ORIGINAL ARTICLE
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING
Proton Pump Inhibitor Therapy for Peptic Ulcer Bleeding:
Cochrane Collaboration Meta-analysis of Randomized Controlled Trials
GRIGORIS I. LEONTIADIS, MD, PHD; VIRENDER K. SHARMA, MD; AND COLIN W. HOWDEN, MD
OBJECTIVE: To evaluate the efficacy of proton pump inhibitors
(PPIs) in treating peptic ulcer bleeding.
MATERIAL AND METHODS: We searched the MEDLINE, EMBASE,
CENTRAL, Cochrane Library, and metaRegister of Controlled Trials
databases and published proceedings of major meetings through
November 2004 for randomized controlled trials that compared
oral or intravenous PPIs with placebo or a histamine2-receptor
antagonist for peptic ulcer bleeding. Pharmaceutical companies
and relevant experts were contacted. Data extraction and assess-
ment of study validity were performed independently in duplicate.
Assessed outcomes were 30-day all-cause mortality, rebleeding,
surgery, and repeated endoscopic treatment. Influence of study
characteristics on outcomes was examined by subgroup analyses
and meta-regression.
RESULTS: We included 24 trials (4373 participants). Statistical
heterogeneity was evident only for rebleeding. Treatment with
PPIs had no significant effect on mortality (odds ratio [OR], 1.01;
95% confidence interval [CI], 0.74-1.40; number needed to treat
[NNT], incalculable) but significantly reduced rebleeding (OR,
0.49; 95% CI, 0.37-0.65; NNT, 13) and the need for surgery (OR,
0.61; 95% CI, 0.48-0.78; NNT, 34) and repeated endoscopic
treatment (OR, 0.32; 95% CI, 0.20-0.51; NNT, 10). Results were
similar when analysis was confined to trials with adequate alloca-
tion concealment. Treatment with PPIs significantly reduced mor-
tality in Asian trials (OR, 0.35; 95% CI, 0.16-0.74; NNT, 34) and in
patients with active bleeding or a nonbleeding visible vessel (OR,
0.53; 95% CI, 0.31-0.91; NNT, 50).
CONCLUSIONS: In ulcer bleeding, PPIs reduce rebleeding and the
need for surgery and repeated endoscopic treatment. They im-
prove mortality among patients at highest risk.
Mayo Clin Proc. 2007;82(3):286-296
CI = confidence interval; H2RA = histamine2-receptor antagonist; IV =
intravenous; NNT = number needed to treat; OR = odds ratio; PPI =
proton pump inhibitor; RCT = randomized controlled trial
I n our original Cochrane Collaboration systematic review
and meta-analysis, we found that proton pump inhibi-
tor (PPI) treatment reduced rates of rebleeding and surgi-
cal intervention after peptic ulcer bleeding but did not
reduce all-cause mortality.1,2 The Cochrane Collabora-
tion aims to provide regularly updated, evidence-based
summaries of treatment effect. Therefore, we have up-
dated our original review with the inclusion of more
recent randomized controlled trials (RCTs). Our aim was
to evaluate the overall benefit of PPI treatment in the
management of ulcer bleeding since this is an impor-
tant clinical issue with major associated health care costs.
This article is based on a regularly updated Cochrane
review.3
286 Mayo Clin Proc. • March 2007;82(3):286-296
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
MATERIAL AND METHODS
We included trials that compared intravenous or oral PPIs
with placebo or a histamine2-receptor antagonist (H2RA) in
patients with ulcer bleeding. All trials had to confirm endo-
scopically the diagnosis of active or recent bleeding from
peptic ulcer, use a concurrent control group, apply con-
comitant therapy equally to both intervention arms, and
report 1 or more of the following: mortality, rebleeding,
and surgical intervention.
Our primary outcome measure was mortality from any
cause within 30 days of randomization. Secondary out-
comes included rebleeding (as defined by the original au-
thors), surgery, and further endoscopic treatment within 30
days of randomization.
We identified trials by searching the CENTRAL,
Cochrane Library (Issue 4, 2004), MEDLINE (1966 to
November 2004), and EMBASE (1980 to November 2004)
databases with no language restrictions. The search used a
combination of subject headings and text words relating to
the use of PPIs for the treatment of bleeding ulcers; the
standard Cochrane search strategy filter for identifying
RCTs was applied (Appendix).4 Reference lists from trials
From the Division of Gastroenterology, 2nd Department of Internal Medicine,
Democritus University of Thrace School of Medicine, Alexandroupolis, Greece
(G.I.L.); Division of Gastroenterology, Mayo Clinic College of Medicine,
Scottsdale, Ariz (V.K.S); and Division of Gastroenterology, Northwestern Uni-
versity Feinberg School of Medicine, Chicago, Ill (C.W.H.).
Dr Leontiadis receives grants to attend meetings from AstraZeneca Pharma-
ceuticals LP, Janssen-Cilag, GlaxoSmithKline, and Sanofi-Aventis. Dr Sharma
receives speaking honoraria and research support from TAP Pharmaceutical
Products Inc and is a consultant for Astra-Zeneca Pharmaceuticals LP. Dr
Howden is a consultant for TAP Pharmaceutical Products Inc, Altana,
Santarus, Inc, Schwarz Pharma, NEGMA-GILD, Allergan Inc, Novartis Con-
sumer Health, Takeda Pharmaceuticals America, Inc, and Valeant Pharmaceu-
ticals International; receives speaking honoraria from TAP Pharmaceutical
Products Inc, Santarus, Inc, Takeda (Japan), Takeda (Austria), Abbott
(Canada), Wyeth (US), Wyeth (Ireland), NEGMA-GILD, and AstraZeneca Phar-
maceuticals LP, and receives research support from AstraZeneca Pharmaceu-
ticals LP.
Data were presented in part at the 106th Annual Meeting of the American
Gastroenterological Association, Chicago, Ill, May 18, 2005, and have been
published in abstract form (Leontiadis et al. Gastroenterology. 2005;
128[suppl 2]:A-639. Abstract W1568). This report is based on a review3
published online in The Cochrane Library (www.thecochranelibrary.com),
which will be updated regularly.
Individual reprints of this article are not available. Address correspondence to
Colin W. Howden, MD, Division of Gastroenterology, Northwestern University,
Feinberg School of Medicine, 676 N St Clair St, Suite 1400, Chicago, IL
60611 (e-mail: c-howden @northwestern.edu).
© 2007 Mayo Foundation for Medical Education and Research
• www.mayoclinicproceedings.com

