The aim of this study was to determine the risk of kidney cancer in 2 cohorts defined on the basis of hospital discharge diagnoses associated with analgesic or diuretic use during the period 1965 to 1983. Patients were followed up through 1984 for cancer incidence. After excluding cancers in the first year of observation, 161 kidney cancers were observed vs. 138 expected among 54,662 patients in the analgesics cohort. The relative risk was higher for women than for men. When examined by sub-site within the kidney, risk for cancer of the renal pelvis was similar in magnitude to that for the renal parenchyma. Among 115,616 patients in the diuretics cohort, 278 kidney cancers occurred vs. 209 expected. The risk for women was higher than for men. This elevation in risk was confined to cancer of the renal parenchyma, with no significantly increased risk seen for cancer of the renal pelvis. Although we observed little excess risk among members of the analgesics cohort, the significantly elevated risk among patients using diuretics supports a number of recent studies, but inability to adjust for confounding factors such as obesity preclude drawing any conclusion regarding diuretics. Further research is warranted to assess in detail the relationship between diuretic use and cancer of the renal parenchyma.