Omeprazole and sucralfate in the treatment of NSAID-induced gastric and duodenal ulcer


Porro Gastrointestinal Unit, L. Sacco University Hospital, Via GB Grassi, 74, 20157 Milan, Italy.



To establish the healing efficacy of two drugs, omeprazole and sucralfate, when given to patients who had developed gastric or duodenal ulcer while undergoing chronic treatment with non-steroidal anti-inflammatory drugs (NSAIDs).


Ninety-eight patients with arthritis or arthrosis and NSAID-related gastric or duodenal ulcer were admitted to the endoscopic, single-blind study. They were randomized to receive either omeprazole 20 mg o.m. or sucralfate 2 g b.d. for 4–8 weeks. The patients continued to receive the same NSAID during the trial. Upper gastrointestinal endoscopy was performed at entry and after 4 or 8 weeks.


Eighty-eight patients completed the 4-week study, but only 81 were available for final analysis at 8 weeks. Omeprazole was significantly superior to sucralfate in inducing gastric ulcer healing after both 4 (87 vs. 52%, = 0.007) and 8 weeks (100 vs. 82%, = 0.04). No statistically significant difference in duodenal ulcer healing rates emerged between the two groups either at 4 (79 vs. 55%) or 8 weeks (95 vs. 73%). The healing rates in patients with combined gastric and duodenal ulcer were 67 vs. 33% after 4 weeks and 67 vs. 67% after 8 weeks of treatment. The percentages of asymptomatic patients were similar in the two treatment groups both at 4 (70 vs. 73%) and 8 weeks (70 vs. 75%). H. pylori infection did not influence healing rates, but significantly more H. pylori-positive patients healed with omeprazole.


The results of this study show that omeprazole is superior to sucralfate in healing NSAID-induced gastroduodenal ulcer in patients who continue to take anti-inflammatory drugs. The good results observed were unrelated to H. pylori status.


Gastroduodenal mucosal damage is known to be the main undesirable side-effect of treatment with non-steroidal anti-inflammatory drugs. Mucosal erosions or ulcers have been reported in more than 30% of patients taking NSAIDs while dyspepsia occurs in up to 60%.1, 2 The problem may sometimes be so serious as to call for suspension of NSAID treatment,3 with negative effects on the state of the underlying disease. Relatively few and sometimes conflicting data are available at present4[5][6][7][8][9]–10 regarding the role of gastric mucosal cytoprotective agents or antisecretory drugs in the treatment of NSAID-related gastroduodenal ulcers.

The aim of our study was to evaluate the healing and analgesic effect of omeprazole compared to a cytoprotective agent, sucralfate, in arthritic patients with NSAID-induced gastroduodenal ulcers where it was not possible to suspend the offending drug.



During recent years, in collaboration with the Rheumatology Unit of our hospital, we have implemented a programme of endoscopic evaluation of NSAID-induced gastroduodenal lesions in patients suffering from rheumatoid arthritis, or osteoarthritis, who were candidates for long-term treatment with NSAIDs. All patients, before starting NSAID treatment, were submitted to upper GI endoscopy. Further endoscopic controls were repeated after 3 and 6 months in asymptomatic cases or sooner in the event of the onset of painful dyspepsia and/or complications. All endoscopies were performed by the same physician (G.B.P).

Ninety-eight consecutive patients (75 females, 23 males; mean age 55.5 ± 22 years) were recruited for the trial. All patients were found to be completely normal at basal endoscopy, but later presented gastric or duodenal ulcer or both during treatment with NSAIDs.

Patients were excluded if they had gastric surgery, gastrointestinal malignancy, recent acute upper gastrointestinal bleeding and severe renal impairment. Patients who had taken anti-ulcer drugs in the last 6 months or were taking anticoagulant or prednisone> 10 mg/day were also excluded.

An ulcer was defined as an excavated mucosal break of 5 mm or more in diameter. Ulcer dimensions were measured using standard Olympus biopsy forceps, with the fully open spoon equivalent to 5 mm.

Study design

All patients gave their informed consent and the study was approved by the local Ethical Committee. While continuing the offending drug at the same dosage, the patients were randomized, according to a single-blind protocol, to omeprazole 20 mg o.m. (n = 50) or sucralfate 2 g b.d. (n = 48).

Endoscopy was performed after 4 or 8 weeks or in the event of complications or antacid-resistant painful dyspepsia. The endoscopist was unaware of the drugs being taken. The study end-point was reached if complete re-epithelialization of the ulcer crater occurred.

The following dyspeptic symptoms were evaluated at each visit: epigastric pain, heartburn, nausea and vomiting, and these were graded as none, mild, moderate or severe.

Patients were invited not to change their usual dietary or other habits (smoking, alcohol) during the treatment period.

Helicobacter pylori status

A serum sample was obtained from each patient at entry for evaluation of IgG antibody to H. pylori, by means of a fluorescence enzyme-immunoassay test.

