1046Physical activity and sitting time in relation to breast cancer risk: A Mendelian randomization analysis

Abstract

Background: Greater physical activity (PA) and less sitting time (ST) have been associated with lower breast cancer (BC) risk in observational studies, but causality is undetermined. Mendelian randomization (MR) investigates causality by using genotype to simulate randomized trials within observational data. A recent summary-data MR study assessed PA and BC by estrogen-receptor (ER) status. We aimed to examine PA and ST in relation to invasive BC risk overall and by molecular subtype. Methods: We performed two-sample MR using individual-level Breast Cancer Association Consortium data from 125,684 women (69,869 cases), and summary-level UK Biobank genetic data for the exposure (n=91,105). We used imputed single nucleotide polymorphisms (SNPs) previously associated (p<5x10-8) with accelerometer-measured PA (n-SNPs=5) or ST (n=6). We estimated odds ratios (OR) from inverse-variance weighted MR. Results: Greater PA was associated with lower risk of BC (OR=0.48;95%CI=0.30-0.78 per standard deviation [SD] in activity). Estimates were inverse for all subtypes, particularly ER+ (OR=0.45; 95%CI=0.25-0.83), PR+(OR=0.43;95%CI=0.22- 0.85), and HER2+(OR=0.48; 95%CI=0.26-0.89). Greater ST was associated with higher risk of hormone-receptor-negative tumours, particularly triple-negative (ER-/PR-/HER2-) (OR=2.04; 95%CI=1.06-3.93 per SD in ST). Estimates overall and for most other subtypes were elevated but had wide CIs crossing null. Sensitivity analyses (weighted-median MR and MR-Egger) did not suggest strong pleiotropic effects for either exposure. Conclusions: Greater PA and lower ST reduces risk of all breast cancers, and hormone-receptor negative tumours, respectively. More widespread adoption of active lifestyles will lower burden from the most common cancer in women. Key messages: More active and less sedentary lifestyles reduce risk of breast cancer.