levels 0, 0.5, 1, 2, and 4. Linearity was suggested only by 8 of the 20 simulations. Four simulations did not reveal the carcinogenicity at all. Three exhibited thresholds, two showed a nonmonotonic behavior with a decrease at low dose, followed by a significant increase at high dose ("hormesis"). Case #2: Logarithmic representation of the dose axis transforms a straight line into a sublinear (up-bent) curve, which can be misinterpreted to indicate a threshold. This is most pronounced if the dose scale includes a wide low dose range. Linear regression of net tumor incidences and intersection with the dose axis results in an apparent threshold, even with an underlying true linear dose-incidence relationship. Case #3: Nonlinear shapes of dose-cancer incidence curves are rarely seen with epidemiological data in humans. The discrepancy to data in rodents may in part be explained by a wider span of individual susceptibilities for tumor induction in humans due to more diverse genetic background and modulation by co-carcinogenic lifestyle factors. Linear extrapolation of a human cancer risk could therefore be appropriate even if animal bioassays show nonlinearity.

Abnormal folate metabolism and mutation in the methylenetetrahydrofolate reductase gene may be maternal risk factors for Down syndrome.

BACKGROUND: Down syndrome, or trisomy 21, is a complex genetic disease resulting from the presence of 3 copies of chromosome 21. The origin of the extra chromosome is maternal in 95% of cases and is due to the failure of normal chromosomal segregation during meiosis. Although advanced maternal age is a major risk factor for trisomy 21, most children with Down syndrome are born to mothers <30 y of age. OBJECTIVE: On the basis of evidence that abnormal folate and methyl metabolism can lead to DNA hypomethylation and abnormal chromosomal segregation, we hypothesized that the C-to-T substitution at nucleotide 677 (677C-->T) mutation of the methylenetetrahydrofolate reductase (MTHFR) gene may be a risk factor for maternal meiotic nondisjunction and Down syndrome in young mothers. DESIGN: The frequency of the MTHFR 677C-->T mutation was evaluated in 57 mothers of children with Down syndrome and in 50 age-matched control mothers. Ratios of plasma homocysteine to methionine and lymphocyte methotrexate cytotoxicity were measured as indicators of functional folate status. RESULTS: A significant increase in plasma homocysteine concentrations and lymphocyte methotrexate cytotoxicity was observed in the mothers of children with Down syndrome, consistent with abnormal folate and methyl metabolism. Mothers with the 677C-->T mutation had a 2.6-fold higher risk of having a child with Down syndrome than did mothers without the T substitution (odds ratio: 2.6; 95% CI: 1.2, 5.8; P < 0.03). CONCLUSION: The results of this initial study indicate that folate metabolism is abnormal in mothers of children with Down syndrome and that this may be explained, in part, by a mutation in the MTHFR gene.

Alternative tests: carcinogenesis as an example.

Acceptance of new tests that are alternatives to currently used toxicology tests is a topic of considerable importance in the field of toxicology. Carcinogenicity testing today normally includes 2-year studies in rats and mice of both sexes, following widely accepted procedures for husbandry; selection of dose levels; pathology and toxicity observations; and statistical interpretation of tumor data. These studies are usually preceded by tests for genetic toxicity and subchronic toxicity studies to select dose levels for the 2-year studies. Although these data are used for quantitative risk assessment, the mechanistic basis for effects is usually unknown. The series of studies is very expensive and requires 5 years or more to conduct. Alternative approaches are being developed that would provide more mechanistic information and hopefully would permit decisions to be made about carcinogenic potential without the need to conduct 2-year studies in rats and mice of both sexes. Decisions could be based on a profile of data rather than on the result of one test. Procedures for regulatory acceptance of new approaches for carcinogenicity testing are critical to future progress.

Quantitative risk analysis for quantal reproductive and developmental effects.

Animal experiments are generally conducted at higher dose levels than anticipated human dose levels in order to elicit otherwise subtle changes in reproduction or developmental effects with relatively few animals. Based on animal data, regulatory strategy generally has been to postulate a no-observed-effect level (NOEL) for toxic effects and to divide this by a safety factor, usually 100, to establish acceptable levels for humans. Various authors have discussed the shortcomings of using NOEL and have suggested the use of an estimable effect level determined from a dose-response curve fitted to bioassay data, e.g., the dose at which 1% of the animals are adversely affected, and employing some form of conservative low dose extrapolation to control risks at lower doses. In this paper, 10 sets of bioassay data on fetal mortality or anomalies were used to compare the estimated upper limits of risk estimated at the NOEL/100 and the lower 95% confidence limit estimate of the dose producing adverse effects in 1% of the embryonic implants or fetuses divided by 100 (LED01/100). The latter quantity is expected to result in a risk (proportion affected) of less than 10(-4) (1 in 10,000). The estimated upper limits of risk associated with the NOEL/100 were from 2 x 10(-4) to 6 x 10(-4) for the 10 data sets investigated.

