Table 4.
Risk of bias
| Reference | Randomization method | Allocation concealment | Incomplete outcome assessment | Selective reporting |
|---|---|---|---|---|
| CEASL30 | Low | Low | Low | Low |
| Ohlsson et al.17 | Unclear | Unclear | Low | Low |
| Pietra et al.22 | Unclear | Unclear | Low | Low |
| Rodriguez-Moranta et al.2 | Low | Low | Low | Low |
| Mäkelä et al.16 | Unclear | Unclear | Low | Low |
| GILDA4 | Unclear | Low | Low | Low |
| FACS3 | Unclear | Unclear | Low | Low |
| Kjeldsen et al.20 | Unclear | Unclear | High* | Low |
| Wang et al.27 | Unclear | Low | Unclear | Low |
| Wattchow et al.24 | Unclear | Low | Low | Low |
| Augestad et al.25 | Low | Low | Low | Low |
| Schoemaker et al.21 | Low | High† | Unclear | Low |
| Secco et al.23 | High‡ | Unclear | High§ | Low |
| CEAwatch31 | Low | Low | Low | Low |
*
Groups not balanced (290 : 307); drop-outs may not be included in assessment.
†
Cards not in envelopes; groups not balanced (167 : 158).
‡
Method unclear; groups not balanced (108 : 84 and 84 : 61).
§
Patients excluded from survival rather than censored if lost to follow-up. COLOFOL has not reported so cannot be assessed; the methods of assignment were not described by Barillari et al.18, and this did not appear to be a true randomized comparison.