and review articles retrieved by electronic searching were
hand searched to identify further relevant trials. Abstracts
from Digestive Disease Week, United European Gastroen-
terology Week, and the meetings of the American College of
Gastroenterology, World Congress of Gastroenterology, and
British Society of Gastroenterology up to November 2004
were also hand searched. By a post hoc decision, registers of
controlled trials (active and archived registers by metaRegister
of Controlled Trials: www.controlled-trials.com/mrct/) were
searched for unpublished and/or ongoing trials in November
2004. Experts in the field and pharmaceutical companies
that market PPIs in the United States or Europe were
contacted for additional published or unpublished data.
Two reviewers (G.I.L., C.W.H.) independently checked
all identified trials for fulfillment of inclusion criteria. We
extracted data and assessed methodological quality, used
predesigned data extraction and validity assessment forms,
and resolved any disagreements by consensus. We ob-
tained the full text of all relevant trials and translations of
publications in languages other than English or French.
When necessary, we contacted original authors for further
information. If multiple publications of the same patient
groups were retrieved, we included only the most recent
version. Methodological quality assessment emphasized
concealment of allocation in accordance with Cochrane
Collaboration methods (grade A indicates adequate
concealment; grade B, uncertain; grade C, inadequate;
and grade D, not randomized).4 Data were extracted on
primary and secondary outcomes by treatment group;
method of randomization; inclusion and exclusion crite-
ria; details of all therapeutic interventions, including dose,
duration, mode of administration, and control treatment;
and details of any cointerventions, including endoscopic
treatment. We also noted age, sex, and ethnicity of pa-
tients; numbers of patients with comorbid conditions; site
of bleeding ulcer (duodenal or gastric); signs of recent
hemorrhage at baseline endoscopy (spurting, oozing, non-
bleeding visible vessel, and adherent clot); degree of blind-
ing of outcomes assessors, patients, and investigators;
numbers of patients withdrawn with reasons; and adverse
reactions to treatment.
We assessed heterogeneity statistically with the χ2 and
I tests4 and considered possible explanations for clinical
2
heterogeneity. We performed a meta-analysis of outcomes
when appropriate by the Mantel-Haenszel method and re-
ported pooled outcomes as odds ratios (ORs) with 95%
confidence intervals (CIs). When heterogeneity was statis-
tically significant (P<.10 for the χ2 test), we used the ran-
dom-effects model; otherwise, we applied a fixed-effect
model. We also derived pooled values for number needed
to treat (NNT) with 95% CIs. We reported crude pooled
rates for outcomes because these are of interest to clini-
Mayo Clin Proc. • March 2007;82(3):286-296
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING
cians. However, these are not weighted and therefore less
precise than the calculated OR and NNT values. We inves-
tigated the presence of publication bias visually with funnel
plots (SE of the logOR against OR) and statistically with the
Egger linear regression test.5 We concluded that there was
asymmetry when P<.10 and presented intercepts with 90%
CIs.
Predetermined subgroup analyses examined the influ-
ence of trial methodological quality, type of control treat-
ment (placebo or H2RA), initial use of endoscopic hemo-
static treatment, site of ulcer, severity of signs of recent
hemorrhage at initial endoscopy, route of PPI administra-
tion, PPI dose, and geographical location of trials. By
predetermined sensitivity analyses, we examined whether
pooled-effect estimates were robust to the exclusion of any
individual trial. We further assessed the influence of these
factors (including that of the actual PPI used) on treatment
effect (logOR) and on the heterogeneity of the analyses with
meta-regression analysis (random-effects model, within-
study variance estimated with the restricted maximum-
likelihood method) when sufficient data were available.
Analyses were performed with Cochrane Collaboration
software (RevMan, version 4.2.8 for Windows, Nordic
Cochrane Centre, Copenhagen, Denmark, 2003). The
Egger test and meta-regression were performed with Stata
statistical software (Intercooled Stata 8.2 for Windows,
StataCorp LP, College Station, Tex).
RESULTS
Our initial search identified 181 articles; reference lists
from these articles and conference proceedings identified
an additional 16 articles (Figure 1). We identified no addi-
tional trials by searching online registers of controlled trials
or by contacting experts in the field or pharmaceutical
companies that market PPIs. After reviewing the abstracts
of these articles, we excluded 143 as irrelevant and re-
trieved the full text for the remaining 54. Of these, 30 did
not meet eligibility criteria. The remaining 24 (4373 partici-
pants) were included in the systematic review6-29; details of
design, participants, interventions, and reported outcomes
are available online at www.mayoclinicproceedings.com
linked to this article. Five trials were available only as ab-
stracts6-10; 19 were full publications.11-29 Two were pub-
lished in French,13,17 1 in Spanish,14 1 in Chinese,29 and 20 in
English. Of the 19 full publications, only 1 was not indexed
in MEDLINE.29 Additional unpublished information was
obtained from the investigators for 2 trials.9,25
The main results are summarized in Table 1; forest plots
are shown in Figures 2 through 5. Treatment with PPIs did
not significantly affect overall mortality (OR, 1.01; 95%
CI, 0.74-1.40; NNT, incalculable) but significantly reduced
• www.mayoclinicproceedings.com 287
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING
197 Potentially relevant trials identified and screened for retrieval
54 Trials retrieved for more detailed evaluation
24 Potentially appropriate trials to be included in meta-analysis
24 Trials included in meta-analysis
Trials with usable information by outcome
20 Mortality
21 Rebleeding
19 Surgical intervention
7 Further endoscopic hemostatic treatment
after randomization
FIGURE 1. Flow diagram of search results.
rebleeding (OR, 0.49; 95% CI, 0.37-0.65; NNT, 13; 95%
CI, 10-25), surgical intervention (OR, 0.61; 95% CI, 0.48-
0.78; NNT, 34; 95% CI, 20-50), and requirement for fur-
ther endoscopic treatment after randomization (OR, 0.32;
95% CI, 0.20-0.51; NNT, 10; 95% CI, 8-17). These results
did not change substantively when, by sensitivity analysis,
any individual trial was excluded or when the analysis was
confined to the 12 trials with adequate concealment of
allocation.9,10,12,16,18,19,21,22,24-26,28
Visual inspection of funnel plots suggested possible
publication bias in missing small “negative” trials. How-
ever, the Egger test confirmed funnel plot asymmetry for
only surgical intervention (coefficient for bias, –0.920;
90% CI, –1.727 to –0.113; P=.06).
288 Mayo Clin Proc. • March 2007;82(3):286-296
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
143 Trials excluded as clearly not relevant
Main reasons: nonrandomized; control treatment
neither placebo nor histamine2-receptor
antagonist; only pH outcomes assessed
30 Trials excluded
7 Duplicate publications
5 No separate data on outcomes for patients with
bleeding from peptic ulcer
10 Control treatment neither placebo nor
histamine2-receptor antagonist
6 Not reporting any of the outcomes predefined
in the review
2 Use of a historical control group
0 Trials excluded
Subgroup analyses are summarized in Table 2. Treat-
ment with PPIs significantly reduced mortality among trials
conducted in Asia (OR, 0.35; 95% CI, 0.16-0.74).19,21,24-29
Mortality was also lower among trials that confined in-
clusion to patients with active bleeding or a nonbleeding
visible vessel or that reported outcomes separately for such
patients (OR, 0.53; 95% CI, 0.31-0.91).9-11,16-19,21,24-26,28
Among such high-risk patients, the reduction of mortality
by PPI treatment remained significant when the analysis
was confined to the 7 trials that consistently used initial
endoscopic hemostatic treatment (OR, 0.54; 95% CI, 0.30-
0.96)9,10,16,24-26,28 but was nonsignificant among trials that
did not consistently do so (OR, 0.51; 95% CI, 0.12-
2.12).11,17-19,21 The reduction in mortality among patients
• www.mayoclinicproceedings.com
 PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING

TABLE 1. Summary Results for Principal Outcomes in Patients With Peptic Ulcer Bleeding*
Unweighted
pooled rates
No. of patients (%)
Outcome at 30 days No. of
after randomization trials PPI Control PPI Control Heterogeneity OR (95% CI) NNT (95% CI)
Mortality6,7,9,11-19,22-29 20 2020 2062 3.91 3.83 No (P=.24) 1.01 (0.74-1.40); Not calculable
0.89 (0.47-1.68)†
Rebleeding6-10,12-16,18-28 21 2098 2126 10.6 17.3 Yes (P=.04) 0.49 (0.37-0.65); 13 (10-25);
0.46 (0.31-0.67)† 12 (7-20)†
Surgical intervention6,9-19,22-28 19 1995 2039 6.1 9.3 No (P=.45) 0.61 (0.48-0.78); 34 (20-50);
0.64 (0.49-0.84)† 34 (20-100)†
Further EHT after 7 468 471 5.6 15.7 No (P=.25) 0.32 (0.20-0.51); 10 (8-17);
randomization15,16,22-25,28 0.30 (0.19-0.49)† 10 (8-15)†
*CI = confidence interval; EHT = endoscopic hemostatic treatment; NNT = number needed to treat; OR = odds ratio; PPI = proton pump inhibitor.
†Results for analyses confined to trials with adequate concealment of allocation.

with high-risk endoscopic signs also remained significant
in a post hoc sensitivity analysis when we further confined
the analysis to 4 trials that had consistently applied initial
endoscopic hemostatic treatment and also used high-dose
PPI treatment (OR, 0.46; 95% CI, 0.24-0.90).9,10,24,25 We
defined high dose a priori as the equivalent of an intrave-
nous (IV) bolus of 80 mg of omeprazole followed by an 8-
mg/h IV infusion for 72 hours. There was no effect on
mortality in the analysis of the remaining 3 trials that used
lower-dose IV or oral PPI treatment (OR, 1.01; 95% CI,
0.26-3.83) (Table 3).16,26,28
Meta-regression analysis showed that geographical lo-
cation of trials was the only characteristic that was signifi-
cantly associated with treatment effect on mortality,
rebleeding, or surgery. Treatment effect was higher among
trials conducted in Asia19,21,24-29 compared with those from

No. in No. in OR OR
Study PPI group control group (95% CI) (95% CI)
Oral PPI
Michel et al,13 1994 2/38 1/37 2.00 (0.17-23.05)
Khuroo et al,19 1997 2/110 6/110 0.32 (0.06-1.63)
Coraggio et al,23 1998 3/24 2/24 1.57 (0.24-10.37)
26
Javid et al, 2001 1/82 2/84 0.51 (0.05-5.69)
28
Kaviani et al, 2003 0/71 1/78 0.36 (0.01-9.01)
Subtotal (95% CI) 325 333 0.67 (0.28-1.64)
Total events: 8 (PPI), 12 (control)
Test for heterogeneity: χ = 2.54; df=4 (P=.64); I =0%
2 2
Test for overall effect: z=0.87 (P=.39)
Intravenous PPI
Brunner & Chang,11 1990 1/19 1/20 1.06 (0.06-18.17)
Daneshmend et al,12 1992 23/246 13/257 1.94 (0.96-3.91)
Pérez Flores et al,14 1994 0/38 0/43 Not estimable
Desprez et al,6 1995 7/38 7/38 1.00 (0.31-3.19)
Lanas et al,15 1995 2/28 2/23 0.81 (0.10-6.23)
Villanueva et al,16 1995 3/45 1/41 2.86 (0.29-28.62)
17
Cardi et al, 1997 0/21 0/24 Not estimable
Hasselgren et al,18 1997 11/159 1/163 12.04 (1.54-94.40)
Schaffalitzky et al,22 1997 10/130 11/135 0.94 (0.38-2.29)
24
Lin et al, 1998 0/50 2/50 0.19 (0.01-4.10)
Fried et al,7 1999 1/66 1/67 1.02 (0.06-16.58)
Lau et al,25 2000 5/120 12/120 0.39 (0.13-1.15)
Sheu et al,27 2002 0/86 2/89 0.20 (0.01-4.28)
Xuan,29 2003 0/31 0/33 Not estimable
Barkun et al,9 2004 8/618 14/626 0.57 (0.24-1.38)
Subtotal (95% CI) 1695 1729 1.08 (0.77-1.52)
Total events: 71 (PPI), 67 (control)
Test for heterogeneity: χ2 = 16.59; df=11 (P=.12); I 2=33.7%
Test for overall effect: z=0.44 (P=.66)

Total (95% CI) 2020 2062 1.01 (0.74-1.40)

Total events: 79 (PPI), 79 (control)
Test for heterogeneity: χ2 = 19.63; df=16 (P=.24); I 2=18.5%
Test for overall effect: z=0.09 (P=.93)
0.01 0.1 1 10 100
Favors PPI Favors control

FIGURE 2. Forest plot of the odds ratios (ORs) and 95% confidence intervals (CIs) of individual trials and pooled data for
mortality; subgroup analysis according to route of proton pump inhibitor (PPI) administration.

Mayo Clin Proc. • March 2007;82(3):286-296 • www.mayoclinicproceedings.com 289

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING

No. in No. in OR OR
Study PPI group control group (95% CI) (95% CI)
Oral PPI
Michel et al,13 1994 8/38 11/37 0.63 (0.22-1.80)
Khuroo et al,19 1997 10/110 37/110 0.20 (0.09-0.42)
Coraggio et al,23 1998 5/24 5/24 1.00 (0.25-4.03)
26
Javid et al, 2001 6/82 18/84 0.29 (0.11-0.77)
28
Kaviani et al, 2003 2/71 9/78 0.22 (0.05-1.07)
Subtotal (95% CI) 325 333 0.32 (0.20-0.50)
Total events: 31 (PPI), 80 (control)
Test for heterogeneity: χ = 5.97; df=4 (P=.20); I =33.0%
2 2
Test for overall effect: z=4.92 (P<.001)
Intravenous PPI
Daneshmend et al,12 1992 58/246 70/257 0.82 (0.55-1.23)
Pérez Flores et al,14 1994 0/38 1/43 0.37 (0.01-9.30)
Desprez et al,6 1995 0/38 3/38 0.13 (0.01-2.64)
15
Lanas et al, 1995 6/28 9/23 0.42 (0.12-1.45)
Villanueva et al,16 1995 11/45 9/41 1.15 (0.42-3.14)
Hasselgren et al,18 1997 5/159 4/163 1.29 (0.34-4.90)
20
Labenz et al, 1997 3/20 2/20 1.59 (0.24-10.70)
Lin et al,21 1997 4/26 5/13 0.29 (0.06-1.36)
Schaffalitzky et al,22 1997 9/130 17/135 0.52 (0.22-1.20)
24
Lin et al, 1998 2/50 12/50 0.13 (0.03-0.63)
Fried et al,7 1999 6/66 10/67 0.57 (0.19-1.67)
Lau et al,25 2000 8/120 27/120 0.25 (0.11-0.57)
Duvnjak et al,8 2001 1/31 4/31 0.23 (0.02-2.14)
Sheu et al,27 2002 5/86 13/89 0.36 (0.12-1.06)
9
Barkun et al, 2004 68/618 89/626 0.75 (0.53-1.05)
Jensen et al,10 2004 5/72 12/77 0.40 (0.13-1.21)
Subtotal (95% CI) 1773 1793 0.62 (0.50-0.75)
Total events: 191 (PPI), 287 (control)
Test for heterogeneity: χ2 = 20.10; df=15 (P=.17); I 2=25.4%
Test for overall effect: z=4.75 (P<.001)
Total (95% CI) 2098 2126 0.49 (0.37-0.65)
Total events: 222 (PPI), 367 (control)
Test for heterogeneity: χ = 32.58; df=20 (P=.04); I =38.6%
2 2
Test for overall effect: z=6.44 (P<.001)
0.01 0.1 1 10 100
Favors PPI Favors control

FIGURE 3. Forest plot of the odds ratios (ORs) and 95% confidence intervals (CIs) of individual trials and pooled data for
rebleeding; subgroup analysis according to route of proton pump inhibitor (PPI) administration.