Carcinogenic risk assessment: comparison of estimated safe doses for rats and mice.

Data from the National Cancer Institute/National Toxicology Program (NCI/NTP) carcinogenesis bioassays were examined to compare cancer risks in rats and mice. Only those bioassays where chemicals were administered orally were used. The ratios for rats to mice of the virtually safe dose (VSD) levels associated with a risk of 10(-6) were compared. Comparisons of the ratios were made for those chemicals that NCI/NTP determined to be carcinogenic in at least one species and that showed a dose response trend in the same sex at the same tissue/organ site in the other species. In all, 69 comparisons from 38 carcinogens were performed. The overall geometric mean of the VSD ratios is 1.27 in terms of concentration (ppm); the mean and the standard deviation in logarithm are 0.24 and 1.83, respectively. The VSD ratios vary from 1:51 to 49:1. Without the restriction of the same sex and site, the geometric mean of the minimum VSDs is 1.38, and the standard deviation in logarithm is 1.79. By directly comparing the VSDs for rats and mice (as they are performed for risk assessment), this study showed a probability of 0.10 that the ratio of VSDs is greater than 10, and the ratio is greater than 20 with a probability of 0.05 when a chemical is carcinogenic in both species.

A model-free approach to low-dose extrapolation.

Estimates of risk associated with exposure to low levels of carcinogenic substances present in the environment are generally obtained by linear extrapolation from higher exposure levels at which risks can be estimated directly. In this paper, we examine the scientific basis for the assumption of low-dose linearity in carcinogenic risk assessment and the different statistical methods that have been proposed for linear extrapolation. A model-free approach to linear extrapolation is described and illustrated using epidemiological data on radiation carcinogenesis. The statistical properties of this method are empirically assessed using 572 selected sets of bioassay data.

Application of biomarkers to risk assessment.

Due to difficulties in conducting epidemiological studies, most estimates of cancer risk are based on data from animal bioassays. Extrapolation of cancer risk estimates in animals to humans requires an assumption of equal potency across species based on the average daily dose. The purpose of this paper is to examine the ability to predict tumor incidence across species from DNA adduct concentrations resulting from exposure to carcinogens. A 100-fold range of structurally diverse adduct concentrations corresponding to the same tumor incidence raises questions about quantitative predictability across chemical classes and across species. Differences in adduct structure, mutagenic efficiency, adduct repair rates, and cellular proliferation could account for some of the differences. For specific carcinogen-DNA adducts, the steady-state levels associated with a 50% tumor incidence appear to vary over a narrower range. An equal incidence of liver tumors was obtained at equal concentrations of aflatoxin B1-DNA adducts for rats and trout. A 2- to 3-fold range of 4-aminobiphenyl-DNA adduct concentrations between mice and dogs appears to be associated with nearly equal bladder tumor incidence, on the basis of limited data. In humans, a 5-fold higher concentration of a 4-aminobiphenyl-DNA adduct in bladders of smokers than of nonsmokers is compatible with the relative risk of bladder cancer due to smoking. DNA adduct concentrations certainly can be used to improve quantification of chemical exposures for epidemiological studies. Although promising, more data are needed to judge the usefulness of DNA adduct concentrations to predict cancer incidence across species.

U.S. Food and Drug Administration perspective of the inclusion of effects of low-level exposures in safety and risk assessment.

