The genetic interplay between body mass index, breast size and breast cancer risk: a Mendelian randomization analysis

Abstract

BackgroundEvidence linking breast size to breast cancer risk has been inconsistent, and its interpretation is often hampered by confounding factors such as body mass index (BMI). Here, we used linkage disequilibrium score regression and two-sample Mendelian randomization (MR) to examine the genetic associations between BMI, breast size and breast cancer risk.MethodsSummary-level genotype data from 23andMe, Inc (breast size, n = 33 790), the Breast Cancer Association Consortium (breast cancer risk, n = 228 951) and the Genetic Investigation of ANthropometric Traits (BMI, n = 183 507) were used for our analyses. In assessing causal relationships, four complementary MR techniques [inverse variance weighted (IVW), weighted median, weighted mode and MR-Egger regression] were used to test the robustness of the results.ResultsThe genetic correlation (rg) estimated between BMI and breast size was high (rg = 0.50, P = 3.89x10−43). All MR methods provided consistent evidence that higher genetically predicted BMI was associated with larger breast size [odds ratio (ORIVW): 2.06 (1.80–2.35), P = 1.38x10−26] and lower overall breast cancer risk [ORIVW: 0.81 (0.74–0.89), P = 9.44x10−6]. No evidence of a relationship between genetically predicted breast size and breast cancer risk was found except when using the weighted median and weighted mode methods, and only with oestrogen receptor (ER)-negative risk. There was no evidence of reverse causality in any of the analyses conducted (P > 0.050).ConclusionOur findings indicate a potential positive causal association between BMI and breast size and a potential negative causal association between BMI and breast cancer risk. We found no clear evidence for a direct relationship between breast size and breast cancer risk.

Body size and composition and site-specific cancers in UK Biobank: a Mendelian randomisation study

“…In this MR study, raised BMI was associated with sex-specific cancers with increased risk of uterine cancer, and decreased risk of breast, prostate and testicular cancer. BMI and breast cancer has been extensively studied in previous MR studies [28,[38][39][40]. A large MR study based on data from the BCAC and DRIVE consortia of 46,325 cases of breast cancer found that 84 BMI-related SNPs were associated with reduced breast cancer risk in both pre-and post-menopausal women [39], consistent with the findings of our study.…”

Population Genetic Considerations for Using Biobanks as International Resources in the Pandemic Era and Beyond

“…Many association studies ignore population stratification or implicitly assume its redundancy if the data were collected from continental groups (e.g., Ooi et al, 2019). Groups are assigned either by self-identified ethnicity or inferred by comparison to the Hapmap or 1000 Genomes populations, and analyse each cluster independently (e.g., Ooi et al, 2019).…”

“…Many association studies ignore population stratification or implicitly assume its redundancy if the data were collected from continental groups (e.g., Ooi et al, 2019). Groups are assigned either by self-identified ethnicity or inferred by comparison to the Hapmap or 1000 Genomes populations, and analyse each cluster independently (e.g., Ooi et al, 2019). This approach is implausible, not accounting for the existence of fine-scale structure everywhere (Lawson et al, 2020), and cannot be applied to more admixed populations, which is important where recent massive migrations have occurred, such as in the Americas.…”

“…In an analysis which is limited to summary data (e.g. Ooi et al, 2019), population stratification bias is difficult to identify, and the analysis is often run without adjustment for possible population differences. Sometimes, the homogeneity of the dataset is assumed due to the continental affiliation of the cohort (e.g., in the case of (Ooi et al, 2019), (Pickrell et al, 2016) analysed third party summary statistics calculated for "Europeans").…”

“…Ooi et al, 2019), population stratification bias is difficult to identify, and the analysis is often run without adjustment for possible population differences. Sometimes, the homogeneity of the dataset is assumed due to the continental affiliation of the cohort (e.g., in the case of (Ooi et al, 2019), (Pickrell et al, 2016) analysed third party summary statistics calculated for "Europeans"). Linkage Disequilibrium (LD) score regression (Loh et al, 2015) can estimate the sample overlap between summary statistics, but this is reliant on relatively large samples and often not used in MR pipelines.…”