Europe, North America, and South Africa6-18,20,22,23 (Table 3).
Meta-regression analysis was not performed for the outcome
of repeated endoscopic hemostatic treatment because of the
small number of trials reporting that outcome.
We could not reach a confident conclusion regarding the
effect of PPI treatment according to ulcer site because of
the limited number of trials that reported such informa-
tion. Only 2 trials provided this information,12,17 although
we obtained additional unpublished data for another 2
trials.9,25 Within-trial comparisons yielded contradictory
results.
DISCUSSION
These results extend and update our previous analyses.1,2
Treatment with PPIs consistently reduces the rate of
rebleeding after an episode of ulcer bleeding and also re-
duces the requirement for surgical treatment. New findings
that add to existing knowledge are that PPI treatment re-
duces all-cause mortality among patients with major endo-
scopic stigmas and reduces the requirement for repeated
endoscopic hemostatic treatment. Trials conducted in Asia
demonstrate some important differences in that PPI treat-
ment reduces all-cause mortality among all patients with
ulcer bleeding and produces quantitatively higher reduc-
tions in rebleeding and surgical intervention than are seen
elsewhere. We previously discussed the likely explanations
for those findings.30 Patients in the Asian RCTs had a mean
age of 57 years, whereas those in the non-Asian RCTs had
a mean age of 66 years. Therefore, Asian patients may have
had less comorbidity, although a detailed analysis of this
was not practicable from the information available. Fur-
thermore, PPI treatment in Asian patients may produce a
more profound reduction in acid secretion because of a
lower parietal cell mass, a higher prevalence of Helico-
bacter pylori infection, and a higher proportion of geneti-
cally determined slow metabolizers of PPIs.
In a previous meta-analysis, we showed that PPI treat-
ment also had a marginally significant beneficial effect on
blood transfusion requirements in patients with peptic ulcer
bleeding (weighted mean difference, –0.6 U of blood; 95%
CI, –1.1 to 0.0; P=.05).31 However, this finding should be
regarded with caution because of the statistically signifi-
cant heterogeneity among trials and the differences among

290 Mayo Clin Proc. • March 2007;82(3):286-296 • www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
 PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING

No. in No. in OR OR
Study PPI group control group (95% CI) (95% CI)
Oral PPI
Michel et al,13 1994 5/38 9/37 0.47 (0.14-1.57)
Khuroo et al,19 1997 8/110 26/110 0.25 (0.11-0.59)
Coraggio et al,23 1998 5/24 5/24 1.00 (0.25-4.03)
Javid et al,26 2001 2/82 7/84 0.28 (0.06-1.37)
28
Kaviani et al, 2003 1/71 1/78 1.10 (0.07-17.92)
Subtotal (95% CI) 325 333 0.38 (0.22-0.66)
Total events: 21 (PPI), 48 (control)
Test for heterogeneity: χ = 3.57; df=4 (P=.47); I =0%
2 2
Test for overall effect: z=3.42 (P<.001)
Intravenous PPI
Brunner & Chang,11 1990 1/19 4/20 0.22 (0.02-2.20)
Daneshmend et al,12 1992 45/246 50/257 0.93 (0.59-1.45)
Pérez Flores et al,14 1994 1/38 1/43 1.14 (0.07-18.79)
Desprez et al,6 1995 5/38 7/38 0.67 (0.19-2.34)
15
Lanas et al, 1995 1/28 5/23 0.13 (0.01-1.24)
Villanueva et al,16 1995 9/45 9/41 0.89 (0.31-2.51)
Cardi et al,17 1997 1/21 2/24 0.55 (0.05-6.54)
18
Hasselgren et al, 1997 7/159 17/163 0.40 (0.16-0.98)
Schaffalitzky et al,22 1997 14/130 18/135 0.78 (0.37-1.65)
Lin et al,24 1998 0/50 0/50 Not estimable
Lau et al,25 2000 3/120 9/120 0.32 (0.08-1.20)
Sheu et al,27 2002 1/86 4/89 0.25 (0.03-2.28)
Barkun et al,9 2004 12/618 13/626 0.93 (0.42-2.06)
10
Jensen et al, 2004 0/72 2/77 0.21 (0.01-4.41)
Subtotal (95% CI) 1670 1706 0.69 (0.52-0.91)
Total events: 100 (PPI), 141 (control)
Test for heterogeneity: χ2 = 9.92; df=12 (P=.62); I 2=0%
Test for overall effect: z=2.65 (P=.008)

Total (95% CI) 1995 2039 0.61 (0.48-0.78)

Total events: 121 (PPI), 189 (control)
Test for heterogeneity: χ2 = 17.02; df=17 (P=.45); I 2=0.1%
Test for overall effect: z=3.96 (P<.001)
0.01 0.1 1 10 100
Favors PPI Favors control

FIGURE 4. Forest plot of the odds ratios (ORs) and 95% confidence intervals (CIs) of individual trials and pooled data for
surgical intervention; subgroup analysis according to route of proton pump inhibitor (PPI) administration.

trials for the precise criteria for transfusion. Since none of
the additional trials included in the current updated meta-
analysis reported transfusion requirements, the conclusions
regarding transfusion requirements should remain un-
changed.
For patients who have had significant ulcer bleeding
and have major endoscopic stigmas, high-dose PPI treat-
ment likely improves outcomes through prolonged and
marked elevation of intragastric pH. It has been proposed
that an intragastric pH of 6 or higher is desirable after an
episode of ulcer bleeding so that platelet function and
coagulation mechanisms are promoted and fibrinolysis is
inhibited through inhibition of peptic activity.32 Logi-
cally, these effects should lead to stabilization of a clot

No. in No. in OR OR
Study PPI group control group (95% CI) (95% CI)
Lanas et al,15 1995 0/28 1/23 0.26 (0.01-6.77)
Villanueva et al,16 1995 6/45 5/41 1.11 (0.31-3.95)
Schaffalitzky et al,22 1997 7/130 18/135 0.37 (0.15-0.92)
Coraggio et al,23 1998 2/24 2/24 1.00 (0.13-7.75)
Lin et al,24 1998 1/50 12/50 0.06 (0.01-0.52)
Lau et al,25 2000 8/120 27/120 0.25 (0.11-0.57)
Kaviani et al,28 2003 2/71 9/78 0.22 (0.05-1.07)
Total (95% CI) 468 471 0.32 (0.20-0.51)
Total events: 26 (PPI), 74 (control)
Test for heterogeneity: χ = 7.84; df=6 (P=.25); I =23.4%
2 2
Test for overall effect: z=4.83 (P<.001)
0.01 0.1 1 10 100
Favors PPI Favors control

FIGURE 5. Forest plot of the odds ratios (ORs) and 95% confidence intervals (CIs) of individual trials and pooled data for
further endoscopic hemostatic treatment after randomization. PPI = proton pump inhibitor.