A brief overview is provided of some of the general safety and risk assessment procedures used by the different centers of the U.S. Food and Drug Administration (U.S. FDA) to evaluate low-level exposures. The U.S. FDA protects public health by regulating a wide variety of consumer products including foods, human and animal drugs, biologics, and medical devices under the federal Food, Drug, and Cosmetic Act. The diverse legal and regulatory standards in the act allow for the consideration of benefits for some products (e.g., drugs) but preclude them from others (e.g., food additives). When not precluded by statutory mandates (e.g., Delaney prohibition), the U.S. FDA considers both physiologic adaptive responses and beneficial effects. For the basic safety assessment paradigm as presently used, for example in the premarket approval of food additives, the emphasis is on the identification of adverse effects and no observed adverse effect level(s) (NOAEL). Generally, the NOAEL is divided by safety factors to establish an acceptable exposure level. This safety assessment paradigm does not preclude the consideration of effects whether they are biologically adaptive or beneficial at lower dose levels. The flexibility to consider issues such as mechanisms of action and adaptive and beneficial responses depends on the product under consideration. For carcinogenic contaminants and radiation from medical devices, the U.S. FDA considers the potential cancer risk at low exposure levels. This generally involves downward extrapolation from the observed dose-response range. The consideration of adverse effects of other toxicologic end points (e.g., reproductive, immunologic, neurologic, developmental) associated with low exposure levels is also becoming more of a reality (e.g., endocrine disrupters). The evaluation of the biologic effects of low-level exposures to toxic substances must include whether the effect is adverse or a normal physiologic adaptive response and also determine the resiliency of a physiologic system. The public health mandate of the U.S. FDA includes an active research program at the National Center for Toxicological Research and the other U.S. FDA centers to support the regulatory mission of the U.S. FDA. This includes the development of knowledge bases, predictive strategies, and toxicologic studies to investigate effects at the lower end of the dose-response range. Because of the wide diversity of legal and regulatory standards for various products regulated by the U.S. FDA agency-wide safety and risk assessment procedures and policies generally do not exist.

Upregulation of apoptosis with dietary restriction: implications for carcinogenesis and aging.

The maintenance of cell number homeostasis in normal tissues reflects a highly regulated balance between the rates of cell proliferation and cell death. Under pathologic conditions such as exposure to cytotoxic, genotoxic, or nongenotoxic agents, an imbalance in these rates may indicate subsequent risk of carcinogenesis. Apoptotic cell death, as opposed to necrotic cell death, provides a protective mechanism by selective elimination of senescent, preneoplastic, or superfluous cells that could negatively affect normal function and/or promote cell transformation. The relative efficiency or dysfunction of the cell death program could therefore have a direct impact on the risk of degenerative or neoplastic disease. Dietary restriction of rodents is a noninvasive intervention that has been reproducibly shown to retard tumor development and most physiologic indices of aging relative to ad libitum-fed animals. As such, it provides a powerful model in which to study common mechanistic processes associated with both aging and cancer. In a recent study we established that chronic dietary restriction (DR) induces an increase in spontaneous apoptotic rate and a decrease in cell proliferation rate in hepatocytes of 12-month-old B6C3F1 DR mice relative to ad libitum (AL)-fed mice. This diet-induced shift in cell death/proliferation rates was associated with a marked reduction in subsequent development of spontaneous hepatoma and a marked increase in disease-free life span in DR relative to AL-fed mice. These results suggest that total caloric intake may modulate the rates of cell death and proliferation in a direction consistent with a cancer-protective effect in DR mice and a cancer-promoting effect in AL mice. To determine whether the increase in spontaneous apoptotic rate was maintained over the life span of DR mice, apoptotic rates were quantified in 12-, 18-, 24- and 30-month-old DR and AL mice. The rate of apoptosis was elevated with age in both diet groups; however, the rate of apoptosis was significantly and consistently higher in DR mice regardless of age. In double-labeling experiments, an age-associated increase in the glutathione S-transferase-II expression in putative preneoplastic hepatocytes in AL mice was rapidly reduced by apoptosis upon initiation of DR. Thus, intervention that promote a low-level increase in apoptotic cell death may be expected to protect genotypic and phenotypic stability with age. If during tumor promotion an adaptive increase in apoptosis effectively balances the dysregulated increase proliferation, the risk of permanent genetic error and carcinogenesis would be minimized.

No threshold dose for estradiol-induced sex reversal of turtle embryos: how little is too much?

Risk assessments for nongenotoxic chemicals assume a threshold below which no adverse outcomes are seen. However, when an endogenous chemical, such as 17ss-estradiol (E2), occurs at a concentration sufficient to cause an effect, the threshold is already exceeded. Under these circumstances, exogenous estradiol is not expected to provide a threshold dose. This principle is demonstrated for E2 in the red-eared slider, a turtle with temperature-dependent sex determination. In this species, gonadal sex is determined by egg incubation temperature; female development requires endogenous estrogen produced by elevated temperature. While normal production of females by endogenous estrogens is not an adverse effect, exogenous estrogens can sex reverse presumptive males, which can be an adverse effect. A large dose-response study was conducted using seven doses and a vehicle control (starting n = 300/group); a single E2 dose was applied to the eggshell of recently laid eggs. Animals were sexed after hatching. The incubation temperature chosen, 28.6 degrees C, generates a minority of females. Thus, the criteria for testing the threshold hypothesis were met, i.e., there is evidence that there is endogenous estrogen and that it generates an irreversible response. The lowest E2 dose tested, 400 pg/egg (40 ng/kg), sex reversed 14.4% of the animals, demonstrating very low dose sensitivity. The data were fit with a modified Michaelis-Menten equation, which provided an estimate of 1.7 ng/egg for endogenous estradiol. The median effective dose (ED50) was 5.0 +/- 2.0 ng/egg (95% confidence limits), of which 1.7 ng/egg was endogenous estradiol and 3.3 ng/egg came from the applied estradiol. There was no apparent threshold dose for E2. A smaller replication confirmed these results. These results provide a simple biologically based dose-response model and suggest that chemicals which act mechanistically like E2 may also show no threshold dose. If so, even low environmental concentrations of such chemicals may carry risk for sex reversal.

