Associations between psoriasis and risk of 33 cancers: a Mendelian randomization study

Abstract

Background: Several observational studies have reported epidemiologic associations between psoriasis and risk of some cancers, but systematic evidence is lacking. Our aim was to comprehensively estimate the association between psoriasis and the risk of 33 common cancers using systematical Mendelian randomization based on genetic data.

Method: Forty-nine independent single-nucleotide polymorphisms (SNPs) significantly associated with psoriasis were extracted as instrumental variables from a large-scale meta-analysis study of genome-wide association study (GWAS) for psoriasis. Outcome GWAS data were obtained from the FinnGen consortium (n = 500,348), UK Biobank (n = 420,531), and other large-scale cancer datasets. The inverse-variance weighted (IVW) was used as the primary method to infer the association between psoriasis and risk of cancer, and finally the results from multiple databases were pooled by meta-analysis.

Results: In the UK Biobank, genetically predicted psoriasis had a suggestive association with colon (OR = 1.055, 95%CI: 1.001-1.113, P = 0.046) and uterine corpus cancer (OR = 0.922, 95%CI: 0.852-0.997, P = 0.042). In the FinnGen consortium, psoriasis had a suggestive association with vulvar cancer (OR = 1.182, 95%CI: 1.023-1.366, P = 0.024), uterine corpus cancer (OR = 0.937, 95%CI: 0.883-0.993, P = 0.028), and prostate cancer (OR = 0.973, 95%CI: 0.948-0.999, P = 0.045). In an additional large-scale cancer dataset, psoriasis also showed a suggestive association with prostate cancer (OR = 0.968, 95%CI:0.942-0.995, P = 0.020). The meta-analysis confirmed the suggestive association of psoriasis with uterine corpus (OR = 0.931, 95% CI: 0.889-0.976, P = 0.003) and prostate cancer (OR = 0.976, 95% CI: 0.955-0.997, P = 0.023). Whereas the effect of psoriasis on colon and vulvar cancer was not in the same direction across different populations. Furthermore, no association between genetically predicted psoriasis and other cancers were observed.

Conclusions: This comprehensive MR study suggests that psoriasis may be a potential protective factor for uterine corpus cancer in women and prostate cancer in men.

Keywords: Association; Cancer; FinnGen consortium; Mendelian randomization; Psoriasis; UK biobank.

Fig. 1

Research Design of Mendelian randomization. The outcome data for this Mendelian randomization study were obtained from UK Biobank, FinnGen Consortium, and other large-scale cancer databases. Assumption 1 (relevance assumption): There is a strong association between genetic variants and exposure factors. Assumption 2 (independence assumption): Genetic variants are independent of confounders affecting “exposure and outcome”. Assumption 3 (exclusion restriction assumption): Genetic variation can only contribute to outcome through exposure, but not through other pathways

Fig. 2

Forest plot of MR analysis results in the UK Biobank and the FinnGen. Abbreviations: OR: odds ratio; CI: confidence interval

Fig. 3

Forest plot of the pooled results of the meta-analysis from the UK biobank and the FinnGen. Cancer types are listed in ascending order of effect size (OR). Abbreviations: OR: odds ratio; CI: confidence interval