Anti-oestrogen medication use in breast cancer patients and subsequent risk of gastrointestinal cancers: a two country pooled analysis


Session type:


Úna McMenamin1,Harlinde De Schutter2,Michael B Cook3,Stuart McIntosh4,Brian T Johnston5,Chris Cardwell1
1Queen's University Belfast, Belfast, UK,2Belgian Cancer Registry, Saint-Josse-ten-Noode, Belgium,3National Cancer Institute, Bethesda, US,4Belfast Health & Social Care Trust, Belfast, UK,5Belfast Health & Social Care Trust



There is a strong male predominance in the incidence of oesophageal and stomach cancer, and to a lesser extent, colorectal cancer, suggesting that oestrogens may be protective against the development of these cancers. In a pooled analysis of two large European breast cancer cohorts, we aimed to evaluate associations between anti-oestrogen medication use and risk of upper GI (oesophageal or stomach) and colorectal cancer.


Breast cancer patients were identified from the Belgian Cancer Registry (2004-2014) and the Scottish Cancer Registry (2009-2017). Linkages to national health insurance records (Belgium) and the Prescribing Information System (Scotland) provided detailed information on anti-oestrogen medication use, including tamoxifen and aromatase inhibitors. The primary outcome was GI cancer (UGI or colorectal). Time-dependent Cox regression was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for hormone therapy use and GI cancer risk, adjusting for potential confounders. Fixed effects meta-analysis was used to pool results between the two cohorts.


A total of 113,771 breast cancer patients (Belgium: 87,166, Scotland: 26,605), were included. During follow-up, 701 GI cancers occurred including 151 UGI and 550 colorectal. Overall, no association was shown for anti-oestrogen medication use and risk of GI cancer in pooled analysis (adjusted HR: 0.92, 95% CI 0.74, 1.15). In analysis by cancer site, findings were similar for colorectal cancer (pooled adjusted HR: 0.84, 95% CI 0.66, 1.07) and although HRs were raised for UGI cancer, results were non-significant (pooled adjusted HR: 1.47, 95% CI 0.85, 2.53). Similar results were obtained in dose-response analysis.


In a two-country study of breast cancer patients, we found little evidence of an association between anti-oestrogen medications and risk of GI cancers of the UGI or colorectum. Given the relatively small number of UGI cancer cases, further studies, preferably through pooled analyses of multiple cohorts, are needed to validate our findings.