Tumors and DNA adducts in mice exposed to benzo[a]pyrene and coal tars: implications for risk assessment.

Current methods to estimate the quantitative cancer risk of complex mixtures of polycyclic aromatic hydrocarbons (PAH) such as coal tar assume that overall potency can be derived from knowledge of the concentration of a few carcinogenic components such as benzo[a]pyrene (B[a]P). Genotoxic damage, such as DNA adducts, is thought to be an essential aspect of PAH-induced tumorigenesis and could be a biomarker for exposure useful for estimating risk. However, the role of B[a]P and the relationship of adduct formation in tumorigenesis have not been tested rigorously in models appropriate for human health risk assessment. Therefore, we directly compared tumor induction and adduct formation by B[a]P and coal tars in several experimental protocols, including one broadly accepted and used by regulators. We found that B[a]P content did not account for tumor incidences after exposure to coal tars. DNA adducts were found in both tumors and tumor-free tissue and tumor outcomes were not predicted by either quantitation of total DNA adducts or by the DNA adduct formed by B[a]P. These data suggest that risk assessments based on B[a]P content may not predict accurately risk to human health posed by environmental PAH.

Threshold analysis of selected dose-response data for endocrine active chemicals.

Using a biologically relevant mathematical model, the Michaelis-Menten equation, we examined published data from endocrine active chemicals for evidence of no-threshold dose-response curves. Data were fit to a modified Michaelis-Menten equation which accounted for total background response. Subsequently, the data sets were analyzed using non-linear regression in order to estimate the four parameters of interest (non-hormone controlled background (Bnh), maximum response (Rmax), endogenous hormone level (D0), and the dose at which a half-maximal response was observed (ED50)) and to determine the fit to the fully modified Michaelis-Menten equation. Subsequently, response data were adjusted to account for Bnh and then normalized to Rmax, while dose data were adjusted to account for D0 and then normalized to the ED50. This data set was combined into a single, composite data set and fit to the fully modified Michaelis-Menten equation. We examined 31 data sets (24 endpoints) from studies on 9 different chemical/hormone treatments. Twenty-six of the data sets fit the modified Michaelis-Menten equation with high multiple correlation coefficients (r>0.90). The normalized data demonstrated a good fit to the modified Michaelis-Menten equation. These results indicate that a variety of biological responses fit the modified Michaelis-Menten equation, which does not have a threshold dose term.

Risk assessment strategies for neuroprotective agents.

Neurotoxicity may be defined as any adverse effect on the structure or function of the central and/or peripheral nervous system by a biological, chemical, or physical agent. Neurotoxic effects may be permanent or reversible, produced by neuropharmacological or neurodegenerative properties of a neurotoxicant, or the result of direct or indirect actions on the nervous system. A multidisciplinary approach is necessary to assess neurotoxicity because of the complexity and diverse functions of the nervous system. Many of the relevant effects can be measured directly by neurochemical, neurophysiological, and neuropathological techniques, whereas, others must be inferred from observed behavior. Some neurotoxicological data can be derived directly from humans. Neurotoxicity in humans is most commonly measured by relatively noninvasive neurophysiologic and neurobehavioral methods that assess cognitive, affective, sensory, and motor function. For most toxicological assessments, however, it is necessary to rely on information derived from animal models. There are many approaches that can be used to assess neurotoxicity, including whole animal (in vivo) and tissue/cell culture (in vitro) testing. Neurotoxicity can be described at multiple levels of organization, including neurochemical, anatomical, physiological, and behavioral. An important aspect of neurotoxic endpoint evaluation involves risk assessment procedures. Risk assessment may be defined as an empirically-based process used to determine the probability that adverse or abnormal effects are associated with exposure to a chemical, physical or biological agent. Risk management, on the other hand, is the process that applies information obtained through the risk assessment process to determine whether the assessed risk should be reduced and, if so, to what extent. For chemicals such as neuroprotective agents and other drugs designed to provide therapeutic benefits, information concerning these benefits is considered during the risk management phase. The risk assessment process usually involves four steps: hazard identification, dose-response assessment, exposure assessment, and risk characterization. Neurotoxicity risk assessment models of the future may well include biomarkers of both effect and exposure as well as biologically-based mechanistic and pharmacokinetic considerations derived from both epidemiologic and experimental data.

