|
|
| |||||||||||||||||||||||||||||||||
Background Many patients who have had upper gastrointestinal bleeding continue to take low-dose aspirin for cardiovascular prophylaxis or other nonsteroidal antiinflammatory drugs (NSAIDs) for musculoskeletal pain. It is uncertain whether infection with Helicobacter pylori is a risk factor for bleeding in such patients.
Methods We studied patients with a history of upper gastrointestinal bleeding who were infected with H. pylori and who were taking low-dose aspirin or other NSAIDs. We evaluated whether eradication of the infection or omeprazole treatment was more effective in preventing recurrent bleeding. We recruited patients who presented with upper gastrointestinal bleeding that was confirmed by endoscopy. Their ulcers were healed by daily treatment with 20 mg of omeprazole for eight weeks or longer. Then, those who had been taking aspirin were given 80 mg of aspirin daily, and those who had been taking other NSAIDs were given 500 mg of naproxen twice daily for six months. The patients in each group were then randomly assigned separately to receive 20 mg of omeprazole daily for six months or one week of eradication therapy, consisting of 120 mg of bismuth subcitrate, 500 mg of tetracycline, and 400 mg of metronidazole, all given four times daily, followed by placebo for six months.
Results We enrolled 400 patients (250 of whom were taking aspirin and 150 of whom were taking other NSAIDs). Among those taking aspirin, the probability of recurrent bleeding during the six-month period was 1.9 percent for patients who received eradication therapy and 0.9 percent for patients who received omeprazole (absolute difference, 1.0 percent; 95 percent confidence interval for the difference, –1.9 to 3.9 percent). Among users of other NSAIDs, the probability of recurrent bleeding was 18.8 percent for patients receiving eradication therapy and 4.4 percent for those treated with omeprazole (absolute difference, 14.4 percent; 95 percent confidence interval for the difference, 4.4 to 24.4 percent; P=0.005).
Conclusions Among patients with H. pylori infection and a history of upper gastrointestinal bleeding who are taking low-dose aspirin, the eradication of H. pylori is equivalent to treatment with omeprazole in preventing recurrent bleeding. Omeprazole is superior to the eradication of H. pylori in preventing recurrent bleeding in patients who are taking other NSAIDs, such as naproxen.
One approach is concurrent therapy with proton-pump inhibitors. Recent epidemiologic data suggest that this treatment reduces the risk of bleeding in patients taking low-dose aspirin3 or other NSAIDs.3,5 However, the long-term cost of this therapy would be substantial.
Another approach is the eradication of Helicobacter pylori infection.6,7 It is uncertain whether H. pylori infection is a risk factor for bleeding ulcers in people who are taking aspirin or other NSAIDs. There are data to suggest that H. pylori increases,8 has no effect on,9,10 or even decreases11 the risk of bleeding among users of aspirin or other NSAIDs. It is not known whether eradicating the infection would substantially reduce the risk of bleeding ulcers in those taking low-dose aspirin or other NSAIDs and would thus obviate the need for acid-suppressive therapy.
We hypothesized that eradicating H. pylori infection was equivalent to maintenance therapy with omeprazole in terms of preventing recurrent upper gastrointestinal bleeding in people with H. pylori infection who were taking low-dose aspirin or other NSAIDs.
Methods
Patients
This study was a randomized comparison of the eradication of H. pylori and treatment with omeprazole for the prevention of recurrent upper gastrointestinal bleeding in patients with H. pylori infection who were taking low-dose aspirin for coronary heart disease or stroke and patients who were infected with H. pylori who were taking NSAIDs other than aspirin for arthritis. The protocol was approved by the ethics committee at the Chinese University of Hong Kong, and all participants provided written informed consent.
