Bladder cancer provides the most definitive example for an association between environmental agents and cancer. However, in the absence of industrial occupational exposure, the primary carcinogen is rarely identified, and the mechanisms involved in cancer formation are poorly understood. The environmental procarcinogen hypothesis of tumour pathogenesis proposes that many carcinogens require metabolic activation by drug metabolizing enzymes to form the proximate carcinogen. A balance of exposure to the carcinogen, the activity of the enzymes involved in either formation of proximate carcinogen, or production of non-toxic metabolites, will determine tumour risk. We have used mephenytoin, debrisoquine and dapsone as selective probes for the phenotypic measures of activity of CYP2C19, CYP2D6, and CYP3A4, respectively. Within subject reproducibility of phenotypic measures, and the lack of cross-inhibition when the three drugs are given in a concurrent cocktail, have been confirmed. We have applied the cocktail drug approach in two, non-overlapping series of cases with bladder cancer and matched controls. In both series, patients with aggressive bladder cancer (GIII histopathology) had a history of excess alcohol intake, an under-representation of poor metabolizers of debrisoquine, a significant mean reduction in dapsone recovery ratio, but no difference in mephenytoin phenotype. Collectively, these observations involving multiple routes of drug metabolism support the procarcinogen environmental hypothesis for bladder cancer and suggest that measurement of activity of selected individual drug metabolizing enzymes involved in the pathogenesis of this tumour can be used to identify subjects at high risk of developing bladder cancer.