Effect of lipids on IGF axis and subsequent breast cancer risk: observational and causal estimates from independent sources


Session type:


Vanessa Tan1,Caroline Bull1,Nicholas Timpson1
1University of Bristol



Circulating lipids have been associated with breast cancer (BCa), however, the mechanisms underpinning this association are not fully understood.  Epidemiological evidence suggests that pharmacological manipulation of circulating lipids alters the abundance of serum insulin-like growth factors (IGFs), which have been linked to BCa risk in observational studies.


Mendelian randomization (MR), observational and recall-by-genotype analyses were conducted to investigate the relationship between circulating lipids, IGFs and BCa risk, using multiple independent lines of evidence.


Using MR, low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) were positively associated with BCa (OR 1.07; 95% CI 1.02 to 1.12 and OR 1.06; 95% CI 1.01 to 1.11, respectively, per SD increase in lipid trait). Triglycerides (TG) were inversely associated with BCa (OR 0.94; 95% CI 0.88 to 1.00, per SD increase). Observationally, LDL-C, HDL-C and TG associated with IGF-I: 0.03 (95% CI 0.004 to 0.06), -0.05 (95% CI -0.08 to -0.02) and -0.06 (95% CI -0.09 to -0.04) SD change in IGF-I per SD increase in lipid trait, respectively. Using MR, TG was inversely associated with IGF-I (-0.13 SD; 95% CI -0.22 to -0.03 per SD increase in TG). The recall by genotype analysis supported evidence for an inverse relationship between TG and IGF-I. Using MR, IGF-I was positively associated with BCa (OR 1.10; 95% CI 1.02 to 1.18, per SD increase).


This study builds on evidence linking circulating lipid traits with BCa and provides novel evidence that TG may alter circulating levels of IGF-I and subsequent BCa risk.