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


Meta-analysis: Misoprostol reduces NSAID-induced gastrointestinal mucosal injury

ACP J Club. 1997 Mar-Apr;126:36. doi:10.7326/ACPJC-1997-126-2-036

Source Citation

Koch M, Dezi A, Ferrario F, Capurso L. Prevention of nonsteroidal anti-inflammatory drug-induced gastrointestinal mucosal injury.A meta-analysis of randomized controlled clinical trials. Arch Intern Med. 1996 Nov 11;156:2321-32.



To compare the effectiveness of histamine type 2 (H2) blockers with misoprostol in preventing gastrointestinal mucosal injury induced by nonsteroidal anti-inflammatory drugs (NSAIDs).

Data sources

Randomized controlled trials were identified by searching MEDLINE (1970 to Dec 1994) using the keywords anti-inflammatory agents, nonsteroidal; stomach diseases, duodenal diseases, or peptic ulcer; prevention; randomized controlled trials; histamine H2 receptor antagonists; and misoprostol. Additional trials were identified by scanning bibliographies of review articles.

Study selection

Studies were selected if they compared H2 blockers or misoprostol with placebo in patients with osteoarthritis or rheumatoid arthritis or in healthy participants and if endoscopy had been done before and during NSAID treatment; patients had no evidence of ulcer at the first endoscopic evaluation; and NSAID therapy had been received for ≥ 5 days.

Data extraction

Data were extracted on baseline risk for gastrointestinal mucosal injury associated with short-term (< 2 weeks) and long-term (> 4 weeks) NSAID treatment, the number of patients who developed either gastric or duodenal ulcers or gastric or duodenal lesions (defined as 1 ulcer or > 5 erosions), and methodologic quality.

Main results

24 trials involving 4325 patients met the selection criteria. Use of misoprostol reduced the risk for gastric ulcer compared with placebo for short-term NSAID treatment (P < 0.05). {The 13% absolute risk reduction (ARR) means that 8 patients would need to be treated (NNT) with short-term misoprostol (rather than placebo) to prevent 1 additional patient from developing a gastric ulcer, 95% CI 4 to 112.}* The results were similar for long-term prevention (P < 0.001, {ARR 8%; NNT 12, CI 6 to 100}*). Short-term and long-term H2 blockers did not reduce the risk for gastric ulcers. Both misoprostol and H2 blockers reduced the risk for duodenal ulcers compared with placebo in patients who received long-term NSAID treatment (P < 0.001 and P = 0.04, respectively, {ARR 3%; NNT 30, CI 18 to 1000 for misoprostol; ARR 2%; NNT 42, CI 22 to 500 for H2 blockers}*). However, misoprostol and H2 blockers did not reduce the risk for duodenal ulcer in patients who received short-term NSAID treatment (ARR 2.0%, CI -1.6% to 5.7% for misoprostol; ARR 1%, CI -5% to 3% for H2 blockers). Misoprostol reduced the risk for gastric and duodenal lesions for both long-term and short-term NSAID therapy (P < 0.001 for all). No differences were shown between H2 blockers and placebo for the prevention of gastric and duodenal lesions.


Misoprostol, but not H2 blockers, reduces the risk for gastric ulcers. Both misoprostol and H2 blockers prevent duodenal ulcers in long-term NSAID treatment.

Source of funding: Not stated.

For article reprint: Dr. M. Koch, Department of Digestive Diseases and Nutrition, General Hospital San Filippo Neri, Via Martinotti 20, 00135 Rome, Italy. FAX 39-6-338-0217.

*Numbers calculated from data in article


The widespread use of NSAIDs has led to increased recognition of their adverse effects on the upper gastrointestinal tract. NSAIDs cause dyspepsia in many patients, and a small number develop clinically significant ulcers and complications. The prevention of adverse events from NSAIDs includes the use of low-risk agents and co-prescription with prostaglandins or acid inhibitors (1).

The meta-analysis by Koch and colleagues examines the ability of H2 blockers and misoprostol to reduce NSAID-induced mucosal injury that was measured by endoscopy. This type of injury, which has uncertain clinical importance, was used as an outcome measure instead of more relevant outcomes, such as reduction in serious ulcer complications. As shown by this meta-analysis, misoprostol reduces both gastric and duodenal ulcers. More important, although misoprostol has been shown to reduce serious adverse events, the NNT is 264 to prevent a single complication (2). This leads us to consider whether misoprostol is cost-effective and whether we need to target high-risk patients for therapy. Koch and colleagues show that it is much easier to prevent clinically silent endoscopic ulcers.

This meta-analysis confirms the ability of H2 blockers, when given in traditional ulcer-healing doses, to prevent NSAID-induced duodenal ulcers only. As a result, the stomach, which is the predominant site of NSAID ulcers, is left vulnerable to injury. Recent data, however, have shown considerable promise for potent acid suppression using high-dose H2 blockers or proton pump inhibitors (3). Antisecretory drugs are effective in reducing NSAID-induced symptoms—a relative advantage over misoprostol. Therefore, potent acid suppression, such as the use of proton pump inhibitors, may provide an alternative approach to current first-line therapy with oral prostaglandin therapy. The results of direct comparative studies are eagerly awaited.

James M. Scheiman, MD
University of MichiganAnn Arbor, Michigan, USA


1. Scheiman JM. Gastroenterol Clin North Am. 1996;25:279-98.

2. Silverstein FE, Graham DY, Senior JR, et al. Ann Intern Med. 1995;123:241-9.

3. Ekstrom P, Carling L, Wetterhus S, et al. Scan J Gastroenterol. 1996;31: 753-8.