Mayo Clin Proc. • March 2007;82(3):286-296 • www.mayoclinicproceedings.com 291

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING

TABLE 2. Summary Results for Subgroup Analyses in Patients With Peptic Ulcer Bleeding*
Unweighted
pooled rates
(%)
Subgroup analyses and outcomes PPI Control Heterogeneity OR (95% CI) NNT (95% CI)
Analysis according to type of control treatment
H2RA as control treatment (17 trials)6-11,13-17,20,21,23,24,27,29
Mortality 2.5 3.0 No (P=.87) 0.82 (0.48-1.41) Not calculable
Rebleeding 10.7 15.7 No (P=.48) 0.61 (0.48-0.78) 20 (13-34)
Surgical intervention 3.9 5.3 No (P=.45) 0.73 (0.47-1.13) Not calculable
Placebo as control treatment (7 trials)12,18,19,22,25,26,28
Mortality 5.7 4.9 Yes (P=.02) 0.96 (0.43-2.15) Not calculable
Rebleeding 10.7 19.2 Yes (P=.005) 0.41 (0.23-0.72) 10 (6-50)
Surgical intervention 8.7 13.5 Yes (P=.09) 0.52 (0.32-0.84) 20 (13-100)
Analysis according to route of PPI administration
IV PPIs (19 trials)6-12,14-18,20-22,24,25,27,29
Mortality 4.2 3.9 No (P=.12) 1.08 (0.77-1.52) Not calculable
Rebleeding 10.8 16.0 No (P=.17) 0.62 (0.50-0.75) 20 (13-34)
Surgical intervention 6.0 8.3 No (P=.62) 0.69 (0.52-0.91) 50 (25-100)
Oral PPIs (5 trials)13,19,23,26,28
Mortality 2.5 3.6 No (P=.64) 0.67 (0.28-1.64) Not calculable
Rebleeding 9.5 24.0 No (P=.20) 0.32 (0.20-0.50) 7 (5-12)
Surgical intervention 6.5 14.4 No (P=.47) 0.38 (0.22-0.66) 13 (8-25)
Analysis according to dose of PPI
High-dose IV PPI as active treatment
(6 trials)†9,10,18,22,24,25
Mortality 3.2 3.7 Yes (P=.04) 0.82 (0.33-2.06) Not calculable
Rebleeding 8.4 13.7 Yes (P=.04) 0.47 (0.28-0.82) 14 (8-50)
Surgical intervention 3.1 5.0 No (P=.43) 0.61 (0.40-0.93) 50 (34-100)
Non–high-dose PPI as active treatment
(18 trials)6-8,11-17,19-21,23,26-29
Mortality 4.8 4.0 No (P=.74) 1.22 (0.77-1.94) Not calculable
Rebleeding 13.2 21.6 No (P=.11) 0.53 (0.41-0.68) 13 (8-25)
Surgical intervention 10.0 15.0 No (P=.36) 0.61 (0.45-0.82) 20 (13-50)
Analysis by prerandomization EHT
Without consistent prerandomization EHT
(9 trials)6,11,12,14,15,17,19,21,29
Mortality 6.6 5.3 No (P=.34) 1.25 (0.75-2.09) Not calculable
Rebleeding 16.0 25.8 Yes (P=.03) 0.38 (0.18-0.81) 10 (5-50)
Surgical intervention 12.4 18.4 No (P=.12) 0.62 (0.44-0.88) 17 (9-50)
Consistent prerandomization EHT
(15 trials)7-10,13,16,18,20,22-28
Mortality 3.0 3.3 No (P=.26) 0.89 (0.59-1.34) Not calculable
Rebleeding 9.4 14.7 No (P=.13) 0.60 (0.48-0.74) 17 (10-34)
Surgical intervention 3.9 6.2 No (P=.75) 0.60 (0.43-0.85) 50 (25-100)
Analysis according to dose of PPI among trials with consistent prerandomization EHT
Consistent prerandomization EHT and use of
high-dose IV PPI treatment (6 trials)†9,10,18,22,24,25
Mortality 3.2 3.7 Yes (P=.04) 0.82 (0.33-2.06) Not calculable
Rebleeding 8.4 13.7 Yes (P=.04) 0.47 (0.28-0.82) 14 (8-50)
Surgical intervention 3.1 5.0 No (P=.43) 0.61 (0.40-0.93) 50 (34-100)
Consistent prerandomization EHT and use of
non–high-dose PPI treatment (9 trials)†7,8,13,16,20,23,26-28
Mortality 2.4 2.4 No (P=.80) 1.00 (0.42-2.35) Not calculable
Rebleeding 10.2 17.2 No (P=.42) 0.52 (0.35-0.78) 14 (8-33)
Surgical intervention 6.6 9.9 No (P=.71) 0.59 (0.33-1.05) Not calculable
Analysis confined to patients with prerandomization endoscopic findings of active bleeding or NBVV (12 trials)9-11,16-19,21,24-26,28
Mortality 1.9 3.6 No (P=.87) 0.53 (0.31-0.91) 50 (34-100)
Rebleeding 10.6 18.1 Yes (P=.03) 0.41 (0.26-0.64) 10 (6-20)
Surgical intervention 3.3 6.2 No (P=.52) 0.49 (0.32-0.74) 34 (20-100)
(Continued)

292 Mayo Clin Proc. • March 2007;82(3):286-296 • www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
 PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING

TABLE 2. Continued*
Unweighted
pooled rates
(%)
Subgroup analyses and outcomes PPI Control Heterogeneity OR (95% CI) NNT (95% CI)
Analysis by prerandomization EHT in patients with prerandomization endoscopic findings of active bleeding or NBVV
Prerandomization endoscopic findings of active
bleeding or a NBVV; consistent prerandomization
EHT (7 trials)9,10,16,24-26,28
Mortality 1.8 3.3 No (P=.72) 0.54 (0.30-0.96) 50 (34-100)
Rebleeding 9.9 16.4 Yes (P=.03) 0.43 (0.25-0.74) 12 (7-25)
Surgical intervention 2.6 3.8 No (P=.61) 0.65 (0.39-1.08) Not calculable
Prerandomization endoscopic findings of active
bleeding or a NBVV; without consistent
prerandomization EHT (5 trials)11,17-19,21
Mortality 3.5 6.5 No (P=.57) 0.51 (0.12-2.12) Not calculable
Rebleeding 19.4 46.8 No (P>.99) 0.29 (0.13-0.63) 4 (2-10)
Surgical intervention 8.7 24.6 No (P=.81) 0.27 (0.12-0.57) 6 (4-14)
Post hoc analysis: prerandomization endoscopic
findings of active bleeding or a NBVV; consistent
prerandomization EHT; high-dose IV PPI
(4 trials)†9,10,24,25
Mortality 1.6 3.5 No (P=.73) 0.46 (0.24-0.90) 50 (34-100)
Rebleeding 9.7 16.0 Yes (P=.02) 0.37 (0.17-0.81) 10 (6-34)
Surgical intervention 1.7 2.7 No (P=.37) 0.63 (0.33-1.21) Not calculable
Post hoc analysis: prerandomization endoscopic
findings of active bleeding or a NBVV; consistent
prerandomization EHT; non–high-dose PPI
(IV or oral) (3 trials)16,26,28†
Mortality 2.4 2.3 No (P=.49) 1.01 (0.26-3.83) Not calculable
Rebleeding 11.4 18.5 No (P=.15) 0.54 (0.29-1.01) Not calculable
Surgical intervention 7.2 9.2 No (P=.58) 0.68 (0.30-1.55) Not calculable

