The main epidemiological evidence of an association between cadmium exposure and human cancer comes from studies on occupational groups such as smelter and nickel-cadmium battery workers. Lung and prostate are the sites more frequently reported to be at increased risk of cancer, but the various epidemiological studies do not enable any definite conclusion to be reached. Methodological problems in the interpretation of these studies include: (i) the assessment of exposure to cadmium; (ii) the dose-response relationship between indicators of cadmium exposure and cancer risk; (iii) concomitant exposure to occupational carcinogens, mainly nickel and arsenic; (iv) confounding by non-occupational risk factors, such as smoking and diet; and (v) interaction between cadmium and other exposures. The confounding effect of smoking and diet is particularly interesting since they may act both as additional uncontrolled sources of cadmium and as independent cancer risk factors.