Statistical analysis

The differences between the two treatment groups were compared statistically by means of the χ2 test, Fisher’s exact test, and Student’s t-test, where indicated. 95% confidence intervals for the difference in the healing rate percentage were also calculated.

The size of the population entered in the study was calculated assuming that the healing rate after omeprazole is superior by at least 30% to that oberved with sucralfate and considering a significant difference (α = 0.01) with a power (β) of 75%.


Table 1 shows the demographic and clinical characteristics of the patients at entry. Patients in both groups were comparable as to age, sex, smoking and drinking habits, number of gastric or duodenal ulcers or both, type of arthritic condition, and prevalence of ulcer symptoms. Table 2 lists the various NSAIDs used specifically for rheumatological treatment and the corresponding numbers of gastric or duodenal ulcers that were induced.

Table 1.  . Demographic and clinical characteristics of the patients entered into the trial (no significant differences between groups) Thumbnail image of
Table 2.  . Type of NSAIDs received by patients participating in the study and number of peptic ulcers developed following administration of each drug Thumbnail image of

A past history of peptic ulcer was observed in 30% of the patients while H. pylori infection was present in 56/98 (57%) of the cases.

In particular, H. pylori infection was present in 35/59 (59%) of subjects with gastric ulcer, in 21/33 (64%) of those who developed a duodenal ulcer, and in none of the six patients with both gastric and duodenal ulcer.

Eighty-eight out of 98 patients completed the study at 4 weeks. Five patients in the omeprazole group and three in the sucralfate group were withdrawn for non-medical reasons; two patients in the sucralfate group stopped taking NSAIDs due to epigastric pain. After 8 weeks, two patients in the sucralfate group interrupted the anti-ulcer treatment due to the onset of vomiting, while five failed to return for non-medical reasons.

The 8-week study was thus completed by 81 patients: 45 in the omeprazole group and 36 in the sucralfate group ( Figure 1).

Figure 1.

. Cumulative number of dropouts and medical reasons for withdrawal from the study.

The clinical-endoscopic outcome according to treatment is reported in Table 3. Considering the total population of gastric ulcers (GU), duodenal ulcers (DU) and combined gastric and duodenal ulcers (GU + DU), after 4 weeks 37/45 patients (82%) in the omeprazole group were healed compared to 22/43 (51%) of those in the sucralfate group (= 0.004). When gastric ulcers were considered separately, healing was observed in 20/23 (87%) of those treated with omeprazole compared to 15/29 (52%) of those treated with sucralfate (= 0.007). No significant differences were observed between the two groups as regards healing rates in DU patients (79 vs. 55%). The healing rates in patients with both DU and GU were, respectively, 67% and 33%.

Table 3.  . Endoscopic results and analgesic effects of omeprazole and sucralfate after 4 and 8 weeks of treatment (95% CI for the observed differences in percentage healing between treatment) Thumbnail image of

After 8 weeks of treatment, the healing rates in the population as a whole rose to 96% (43/45 patients) in the omeprazole group vs. 78% (28/36 patients) in the sucralfate group (= 0.01)

Omeprazole again proved significantly superior to sucralfate in healing GU (100 vs. 81%, = 0.04), whereas there was no statistically significant difference between the two groups as regards DU (95 vs. 73%). The percentage of healing in both DU and GU patients was 67%.

Table 4 shows the healing rates of combined gastric and duodenal ulcers according to H. pylori status. The two groups of treatment showed no difference in success percentages between infected and uninfected patients, however, omeprazole proved significantly more effective than sucralfate both at 4 (81 vs. 48% of patients healed) and 8 weeks (96 vs. 71%) only in H. pylori-positive patients.

Table 4.  . Per cent of healing according to type of treatment and H. pylori status Thumbnail image of

Regarding the analgesic effect, omeprazole proved comparable to sucralfate in inducing symptom relief in healed patients, both at 4 weeks (70 vs. 73%, respectively) and after 8 weeks of treatment (70 vs. 75%, respectively).


The appearance of gastric or duodenal ulcer is a fairly frequent event during chronic NSAID therapy,1, 11 and so is the risk of serious complications.12[13]–14 In clinical practice, when faced with NSAID-induced ulcers, the standard policy has been to interrupt anti-inflammatory therapy. While this practice may speed up the healing of mucosal lesions, such a policy is not always possible.

Reports in the literature have already shown that it is possible to achieve healing of NSAID-induced ulcers whilst continuing anti-inflammatory treatment.4[5][6][7][8][9]–10 In the present controlled study, we compared the respective capacities of omeprazole and sucralfate to heal NSAID-induced ulcers in rheumatic patients continuing NSAIDs at the same dosage which initially caused the ulcer.