Analysis according to geographical location of trials

Trials conducted in Asia (8 trials)19,21,24-29
Mortality 1.5 4.4 No (P=.99) 0.35 (0.16-0.74) 34 (20-100)
Rebleeding 6.8 22.2 No (P=.95) 0.24 (0.16-0.36) 7 (5-9)
Surgical intervention 2.9 9.2 No (P=.91) 0.29 (0.16-0.53) 7 (6-13)
Trials conducted elsewhere (16 trials)6-18,20,22,23
Mortality 4.8 3.6 No (P=.37) 1.36 (0.94-1.96) Not calculable
Rebleeding 11.9 15.5 No (P=.85) 0.72 (0.58-0.89) 25 (17-100)
Surgical intervention 7.2 9.4 No (P=.77) 0.73 (0.55-0.95) 34 (14-100)
*CI = confidence interval; EHT = endoscopic hemostatic treatment; H2RA = histamine2-receptor antagonist; IV = intravenous; NBVV =
nonbleeding visible vessel; NNT = number needed to treat; OR = odds ratio; PPI = proton pump inhibitor.
†High-dose PPI treatment is defined a priori as the equivalent of omeprazole, 80-mg IV bolus, followed by 8-mg/h IV infusion for 72 hours.

over an ulcer base, which would be important after a major
bleeding episode. However, oral PPI treatment is likely to
be adequate for patients with more minor episodes of
bleeding and absence of major endoscopic stigmas. In
those patients, PPI treatment does not need to achieve
such a high intragastric pH because clot stabilization over
low-risk ulcers is unnecessary. The main role of PPI
therapy in those patients is to initiate the ulcer healing
process. Thus, PPI treatment after ulcer bleeding may
achieve 1 of 2 goals depending on the severity of the
bleed. High-dose PPI treatment might be considered
prohemostatic in patients with major endoscopic stigmas,
whereas regular to high-dose (usually oral) PPI treatment
likely promotes ulcer healing among patients without high-
risk endoscopic stigmas.
Previous meta-analyses have addressed the effects of
PPI treatment in patients with nonvariceal upper gas-
trointestinal tract bleeding in general33 or peptic ulcer
bleeding in particular.34,35 Although these meta-analyses
and our initial report2 have some methodological differ-
ences and different conclusions regarding the effect of PPIs
on mortality, the conclusion that PPI treatment reduces
rates of ulcer rebleeding and surgical intervention is highly
consistent. A multidisciplinary consensus group critically
appraised evidence on nonvariceal upper gastrointestinal
tract bleeding and recommended high-dose IV PPI treat-

Mayo Clin Proc. • March 2007;82(3):286-296 • www.mayoclinicproceedings.com 293

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING
TABLE 3. Results of Meta-regression Analysis of the Influence of
the Geographical Location of Trials (Asia vs Elsewhere)
on the Treatment Effect*
Outcome Coefficient (95% CI) Constant P value
Mortality –1.91 (–2.18 to –0.21) 0.23 .02
Rebleeding –1.14 (–1.59 to –0.69) 0.29 .001
Surgical
intervention –0.96 (–1.63 to –0.29) 0.28 .01
*Results are expressed in logarithmic form (meta-regression analysis was
not performed for the outcome of further endoscopic hemostatic treat-
ment because of the small number of trials reporting that outcome). CI =
confidence interval.
ment (IV bolus followed by continuous IV infusion) after
successful endoscopic therapy.36 However, neither the con-
sensus report nor any of these meta-analyses considered
studies published later than August 2003; therefore, they
did not evaluate those studies that included North Ameri-
can patients.9,10
As with any systematic review, publication bias is a
possibility. In particular, there is a concern that studies find-
ing no benefit from PPI treatment may have been rejected for
publication or may not have been submitted for publication.
However, our analyses only suggested publication bias for
the secondary outcome of surgical intervention.
Another potential criticism of our approach is the large
number of subgroup analyses performed. However, pa-
tients with ulcer bleeding represent a heterogeneous popu-
lation, and trials of PPI treatment have been designed dif-
ferently (with respect to routes of drug administration or
the control treatment used). The only subgroup analysis to
be undertaken post hoc concerned the separate analysis of
trials that used high-dose IV PPIs (as opposed to low-dose
IV or oral PPIs) and that consistently applied endoscopic
hemostatic treatment before randomization and included
patients with high-risk endoscopic stigmas (Table 2).
By an a priori decision in our main analysis, we pooled
trials that had used either H2RA or placebo as control
treatment. This decision was based on evidence that there is
little if any benefit of H2RA treatment over placebo for
peptic ulcer bleeding.37,38 Nevertheless, we performed sepa-
rate subgroup analyses for each comparator treatment (Table
2). These analyses showed only a possibly different effect
on surgical intervention rates; these rates were reduced
when PPI treatment was compared with placebo but not
when compared with H2RA treatment. However, meta-
regression analysis detected no overall effect of the type of
control treatment used. Overall, if the effect of H2RA treat-
ment is considered a confounder, it would be weak and,
most importantly, would tend merely to reduce the ob-
served treatment effect.
Another potential concern with any meta-analysis is
heterogeneity among included trials. We found statistically
294 Mayo Clin Proc. • March 2007;82(3):286-296
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
significant heterogeneity among the 21 trials that reported
rebleeding rates. Of note, significant heterogeneity was
evident only by the χ2 test (P=.04) and not by the I2 test
(I2=38.6) (Figure 3). Clinical heterogeneity was consider-
able among these trials regarding baseline endoscopic
signs of recent hemorrhage, the use of endoscopic hemo-
static treatment, the dose and route of administration
of PPIs, and the type of control treatment used. However,
we attempted to address these issues in our prespecified
subgroup analyses. For the secondary outcome of re-
bleeding, the only subgroup analyses that resulted in 2
statistically homogeneous groups of trials were those con-
cerning route of PPI administration and geographical lo-
cation. Moreover, the separation of trials according to
geographical location further decreased heterogeneity for
the outcomes of mortality and surgery. Meta-regression
analysis suggested enhanced PPI efficacy in Asian trials.
Thus, this difference in PPI efficacy between Asian and
non-Asian trials is the most likely explanation for the
statistical heterogeneity in the analysis for the outcome of
rebleeding.
Future pharmacodynamic and clinical outcomes stud-
ies should aim to determine the optimal IV (or oral) PPI
doses to be used in European and North American pa-
tients with peptic ulcer bleeding. Subsequently, a large
multicenter trial that addresses mortality as the primary
end point should evaluate that dose in Europe and North
America.
Since our literature search and analysis, 4 additional
trials have been reported. Hsu et al39 compared IV bolus
pantoprazole with IV bolus ranitidine in 102 patients in
Taiwan. Khoshbaten et al40 compared oral omeprazole
and oral cimetidine in 80 patients in Iran. Lin et al41 from
Taiwan compared 2 regimens of IV omeprazole with
IV cimetidine in 200 patients. Zargar et al42 from India
compared high-dose IV infusion pantoprazole with pla-
cebo. All 4 trials found a statistically significant reduction
in rebleeding with PPI treatment but no demonstrable
effect on mortality. These trials were not included in our
meta-analysis. However, since their results are consistent
with our findings and since they are relatively small stud-
ies, it is unlikely that they would substantially affect our
results.
Also subsequent to our literature search, other trials
have been undertaken, although results are unknown. One
trial (ClinicalTrials.gov identifier: NCT00279123) in Hong
Kong with an unknown number of patients compared 2
regimens of IV pantoprazole with placebo; it was recently
completed but has not yet been published. Two other trials are
ongoing. One (ClinicalTrials.gov identifier: NCT00247130)
has an expected enrollment of 400 patients and will compare
IV omeprazole with IV ranitidine in Japan. The other
• www.mayoclinicproceedings.com

