Abstract
Data from a population-based case-control study of incident leukemia and non-Hodgkin's lymphoma among adult men in Iowa and Minnesota were used to evaluate risk associated with hair dye use. The relative risk for ever using hair dyes was 1.8 (95% confidence interval [CI] = 1.1-2.7) among leukemia patients, and 2.0 (CI = 1.3-3.0) among cases with non-Hodgkin's lymphoma. There was a suggestion of increased risk with extent of hair dye use. Given the widespread use of hair coloring products, these observations deserve more detailed evaluation in populations where the exposure is relatively common.