Table 5.
Applied example: effect of three cytokines on stroke risk
| MCP‐1 | MCP‐3 | Eotaxin‐1 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Method | Pruning | Variants/PCs | Cond number | Estimate (SE) | p Value | Estimate (SE) | p Value | Estimate (SE) | p Value |
| All stroke | |||||||||
| MV‐IVW | 0.1 | 20 | 27.7 | 0.091 (0.045) | 0.046 | −0.062 (0.046) | 0.18 | 0.062 (0.082) | 0.45 |
| 0.4 | 75 | 1224 | 0.057 (0.035) | 0.11 | −0.014 (0.024) | 0.55 | 0.110 (0.050) | 0.028 | |
| 0.6 | 151 | 17762 | −0.038 (0.022) | 0.09 | −0.014 (0.017) | 0.41 | 0.040 (0.029) | 0.17 | |
| MV‐IVW‐PCA | 30 | 24.9 | 0.075 (0.041) | 0.071 | −0.029 (0.027) | 0.29 | 0.000 (0.063) | 0.99 | |
| Cardioembolic stroke | |||||||||
| MV‐IVW | 0.1 | 19 | 22.7 | 0.270 (0.095) | 0.005 | 0.151 (0.104) | 0.14 | −0.174 (0.169) | 0.30 |
| 0.4 | 70 | 870 | 0.141 (0.073) | 0.053 | 0.003 (0.051) | 0.96 | −0.132 (0.107) | 0.21 | |
| 0.6 | 145 | 15,790 | 0.089 (0.046) | 0.056 | 0.040 (0.036) | 0.27 | 0.019 (0.062) | 0.76 | |
| MV‐IVW‐PCA | 29 | 25.9 | 0.254 (0.089) | 0.004 | −0.018 (0.065) | 0.78 | −0.108 (0.145) | 0.46 | |
Note: Estimates (standard errors, SE) and p values from MV‐IVW and MV‐IVW‐PCA methods. Variants/PCs indicates the number of genetic variants (MV‐IVW method) or principal components (PCs, MV‐IVW‐PCA method) included in the analysis. Cond number indicates the condition number of the variance‐covariance matrix ; larger numbers signal worse problems due to ill‐conditioning. Estimates represent log odds ratios per 1 standard deviation increase in the cytokine.