| Method1 | # IV | Exp. var.2 (%) | OR | 95% CI | p-val | Q p-val3 | intercept p-val4 |
---|
Main analysis |
Two-sample MR | IVW | 98 | 7.1 | 0.974 | 0.970 | 0.978 | 3.1 × 10–40 | > 0.99 | |
Sensitivity analyses |
Two-sample MR | Egger | 98 | 7.1 | 0.983 | 0.950 | 1.017 | 0.33 | > 0.99 | 0.60 |
Two-sample MR | WM | 98 | 7.1 | 0.975 | 0.969 | 0.980 | 1.1 × 10–20 | |
HGDP as reference5 | IVW | 95 | 6.6 | 0.974 | 0.970 | 0.978 | 6.1 × 10–37 | > 0.99 |
Stratified analyses |
BC diagnosed ≤ 45 y6 | IVW | 106 | 7.7 | 0.978 | 0.972 | 0.984 | 1.2 × 10–13 | > 0.99 | |
Familial BC7 | IVW | 104 | 7.5 | 0.958 | 0.952 | 0.964 | 2.0 × 10–41 | 0.98 |
ER-positive BC8 | IVW | 94 | 6.8 | 0.968 | 0.963 | 0.972 | 7.1 × 10–43 | > 0.99 |
ER-negative BC | IVW | 111 | 8.1 | 0.988 | 0.981 | 0.995 | 1.1 × 10–03 | > 0.99 |
Triple-negative BC9 | IVW | 110 | 7.9 | 0.993 | 0.983 | 1.003 | 0.19 | > 0.99 |
- Bold values indicate that
the probability value is below 0.05
- 1IVW: Inverse variance weighted; Egger: MR-Egger regression; WM: weighted median estimates
- 2Exp. var.: Variance explained by the instrumental variables
- 3Cochran’s Q statistic p-values higher than 0.05 are suggestive of no instrument heterogeneity as a proxy for pleiotropy
- 4MR-Egger intercept p-values higher than 0.05 are consistent with no horizontal pleiotropy
- 5Individuals from the Human Genome Diversity Project were used instead of the 29 Native Americans from Colombia as the reference panel for estimation of global ancestry proportions
- 6Breast cancer diagnosed before the age of 45 years
- 7Breast cancer patients with first-degree relatives affected by breast and/or ovarian cancer
- 8Oestrogen receptor-positive breast cancer
- 9Oestrogen receptor-negative, progesterone receptor-negative and human epidermal growth factor receptor-negative breast cancer