Keratinocyte carcinoma (KC), including basal cell carcinoma (BCC) and squamous cell carcinoma (SCC), is the most commonly diagnosed cancer in the US. Associations between natural hair color and KC risk and the effect magnitudes have not been thoroughly examined through large-scale studies1,2. Associations for site-specific KC are unclear. We evaluated the associations based on 205,410 Whites from three prospectively administered cohorts, including 78,879 women from Nurses’ Health Study (NHS), 89,516 women from NHS-II, and 37,015 men from Health Professionals Follow-up Study (HPFS)3,4. Information on natural hair color in early adulthood (age 21 (NHS) or 18 (NHS-II and HPFS)) was assessed in 1982 (NHS), 1991 (NHS-II) or 1988 (HPFS). Diagnoses of KC were self-reported biennially. Pathological reports were reviewed to confirm SCC diagnosis and to obtain information on tumor histology and location. Medical information was not retrieved for BCC cases. However, a high accuracy of self-reported BCC was shown in our cohorts previously4. Cox proportional hazards models were performed to estimate the hazard ratios (HRs) with 95% confidence interval (CIs) of hair color associated with BCC, SCC overall, invasive SCC, SCC in situ, and SCC by body sites.
During follow-up of NHS (1982–2012, median 26.3 years), NHS-II (1991–2011, median 19.5 years), and HPFS (1988–2012, median 16.6 years), 5,204 incident SCCs and 33,799 incident BCCs were identified (Supplementary Table S1). Compared with light brown-haired participants, multivariate-adjusted HR (95%CI) of developing SCC was 0.76(0.63–0.93) for black, 0.92(0.85–1.01) for dark brown, 1.08(1.00–1.18) for blonde, and 1.45(1.28–1.64) for red-haired participants in three cohorts combined. Associations appeared similar for invasive and in situ SCC. Individuals with black (HR=0.83, 95%CI:0.73–0.94) or dark brown hair (HR=0.89, 95%CI:0.87–0.92) had a decreased risk of BCC, while blonde (HR=1.11, 95%CI:1.06–1.16) or red-haired individuals (HR=1.28, 95%CI:1.09–1.51) had a higher risk of BCC (Table 1, Supplementary Table S2).
Table 1.
Pooled analysis of risk of squamous cell carcinoma (SCC) and basal cell carcinoma (BCC) according to hair color in the Nurses’ Health Study (NHS, 1982–2012), NHS II (1991–2011) and Health Professionals Follow-up Study (HPFS, 1988–2012)
| Black | Dark Brown | Light Brown | Blonde | Red | ||
|---|---|---|---|---|---|---|
| Pooled analysis | Person-Year | 151,010 | 1,843,543 | 1,696,618 | 582,213 | 160,146 |
| SCC in Total | No. of Cases | 182 | 1,990 | 1,973 | 736 | 323 |
| Age-adjusted HR (95% CI) | 0.66 (0.52–0.83) | 0.86 (0.78–0.94) | 1.00 (REF) | 1.17 (1.07–1.27) | 1.82 (1.61–2.05) | |
| Multivariate-adjusted HR (95% CI) a | 0.76 (0.63–0.93) | 0.92 (0.85–1.01) | 1.00 (REF) | 1.08 (1.00–1.18) | 1.45 (1.28–1.64) | |
| Invasive SCC | No. of cases | 124 | 1,237 | 1,236 | 466 | 201 |
| Age-adjusted HR (95% CI) | 0.65 (0.53–0.78) | 0.85 (0.79–0.92) | 1.00 (REF) | 1.19 (1.07–1.32) | 1.81 (1.56–2.10) | |
| Multivariate-adjusted HR (95% CI) a | 0.76 (0.62–0.92) | 0.92 (0.85–0.99) | 1.00 (REF) | 1.10 (0.99–1.23) | 1.43 (1.23–1.67) | |
| SCC in Situ | No. of Cases | 58 | 170 | 753 | 270 | 122 |
| Age-adjusted HR (95% CI) | 0.71 (0.48–1.04) | 0.89 (0.68–1.18) | 1.00 (REF) | 1.14 (0.99–1.32) | 1.84 (1.51–2.23) | |
| Multivariate-adjusted HR (95% CI) a | 0.80 (0.56–1.15) | 0.95 (0.73–1.23) | 1.00 (REF) | 1.06 (0.92–1.22) | 1.49 (1.22–1.82) | |
| BCC | No. of Cases | 1,170 | 12,620 | 13,124 | 5,032 | 1,853 |
| Age-adjusted HR (95% CI) | 0.74 (0.66–0.83) | 0.84 (0.81–0.87) | 1.00 (REF) | 1.18 (1.10–1.26) | 1.48 (1.19–1.83) | |
| Multivariate-adjusted HR (95% CI) a | 0.83 (0.73–0.94) | 0.89 (0.87–0.92) | 1.00 (REF) | 1.11 (1.06–1.16) | 1.28 (1.09–1.51) |
Stratifying by age and questionnaire cycle, and adjusting for smoking pack-years, alcohol intake (0, 0.1–4.9, 5.0–9.9, 10.0–19.9, or ≥20.0 g/d), body mass index (<18.5, 18.5–24.9, 25.0–29.9, 30.0–34.9, or ≥35.0 kg/m^2), UV exposure at residence, family history of melanoma (yes or no), tendency to sunburn as a child or adolescent (none or redness only, burn, painful burn, or painful burn with blisters), number of severe or blistering sunburns, and number of moles on the extremity. Meta-analyses for different cohorts were performed using the random-effect model. HR: hazard ratio; CI: confidence interval.
Analyses by body sites yielded a higher HR for SCC at limbs (HR=1.96, 95%CI:1.63–2.37) associated with red hair than SCC at head and neck or trunk (P-heterogeneity=0.01). Associations with SCC at limbs was consistent for invasive and in situ SCC (Table 2, Supplementary Table S3).
Table 2.
Pooled analysis of risk of squamous cell carcinoma (SCC) by different body site according to hair color in the Nurses’ Health Study (NHS), the NHS II and the Health Professionals Follow-up Study (HPFS)
| Black/Dark Brown | Light Brown | Blonde | Red | ||
|---|---|---|---|---|---|
| SCC in Total | Person-Year | 1,994,666 | 1,696,729 | 582,249 | 160,163 |
| Head and Neck | No. of Cases | 998 | 873 | 321 | 112 |
| Age-adjusted HR (95% CI) | 0.85 (0.77–0.93) | 1.00 (REF) | 1.14 (1.00–1.30) | 1.47 (1.14–1.90) | |
| Multivariate-adjusted HR (95% CI)a | 0.92 (0.84–1.01) | 1.00 (REF) | 1.05 (0.92–1.20) | 1.13 (0.90–1.42) | |
| Trunk | No. of Cases | 300 | 271 | 112 | 42 |
| Age-adjusted HR (95% CI) | 0.87 (0.73–1.02) | 1.00 (REF) | 1.28 (1.03–1.60) | 1.71 (1.23–2.37) | |
| Multivariate-adjusted HR (95% CI)a | 0.92 (0.78–1.09) | 1.00 (REF) | 1.19 (0.95–1.49) | 1.41 (1.00–1.97) | |
| Limbs | No. of Cases | 724 | 691 | 263 | 150 |
| Age-adjusted HR (95% CI) | 0.83 (0.73–0.95) | 1.00 (REF) | 1.21 (1.05–1.40) | 2.39 (2.00–2.86) | |
| Multivariate-adjusted HR (95% CI)a | 0.90 (0.80–1.02) | 1.00 (REF) | 1.12 (0.97–1.29) | 1.96 (1.63–2.37) | |
| P for Heterogeneity | 0.88 | - | 0.8 | 0.01 | |
| Invasive SCC | Person-Year | 1,995,290 | 1,697,273 | 582,457 | 160,267 |
| Head and Neck | No. of Cases | 677 | 573 | 209 | 72 |
| Age-adjusted HR (95% CI) | 0.86 (0.77–0.97) | 1.00 (REF) | 1.12 (0.96–1.32) | 1.42 (1.11–1.82) | |
| Multivariate-adjusted HR (95% CI)a | 0.94 (0.84–1.05) | 1.00 (REF) | 1.04 (0.88–1.22) | 1.09 (0.85–1.41) | |
| Trunk | No. of Cases | 163 | 170 | 70 | 26 |
| Age-adjusted HR (95% CI) | 0.75 (0.60–0.93) | 1.00 (REF) | 1.28 (0.97–1.70) | 1.70 (1.12–2.58) | |
| Multivariate-adjusted HR (95% CI)a | 0.80 (0.65–1.00) | 1.00 (REF) | 1.19 (0.89–1.58) | 1.32 (0.86–2.04) | |
| Limbs | No. of Cases | 449 | 430 | 167 | 96 |
| Age-adjusted HR (95% CI) | 0.82 (0.72–0.94) | 1.00 (REF) | 1.26 (1.05–1.51) | 2.48 (1.98–3.11) | |
| Multivariate-adjusted HR (95% CI)a | 0.89 (0.78–1.02) | 1.00 (REF) | 1.16 (0.97–1.39) | 2.03 (1.61–2.57) | |
| P for Heterogeneity | 0.76 | - | 0.89 | 0.07 | |
| SCC in Situ | Person-Year | 1,995,826 | 1,697,776 | 582,654 | 160,336 |
| Head and Neck | No. of Cases | 321 | 300 | 112 | 40 |
| Age-adjusted HR (95% CI) | 0.74 (0.50–1.10) | 1.00 (REF) | 1.17 (0.94–1.46) | 1.54 (0.83–2.57) | |
| Multivariate-adjusted HR (95% CI)a | 0.80 (0.53–1.20) | 1.00 (REF) | 1.07 (0.86–1.33) | 1.21 (0.69–2.13) | |
| Trunk | No. of Cases | 137 | 101 | 42 | 16 |
| Age-adjusted HR (95% CI) | 1.07 (0.80–1.43) | 1.00 (REF) | 1.29 (0.89–1.86) | 1.81 (1.06–3.08) | |
| Multivariate-adjusted HR (95% CI)a | 1.11 (0.85–1.44) | 1.00 (REF) | 1.21 (0.84–1.75) | 1.60 (0.93–2.78) | |
| Limbs | No. of Cases | 275 | 261 | 96 | 54 |
| Age-adjusted HR (95% CI) | 0.86 (0.72–1.02) | 1.00 (REF) | 1.13 (0.85–1.51) | 2.29 (1.70–3.08) | |
| Multivariate-adjusted HR (95% CI)a | 0.91 (0.76–1.08) | 1.00 (REF) | 1.06 (0.80–1.40) | 1.90 (1.40–2.59) | |
| P for Heterogeneity | 0.04 | - | 0.84 | 0.20 |
Stratifying by age and questionnaire cycle, and adjusting for smoking pack-years, alcohol intake (0, 0.1–4.9, 5.0–9.9, 10.0–19.9, or ≥20.0 g/d), body mass index (<18.5, 18.5–24.9, 25.0–29.9, 30.0–34.9, or ≥35.0 kg/m^2), UV exposure at residence, family history of melanoma (yes or no), tendency to sunburn as a child or adolescent (none or redness only, burn, painful burn, or painful burn with blisters), number of severe or blistering sunburns, and number of moles on the extremity. Meta-analyses for different cohorts were performed using the random-effect model. HR: hazard ratio; CI: confidence interval.
Based on three large prospective cohorts, we found that red-haired individuals were associated with increased risk of KC. The associations appeared stronger for SCC at limbs than other sites. We were able to determine Fitzpatrick skin type in NHS only (Supplementary Table S1). A secondary analysis further adjusting for Fitzpatrick skin type in NHS didn’t materially change our results, suggesting the association of red hair with KC independent from Fitzpatrick skin type (Supplementary Table S4–S5).
Skin cancer at different body sites may have distinct etiology. However, we acknowledge the limitation as an epidemiologic study. Studies are warranted to elucidate underlying mechanisms for the observed site-specific associations in SCC. In addition, we did not evaluate associations between hair color and site-specific BCC because of the limitation that medical records were not retrieved for BCC, which requires further efforts in the future. Our findings may further inform dermatologists and general practitioners on the increased KC risk, particularly at limb, for red-haired individuals.
Supplementary Material
Funding sources:
The work was supported by the National Institute of Health grants for Nurses’ Health Study (P01 CA87969 and UM1 CA186107), Nurses’ Health Study II (U01 CA176726) and Health Professionals Follow-up Study (U01 CA167552). The funding sources had no role in study design and conduct; in the collection, management, analysis, and interpretation of data; in the preparation, review, or approval of the report; or in the decision to submit the article for publication.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflicts of Interest: None declared.
References
- 1.Khalesi M, Whiteman DC, Tran B, Kimlin MG, Olsen CM, Neale RE. A meta-analysis of pigmentary characteristics, sun sensitivity, freckling and melanocytic nevi and risk of basal cell carcinoma of the skin. Cancer epidemiology. 2013;37(5):534–543. [DOI] [PubMed] [Google Scholar]
- 2.Green A, Battistutta D, Hart V, Leslie D, Weedon D. Skin cancer in a subtropical Australian population: incidence and lack of association with occupation. The Nambour Study Group. Am J Epidemiol. 1996;144(11):1034–1040. [DOI] [PubMed] [Google Scholar]
- 3.van Dam RM, Huang Z, Rimm EB, et al. Risk factors for basal cell carcinoma of the skin in men: results from the health professionals follow-up study. Am J Epidemiol. 1999;150(5):459–468. [DOI] [PubMed] [Google Scholar]
- 4.Li WQ, Cho E, Weinstock MA, Mashfiq H, Qureshi AA. Epidemiological Assessments of Skin Outcomes in the Nurses’ Health Studies. Am J Public Health. 2016;106(9):1677–1683. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.