RESUMEN
Objective: In this study, we aimed to investigate whether age first had sexual intercourse (AFSI) and lifetime number of sexual partners (LNSP) have a direct causal effect on cervical cancer by Mendelian randomization (MR) analysis. Methods: Four approaches were used for MR Analysis, including MR-Egger, weighted method, weighted median, and inverse variance weighted (IVW). MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) as well as MR-Egger regression analysis were conducted to detect whether there was pleiotropy between IVs and outcome, and the outlier SNPs can be detected by MR-PRESSO. The presence or absence of heterogeneity among IVs was suggested according to Cochran's Q statistic. Leave-one-out sensitivity analysis was performed to identify and remove SNPs which could independently change the results. We corrected the results using Bonferroni correction. Results: From the results of IVW, AFSI had a negative effect on cervical cancer (OR = 0.996, 95 % CI: 0.995, 0.998 P = 1.70E-07), which still persisted after Bonferroni correction. However, no causal effect of LNSP on cervical cancer was found according to the IVW results (OR = 1.003, 95 % CI: 1.000, 1.007, P = 0.071). From the results of MR-PRESSO and MR-Egger, no SNP with horizontal pleiotropy between cervical cancer was detected and no SNP was identified as an outlier SNP. Cochran's Q statistic suggested that no heterogeneity existed among IVs of AFSI and LNSP. According to Leave-one-out analysis, the results of MR did not change after excluding any single IV. Conclusion: This MR study reveals that early AFSI has a causal effect on cervical cancer.
