Table 2.
Results from the best fitting models for Mendelian randomization
| Exposure phenotype | Outcome phenotype | Number of SNPs | Model | Beta | SE | P-value |
|---|---|---|---|---|---|---|
| CH | IOPcc | 144 | IVW | −0.128 | 0.024 | <1E-3 |
| Egger | 0.048 | 0.065 | 0.462 | |||
| Egger intercept | −0.011 | 0.004 | 0.014 | |||
| CRF | IOPcc | 166 | IVW | 0.31 | 0.021 | <1E-3 |
| Egger | 0.005 | 0.054 | 0.928 | |||
| Egger intercept | 0.018 | 0.004 | <1E-3 | |||
| IOPcc | CH | 99 | IVW | −0.154 | 0.009 | <1E-3 |
| Egger | −0.153 | 0.032 | <1E-3 | |||
| Egger intercept | 0.002 | 0.004 | 0.537 | |||
| IOPcc | CRF | 96 | IVW | 0.137 | 0.01 | <1E-3 |
| Egger | 0.12 | 0.031 | <1E-3 | |||
| Egger intercept | 0.002 | 0.004 | 0.54 |
For the model, IVW is the penalized inverse weighted variance, Egger is the penalized Egger model and Egger intercept is the corresponding intercept results for the given Egger results shown. Number of SNPs is the number of uncorrelated SNPs used as genetic instruments for each test.