Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Endoscopic therapy reduces mortality and morbidity in acute nonvariceal upper gastrointestinal hemorrhage

ACP J Club. 1992 July-Aug;117:14. doi:10.7326/ACPJC-1992-117-1-014

Source Citation

Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic therapy for acute nonvariceal upper gastrointestinal hemorrhage: a meta-analysis. Gastroenterology. 1992 Jan;102:139-48.



To evaluate endoscopic therapy for acute nonvariceal upper gastrointestinal hemorrhage by means of a review of the literature.

Data sources

All MEDLINE and EMBASE files were searched to 1991 using the terms hemorrhage (gastrointestinal), endoscopy, and clinical trials. SCISEARCH, reference lists, and other information services were used to identify further published and unpublished studies.

Study selection

Inclusion criteria were randomized controlled trials that compared ≥ 1 endoscopic therapy with no treatment for patients with acute nonvariceal upper gastro-intestinal bleeding. The end points were rebleeding, need for surgery, and mortality. 2 independent assessors unanimously selected 30 trials from 298 articles. All studies were published (20 in peer-reviewed journals).

Data extraction

2 reviewers independently abstracted data describing the lesions, patient characteristics, interventions, outcome meas-urements, and methods of the studies. 80% of the study authors provided further information and data verification. 3 endoscopic treatments were included: laser therapy, thermal-contact devices (monopolar and bipolar electrocoagulation and heater probe), and injection therapy.

Main results

Pooled results showed that endoscopic therapy reduced the risk for rebleeding (common odds ratio [OR] 0.38, 95% CI 0.32 to 0.45); mortality (OR 0.55, CI 0.40 to 0.76); and rates for surgery (OR 0.36, CI 0.28 to 0.45). Each form of endoscopic therapy reduced the risk for each end point; for thermal contact and injection therapy, however, the reduction in mortality was not significant. Significant heterogeneity was noted among the studies in 5 of the 12 comparisons. The benefits of therapy were confirmed in 9 studies with the highest methodologic quality. Data for patients bleeding from peptic ulcers were obtained from 22 trials. In this subgroup, endoscopic therapy reduced risks for rebleeding (OR 0.57, CI 0.39 to 0.85); for death (OR 0.40, CI 0.31 to 0.48); and for surgery (OR 0.37, CI 0.27 to 0.46). Among high-risk patients (those with actively bleeding or nonbleeding visible vessels) evaluated in 16 to 21 trials, risk for rebleeding (OR 0.23, CI 0.15 to 0.27); death (OR 0.62, CI 0.38 to 0.98); and surgery (OR 0.26, CI 0.17 to 0.32) was reduced by endoscopic therapy. Rebleeding was not reduced by endoscopic therapy in patients with adherent clots or flat pigmented spots.


Endoscopic therapy reduces morbidity and mortality in patients with acute nonvariceal upper gastrointestinal hemorrhage.

Source of funding: No external funding.

Address for article reprint: Dr. L.A. Laine, Division of Gastroenterology and Liver Diseases, Department of Medicine, University of Southern California School of Medicine, 2025 Zonal Avenue, Los Angeles, CA 90033, USA.


The study by Cook and colleagues is a good meta-analysis of a difficult subject. The study gives a reason for a gastrointestinal endoscopist to do upper gastrointestinal endoscopy at midnight and to be able to intervene therapeutically if a suitable lesion is found. Previous studies showed that early endoscopy did not affect mortality, recurrence of bleeding, or duration of hospital stay; and it was concluded that there was no benefit (1, 2). This finding, however, may have been true only for diagnostic upper endoscopy.

This study clearly shows that several kinds of endoscopic therapy (of which laser therapy is the most expensive) are effective in reducing rates of further bleeding, need for surgery, and mortality in those patients who are actively bleeding or who have visible vessels. The results would be even more impressive if only patients with actively spurting vessels were included in the analysis. Because rebleeding was not reduced in patients with pigmented spots or adherent clots, it is good advice to "not disturb the clot" at the time of endoscopy.

Nonvariceal causes of bleeding may include Mallory-Weiss tear, the Dieulafoy ulcer, peptic ulcer, and other conditions. An analysis of the data from 20 trials including only patients with peptic ulcers found that the results were similar to those of the overall analysis; thus, the conclusions are probably applicable to all causes of nonvariceal upper gastrointestinal bleeding.

The complications from endoscopic therapy were rare. We do not know, however, how many patients had iatrogenic lesions in the esophagus, stomach, or duodenum. The extent of mucosal injury by monopolar electrocoagulation differs in the esophagus, stomach, and duodenum (3).

Nirmal S. Mann, MD, DSc
Texas A & M UniversityTemple, Texas, USA


1. Peterson WL, Barnett CC, Smith HJ, et al. N Engl J Med. 1981;304:925-9

2. Erickson RA, Glick ME. Dig Dis Sci. 1986;31:760-8.

3. Mann SK, Mann NS. Am J Gastroenterol. 1979;71:568-71.