Skip navigation

NSAID-induced GI Injury


Bandolier 25 reported on the protective effects of misoprostol on NSAID-induced gastrointestinal (GI) bleeding shown in a large randomised study. It is useful to see those results largely confirmed in a superb meta-analysis of randomised trials of the ability of misoprostol and histamine antagonists to prevent NSAID-induced GI lesions [1]. The authors, from Rome, used techniques readers of Bandolier are familiar with, including L'Abbé plots, as well as giving NNTs for different levels of baseline risk from odds ratios, as demonstrated in Bandolier 36 .

NSAID risks

Although NSAIDs are commonly prescribed by family physicians, their use triples the risk of developing severe GI adverse events. This meta-analysis assessed the effectiveness of misoprostol or histamine antagonists as preventive agents when coadministered with NSAIDs.

Systematic review

Clinical trials published between January 1970 through December 1994 were identified by searching both MEDLINE and relevant review articles. Inclusion criteria were:

  • endoscopy performed before NSAID treatment
  • no evidence of ulcer initially
  • random allocation with a placebo arm
  • NSAIDs given for at least 5 days

A variety of NSAIDs were used and drug dosages were not specified. Histamine antagonists studied were ranitidine, cimetidine and nizatidine. Daily misoprostol dosages ranged from 200-800 micrograms. Outcomes were examined with respect to short term (< 2 weeks) or long term (>4 weeks) exposure to NSAIDs. The main endpoints measured were the number of subjects in which gastric ulcers, gastric lesions (> 5 erosions or 1 ulcer), duodenal ulcers, and duodenal lesions appeared. Ulcer disease is the key outcome, as the clinical significance and natural history of erosions is unclear.

Data for analysis

The search found 24 articles describing 4325 patients. About 16 percent were volunteers, and the rest were patients with acute musculoskeletal disease, osteoarthritis or rheumatoid arthritis. Analysis was by rate difference between patients on NSAIDs treated with histamine antagonist or placebo.


Gastric ulcers

The weighted average baseline risks for gastric ulcers were 3.6% and 6.8% with <2 weeks and >4 weeks use of NSAIDs. For gastric lesions the weighted average baseline risks were 53% and 27% respectively.

Misoprostol use significantly reduced the rate of gastric ulcers both in short-term (rate difference -13%; 95%CI, -26% to -1%) and long-term (rate difference -8%; 95% CI, -18% to -1%) NSAID treatment. This translates into NNTs of 8 for <2 weeks treatment and 12.5 for >4 weeks treatment. Histamine antagonists were not found to reduce the rate of gastric ulcers with either short term or long term trials.

Duodenal ulcers

The weighted average baseline risks for duodenal ulcers were 3.0% and 4.0% with <2 weeks and >4 weeks use of NSAIDs. For duodenal lesions the weighted average baseline risks were 11% and 12% respectively.

Both histamine antagonists and misoprostol significantly reduced the risk in the long term trials, but neither result was clinically significant (NNT = 42 and 29, respectively), and neither agent had an effect on short term trials.

Results were not modified significantly whether normal subjects or patients were studied, although there were no studies of long term prevention with normal subjects. Studies with misoprostol had patients with a higher risk in the control groups than with histamine antagonist comparisons. The authors addressed this by calculating NNTs with baseline risks between 3 and 40% for gastric ulcer. For instance, with a baseline risk of 3% the NNT for misoprostol to prevent one ulcer over two weeks of NSAID exposure compared with placebo was 35. At a baseline risk of 40% it was 3.


This study was an overview of lesion prevention. Lesions are not the same as clinically significant GI bleeding caused by NSAIDs. The large randomised study which had clinical effects as an outcome [2] indicated that misoprostol use may be usefully considered in patients at high risk for complications on NSAID therapy - specifically, patients older than 75 years of age or with a history of peptic ulcer disease or upper GI tract bleeding. The table of NNTs with differing baseline risk emphasises this advice. The drugs should be coadministered as soon as possible, because the risk of erosive lesions is high from the start of NSAID use.


  1. Koch M, Dezi A, Ferrario F, Capurso L. Prevention of nonsteroidal anti-inflammatory drug-induced gastrointestinal mucosal injury. Archives of Internal Medicine 1996 156: 2321-32.
  2. FE Silverstein, DY Graham, JR Senior et al. Misoprostol reduces serious gastrointestinal complications in patients with rheumatoid arthritis receiving nonsteroidal anti-inflammatory drugs. Annals of Internal Medicine 1995 123: 241-9.

previous or next story in this issue