Upper bound risk estimates for mixtures of carcinogens

• Published in: Toxicology Vol. 105; no. 2; pp. 199 - 208
• Main Authors: Kodell, Ralph L., Ahn, Hongshik, Chen, James J., Springer, Janet A., Barton, curtis N., Hertzberg, Richard C.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Shannon Elsevier Ireland Ltd 28-12-1995; Amsterdam Elsevier Science
• Subjects: Additivity; Animals; Biological and medical sciences; Bootstrap; Carcinogenesis, carcinogens and anticarcinogens; Carcinogenicity Tests; Carcinogens - toxicity; Chemical agents; Cocarcinogenesis; Confidence Intervals; Conservatism; Data Interpretation, Statistical; Dose-Response Relationship, Drug; Drug Interactions; Likelihood Functions; Likelihood ratio; Mathematical Computing; Medical sciences; Monte Carlo; Monte Carlo Method; Multistage model; Probability; Risk Assessment; Tumors; Additivity; Likelihood ratio; Bootstrap; Conservatism; Monte Carlo; Multistage model; Carcinogen; Monte Carlo method; Poison interaction; Investigation method; Toxicity; Chemical compound; Toxic association; Mixture
• ISSN: 0300-483X; 1879-3185
• DOI: 10.1016/0300-483X(95)03213-Y

The excess cancer risk that might result from exposure to a mixture of chemical carcinogens usually is estimated with data from experiments conducted on individual chemicals. An upper bound on the total excess risk is estimated commonly by summing individual upper bound risk estimates. The degree to which this approach might overstate the true risk associated with the mixture has not been evaluated previously. This paper reports the results of a Monte Carlo simulation study on the degree of reduction in conservatism that might be achieved using alternative methods for calculating mixture upper bounds. An unexpected finding is that for chemicals that exhibit strongly linear dose-response relationships, the summing of multistage-model-based upper bounds on excess risk can be anti-conservative, that is, it can provide less than the nominal 100(1 — α)% coverage.