Table 2.
Meta-analysis of multivariable relative risks of association between egg consumption and cardiovascular disease using random effects models
| Outcome | No of risk estimates | No of participants | No of events | Pooled relative risk (95% CI) | I2 (%) |
|---|---|---|---|---|---|
| Cardiovascular disease* | |||||
| One egg per day increase | 33 | 1 720 108 | 139 195 | 0.98 (0.93 to 1.03) | 62.3 |
| High v low intake | 34 | 1 730 088 | 139 259 | 0.99 (0.93 to 1.06) | 52.9 |
| Coronary heart disease | |||||
| One egg per day increase | 21 | 1 411 261 | 59 713 | 0.96 (0.91 to 1.03) | 38.2 |
| High v low intake | 22 | 1 421 241 | 59 777 | 0.97 (0.91 to 1.04) | 42.1 |
| Stroke† | |||||
| One egg per day increase | 22 | 1 059 315 | 53 617 | 0.99 (0.91 to 1.07) | 71.5 |
| High v low intake | 22 | 1 059 315 | 53 617 | 0.96 (0.88 to 1.06) | 53.0 |
| Cardiovascular disease composite‡ | |||||
| One egg per day increase | 17 | 940 464 | 112 377 | 1.01 (0.93 to 1.11) | 76.6 |
| High v low intake | 17 | 940 464 | 112 377 | 1.05 (0.94 to 1.17) | 67.3 |
| People with type 2 diabetes | |||||
| Cardiovascular disease*: | |||||
| One egg per day increase | 9 | 28 608 | 3663 | 1.25 (0.99 to 1.59) | 64.6 |
| High v low intake | 10 | >28 608§ | 4774 | 1.40 (1.00 to 1.97) | 65.1 |
Cardiovascular disease meta-analysis includes risk estimates from all included studies. The risk estimate for total cardiovascular disease events was used when available. For studies that reported risk estimates for coronary heart disease and stroke, but not for total cardiovascular disease events, risk estimates for coronary heart disease and stroke were pooled using fixed effect meta-analysis, and pooled estimates were used in cardiovascular disease meta-analysis.
For stroke, risk estimate for total stroke was used when available. For studies that reported risk estimates for ischemic stroke and hemorrhagic stroke, but not for total stroke, risk estimates for ischemic stroke and hemorrhagic stroke were pooled using fixed effect meta-analysis, and pooled estimates were used in stroke meta-analysis.
Cardiovascular disease composite meta-analysis includes only risk estimates for total cardiovascular disease events from studies that reported risk estimates for total cardiovascular disease events.
In one study, the number of participants with type 2 diabetes was not provided.