Skip to main content
. 2005 Mar 12;330(7491):568. doi: 10.1136/bmj.38356.641134.8F

Table 3.

Summary results of subgroup analyses in patients with ulcer bleeding

Pooled rates (%)
Subgroup analyses and outcomes PPI Control Heterogeneity (P value) Odds ratio (95% CI) Number needed to treat (95% CI)
H2 receptor antagonist as control (14 trials)
Mortality 4.2 4.2 No (0.89) 1.02 (0.51 to 2.05) Not calculable
Rebleeding 10.4 17.6 No (0.51) 0.53 (0.35 to 0.78) 13 (8 to 25)
Surgical intervention 7.9 11.1 No (0.36) 0.68 (0.38 to 1.20) 25 (13 to 100)
Placebo as control (7 trials)
Mortality 5.7 4.9 Yes (0.02) 0.96 (0.43 to 2.15) Not calculable
Rebleeding 10.7 19.2 Yes (0.005) 0.41 (0.23 to 0.72) 11 (8 to 20)
Surgical intervention 8.7 13.5 Yes (0.09) 0.52 (0.32 to 0.84) 20 (13 to 50)
IV PPI versus IV placebo or IV H2 receptor antagonist (16 trials)



Mortality 6.0 5.0 No (0.15) 1.22 (0.84 to 1.78) Not calculable
Rebleeding 10.9 17.1 No (0.15) 0.57 (0.44 to 0.73) 17 (11 to 25)
Surgical intervention 9.0 12.6 No (0.55) 0.67 (0.50 to 0.90) 25 (17 to 100)
“High dose” IV PPI versus IV placebo or IV H2 receptor antagonist (4 trials)



Mortality 5.7 5.6 Yes (0.02) 0.98 (0.25 to 3.77) Not calculable
Rebleeding 5.2 12.8 Yes (0.09) 0.39 (0.18 to 0.87) 11 (5 to 50)
Surgical intervention 5.2 9.4 No (0.36) 0.53 (0.31 to 0.89) 25 (14 to 100)
IV PPI in doses other than “high dose” versus IV placebo or IV H2 receptor antagonist (12 trials)


Mortality 6.3 4.5 No (0.76) 1.42 (0.85 to 2.38) Not calculable
Rebleeding 16.1 20.3 No (0.53) 0.67 (0.50 to 0.91) 17 (10 to 100)
Surgical intervention 12.3 15.3 No (0.60) 0.76 (0.53 to 1.08) Not calculable
Oral PPI treatment versus oral placebo or H2 receptor antagonist (5 trials)



Mortality 2.5 3.6 No (0.64) 0.67 (0.28 to 1.64) Not calculable
Rebleeding 9.5 24.0 No (0.20) 0.32 (0.20 to 0.50) 7 (5 to 11)
Surgical intervention 6.5 14.4 No (0.47) 0.38 (0.22 to 0.66) 13 (8 to 25)
Routine EHT before randomisation (13 trials)




Mortality 4.1 4.1 No (0.24) 1.01 (0.64 to 1.61) Not calculable
Rebleeding 8.6 15.0 No (0.11) 0.52 (0.39 to 0.70) 14 (11 to 25)
Surgical intervention 5.4 9.3 No (0.73) 0.53 (0.35 to 0.79) 25 (14 to 50)
No routine EHT before randomisation (8 trials)




Mortality 7.0 5.6 No (0.34) 1.25 (0.75 to 2.09) Not calculable
Rebleeding 16.0 25.8 Yes (0.03) 0.38 (0.18 to 0.81) 9 (5 to 50)
Surgical intervention 12.8 18.6 No (0.16) 0.64 (0.46 to 0.89) 17 (10 to 50)
Routine EHT before randomisation and use of high dose IV PPI treatment (4 trials)


Mortality 5.7 5.6 Yes (0.02) 0.98 (0.25 to 3.77) Not calculable
Rebleeding 5.2 12.8 Yes (0.09) 0.39 (0.18 to 0.87) 11 (8 to 33)
Surgical intervention 5.2 9.4 No (0.36) 0.53 (0.31 to 0.89) 25 (17 to 100)
Routine EHT before randomisation and use of lower dose IV or oral PPI treatment (9 trials)


Mortality 2.4 2.4 No (0.80) 1.00 (0.42 to 2.35) Not calculable
Rebleeding 10.2 17.2 No (0.42) 0.52 (0.35 to 0.78) 14 (8 to 33)
Surgical intervention 6.6 9.9 No (0.71) 0.59 (0.33 to 1.05) Not calculable
Endoscopic findings of active bleeding or NBVV before randomisation (10 trials)



Mortality 2.8 5.5 No (0.79) 0.51 (0.26 to 1.01) Not calculable
Rebleeding 10.5 24.7 No (0.23) 0.33 (0.22 to 0.50) 7 (5 to 13)
Surgical intervention 5.6 12.2 No (0.69) 0.39 (0.23 to 0.65) 14 (10 to 33)
Active bleeding or NBVV; routine EHT before randomisation (5 trials)



Mortality 2.7 5.2 No (0.59) 0.51 (0.23 to 1.12) Not calculable
Rebleeding 11.6 25.7 Yes (0.09) 0.35 (0.17 to 0.72) 8 (6 to 17)
Surgical intervention 4.5 7.3 No (0.56) 0.54 (0.27 to 1.08) Not calculable
Active bleeding or NBVV; without routine EHT before randomisation (5 trials)



Mortality 3.5 6.5 No (0.57) 0.51 (0.12 to 2.12) Not calculable
Rebleeding 19.4 46.8 No (1.00) 0.29 (0.13 to 0.63) 4 (2 to 10)
Surgical intervention 8.7 24.6 No (0.81) 0.27 (0.12 to 0.57) 6 (4 to 14)

PPI=proton pump inhibitor, IV=intravenous, EHT=endoscopic haemostatic therapy, NBVV=non-bleeding visible vessel.