Table 3. Stratified analyses of pooled relative risks (95% CI) of AF for HbA1c levels as categorical variable.
| Stratified analyses | No. of studies | Pooled RR(95% CI) | Heterogeneity test | |
|---|---|---|---|---|
| P value | I2(%) | |||
| style | ||||
| prospective | 6 | 1.09(1.00,1.18) | 0.37 | 6.70% |
| retrospective | 1 | 0.37(0.20,0.67) | . | . |
| Country | ||||
| America | 4 | 1.09(1.01,1.17) | 0.44 | 0.00% |
| Not America | 3 | 0.57(0.18,1.77) | 0.03 | 72.30% |
| Publication year | ||||
| before 2011 | 2 | 0.73(0.12,4.59) | 0.10 | 62.20% |
| After 2012 | 5 | 0.99(0.83,1.18) | <0.01 | 73.20% |
| Sample size | ||||
| <3000 | 3 | 0.57(0.18,1.77) | 0.03 | 72.30% |
| ≥3000>ze7H耂 ≥96m>ez7I耂 ≥6>ez71.13(1.09,1.18)J耀1.13(1.09,1.18)70. |
4 | 1.09(1.01,1.17) | 0.44 | 0.00% |
| Follow-up duration | ||||
| <96m | 3 | 0.95(0.56,1.63) | 0.24 | 29.70% |
| ≥96m>ez7I耂 ≥6>ez71.13(1.09,1.18)J耀 1.13(1.09,1.18)70.83( |
3 | 1.10(1.02,1.19) | 0.48 | 0.00% |
| age | ||||
| <63y | 5 | 0.99(0.83,1.18) | <0.01 | 73.20% |
| ≥63y | 2 | 0.73(0.12,4.59) | 0.10 | 62.20% |
| quintiles | ||||
| 2 | 3 | 0.95(0.56,1.63) | 0.24 | 29.70% |
| 3 | 2 | 0.68(0.22,2.11) | <0.01 | 92.60% |
| 4 | 2 | 1.05(0.95,1.17) | 0.96 | 0.00% |
| prime effect size | ||||
| OR | 1 | 0.37(0.20,0.67) | . | . |
| RR | 3 | 0.95(0.56,1.63) | 0.24 | 29.70% |
| HR | 3 | 1.10(1.02,1.19) | 0.48 | 0.00% |
| Adjustment for CABG | ||||
| Yes | 1 | 1.49(0.61,3.64) | . | . |
| No | 6 | 0.98(0.82,1.17) | <0.01 | 70.70% |
| Study quality | ||||
| <6 | 4 | 0.69(0.35,1.36) | 0.02 | 71.00% |
| ≥6 | 3 | 1.10(1.02,1.19) | 0.48 | 0.00% |
Note: I2 is interpreted as the proportion of total variation across studies that is due to heterogeneity rather than chance; CABG, coronary artery bypass grafting.