Table 2. Multi-variant, pleiotropy robust MR methods.

Method	Robust to pleiotropy by …	CHD Causal OR (95% CI)	MI Causal OR (95% CI)
MR Egger	Intercept-based adjustment for global effect of pleiotropy	0.98 (0.70, 1.39)	0.89 (0.63, 1.26)
Weighted MBE (phi = 1)	Assuming causal effect is most common shared effect across variants	1.09 (0.89, 1.33)	0.98 (0.79, 1.21)
Weighted Median	Assuming most (≥ 50%) genetic instruments are unaffected by pleiotropy	1.12 (0.91, 1.37)	1.03 (0.82, 1.29)
MR PRESSO	Assuming <50% of genetic instruments have horizontal pleiotropy	1.18 (0.98, 1.42)	1.17 (0.98, 1.40)

To further examine potential effects of pleiotropy we ran several multi-variant, pleiotropy robust models including MR Egger, Weighted Mode Based Estimator (MBE), Weighted Median, and MR PRESSO. Each uses a different means to account for pleiotropy and has different assumptions used to estimate the causal effect in the presence of pleiotropy. Odds ratios are per 1 g/L increase in genetically determined fibrinogen. CHD = coronary heart disease; CI = confidence interval; MI = myocardial infarction; MR = mendelian randomization