Evidence Table 11

Randomized controlled trials of proton pump inhibitors for prevention of nonsteroidal anti-inflammatory drug-induced ulcer

Population settingDiagnosisEligibility criteriaInterventionsControlOther MedicationsDefinition of Treatment Failure/SuccessOutcomes Reported (Results)Adverse EffectsQuality Rating
Lai et al. 2002123 patients, double blind, ITT.
Hong Kong, mean age 70 (range 18–80), female 28%, race NR. 245
screened, 171 eligible by H pylori, 127 treated, 4 H pylori uneradicated.
History of cerebrovascular accident (52%) or heart disease (48%) - endo revealed gastric (74%), duodenal (21%) or gastroduodenal (5%) ulcer. History of stroke or ischemic heart disease requiring long-term aspirin therapy;
Ulcer developed after at least one month low- dose aspirin therapy;
H pylori infection;
Ulcer and H pylori successfully eradicated during initial healing phase of study;
No esophagitis, history of ulcer surgery, comcomitant treatment with NSAIDs, corticosteroids or anticoagulant agents, active cancer, or allergic to study drugs.
30 mg (l) + 100 mg aspirin bid for median 12 monthsMatching placebo + 100 mg aspirin bidAntacid permitted, advised to avoid other NSAIDs if possiblePrimary endpoint: recurrence of ulcer complications (bleeding, outlet obstruction, perforation).
Secondary endpoint: recurrence of ulcer.
Clinical Bleeding:
(l) = 0, (pl) = 8 (p<.01)

Ulcer recurrence:
(l) = 1, (pl) = 9 (p=.008)

H pylori recurrence:
(l) = 0, (pl) = 4 (p<.05)
Death: (l) = 1, (pl) = 0

Other adverse effects NR
Graham, 2002US and Canada
Mean age 60
65% female
90% white, 6% black, 4% other.
No H pylori; reason for long- term NSAID use not reported, previous GI disease: 59% reflux esophagitis, 50% duodenal ulcer, 99% gastric ulcer.Age 18 or older, h/o endoscopically-documented gastric ulcer with or without coexisting duodenal ulcer or GI bleeding, and treatment with stable, full therapeutic doses of an NSAID (except nabumetone or aspirin >1300 mg/day) for at least the previous month.lansoprazole 15 or 30 mg for 12 weeksmisoprostol 200 mcg qid for 12 weeks40% ibuprofen, 35% naproxen, 32% diclofenac, 22% aspirin or aspirin combinations, 17% piroxicam, 34% other NSAIDSOccurrence of gastric ulcer (definition of gastric ulcer not specified), included analysis with withdrawals considered treatment failures (having a gastric ulcer).Treatment success:
Free of gastric ulcer by week 12 (per protocol):
(pl):51% (m): 93% (l15): 80% (l30): 82%
Treatment success:
Results when withdrawals classified as treatment failures:
(pl):34% (m): 67% (l15): 69% (l30): 68%
Withdrawals due to adverse events: (pl) 6.7%, (m) 10.4%, (l15) 2.9%, (l30) 7.5%; Higher percentage of treatment related adverse events in misoprostol group (31% (m), 10% (pl), 7% (l15), 16% in (l30); most common diarrhea. One upper GI tract hemorrhage (l15).Fair: randomization and allocation method not reported.
Bianchi Porro
Single center
Mean age 59.9 (range 22–80)
83% female
ethnicity not given
63% rheumatoid arthritis
38% osteoarthritis.
Over age 18, with rheumatoid arthritis or osteoarthritis, treated with effective and constant doses of NSAIDs (diclofenac, ketoprofen, indomethacin) for at least 8 weeks prior to start of study. Lanza endoscopic grade 0,1, or 2.pantoprazole 40 mgplacebo37% diclofenac, 34% ketoprofen, 35% indomethacin.Occurrence of gastric or duodenal ulcers (grade 4, Lanza classification) after 4 and 12 weeks, or patients who discontinued the study due to lack of efficacy leading to discontinuation of the study medication, an adverse event which was assessed by the study investigator as possibly or definitely related to the study medication.Ulcer status assigned (treatment failure):
(p): 13 with endoscopically-proven peptic ulcer, 3 due to lack of efficacy, 2 adverse events
(pl): 9 with endoscopically-proven peptic ulcer (1 with both gastric and duodenal ulcer), 1 lack of efficacy, 2 adverse events.
Endoscopically proven duodenal and/or gastric ulcers:
(p): 13
(pl): 9
4.3% (p) (m) unrelated to treatment, vomiting possibly related, diarrhea definitely related), 5.9% (pl) (diarrhea possibly related, asthenia definitely related), all withdrew for adverse events.Fair/Good:
concealment of allocation not reported
Labenz et al.
2264 patients screened, 832 randomized, 660 analyzed - in 3 countries in central Europe, double blind, not ITT.
Mean age: 55
Male: 38%
Systemic inflammatory disease (24%), noninflammatory disease (73%), mild dyspepsia (42%), Lanza score “0” on study entry (stomach 68%; duodenum 89%).Age >18 years with inflammatory disease of musculoskeletal system requiring NSAID treatment >5 weeks, and H pylori positive.