In the clinical setting the positive role of omeprazole was first suggested by Walan et al.6 in a small subset of patients with gastric ulcer who received continuous treatment with NSAIDs. This observation was confirmed by Hawkey and co-workers9, 10 who showed that, in NSAID users, the proton pump inhibitor 20–40 mg daily is better than misoprostol 800 μg daily for healing and prevention of duodenal ulcers and is superior to ranitidine 300 mg daily for healing and prevention of both gastric and duodenal ulcers.

The role of cytoprotective drugs such as sucralfate15 is less well defined. Manniche et al.5 compared sucralfate 1 g q.d.s. with ranitidine 150 mg b.d. in rheumatic patients with gastric or duodenal ulcers. Similar 9-week healing rates were reported with sucralfate (83%) and ranitidine (84%). There was no statistically significant difference in ulcer healing rates in patients who stopped NSAID treatment compared with those who continued. This observation was confirmed in another small study which showed an improvement in the gastric lesion score.16 To the contrary, a controlled study in 60 patients receiving a fixed daily dose of NSAIDs failed to show a difference between sucralfate and placebo17 in the healing of gastric mucosal lesions.

Our results show that omeprazole is significantly more effective than sucralfate in inducing the healing of gastric ulcers. Where DU patients are concerned, the trend favouring the proton pump inhibitor is indicative, but not statistically significant, probably because of the small number of patients in our study.

The cumulative healing rate after omeprazole 20 mg daily is satisfactory and in line with the data in the literature at 4 and 8 weeks in patients with NSAID-unrelated gastric or duodenal ulcer.18 Unlike that observed with standard doses of H2-antagonists,7 a more potent acid inhibition obtained with proton pump inhibitors or with high doses of famotidine19 appears to overcome the negative interaction of NSAIDs with the proliferative responses at the ulcer margin and speed up the healing of both gastric and duodenal ulcers.

Our data disagree with previous observations20 concerning the not well defined role of omeprazole in the prevention of NSAID-induced gastric lesions. However, in this experience, gastric erosions were also included, for which recent experimental observations suggest a cause other than acid.21

Our study was not aimed toward the evaluation of the relationship between H. pylori and NSAID use as the cause of peptic ulcer or the influence of the microorganism in the natural history of NSAID-related peptic ulcer. We have therefore restricted our research to serological determination of IgG specific antibody under basal conditions. The systemic humoral antibody response to H. pylori is well documented and correlates well with infection.22, 23 The ELISA test has been used widely in epidemiological studies and more recently for monitoring the long-term outcome of H. pylori therapeutic regimens.24 We cannot exclude cases of false positivity since H. pylori serology may remain elevated for long periods (at least 6 months) after eradication of infection. The selection criteria used by us should, however, have significantly reduced this possibility.

Some considerations emerge from our results. Fifty-seven per cent of the patients studied were infected with H. pylori, which is not significantly different from the age-adjusted rate in a control population in a Western country and is consistent with previous studies in rheumatoid arthritis25 and in arthritics of different aetiologies.26 The prevalence of infection is superior to that observed in an acute study by Lanza et al.,27 but is in line with Kim and Graham (50% of positivity for anti H. pylori IgG)28 and Shalcross et al. (61%) in rheumatic patients with NSAID-related chronic gastric or duodenal ulcer.29 However, the fact that in our study more than 40% of these ulcers occurred in uninfected patients, supports the hypothesis that most NSAID-associated gastric or duodenal ulcers develop via a mechanism which does not require the presence of H. pylori or gastritis.

We have also observed that omeprazole is significantly superior to sucralfate in healing gastric and duodenal ulcers only in patients infected by H. pylori. This observation is in accord with Hawkey et al.9 who documented that the proton pump inhibitor appears to be more effective than misoprostol for the healing and prevention of NSAID-associated ulcers in H. pylori-positive patients than H. pylori-negative patients. One may speculate that this may be because the interaction between the microorganism and NSAIDs reduces the gastric acid output and enhances the effectiveness of antisecretory drugs. In effect some anti-inflammatory agents may potentiate the inhibitory effect of H. pylori protein on gastric fundic cyclic-AMP, which in turn mediates acid secretion in vitro.30

The estimation of symptom improvement was also not the major objective of this study, as patients were eligible for entry whether or not symptoms of dyspepsia were present. In addition, symptoms were not monitored daily but were assessed retrospectively at each clinic visit. A complete analgesic effect was achieved with both drugs in a percentage between 70% (with omeprazole) and 75% (with sucralfate) of healed patients. This data, in our opinion, confirms the observations in the literature31 concerning the absence of correlation between clinical manifestation and NSAID-induced mucosal lesions.

In conclusion, the study shows that omeprazole is superior to sucralfate in producing healing of NSAID-induced gastroduodenal ulcers in patients continuing anti-inflammatory therapy. Furthermore, the good results we have observed are unrelated to H. pylori status and allow us to confirm that it is not always necessary to interrupt NSAID treatment or to eradicate H. pylori infection in rheumatic patients who develop peptic ulcer.