Table 2. Results of this meta-analysis.
| Association | Heterogeneity | ||||||
|---|---|---|---|---|---|---|---|
| Num | SMD (95% CI) | P value | Model | I2 (%) | P value | ||
| overall | 9 | −1.08 (−0.136,– 0.08) | P < 0.001 | R | 74.3 | P < 0.001 | |
| Study design | |||||||
| case-control | 6 | −1.388 (−1.751, −1.026) | P < 0.001 | R | 47.6 | P = 0.09 | |
| cohort | 3 | −0.733 (−0.945, −0.520) | P < 0.001 | R | 52.5 | P = 0.122 | |
| Race | |||||||
| Caucasian | 3 | −0.918 (−1.443, −0.393) | P < 0.001 | R | 50.5 | P = 0.133 | |
| Asian | 6 | −1.115 (−1.526, −0.784) | P < 0.001 | R | 80.6 | P < 0.001 | |
| Sample collection | |||||||
| Serum | 3 | −1.113 (−1.472, −0.755) | P < 0.001 | R | 23.5 | P = 0.27 | |
| Blood | 5 | −1.241 (−1.795, −0.678) | P < 0.001 | R | 84.7 | P < 0.001 | |
| Toenail | 1 | −0.889 (−1.081, −0.697) | P < 0.001 | R | 0 | P < 0.001 | |
Num, number; SMD, standardized mean difference; CI, confidence intercal; R, randomeffects model.