Excluded for ulcer or history of ulcer, clotting disorders, prior regular use of NSAIDS (except aspirin <100 mg/day), antibiotics, PPIs, misoprostol, or bismuth salts within 4 weeks; regular use of H2R antagonists, prokinetics or sucralfate; systemic corticosteroids, known or suspected intolerance to study drug, severe concomitant diseases; previous gastric surgery; pregnancy or nursing; and dyspepsia therapy.
OAC-O = omeprazole 40 mg + amoxicillin 2 g +clarithro-mycin
1000 mg for 1 week, then 20 mg ome for 4 weeks.
O-O = 20 mg ome for 5 weeks.
OAC-P = OAC for 1 week, then placebo for 4 weeks.
P-P = placebo for 5 weeks.
NSAID treatment:
diclofenac 100–150 mg, and could add tramadol 200 mg.
Dyspeptic therapy with an antacid.
Primary endpoint: endoscopically proved peptic ulcer.

Secondary endpoints: dyspeptic complaints, signs of gastrointestinal bleeding.
OAC-O vs. O-O vs. OAC-P vs. P-P

Developed peptic ulcers -
Total: 2/173 (1.2%) vs. 0/155 vs. 2/161 (1.2%) vs. 10/171 (5.8%)
- Duodenal: 0/173 vs. 0/155 vs. 2/161(1.2%) vs. 7/171(4.1%)
- Gastric: 2/173 (1.2%)vs. 0/155 vs. 0/161 vs. 3/171 (1.8%)
(Bonferroni p-value significant for all ome groups vs. pla)

Dyspepsia developed requiring therapy:
10.4% vs. 12.3% vs. 10.6% vs. 19.9% (All treatment groups significantly different from pla only group - p-value NR)

Negative H pylori status:
85.3% vs. 21.9% vs. 81.3% vs. 11.8%
201 of 660 patients reported
302 adverse events (no details reported):
OAC-O 31%
O-O 16%
OAC-P 26%
P-P 26%

