Use of permanent hair dyes and bladder-cancer risk

The case-control study by Gago-Dominguez et al.1 raises the possibility that use of permanent hair dye may increase the risk of bladder cancer. In that study, the odds ratio (OR) associated with developing incident bladder cancer among women who ever used permanent hair dye was 1.9 (95% CI = 1.1– 3.3) compared to nonusers and was 3.7 (95% CI = 1.2–11.2) among those with more than 30 years of use. Women who never smoked had a higher relative risk (OR=2.7) compared to women who ever smoked (OR=1.4), although this difference was not statistically significant. The authors suggest that hair dye causes 19% of bladder cancers in women in Los Angeles County, California. We have previously reported no association between the use of permanent hair dye and death from bladder cancer in a prospective study of approximately 600,000 women that participated in Cancer Prevention Study II, a cohort study begun by the American Cancer Society in 1982.2, 3 The age-standardized rate ratio (RR) associated with ever use of hair dye after 7 years of follow-up (based on 84 deaths among never users and 15 deaths among ever users) was 0.56 (95% CI= 0.32–0.99) and after 12 years of follow-up was 1.0 (95% CI=0.7–1.4), based on 154 and 48 deaths among never and ever users, respectively. We now report the results after 16 years of follow-up. The present analysis is based on 547,571 women who were cancer-free (excepting non-melanoma skin cancer) in 1982 and who reported complete information on race, hair dye use and smoking status. After 16 years of follow-up, 131,039 women had died (19.4%) and 543,171 were alive (80.3%). Follow-up of 2,078 women was truncated on September 1, 1988. Information on hair dye use was based on questionnaire responses in 1982 and was not updated during the follow-up. Women were asked “Do you now or have you ever used permanent hair dye?” Those who reported hair dye use were asked the following open-ended questions: (i) “What brand?”, (ii) “What color?”, (iii) “How often applied?” and (iv) “How many years have you used it?” Our analyses also stratified the duration of use into 1–9 years, 10–19 years and 20 or more years according to the baseline report; duration was not allowed to advance during follow-up. Table I presents age-adjusted death rates that are directly standardized to the age distribution of person-years among CPS-II participants during the first 6 years of follow-up. Multivariate-adjusted rate ratios were estimated by using Cox proportional hazards modeling, adjusting for age, race, cigarettes smoked per day (current smokers) or age at quitting (former smokers), education and blue collar occupation. Multivariate-adjusted death rates were derived by multiplying the adjusted ratio for each category of hair dye use by the age-standardized death rate for women who never used hair dye. Because cigarette smoking is an established risk factor for bladder cancers among women4 and was associated with hair dye use in our cohort (37% of ever smokers used hair dye compared to 28% of never smokers), we present results separately for lifelong nonsmokers. The death rate from bladder cancer was similar among women who reported ever using permanent hair dye to that of never users, especially after multivariate adjustment (RR=1.08, 95% CI = 0.84–1.38) and in analyses restricted to lifelong nonsmokers (RR=0.92, 95% CI = 0.61–1.40). Among women who had used permanent hair dyes, no consistent increase was seen in either the death rate from bladder cancer or the rate ratio associated with hair dye use. While it is conceivable that use of permanent hair dye might increase incidence, but not mortality, from bladder cancer, no association with incident bladder cancer was observed in a previous prospective study of registered nurses.5 It may be premature to classify permanent hair dye use as an important cause of bladder cancer among women. Nevertheless, the ongoing debate about the safety of hair dye illustrates the difficulty of accurately measuring and describing the small risks that may or may not be associated with these widely used consumer products.