Combining uncertainty factors in deriving human exposure levels of noncarcinogenic toxicants.

Acceptable levels of human exposure to noncarcinogenic toxicants in environmental and occupational settings generally are derived by reducing experimental no-observed-adverse-effect levels (NOAELs) or benchmark doses (BDs) by a product of uncertainty factors (Barnes and Dourson, Ref. 1). These factors are presumed to ensure safety by accounting for uncertainty in dose extrapolation, uncertainty in duration extrapolation, differential sensitivity between humans and animals, and differential sensitivity among humans. The common default value for each uncertainty factor is 10. This paper shows how estimates of means and standard deviations of the approximately log-normal distributions of individual uncertainty factors can be used to estimate percentiles of the distribution of the product of uncertainty factors. An appropriately selected upper percentile, for example, 95th or 99th, of the distribution of the product can be used as a combined uncertainty factor to replace the conventional product of default factors.

Dose-response trend tests for tumorigenesis, adjusted for body weight.

Several studies have demonstrated a relationship between rodent body weight and tumor incidence for some tissue/organ sites. It is not uncommon for a chemical tested for carcinogenicity to also affect body weight. In such cases, comparisons of tumor incidence may be biased by body-weight differences across dose groups. A simple procedure was investigated for reducing this bias. This procedure divides the animals into a few groups based on body weight. Body weight at 12 months was used, before the appearance of a tumor was likely to affect body weight. Statistics for dose-response trend tests are calculated within body weight strata and pooled to obtain an overall dose-response trend test. This procedure is analogous to that currently used, of stratifying animals, based on their age at the time of removal from a study. Age stratification is used to account for differences in animal age across dose groups, which can affect comparisons of tumor incidence. Several examples were investigated where the high-dose group had reduced body weights and associated reductions in tumor incidence. When the data were analyzed by body-weight strata, some positive dose-response trends for tumor incidence were demonstrated. In one case, the weight-adjusted analysis indicated that a negative dose-response trend in tumor incidence was a real effect, in addition to a body weight reduction. These examples indicate that it is important to consider the effects of body weight changes as low as 10%, and perhaps below, that were caused by chemicals in 2-year bioassays for carcinogenesis. The simple procedure of analyzing tumor incidence within body-weight strata can reduce the bias introduced by weight differences across dose groups.

Dose-response trend tests for tumorigenesis adjusted for differences in survival and body weight across doses.

A relationship between rodent body weight and tumor incidence for some tissue/organ sites has been demonstrated in many studies. It is not uncommon for a chemical tested for carcinogenicity to also affect body weight due to toxicity and/or food consumption. In such cases, comparisons of tumor incidence may be biased by body weight differences across dose groups. A simple procedure was investigated for reducing this bias. This procedure divides the animals into a few groups on the basis of body weight. Body weight at 12 months was used, before the appearance of a tumor was likely to affect body weight. Statistics for dose-response trend tests are calculated within body weight strata and pooled to obtain an overall dose-response trend test. This procedure is analogous to stratifying animals on the basis of age at the time of removal from a study to account for differences in ages of animals across dose groups that can affect comparisons of tumor incidence. In this paper, differences in survival times of animals were adjusted by the Poly-3 technique used by the National Toxicology Program. This technique does not require the assignment of cause of death. Several examples from rodent chronic bioassays were investigated, where the high dose group had reduced body weights and associated reductions in tumor incidence. When we analyzed the data by body weight strata, some positive dose-response trends for tumor incidence were demonstrated. In one case, the body weight adjusted analysis indicated that a negative dose-response trend in tumor incidence was a real effect in addition to a body weight reduction. These examples indicate that it is important to consider the effects of body weight changes as low as 10%, and perhaps less, as possibly being caused by chemicals in 2-year bioassays for carcinogenesis. The simple procedure of analyzing tumor incidence within body weight strata can reduce the bias introduced by body weight differences across dose groups.