Abstract
INTRODUCTION: The controversial relationship between smoking and prostate cancer (PCa) risk prompted us to conduct a cross-sectional study using the National Health and Nutrition Examination Survey (NHANES) database and apply Mendelian randomization (MR) analyses in order to clarify the possible causal effect of smoking on PCa risk. METHODS: Using univariate and multivariate logistic regression methods, a secondary analysis of the pooled 2003-2018 NHANES dataset was performed to explore the association between smoking and PCa risk. Propensity-score matching was used to reduce selection bias. Then, we conducted subsequent MR analysis study to investigate the potential causal effect of smoking on PCa risk, with genetic variants of four exposure factors including the lifetime smoking index, light smoking, smoking initiation, and the amount of smoking per day obtained from genome-wide association studies, and PCa summary statistics obtained from three database populations. Inverse-variance weighting was the primary analytical method, and weighted median and MR-Egger regression were used for sensitivity analyses. The MR results for the three PCa databases were combined using meta-analysis. RESULTS: The study included 16073 NHANES subjects, comprising 554 with PCa and 15519 without PCa. Logistic regression before and after matching did not reveal any significant association. Meta-analysis of the MR results also did not support an association of PCa risk with lifetime smoking index (OR=0.95; 95% CI: 0.83-1.09), light smoking (OR=1.00; 95% CI: 0.95-1.06), smoking initiation (OR=0.99, 95% CI=0.99-1.00), or the amount of smoking per day (OR=1.00; 95% CI: 0.99-1.00) and PCa risk. CONCLUSIONS: There was no evidence for an association between smoking and the risk of PCa. Further studies are needed to determine if there are any associations of other forms of smoking with the risk of PCa at different stages.