Diarrhea more frequent in antibiotic groups:
OAC-O 8.8%
O-O 3.0%
OAC-P 8.4%
P-P 3.3%
Hawkey, 199893 centers in 14 countries
mean age 58 (range 20– 85)
64% female
ethnicity not given
38% rheumatoid arthritis, 47% osteoarthritis, 13% other, 2% combinations.39% gastric ulcer with or without erosions, 20% duodenal ulcer with or without erosions, 4% gastric and duodenal ulcer with or without erosions, 36% erosions only.Patients who successfully healed during treatment phase of study. Age 18 to 85, with any condition requiring continuous treatment with oral or rectal NSAIDS above a predetermined minimal dose (no maximal dose). Minimal (and mean) daily oral doses: 50 mg (129 mg) diclofenac, 100 mg (137 mg) ketoprofen, 500 mg (844 mg) naproxen. By endoscopy, any or all of the following: ulcer, defined as a mucosal break at least 3 mm in diameter with definite depth in the stomach, duodenum, or both, more than 10 gastric erosions, and more than 10 duodenal erosions.omeprazole 20 mgmisoprostol 200 mcg bid or placeboAt baseline (all patients):most common diclofenac (23%), naproxen (22%), ketoprofen (16%).Development of any of the following: an ulcer, more than 10 gastric erosions, more than 10 duodenal erosions, at least moderate symptoms of dyspepsia, or adverse events resulting in the discontinuation of treatment.In remission at 6 months:
( o20):61%(m): 48%(pl): 27%p = 0.001 for (o20) vs (m)
Gastric ulcers at relapse:( o20):13%(m):10%(pl):32%
Duodenal ulcers at relapse:( o20): 3%(m):10%(pl):12%
Withdrawals due to adverse events: (o20): 3.9%, (m): 7.7%, (pl): 1.9%; most common diarrhea (7.6% (o20), 8.4% (m), 4.5% (pl), abdominal pain (5.1% (o20), 4.7% (m), 5.8% (pl). One perforated duodenal ulcer after 31 days of (pl).Fair:
randomization and allocation method not reported, not intention-to- treat.
73 centers in 15 countries; mean age 56 (range 20–80); 69% female; ethnicity not given44% rheumatoid arthritis, 32% osteoarthritis, 6% psoriatic arthritis, 5% anklyosing spondylitisAge 18 to 85, with any condition requiring continuous therapy with NSAIDs above specified therapeutic doses (no maximal dose), and not more than 10 mg prednisolone or equivalent per day. By endoscopy, any or all of the following: ulcers 3 mm of more in diameter, more than 10 erosions in stomach, more than 10 erosions in the duodenum. (Lanza scale)omeprazole 20 mgranitidine 150 mg bidNot reported for maintenance phase. Most common at baseline (including healing phase) diclofenac (29%), indomethacin (23%), naproxen (16%)Remission defined as absence of a relapse of lesions, dyspeptic symptoms, and adverse events leading to the discontinuation of treatment.In remission at 6 months:
(o20): 72%(r): 59%p = 0.004
Any adverse event: (o20): 64%, (r): 58%; withdrawals due to adverse events: 6.1% (o20), 3.2% (ran). Most common arthritis, rheumatoid arthritis, vomiting (2.9% (o20), 2.3% (ran)), abdominal pain (2.9% (o)o, 1.9% (ran)), diarrhea (3.3% (o20), 1.4% (ran)). One bleeding duodenal ulcer after 10 days of (o20).Fair:
randomization and allocation method not reported, not intention-to- treat.
Stupnicki et al.
515 patients, multiple
European countries
Multicenter, double-blind
73% female
median age 64 (range 31–93)
ethnicity not reported
55% erosions at entrance exam; 45% 1–5 erosions; 32% H pylori positive; 41% osteoarthritis, 30% rheumatoid arthritis, 2% spondylitis, 7% spondylosis, 19% multiple disease.Outpatients aged 55 or older receiving or planned to receive continuous NSAID therapy for rheumatoid arthritis, osteoarthritis, arthrosis, spondylosis, or spondylitis, and who experienced gastrointestinal symptoms of at most moderate intensity. No signs of reflux esophagitis (endoscopically-proven). At least one of the following criteria: history of endoscopically proven peptic ulcer (including bleeding and/or perforation) within the last 5 years, or history of repeated gastrointestinal symptoms within the last year, or intake of more than one NSAID (the second NSAID could be dosed below the minimal dose), or regular intake of corticosteroids as concomitant medication, or regular intake of anticoagulants as concomitant medication, or NSAID treatment since maximally 4 weeks, or change of the NSAID drug substance since maximally 4 weeks.pantoprazole 20 mg for 6 monthsmisoprostol 400 mcg for 6 months17% more than one NSAID, 17% corticosteroids, 2% anticoagulantsTherapeutic failure: more than 10 erosions/petechiae in the stomach/duodenum, peptic ulcer, reflux esophagitis, discontinuation of study due to an adverse event assessed as “likely” or “definitely” related to the study medication.; discontinuation of study due to severe gastrointestinal symptoms
Endoscopic failure: more than 10 erosions/petechiae in the stomach/duodenum, peptic ulcer, reflux esophagitis
Symptomatic failure: severe gastrointestinal symptoms
In remission at 3 months:
76% pantoprazole vs 63% misoprostol
In remission at 6 months:
67% pantoprazole vs 52% misoprostol

Remission rates for therapeutic failure (pantoprazole vs misoprostol)
3 months: 93% vs 79% (p<0.001)
6 months: 89% vs 70% (p<0.001)
Remission rates for endoscopic failure (pantoprazole vs misoprostol)
3 months: 98% vs 95% (NS)
6 months: 95% vs 86% (p=0.005)
Remission rates for symptomatic failure (pantoprazole vs misoprostol)
3 months: 99% vs 92% (p=0.005)
6 months: 99% vs 92% (p=0.002)
Withdrawals due to adverse events:
5% pantoprazole vs 13% misoprostol (events assessed by investigator as likely or definitely related to study drug)
3 deaths in pantoprazole group; all assessed as not related to study drug.
serious adverse events: 18 pantoprazole vs 16 misoprostol patients serious adverse events classified as at least ‘likely’ related to study drug: 0 pantoprazole vs 2 misoprostol (hypertensive crisis and diarrhea)
Allocation concealment method not reported, baseline characteristics given for ITT population only.

From: Evidence Tables

Cover of Drug Class Review: Proton Pump Inhibitors
Drug Class Review: Proton Pump Inhibitors: Final Report Update 5 [Internet].
McDonagh MS, Carson S, Thakurta S.
Portland (OR): Oregon Health & Science University; 2009 May.
Copyright © 2009, Oregon Health & Science University, Portland, Oregon.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.