Genetic variants for smoking behaviour and risk of skin cancer

Abstract

Observational studies have suggested that smoking may increase the risk of cutaneous squamous cell carcinoma (cSCC) while decreasing the risks of basal cell carcinoma (BCC), and melanoma. However, it remains possible that confounding by other factors may explain these associations. The aim of this investigation was to use Mendelian randomization (MR) to test whether smoking is associated with skin cancer, independently of other factors. Two-sample MR analyses were conducted to determine the causal effect of smoking measures on skin cancer risk using genome-wide association study (GWAS) summary statistics. We used the inverse-variance-weighted estimator to derive separate risk estimates across genetic instruments for all smoking measures. A genetic predisposition to smoking initiation was associated with lower risks of all skin cancer types, although none of the effect estimates reached statistical significance (OR 95% CI BCC 0.91, 0.82-1.01; cSCC 0.82, 0.66-1.01; melanoma 0.91, 0.82-1.01). Results for other measures were similar to smoking initiation with the exception of smoking intensity which was associated with a significantly reduced risk of melanoma (OR 0.67, 95% CI 0.51-0.89). Our findings support the findings of observational studies linking smoking to lower risks of melanoma and BCC. However, we found no evidence that smoking is associated with an elevated risk of cSCC; indeed, our results are most consistent with a decreased risk, similar to BCC and melanoma.

Figure 1

Two-sample Mendelian randomisation analyses for the associations between genetic predisposition to smoking initiation and risk of developing basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma. OR odds ratio, CI confidence interval. N SNPs number of single nucleotide polymorphisms.

Figure 2

Two-sample Mendelian randomisation analyses for the associations between smoking intensity and risk of developing basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma. OR odds ratio, CI confidence interval. N SNPs number of single nucleotide polymorphisms; NA the MR-PRESSO method did not detect any outlier variant for SCC.

Figure 3

Two-sample Mendelian randomisation analyses for the associations between lifetime smoking index and risk of developing basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma. OR odds ratio, CI confidence interval. N SNPs: number of single nucleotide polymorphisms.