Objective: To investigated the genetic relationship between gestational diabetes mellitus (GDM) and 14 female prevalent cancers. Design: Evidence triangulation from linkage disequilibrium score regression (LDSC), two-sample Mendelian randomization (MR), and colocalization analysis. Setting: Summary-level genome-wide association study (GWAS) data. Population or Sample: All participants were of European ancestry. Methods: Exposure was GDM. The mediators considered were genetically estimated body mass index, hypertension, type 2 diabetes mellitus, fasting glucose, fasting insulin, and insulin-like growth factor 1. Main Outcome Measures: Breast, lung, colorectal, cervical, liver, stomach, pancreatic, ovarian, leukaemia, non-Hodgkin lymphoma, kidney, thyroid, endometrial, brain and central nervous system cancer. Results: The LDSC revealed no significant genetic correlations between GDM and most cancers, except for a nominal association with endometrial cancer (rg = 0.2294, p = 0.0458) that failed replication. Two-sample MR analysis identified potential causal links between GDM and 10 cancers in test datasets, with modest effect sizes, but only colorectal cancer (OR = 1.66), liver cancer (OR = 2.11), and leukemia (OR = 1.01) showed replicable associations. Colocalization analysis was limited for most cancers (PPH4 < 0.8). The integrated genetic evidence offered only weak support for associations with endometrial cancer, without identifying clear mediating pathways. Conclusions: We did not find reproducible genetic evidence of causal associations between GDM and cancer risk. The finding highlighted the need for cautious interpretation of epidemiological observations.
