TABLE 3.
Prespecified subgroup meta-analyses for type 2 diabetes risk, per 1-egg/d increase, using random-effects models1
| Stratification, subcategories | Risk estimates, n | Participants, n | Cases, n | Pooled RR (95% CI) | I 2, % | P-interaction |
|---|---|---|---|---|---|---|
| Region | 0.01 | |||||
| United States | 10 | 339,377 | 28,528 | 1.18 (1.10, 1.27) | 51.3 | |
| Europe | 8 | 179,714 | 10,698 | 0.99 (0.85, 1.15) | 73.5 | |
| Asia | 4 | 70,468 | 2022 | 0.82 (0.62, 1.09) | 59.1 | |
| Sex | 0.68 | |||||
| Men | 9 | 153,935 | 11,998 | 1.01 (0.86, 1.19) | 81.4 | |
| Women | 7 | 316,176 | 21,502 | 1.12 (0.99, 1.26) | 66.4 | |
| Both | 6 | 119,448 | 7748 | 1.07 (0.99, 1.17) | 69.8 | |
| Subjects, n | 0.54 | |||||
| <10,000 | 9 | 31,188 | 3179 | 0.90 (0.71, 1.16) | 67.1 | |
| >10,000 | 13 | 558,371 | 38,069 | 1.13 (1.06, 1.20) | 57.6 | |
| Risk of bias2 | 0.78 | |||||
| Low | 15 | 500,718 | 36,793 | 1.08 (0.99, 1.17) | 75.9 | |
| Unclear to high | 7 | 88,841 | 4455 | 1.03 (0.89, 1.19) | 25.8 | |
| Dietary assessment | 0.89 | |||||
| Baseline only | 11 | 279,005 | 16,358 | 1.05 (0.93, 1.18) | 74.1 | |
| Repeated measurements | 11 | 310,554 | 24,890 | 1.09 (0.99, 1.20) | 65.0 | |
| Adjustment for dietary confounders3 | 0.47 | |||||
| Suboptimal | 10 | 108,114 | 9215 | 1.10 (0.94, 1.29) | 79.8 | |
| Sufficient | 12 | 481,445 | 32,033 | 1.06 (0.99, 1.14) | 50.9 |
1
NOS, Newcastle-Ottawa Scale.
2
Low: NOS score ≥7; unclear to high risk of bias: NOS score <7.
3
According to the comparability criteria for control for secondary confounders of the NOS.