Table 3. Summary of two-arm meta analyses of chest imaging findings between COVID-19 and Non-COVID-19 patients.
| Feature | No. of studies | Cochran’s Q statistic for homogeneity | Significance | Pooled odds ratio | Lower 95% confidence limit | Upper 95% confidence limit |
|---|---|---|---|---|---|---|
| Air bronchogram | 2 | 8.8982 | ** | 1.7278 | 0.1380 | 21.6247 |
| Bilateral lung involvement | 3 | 16.7833 | *** | 5.2572 | 0.8808 | 31.3799 |
| Central involvement | 2 | 2.3047 | ns | 0.1067 | 0.0106 | 1.0705 |
| Consolidation | 3 | 2.1513 | ns | 0.7589 | 0.4792 | 1.2018 |
| Crazy paving pattern | 2 | 1.6045 | ns | 1.5246 | 0.2592 | 8.9672 |
| Ground glass opacity | 4 | 5.7531 | ns | 7.0664 | 3.0037 | 16.6242 |
| Lymphadenopathy | 3 | 1.2547 | ns | 0.3185 | 0.1453 | 0.6982 |
| Ground glass opacity + consolidation | 2 | 0.7161 | ns | 1.2531 | 0.8573 | 1.8321 |
| Pleural effusion | 3 | 13.6684 | ** | 0.2911 | 0.0223 | 3.8004 |
| Peripheral involvement | 2 | 3.2278 | ns | 8.0191 | 0.6149 | 104.5740 |
| Unilateral lung involvement | 2 | 13.3551 | *** | 0.5207 | 0.0230 | 11.7804 |
Degrees of freedom for Q statistic = number of studies less 1; ns, not significant, i.e., P>0.05. **, highly significant, i.e. P < 0.01; ***, very highly significant, i.e., P<0.001.