Abstract
Circulating adipokines and C-reactive protein (CRP) have been linked to breast cancer risk in observational epidemiological studies. The causal nature of these associations is unclear because of the susceptibility of conventional observational designs to residual confounding, reverse causation and other forms of bias. Mendelian randomisation (MR) uses genetic variants as proxies for risk factors to strengthen causal inference in observational settings. We performed a MR analysis to evaluate the causal relevance of six previously reported circulating adipokines [adiponectin, hepatocyte growth factor (HGF), interleukin-6, leptin receptor, plasminogen activator inhibitor-1 and resistin] and CRP in risk of overall and oestrogen receptor-stratified breast cancer in up to 122,977 cases and 105,974 controls of European ancestry. Genetic instruments were constructed from single-nucleotide polymorphisms robustly (p < 5 × 10(-8) ) associated with risk factors in genome-wide association studies. Colocalisation was performed as a sensitivity analysis to examine whether findings reflected shared causal variants or genomic confounding. In MR analyses, there was evidence for an association of HGF with oestrogen receptor-negative cancer (odds ratio per standard deviation increase: 1.17, 95% confidence interval: 1.01-1.35; p = 0.035) but little evidence for associations of other adipokines or CRP with overall or oestrogen receptor-stratified breast cancer. Colocalisation analysis suggested that the association of HGF with oestrogen receptor-negative breast cancer was unlikely to reflect a causal association. Collectively, these findings do not support an important aetiological role of various adipokines or CRP in overall or oestrogen receptor-specific breast cancer risk.