Summary of findings of the network meta-analysis for primary outcome (clinically important upper gastrointestinal bleeding)
| Outcome | CoE | Classification | Intervention | RR (95% CI) | Anticipated absolute effect (95% CI) | SUCRA | Number of studies (participants) |
| Clinically important upper gastrointestinal bleeding | High (moderate to high) | The most effective | Ranitidine+antacids | 0.13 (0.03 to 0.62)* | Not Applicable† | 85.3 | Not applicable† |
| Cimetidine | 0.56 (0.40 to 0.77)* | 79 fewer per 1000 (from 108 fewer to 41 fewer) | 37.9 | 13 (1288) | |||
| Ranitidine | 0.54 (0.38 to 0.76)* | 88 fewer per 1000 (from 118 fewer to 46 fewer) | 40.8 | 12 (871) | |||
| Antacids | 0.48 (0.33 to 0.68)* | 87 fewer per 1000 (from 113 fewer to 54 fewer) | 50.7 | 8 (708) | |||
| Sucralfate | 0.54 (0.39 to 0.75)* | 58 fewer per 1000 (from 78 fewer to 32 fewer) | 40.4 | 8 (698) | |||
| Inferior to the most effective/superior than the least effective | – | – | – | ||||
| Among the least effective | Prostaglandin analogues | 1.26 (0.53 to 2.99)‡ | 27 more per 1000 (from 49 fewer to 206 more) | 6.5 | 1 (58) | ||
| Low (low to very low) | May be the most effective | Omeprazole | 0.33 (0.20 to 0.56)*§¶ | 25 fewer per 1000 (from 30 fewer to 16 fewer) | 69.9 | 3 (371) | |
| May be inferior to the most effective/superior than the least effective | Lansoprazole | 0.10 (0.01 to 0.94)*§ | 90 fewer per 1000 (from 99 fewer to 6 fewer) | 84.2 | 1 (120) | ||
| Esomeprazole | 0.19 (0.06 to 0.63) | Not Applicable† | 80.8 | Not Applicable† | |||
| Pirenzepine | 0.42 (0.19 to 0.93)*§ | 56 fewer per 1000 (from 79 fewer to 7 fewer) | 54.8 | 2 (131) | |||
| May be among the least effective | Bioflavonoids | 0.53 (0.22 to 1.28)‡ | Not applicable† | 43 | Not applicable† | ||
| Pantoprazole | 0.73 (0.36 to 1.47)*‡ | 10 fewer per 1000 (from 24 fewer to 18 more) | 24.8 | 4 (3784) | |||
| Famotidine | 0.52 (0.22 to 1.24)*§** | 53 fewer per 1000 (from 86 fewer to 26 more) | 42.5 | 2 (196) | |||
| Rabeprazole | 0.06 (0.00 to 1.22)†† | Not applicable† | 87.4 | Not applicable† | |||
| Teprenone | 0.54 (0.12,2.45) | Not applicable† | 42.1 | Not applicable† |
*Downgraded one level for risk of bias because the trial(s) included in the analysis was/were at high risk of bias.
†Represents evidence coming only from indirect comparisons.
‡Downgraded one level for imprecision because the credible intervals were wide (included clinical benefit and harms).
§Downgraded one level for imprecision because the sample size was small.
¶Downgraded two levels for incoherence.
**Downgraded one level for inconsistency because there was evidence of statistical heterogeneity.
††Downgraded two levels for imprecision because the credible intervals were very wide (included clinical benefit and harms).
RR, relative risk; SUCRA, surface under the cumulative ranking curve.