EBQ:Omeprazole in Bleeding Peptic Ulcers

Under Review Journal Club Article

Lau JYW, et al. "Effect of Intravenous omeprazole on Recurrent Bleeding after Endoscopic Treatment of Bleeding Peptic Ulcer". NEJM. 2000. 343(5):310-316.
PubMed Full text PDF

Clinical Question

Does high dose intravenous omeprazole reduce the incidence of recurrent bleeding in patients who have undergone endoscopic intervention for bleeding peptic ulcers?

Conclusion

High-dose infusion of omeprazole after endoscopic treatment of bleeding peptic ulcers substantially reduces the risk of recurrent bleeding.

Major Points

• Bleeding recurrence in 30 days was 6.7% (8 patients) in the omeprazole group, and 22.% (27 patients) in the placebo group, with bleeding most commonly occurring in the first 3 days after treatment.
  • Hazard ratio: 3.9 (95% CI 1.7-9.0)
  • Study stopped early due to significant difference between omeprazole and placebo (P<0.001)
• Landmark in vitro studies have shown that high intragastric pH may facilitate platelet aggregation, thus preventing bleeding. While this study did not directly measure intragastric pH, it supports the idea that neutralized pH will increase platelet aggregation.

Study Design

• Randomized, double-blind, placebo-controlled study
• Patients underwent endoscopy within 24 hours after admission
  • Patients with actively bleeeding ulcers or ulcers with nonbleeding visible vessels were treated with local epi injection followed by thermocoagulation
• After hemostasis achieved, patients were randomly assigned in double-blind fashion to receive omeprazole or placebo (details in interventions).
• After infusion, all patients were given 20 mg omeprazole PO daily for 8 weeks
• Primary end point was recurrent bleeding within 30 days after endoscopy
  • Bleeding was considered to have recurred if any of the following occurred: vomiting of fresh blood, shock (SBP <90 mm Hg or HR >110) with melena after stabilization, or a drop of Hb >2 g/dL within 24 hours after a transfusion to 10 g/dL
  • Patients judged to have recurrent bleeding underwent urgent endoscopy
• All patients reevaluated at eight weeks in outpatient clinic
• All investigators remained unaware of the patients' treatment assignment until the study was completed

Population

Patient Demographics

Omeprazole vs. Placebo
Male: 66.7% vs. 66.7%
Age: 64 vs. 67
Hemoglobin (g/dL): 9.4 vs. 9.5
Location of ulcer:

Stomach: 44% vs. 40%

Duodenum: 54% vs. 54%

Stoma: 2% vs. 6%

Endoscopic signs of bleeding:

Spurting hemorrhage: 12% vs. 8%

Oozing hemorrhage: 42% vs. 41%

Nonbleeding visible vessel: 32% vs. 30%

Clot with underlying vessel: 15% vs. 22%

Size of ulcer (cm): 1.2 vs. 1.1
Previous ulcer disease: 32% vs. 38%
Previous ulcer bleeding: 30% vs. 30%
Recent use of H2 antagonist or PPI: 2% vs. 2%
Risk factor of bleeding peptic ulcer:

H.pylori infection: 65% vs. 53%

Use of NSAIDs: 33% vs. 33%

Use of aspirin: 19% vs. 15%

Use of warfarin: 4% vs. 4%

Development of bleeding during hospitalization: 18% vs. 19%
Pts with coexisting illnesses: 25% vs. 30%

Inclusion Criteria

• Patients admitted to Prince of Wales Hospital from May 1998 to July 1999 with bleeding peptic ulcers
• >16 years old
• Patients with successful endoscopic treatment of actively bleeding ulcers or ulcers with nonbleeding visible vessels

Exclusion Criteria

• Unsuccessful endoscopic treatment, requiring immediate surgery
• Ulcers with clean bases or flat pigments on endoscopy, therefore not requiring endoscopic treatment
• Terminal cancer (n=10)
• Moribund from concomitant illnesses (n=9)

Interventions

• After endoscopic treatment, patients were randomly assigned to receive an intravenous infusion of placebo or omeprazole
  • Infusion given as an 80-mg bolus injection followed by a continuous infusion of 8 mg per hour for a period of 72 hours.
  • Identical-appearing vials of omeprazole and placebo were prepared by pharmacy with random numbers in blocks of 80
  • Eligible patients were given the lowest-numbered treatment package
  • Treatment was started in the recovery area of the endoscopy suite and continued in a surgical ward

Outcomes

n=240, 120 patients in each group

• End point data analyzed according to intention-to-treat principle

Primary Outcome

• Bleeding recurred within 30 days after treatment in 8 patients (6.7%) in omeprazole group, as compared with 27 (22.5%) in the placebo group.
• Hazard ratio: 3.9 (95% CI 1.7-9.0)
• Recurrent bleeding most common in first 3 days: 4.2% in omeprazole vs. 20% in placebo

Secondary Outcomes

• Mean number of units transfused after endoscopy was significant smaller in omeprazole group (1.7 vs. 2.4 units, P=0.03)
  • Mean number of units transfused before treatment was similar in the groups (1.0 vs. 1.1, P=0.46)
• Significantly more patients in the omeprazole group had a hospital stay <5 days (46.7% vs. 31.7%, P=0.02)
• No significant difference in patients who underwent surgery (2.5% in omeprazole group vs. 7.5% in placebo group, P=0.14)
• No significant difference in patients who died within 30 days after endoscopy (4.2% in omeprazole group vs. 10% in placebo group, P=0.13)

Criticisms & Further Discussion

Outcomes

• Composite primary outcome used: included vomiting, shock with melena, or drop of Hb>2 g/dL after transfusion to 10 g/dL
  • Hb of 10 g/dL no longer regarded as standard for transfusion threshold. A more conservative threshold of 7 g/dL has been shown to have overall benefit in the literature.
  • Studied end points may not directly translate to the emergency department, as the composite outcomes do not require the patient to be clinically sick.

Study Design

Study intended to recruit 320 patients and to conduct four interim analysis. However, the trial was terminated after the third planned interim analysis with 240 patients because a significant different (P<0.001) between the groups.

• While this study showed a significant difference in an Asian population with this composite outcomes, studies performed in Western, industrialized countries have demonstrated decreased efficacy, and even increased mortality
• Study completed in 1997 to study efficacy of PPI in western patients with peptic ulcers was stopped early due to statistically significant mortality in patients who received PPI

Researchers chose to exclude patients with flat pigments or clean bases seen on endoscopy, due to their low risk of recurrent bleeding

• Past studies have shown that interobsrver agreement in peptic ulcer classification is poor to fair

The researchers propose that decreased intragastric pH leads to decreased re-bleeding, however this was not directly measured with pH monitoring

Further Discussion

• Meta-analysis of PPI initiation prior to endoscopy in patients with upper GI bleeding showed no clinically significant difference between PPI vs. placebo.
  • No significant difference in mortality, re-bleeding, or need for surgery
  • Use of omeprazole likely less important before endoscopy, and may be harmful in Western, industialized countries

Funding

Supported in full by a grant from the Research Grants council of the Hong Kong Special Administration Region

External Links

• SMART EM - PPI
• SMART EM - PPI #2

Sources

Authors

Catie Reynolds, Daniel Ostermayer, Juliana Tolles, Greg Armani