Genetic modulation of environmental exposures associated with common malignancies (lung and bladder) is an attractive mechanism to explain differential susceptibility to tobacco or occupation related carcinogens in the population. Epidemiologic studies to test the hypothesis of such associations and to evaluate evidence for a causal role for genetic factors in the etiology of chemically-induced tumors are challenging and require the close collaboration of epidemiologists, clinicians and laboratory investigators. In this work we review the evidence for an association of three polymorphisms of drug or xenobiotic metabolism with human cancers. Methodologic considerations and data relevant to evaluating a causal role for each polymorphism are considered. Fair to good support for both an association of the acetylation phenotype with occupationally-related bladder cancer and for an association of the debrisoquine metabolic phenotype and lung cancer is found, although in neither case is the evidence completely convincing. Epidemiologic evidence for the association of aryl hydrocarbon hydroxylase and lung cancer is presently problematic because of difficulties in the assay and subsequent confounding factors. DNA based assays are at various stages of development for each of the genotypes and promise to simplify future studies while introducing new methodologic pitfalls. Further studies in all three areas are warranted as each has important implications for the understanding of the carcinogenic process, etiology and the public health aspects of common malignancies.