Postmenopausal Hormone Therapy Is Primarily Associated with Reduced Risk of Colorectal Cancer Arising through the Adenoma-Carcinoma Pathway

Abstract

Abstract Our goal was to evaluate whether the inverse association of postmenopausal hormone therapy (PMH) and colorectal cancer (CRC) differs by molecularly defined CRC tumor subtypes. Methods: We pooled data on tumor markers and PMH use among 8,220 post-menopausal women (3,898 CRC cases and 4,322 controls) from eight observational studies in the Genetics of Epidemiology of Colorectal Cancer Consortium and the Colon Cancer Family Registry. We used multinomial logistic regression to estimate odds ratios (OR) and 95% confidence intervals (CI) for the association of ever versus never PMH use and each tumor subtype compared with controls. We defined subtypes according to microsatellite instability (MSI-high or -low/stable), CpG island methylator phenotype (CIMP positive or negative), oncogenic mutations in BRAF and KRAS, and combinations of these markers that have been linked to specific pathways (adenoma-carcinoma, serrated, alternate). Additionally, we investigated whether associations varied by tumor anatomic location (proximal colon, distal colon, rectum). All models were adjusted for study, age, body mass index, smoking status, and family history of CRC. Wald chi-square tests were used to evaluate whether the association differed by tumor-specific subtypes. Results: Ever use of PMH was associated with a 38% reduction in overall CRC risk (OR 0.62, 95% CI 0.56–0.69). In general, this association was observed regardless of individual markers for MSI, CIMP, BRAF, or KRAS status. However, when taken altogether and grouping cases by pathway, the association was attenuated for tumors arising through the serrated pathway compared with the adenoma-carcinoma pathway (OR 0.81, 95% CI 0.65–1.01; p for difference 0.046). We also observed a weaker association for tumors of the proximal colon compared with the distal colon and rectum (OR 0.71, 95% CI 0.62–0.80; p for difference 0.010). Conclusions: In this large consortium analysis, we observed a strong inverse association between PMH use and overall CRC risk. The association may predominantly reflect a benefit of PMH use for tumors arising through the adenoma-carcinoma pathway and tumors of the distal colon and rectum, as the association was weaker for tumors arising through the serrated pathway and proximal colon tumors.