Personal Use of Hair Dye and the Risk of Certain Subtypes of Non-Hodgkin Lymphoma

Yawei Zhang; Silvia De Sanjose; Paige M. Bracci; Lindsay M. Morton; Rong Wang; Paul Brennan; Patricia Hartge; Paolo Boffetta; Nikolaus Becker; Marc Maynadie; Lenka Foretova; Pierluigi Cocco; Anthony Staines; Theodore Holford; Elizabeth A. Holly; Alexandra Nieters; Yolanda Benavente; Leslie Bernstein; Shelia Hoar Zahm; Tongzhang Zheng


Am J Epidemiol. 2008;167(11):1321-1331. 

In This Article


Results from this large, pooled, InterLymph-based case-control study indicate that personal use of hair dyes may play a role in the risk of NHL mainly for women who started using these products before 1980, particularly for follicular lymphoma and CLL/SLL.

The underlying mechanisms that may explain an association between NHL and hair-dye use, particularly for follicular lymphoma and CLL/SLL, are unknown. Many ingredients used in hair dyes before 1980 were shown to be mutagenic or carcinogenic in bacteria and rodents.[6,24] This may partly explain the observed association with NHL. Chromosome translocations are a hallmark of some lymphoma subtypes. Among follicular lymphomas, more than 75 percent of patients carry t(14:18)(q32;21) translocations, and 50-60 percent of CLL/SLL patients carry chromosomal changes, most frequently a numerical change of chromosome 12 and del.[13q,25] Chromosome alterations produced by chemical carcinogens have been reported in both in vitro and in vivo studies.[26,27] Thus, it is biologically plausible that personal use of hair dye may result in increased risks of follicular lymphoma and CLL/SLL.

Our results also showed that the risk associated with personal hair-dye use was present mainly among women who started using hair dyes before 1980. However, the increased risk was not limited to these women. Although the use of some carcinogenic compounds in hair dyes (such as 2,4-diaminoanisole and several yellow nitro dyes)[28] was discontinued after 1980, NHL risk may be associated with current formulations and with compounds created during the oxidization process. For example, PPD, a major arylamine currently used in hair dyes, is a putative carcinogen and a known agent for allergic contact dermatitis.[7] When oxidized, PPD in bioassays can form a compound called Bandrowski's base, which has been reported to be mutagenic.[29] In addition, an in vitro study showed that PPD can activate dendritic cell function, resulting in increased expression of CD40 protein and major histocompatibility complex class II, as well as stimulation of allogeneic lymphocyte proliferation.[30] Among follicular lymphoma patients, an association between survival and gene expression signatures of nonmalignant tumor-infiltrating immune cells, including T cells, macrophages, and dendritic cells, was reported in a recent study[31] This suggests that PPD-induced overexpression of CD40 protein and major histocompatibility complex class II molecules may be associated with risk of this NHL subtype.

Whereas a biologically plausible explanation for the association between hair-dye use and NHL risk exists, studies of hair-dye use and NHL risk have provided inconsistent results, with some studies showing an increased risk[9,10,11,15,17,18,19] and others showing no association.[12,13,14,16] Several factors may explain these conflicting findings. First, not all studies obtained detailed information on lifetime use of hair dye. As Zahm and Fraumeni[32] emphasized, detailed information on lifetime exposure to hair-dye products, such as duration of use, dates of use, and type and color, is needed to establish a relation between hair-dye use and risk of NHL. The results from this analysis support the importance of these measures in providing additional evidence of a dose-response effect and thus strengthening the plausibility of an association with hair-dye use. Second, as demonstrated in this study, it is necessary to examine the relation between personal hair-dye use and NHL by time period of use. The risk pattern by time period observed in this study is consistent with changes in hair-dye formulations during the past two decades that may have affected the magnitude of the underlying association. Holly et al.[14] first suggested the need to examine the relation between hair-dye use and NHL risk by time period of use, and this is supported by results from Zhang et al.,[17] Morton et al.,[20] and the current pooled analysis. Third, few earlier studies investigated the association between NHL risk and hair-dye use by histologic subtype. As the results from our analyses indicate, risk is likely to differ by subtype and may account for some of the inconsistencies in reported results. Fourth, statistical power was limited in most of the earlier studies, especially for analyses by type of hair dye and by NHL subtype. Fifth, Morton et al.[20] recently reported that genetic polymorphisms may affect the risk of NHL associated with hair-dye exposure. One putative carcinogen, PPD, may lose its reactivity through N-acetylation.[33] An in vitro study showed that acetylated PPD, monoacetyl-PPD, and N,N'-diacetyl-PPD cannot be transformed to the mutagenic Bandrowski's base.[33] Thus, the genetic predisposition for N-acetylation of PPD may alter susceptibility to PPD carcinogenicity among PPD-exposed persons. Population differences in genetic susceptibility to putative carcinogens in hair dyes may partly explain the variation in reported results linking hair-dye use and NHL risk. It also may explain the observed difference in risk of NHL associated with hair-dye use between men and women. Animal studies have shown higher activity of N-acetyltransferase 1, an enzyme responsible for biotransformation of carcinogens, including many arylamine and hydroxylamine xenobiotics, in males than in females,[34] although results from human studies have been inconsistent.[35] Finally, it also is possible that differences in product formulations across countries may have contributed to variation in the reported results.

Results from our analyses showed that risk of CLL/SLL was increased mainly for European women who used hair dyes, not US women, whereas the association with follicular lymphoma was seen in both geographic regions. Although the duration and frequency of hair-dye use in the control group were similar between European women and US women, duration of use among European women with CLL/SLL was greater than that for US women (p values ranged from 0.006 to 0.06 for various hair-dye products). Other factors, such as differences in hair-dye formulations, classification of rare NHL subtypes, or chance variations, also may have contributed to the observed difference in risk pattern between the European and US populations.

Note that although statistical power was adequate to test the overall association, power was reduced for analyses conducted by exposure and NHL subtype, especially for men, in whom the prevalence of hair-dye use was low, and racial minorities, who comprised less than 10 percent of the pooled population.

Information about hair-dye use in these analyses was obtained through participants' recall during in-person interviews. If NHL patients overreported their lifetime use of hair dye in comparison with controls, the observed results could have overestimated the association between hair-dye use and risk of NHL. However, the lack of an association with diffuse large B-cell lymphoma, a major subtype of NHL in Western populations, suggests that recall bias does not play a major role in the observed association between hair-dye use and risk of follicular lymphoma and CLL/SLL. Nondifferential misclassification of exposure status also may exist and would have resulted in underestimation of the association between hair-dye use and NHL risk. Central review of all cases by a study pathologist was not feasible, so it is possible that some disease misclassification occurred for analyses by NHL subtype if classification rules differed between studies. However, it is likely that any disease misclassification was nondifferential, thus biasing the results toward the null hypothesis. Finally, multiple comparisons were made in these analyses, and therefore some of the observed associations may be due to chance.

In summary, the results from this large InterLymph-based pooled analysis indicate that personal use of hair dye may play a role in the risk of NHL, particularly for follicular lymphoma and CLL/SLL. Our study also indicates that although the risk associated with personal hair-dye use was observed mainly among women who started using hair dyes before 1980, the risk was not limited to those women. Future studies are needed to examine the risk of NHL by time period of hair-dye use and by genetic susceptibility.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: