Analysis of the Stockholm Public Health Cohort: Exploring How Ultraviolet Radiation and Other Factors Associate with Skin Cancer.

Objective: The primary aims of the study were to (1) explore the association of skin cancer and four ultraviolet radiation (UVR) indicators (sunbed use, healthcare data on diagnosed melanocytic nevi (MN) and actinic keratosis (AK), and latitude of birthplace), and (2) find factors other than UVR that could explain the increasing trend in incidence of skin cancers, including basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and cutaneous malignant melanoma (CMM).

Methods: This population-based cohort study used self-reported questionnaire data from the Stockholm Public Health Cohort, encompassing 103 questions, merged with data from Swedish national registers. The study population included almost 35,000 Swedish-born people aged 30-66 years in 2014. Binomial logistic regression was employed for analysis. A forward stepwise regression was applied to select significant risk factors among all the factors included. We tentatively tested >30 variables separately for any relationship with each of the three skin cancers. A 5% level of significance was applied. Melanoma in situ and SCC in situ were excluded.

Results: The four UVR-related factors (sunbed use, being diagnosed with AK or MN, birthplace latitude) had a significant association with at least one of the three skin cancers that remained after adjustment including behavioural, social, hereditary, and medical factors. Sunbed use >10 times before age 30 years was related to all three skin cancers; SCC adjusted odds ratio (aOR) 1.66, 95% confidence interval (CI) 1.12-2.47, CMM (aOR 1.57, 95% CI 1.11-2.22), and the clearest dose-response association with BCC (aOR 1.74, 95% CI 1.46-2.06). None of the examined lifestyle factors, except physical activity, had any significant associations with UVR indicators or skin cancer.

Conclusion: We did not find any preventable explanatory cause other than UVR exposure for the increasing incidence of skin cancers. This result remained when adjusting for an array of possible confounders including behavioural, social, hereditary, and medical factors.