A case study of cancer data set combinations for PCBs.

Results of several animal bioassays have demonstrated the carcinogenic potential of polychlorinated biphenyl (PCB) mixtures. Although PCBs are no longer manufactured, cancer risk assessment for PCBs remains an important issue because of continued potential human exposure from many sources. The existing cancer risk estimate for PCBs used by the U.S. EPA is based on liver tumors observed in female Sprague-Dawley rats in a lifetime bioassay. Liver cancer has been observed in other long-term bioassays as well. In this case study, experimental designs and biological characteristics of the data from these studies were evaluated to determine whether a combination of the data sets is scientifically reasonable. A statistical analysis of the data sets based on likelihood ratio theory was used to assess the compatibility of individual data sets to a common multistage dose-response model. The results from these biological and statistical assessments suggest that at least two data sets could be combined to derive a quantitative risk estimate for PCBs. Increased confidence in the quantitative estimate would result from such combination because more data are being used to assess the dose-response relationship.

Bioavailability of cadmium in food and water: a case study on the derivation of relative bioavailability factors for inorganics and their relevance to the reference dose.

Published studies in which rats were exposed to CdCl2 in standard chow or drinking water were analyzed to compare the relative bioavailability of cadmium from the two media. Relative bioavailability was assessed from estimates of the rate of accumulation of cadmium in kidney cortex or liver. Data were grouped into tiers based on study design and reporting of data: Tier 1, identical experimental protocols and dosage can be estimated; Tier 2, very similar or identical protocols and dosage can be estimated; Tier 3, protocols may differ and dosage can be estimated; and Tier 4, protocols may differ and dosages cannot be estimated (but concentration of cadmium in food or water is reported). Tiers were nested, such that Tier 4 contained all relevant studies; Tier 3 included data sets from Tiers 1 and 2; and Tier 2 included the data set from Tier 1. Data within Tiers 1, 2, and 3 were subjected to a linear regression analysis with dosage as the independent variable and tissue accumulation rate as the dependent variable to determine whether bioavailability of cadmium was significantly different based on medium of administration. The results of this analysis show the following: (1) In rats receiving food and drinking water ad libitum, the bioavailability of cadmium in drinking water is not significantly different (P > 0.05) from the bioavailability of cadmium in food when dosages are less than 4 mg/kg body wt/day. (2) Cadmium decreases food and water consumption; therefore, assessments of relative bioavailability should be made based on actual dosage rather than exposure levels. (3) Diet composition and status of the gastrointestinal tract are probably a more important determinant of the bioavailability of cadmium than is the exposure medium. (4) Studies of the effect of total diet composition on bioavailability of cadmium may be more relevant than are studies of the effect of the exposure medium. It is concluded from this analysis that the bioavailability of cadmium in food is not different from that in water when diet is provided ad libitum. Therefore, we recommend that distinct RfDs for cadmium in food and drinking water should not be based on the assumption that the bioavailability of cadmium in drinking water is greater than that of cadmium in food.

Combination of cancer data in quantitative risk assessments: case study using bromodichloromethane.

There are often several data sets that may be used in developing a quantitative risk estimate for a carcinogen. These estimates are usually based, however, on the dose-response data for tumor incidences from a single sex/strain/species of animal. When appropriate, the use of more data should result in a higher level of confidence in the risk estimate. The decision to use more than one data set (e.g., representing different animal sexes, strains, species, or tumor sites) can be made following biological and statistical analyses of the compatibility of the these data sets. Biological analysis involves consideration of factors such as the relevance of the animal models, study design and execution, dose selection and route of administration, the mechanism of action of the agent, its pharmacokinetics, any species- and/or sex-specific effects, and tumor site specificity. If the biological analysis does not prohibit combining data sets, statistical compatibility of the data sets is then investigated. A generalized likelihood ratio test is proposed for determining the compatibility of different data sets with respect to a common dose-response model, such as the linearized multistage model. The biological and statistical factors influencing the decision to combine data sets are described, followed by a case study of bromodichloromethane.

Biological considerations for combining carcinogenicity data for quantitative risk assessment.

Quantitative risk assessments for carcinogens conducted by the U.S. EPA have most frequently been based on results of a bioassay from a single sex/strain/species of animal. In some cases, more than one data set derived from different sexes, strains, or species of animals is suitable for quantitative risk assessment. When there is no apparent difference among the data sets in sensitivity to the carcinogen, use of more of the available data should result in a higher level of confidence in the risk estimate. Several biological factors must be considered before combining data from different animal sexes, strains, species, or tumor sites. The relevance of the animal models, study design and execution, dose selection, and route of administration are factors which influence whether data from separate studies should be combined. The decision to combine data sets is also based on what is known of the mechanism of action of the agent, its pharmacokinetics, any species/sex specificity of the effect, and considerations regarding tumor site specificity. Statistical analysis also indicates whether the data sets may be described by the same multistage model. The evaluation of these factors in the decision to combine or not combine data sets is discussed.