A. Hammermeister Suger, T. A. Harrison, C. B. Haas, H. Chen, S. Fan, M. K. Bolla, Q. Wang, J. Dennis, K. Michailidou, A. M. Dunning, D. F. Easton, M. C. Southey, P. A. Fasching, L. Haeberle, J. Stone, M. Gago-Dominguez, J. E. Castelao, K. Aronson, R. A. Murphy, D. G. Evans, S. Loi, kConFab Investigators, F. J. Couch, S. Yadav, R. L. Milne, C. A. Haiman, C. G. Scott, A. Norman, E. M. John, A. M. Mulligan, V. N. Kristensen, A. H. Eliassen, I. L. Andrulis, J. D. Figueroa, S. Astley, K. Czene, P. D. P. Pharoah, A. C. Antoniou, M. Garcia-Closas, Berrington
Observational studies have identified associations between circulating inflammatory markers (CIMs) and mammographic density (MD) phenotypes, but whether these associations reflect causal effects remains unclear. We utilized two-sample Mendelian randomization (MR) to assess relationships between 60 CIMs and MD phenotypes in individuals of European genetic ancestry by examining associations between genetically predicted CIM concentrations and three MD phenotypes (dense area, DA: N = 13,965; non-dense area, NDA: N = 14,036; percent density, PMD: N = 17,841). Primary analyses were conducted using the inverse variance weighted MR method. No CIM-MD phenotype associations reached statistical significance after correction for multiple testing. At a nominal significance level of p < 0.05, we observed suggestive evidence for eight associations represented by six unique CIMs (MD phenotype/direction of association): CCL23 (DA/+), CD40 (NDA/-), GDNF (DA/+ & PMD/+), MMP-10 (DA/+ & PMD/+), TRANCE (DA/-), and TWEAK (NDA/+). In multivariable MR analyses accounting for body mass index (BMI), we observed six nominal associations represented by six unique CIMs, although most instruments demonstrated weak conditional instrument strength after inclusion of BMI. Overall, our comprehensive analysis of 60 CIMs provides limited evidence supporting broad causal effects of circulating inflammatory markers contributing to variation in MD phenotypes.