(ClinicalTrials.gov identifier: NCT00251979) is a multina-
tional trial in Europe, South America, South Africa, and
Hong Kong that compares IV esomeprazole with placebo
and is expected to enroll 760 patients.
The main clinical implication of our findings is that PPI
treatment should be initiated as soon as practicable in pa-
tients with ulcer bleeding. The route of administration is
best determined by clinical judgment and the presence or
absence of major endoscopic stigmas, such as active bleed-
ing or a nonbleeding visible vessel. Data comparing oral
with IV PPI treatment are scarce. Patients without major
endoscopic stigmas may require only oral PPI treatment;
trials have used higher oral PPI doses than conventionally
recommended for nonurgent indications such as gastro-
esophageal reflux disease. Intravenous administration is
more appropriate for patients with stigmas of recent hemor-
rhage. The exact IV dose of any PPI is, as yet, undefined,
and substantial variation is likely to occur on the basis of
the population being studied.30
CONCLUSIONS
Treatment with PPIs plays an important part in the manage-
ment of patients with ulcer bleeding. It reduces rebleeding
and the requirements for surgical intervention and addi-
tional endoscopic hemostatic treatment. When appropri-
ately combined with endoscopic hemostatic treatment, PPI
treatment also improves mortality among patients with the
highest-risk endoscopic findings.
We thank Iris Gordon, MSc, trial search coordinator, Cochrane
Collaboration, Upper Gastrointestinal and Pancreatic Diseases
Group, University of Leeds, for her help with the literature
search.
REFERENCES
1. Leontiadis GI, McIntyre L, Sharma VK, Howden CW. Proton pump
inhibitor treatment for acute peptic ulcer bleeding. Cochrane Database Syst
Rev. 2004;(3):CD002094.
2. Leontiadis GI, Sharma VK, Howden CW. Systematic review and meta-
analysis of proton pump inhibitor therapy in peptic ulcer bleeding. BMJ. 2005
Mar 12;330:568-570. Epub 2005 Jan 31.
3. Leontiadis GI, Sharma VK, Howden CW. Proton pump inhibitor treat-
ment for acute peptic ulcer bleeding. Cochrane Database Syst Rev. 2006;(1):
CD002094.
4. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews
of Interventions 4.2.5. In: The Cochrane Library. Issue 3. Chichester, UK:
John Wiley & Sons, Ltd; 2005.
5. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis
detected by a simple, graphical test. BMJ. 1997;315:629-634.
6. Desprez D, Blanc P, Bories JM, et al. Acute upper gastrointestinal
hemorrhage: preliminary report of a randomized controlled trial comparing
intravenous administration of ranitidine and omeprazole [abstract]. Gastroen-
terology. 1995;108(suppl):A82.
7. Fried R, Beglinger C, Meier R, et al. Comparison of intravenous
pantoprazole with intravenous ranitidine in peptic ulcer bleeding [abstract].
Gut. 1999;45(suppl 5):A100. Abstract P0104.
Mayo Clin Proc. • March 2007;82(3):286-296
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING
8. Duvnjak M, Supanc V, Troskot B, et al. Comparison of intravenous
pantoprazole with intravenous ranitidine in prevention of rebleeding from
gastroduodenal ulcers [abstract]. Gut. 2001;49(suppl 3):abstract 2379.
9. Barkun A, Racz I, van Rensburg C, et al. Prevention of peptic ulcer
rebleeding using continuous infusion of pantoparzole vs ranitidine: a
multicenter, multinational, randomized, double-blind, parallel-group compari-
son [abstract]. Gastroenterology. 2004;126(suppl 2):A-78. Abstract 609.
10. Jensen DM, Pace SC, Soffer EF, Mack ME, Comer GM. Lower
rebleeding rates in high risk patients treated with IV pantoprazole than IV
ranitidine after endoscopic hemostasis in a randomised controlled US study
[abstract]. Am J Gastroenterol. 2004;99(suppl):S296. Abstract 903.
11. Brunner G, Chang J. Intravenous therapy with high doses of ranitidine
and omeprazole in critically ill patients with bleeding peptic ulcerations of the
upper intestinal tract: an open randomized controlled study. Digestion.
1990;45:217-225.
12. Daneshmend TK, Hawkey CJ, Langman MJS, Logan RF, Long RG,
Walt RP. Omeprazole versus placebo for acute upper gastrointestinal bleeding:
randomised double blind controlled study. BMJ. 1992;304:143-147.
13. Michel P, Duhamel C, Bazin B, et al. Lansoprazole versus ranitidine
dans la prévention des récidives précoces des hémorragies digestives
ulcéreuses gastro-duodénales. Gastroenterol Clin Biol. 1994;18:1102-1105.
14. Pérez Flores R, García Molinero MJ, Herrero Quirós C, et al.
Tratamiento de la hemorragia digestive alta de origen péptico: ranitidina
intravenosa versus omeprazol intravenoso. Rev Esp Enferm Dig. 1994;86:637-
641.
15. Lanas A, Artal A, Blas JM, Arroyo MT, Lopez-Zaborras J, Sainz R.
Effect of parenteral omeprazole and ranitidine on gastric pH and the outcome
of bleeding peptic ulcer. J Clin Gastroenterol. 1995;21:103-106.
16. Villanueva C, Balanzo J, Torras X, et al. Omeprazole versus ranitidine as
adjunct therapy to endoscopic injection in actively bleeding ulcers: a prospec-
tive and randomized study. Endoscopy. 1995;27:308-312.
17. Cardi M, Muttillo IA, Amadori L, et al. Omeprazole versus ranitidine
intraveineux dans le traitement de l’ulcere duodenal hemorragique: une etude
randomisee prospective. Ann Chir. 1997;51:136-139.
18. Hasselgren G, Lind T, Lundell L, et al. Continuous intravenous infusion
of omeprazole in elderly patients with peptic ulcer bleeding: results of a
placebo-controlled multicenter study. Scand J Gastroenterol. 1997;32:328-
333.
19. Khuroo MS, Yattoo GN, Javid G, et al. A comparison of omeprazole and
placebo for bleeding peptic ulcer. N Engl J Med. 1997;336:1054-1058.
20. Labenz J, Peitz U, Leusing C, Tillenburg B, Blum AL, Borsch G.
Efficacy of primed infusions with high dose ranitidine and omeprazole to
maintain high intragastric pH in patients with peptic ulcer bleeding: a prospec-
tive randomised controlled study. Gut. 1997;40:36-41.
21. Lin HJ, Lo WC, Perng CL, Wang K, Lee FY. Can optimal acid suppres-
sion prevent rebleeding in peptic ulcer patients with a non-bleeding visible
vessel: a preliminary report of a randomized comparative study. Hepatogastro-
enterology. 1997;44:1495-1499.
22. Schaffalitzky de Muckadell OB, Havelund T, Harling H, et al. Effect of
omeprazole on the outcome of endoscopically treated bleeding peptic ulcers:
randomized double-blind placebo-controlled multicentre study. Scand J
Gastroenterol. 1997;32:320-327.
23. Coraggio F, Rotondano G, Marmo R, et al. Somatostatin in the prevention
of recurrent bleeding after endoscopic haemostasis of peptic ulcer haemorrhage:
a preliminary report. Eur J Gastroenterol Hepatol. 1998;10:673-676.
24. Lin HJ, Lo WC, Lee FY, Perng CL, Tseng GY. A prospective random-
ized comparative trial showing that omeprazole prevents rebleeding in patients
with bleeding peptic ulcer after successful endoscopic therapy. Arch Intern
Med. 1998;158:54-58.
25. Lau JY, Sung JJ, Lee KK, et al. Effect of intravenous omeprazole on
recurrent bleeding after endoscopic treatment of bleeding peptic ulcers. N Engl
J Med. 2000;343:310-316.
26. Javid G, Masoodi I, Zargar SA, et al. Omeprazole as adjuvant therapy to
endoscopic combination injection sclerotherapy for treating bleeding peptic
ulcer. Am J Med. 2001;111:280-284.
27. Sheu BS, Chi CH, Huang CC, Kao AW, Wang YL, Yang HB. Impact of
intravenous omeprazole on Helicobacter pylori eradication by triple therapy in
patients with peptic ulcer bleeding. Aliment Pharmacol Ther. 2002;16:137-
143.
• www.mayoclinicproceedings.com 295
PROTON PUMP INHIBITOR THERAPY FOR PEPTIC ULCER BLEEDING

