Supplementary Table 4.
Results of sensitivity analyses examining association between MA and CHD.
| Matching method | Caliper | No (%) of CHD patients | OR (95% CI) | p value |
|---|---|---|---|---|
| The nearest-neighbor Mahalanobis matching with replacement | 0.20 | 48 (41.4%) | 1.13 (1.02–1.26) | 0.03 |
| 0.25 | 49 (41.9%) | 1.15 (1.03–1.29) | 0.01 | |
| NA | 51 (43.2%) | 1.15 (1.04–1.30) | 0.01 | |
| The 1: 1 nearest-neighbor matching | 0.20 | 45 (37.2%) | 1.15 (1.03–1.29) | 0.01 |
| 0.25 | 46 (36.8%) | 1.14 (1.03–1.28) | 0.02 | |
| NA | 51 (37.5%) | 1.13 (1.02–1.25) | 0.02 |
NA=matching without a caliper. Matched factors included age, sex, the year of physical examination and medication history (antiplatelet drug and lipid-lowering drug). Models were adjusted for age, sex, the year of physical examination. CI – confidence interval; CHD – coronary heart disease; MA – maximum amplitude; OR – odds ratio.