Table 2. Meta-analyses and subgroup analyses for mortality.
| Patients receiving immunotherapy vs. control patients | N of studies | Pooled OR (95%CI) a | I 2 (%) b | P | P for interaction |
| Overall | 14 | 0.92 (0.68, 1.25) | 4% | 0.61 | |
| Study type | |||||
| Prospective | 3 | 0.67 (0.43, 1.04) | 0% | 0.07 | 0.054 |
| Retrospective | 11 | 1.19 (0.81, 1.73) | 0% | 0.38 | |
| Number of patients | |||||
| <100 | 6 | 0.64 (0.28, 1.49) | 0% | 0.30 | 0.39 |
| >100 | 8 | 0.97 (0.63, 1.49) | 34% | 0.90 | |
| Cancer type | |||||
| Hematological malignancies | 3 | 0.72 (0.27, 1.91) | 0% | 0.51 | 0.48 |
| Solid tumor | 3 | 1.10 (0.69, 1.76) | 36% | 0.68 | |
| Non-specific cancer | 8 | 0.72 (0.39, 1.33) | 8% | 0.29 | |
| Immunotherapy interval before the COVID-19 diagnosis (days) | |||||
| >20 | 8 | 0.98 (0.56, 1.69) | 25% | 0.94 | 0.44 |
| <20 | 2 | 1.23 (0.54, 2.76) | 0% | 0.62 | |
| Number of patients receiving active immunotherapy | |||||
| <10 | 7 | 0.75 (0.27, 2.08) | 22% | 0.58 | 0.66 |
| >10 | 7 | 0.95 (0.70, 1.28) | 0% | 0.73 |
aCalculated by using the random-effect model.
bI 2 means the inconsistency across studies.