28. Kaviani MJ, Hashemi MR, Kazemifar AR, et al. Effect of oral APPENDIX. MEDLINE Search Strategy
omeprazole in reducing re-bleeding in bleeding peptic ulcers: a prospective,
double-blind, randomized, clinical trial. Aliment Pharmacol Ther. 2003;17: 1. randomized controlled trial.pt.
211-216. 2. controlled clinical trial.pt.
29. Xuan JL. Loseco compared with famotidine in the treatment of upper 3. randomized controlled trials.sh.
gastrointestinal bleeding: clinical analysis of 90 cases [in Chinese]. Guangxi 4. random allocation.sh.
Med J. 2003;25:529-531. 5. double blind method.sh.
30. Leontiadis GI, Sharma VK, Howden CW. Systematic review and meta- 6. single-blind method.sh.
analysis: enhanced efficacy of proton-pump inhibitor therapy for peptic ulcer 7. or/1-6
bleeding in Asia—a post hoc analysis from the Cochrane Collaboration. Ali- 8. (animal not human).sh.
ment Pharmacol Ther. 2005;21:1055-1061. 9. 7 not 8
31. Leontiadis GI, Sharma VK, Howden CW. Systematic review and meta- 10. clinical trial.pt.
analysis: proton-pump inhibitor treatment for ulcer bleeding reduces transfu- 11. exp clinical trials/
sion requirements and hospital stay—results from the Cochrane Collaboration. 12. (clin$ adj25 trial$).ti,ab.
Aliment Pharmacol Ther. 2005;22:169-174. 13. ((singl$ or doubl$ or trebl$ or tripl$) adj25 blind$).mp. or mask$.ti,ab.
32. Green FW Jr, Kaplan MM, Curtis LE, Levine PH. Effects of acid and 14. placebos.sh.
pepsin on blood coagulation and platelet aggregation: a possible contributor to 15. placebo$.ti,ab.
prolonged gastroduodenal mucosal hemorrhage. Gastroenterology. 1978;74: 16. random$.ti,ab.
38-43. 17. research design.sh.
33. Khuroo MS, Khuroo MS, Farahat KL, Kagevi IE. Treatment with proton 18. or/10-17
pump inhibitors in acute non-variceal upper gastrointestinal bleeding: a meta- 19. 18 not 8
analysis. J Gastroenterol Hepatol. 2005;20:11-25. 20. 19 not 9
34. Andriulli A, Annese V, Caruso N, et al. Proton-pump inhibitors and 21. comparative study.sh.
outcome of endoscopic hemostasis in bleeding peptic ulcers: a series of meta- 22. exp evaluation studies/
analyses. Am J Gastroenterol. 2005;100:207-219. 23. follow up studies.sh.
35. Bardou M, Toubouti Y, Benhaberou-Brun D, Rahme E, Barkun AN. 24. prospective studies.sh.
Meta-analysis: proton-pump inhibition in high-risk patients with acute peptic 25. (control$ or prospectiv$).mp. or volunteer$.ti,ab.
ulcer bleeding. Aliment Pharmacol Ther. 2005;21:677-686. 26. or/21-25
36. Barkun AN, Bardou M, Marshall JK, Nonvariceal Upper GI Bleeding 27. 26 not 8
Consensus Conference Group. Consensus recommendations for managing pa- 28. 27 not (9 or 20)
tients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med. 2003; 29. 9 or 20 or 28
139:843-857. 30. exp stomach/
37. Collins R, Langman M. Treatment with histamine H2 antagonists in acute 31. stomach.tw.
upper gastrointestinal hemorrhage: implications of randomized trials. N Engl J 32. gastr$.tw.
Med. 1985;313:660-666. 33. exp duodenum/
38. Levine JE, Leontiadis GI, Sharma VK, Howden CW. Meta-analysis: the 34. duoden$.tw.
efficacy of intravenous H2-receptor antagonists in bleeding peptic ulcer. Ali- 35. peptic$.tw.
ment Pharmacol Ther. 2002;16:1137-1142. 36. exp peptic ulcer/
39. Hsu PI, Lo GH, Lo CC, et al. Intravenous pantoprazole versus ranitidine 37. exp peptic ulcer hemorrhage/
for prevention of rebleeding after endoscopic hemostasis of bleeding peptic 38. (peptic adj5 ulcer$).tw.
ulcers. World J Gastroenterol. 2004;10:3666-3669. 39. (bleed$ adj5 ulcer$).tw.
40. Khoshbaten M, Fattahi E, Naderi N, Khaleghian, Rezailashkajani M. A 40. (rebleed$ adj5 ulcer$).tw.
comparison of oral omeprazole and intravenous cimetidine in reducing com- 41. (recurrent adj5 bleed$ adj5 ulcer$).tw.
plications of duodenal peptic ulcer. BMC Gastroenterol. 2006;6:2. 42. (acute adj5 bleed$ adj5 ulcer$).tw.
41. Lin HJ, Lo WC, Cheng YC, Perng CL. Role of intravenous omeprazole 43. (gastrointestinal adj5 bleed$).tw.
in patients with high-risk peptic ulcer bleeding after successful endoscopic 44. (gastrointestinal adj5 rebleed$).tw.
epinephrine injection: a prospective randomized comparative trial. Am J 45. (gastrointestinal adj5 hemorrhag$).tw.
Gastroenterol. 2006;101:500-505. 46. (gastrointestinal adj5 haemorrhag$).tw.
42. Zargar SA, Javid G, Khan BA, et al. Pantoprazole infusion as adjuvant 47. (ulcer adj5 hemorrhag$).tw.
therapy to endoscopic treatment in patients with peptic ulcer bleeding: pro- 48. (ulcer adj5 haemorrhag$).tw.
spective randomized controlled trial. J Gastroenterol Hepatol. 2006;21:716- 49. exp omeprazole/
721. 50. omeprazole.tw.
51. lansoprazole.tw.
52. pantoprazole.tw.
53. rabeprazole.tw.
54. esomeprazole.tw.
55. (proton adj5 pump adj5 inhibitor).tw.
56. or/30-35
57. or/36-48
58. or/49-55
59. 56 and 57 and 58
60. 59 and 29

296 Mayo Clin Proc. • March 2007;82(3):286-296 • www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.