Schizophrenia and Lung Cancer: Evidence from Mendelian Randomization and Genetic Pleiotropy Analysis.

Abstract

Background: Previous observational studies provided inconsistent results in the relationship between schizophrenia (SCZ) and lung cancer (LUCA), with substantial between-study variance. The causality from SCZ to LUCA remains unknown. The aim was to conduct a Mendelian randomization (MR) analysis to investigate the impact of SCZ on LUCA.

Methods: The SNP-phenotype association data were acquired from genome-wide association studies (GWAS) analysis for each corresponding phenotype, including SCZ (53 386 cases and 77 258 controls), LUCA (29 266 cases and 56 450 controls), lifetime smoking (462 690 participants), and alcoholic drinks per week (2 428 851 participants). Univariable and multivariable MR analysis were conducted to evaluate the potential causal effect of SCZ on LUCA and a mediation approach to quantify the relative contribution of risk factors. Sensitivity and additional analyses were performed to further validate the robustness of the results.

Results: Univariable MR analysis demonstrated that genetically predicted SCZ causally increases the risk of carcinogenesis of LUCA (OR = 1.068, 95% CI = 1.024-1.114, P = .002). Approximately 27.6% (95% CI 9.2-47.3%) of the effect is mediated by lifetime smoking exposure, 5.9% (95% CI 1.6-12.3%) by drinks per week, and 31.4% (95% CI 9.2-56.5%) by both mediators combined. Consistent results were observed during sensitivity analyses and additional analyses.

Conclusion: This study provides genetic evidence for the causal relationship between genetically predicted SCZ and a higher risk of LUCA, which could be reduced by adopting population-level interventions targeting smoking cessation and alcohol reduction.