From May 1995 to January 2000, we screened users of low-dose aspirin (
325 mg per day) or other NSAIDs who presented with upper gastrointestinal bleeding at the Prince of Wales Hospital. The patients underwent endoscopy within 24 hours after presentation to identify the site of bleeding and determine their H. pylori status. The inclusion criteria were endoscopically confirmed ulcers or bleeding erosions, histologically documented H. pylori infection, and long-term use or anticipated long-term use (i.e., use for at least six months) of low-dose aspirin or other NSAIDs. The exclusion criteria were concomitant use of nonaspirin NSAIDs and low-dose aspirin, corticosteroids, or anticoagulants; a history of gastric surgery; or the presence of erosive esophagitis, gastroesophageal varices, gastric-outlet obstruction, renal failure (defined by a serum creatinine level of more than 2.26 mg per deciliter [200 µmol per liter]), terminal illness, or cancer.
All patients were given 20 mg of omeprazole daily for eight weeks to promote ulcer healing. Aspirin and other NSAIDs were withheld during this period. Follow-up endoscopy was performed by one investigator to determine whether the ulcer had healed. Unhealed ulcers were treated with additional courses of omeprazole until complete healing occurred, as determined endoscopically.
Randomization
Eligible patients underwent randomization once the healing of ulcers was confirmed. Patients with coronary heart disease or stroke were given 80 mg of aspirin daily for six months and patients with arthritis were given 500 mg of naproxen twice daily for six months. Randomization was carried out separately in these two groups. In each group, the patients were randomly assigned to receive 20 mg of omeprazole daily for six months or a one-week course of eradication therapy (consisting of 120 mg of bismuth subcitrate, 500 mg of tetracycline, and 400 mg of metronidazole, all of which were given four times daily), followed by an identical-appearing omeprazole placebo. Randomization was carried out through the use of computer-generated lists of random numbers in blocks of 10. Consecutively numbered, sealed packages of drugs were dispensed by a research nurse.
Monitoring
One investigator monitored the patients' hemoglobin levels, serum biochemical values, and symptoms of recurrent bleeding every eight weeks for six months. A research nurse evaluated drug compliance and adverse events. A direct telephone line was provided so that the patients could report any serious adverse events. The final H. pylori status was determined by a carbon-13 urea breath test four weeks after the study medications were stopped. Patients who withdrew early were reassessed at six months for recurrent bleeding.
End Point
The end point was recurrent upper gastrointestinal bleeding within six months, which was defined as hematemesis or melena with ulcers or bleeding erosions confirmed by endoscopy or a decrease in the hemoglobin level of at least 2 g per deciliter in the presence of endoscopically proved ulcers or erosions. An ulcer was defined as a circumscribed mucosal break that was at least 0.5 cm in diameter and had a perceptible depth, and a bleeding erosion was defined as a flat mucosal break of any size in the presence of blood in the stomach. Endoscopy was performed if hematemesis or melena was confirmed by the admitting medical officer. Undocumented hematemesis or melena or the presence of heme-positive stool was not an indication for endoscopy. Members of an independent adjudication committee who were not aware of the patients' treatment assignments performed endoscopy and determined the source of bleeding.
Statistical Analysis
We considered that the effect of the eradication of H. pylori would be equivalent to the effect of treatment with omeprazole if the upper limit of the 95 percent confidence interval of the difference in the probability of recurrent bleeding at six months did not exceed 5 percentage points. The power calculations were applied separately for the aspirin and the naproxen groups. On the assumption that omeprazole would reduce the six-month probability of recurrent bleeding to 1.5 percent, we calculated that 125 patients would be required in the eradication-therapy group and 125 patients in the omeprazole group to give the study a power of 80 percent at a 5 percent significance level with the use of a one-sided equivalence test of proportions (PASS software, version 2000, NCSS, Kaysville, Utah).
Because recurrent bleeding is a serious complication, two planned interim analyses were performed in July 1997 and July 1999 to compare the safety of the two treatments. To terminate the trial if one treatment was markedly inferior to the other, we used a predefined early stopping rule that specified a level of significance of 0.0005 for the first analysis and 0.01 for the second analysis.12 The first interim results did not justify early termination.13 The second interim analysis included data for 150 patients in the naproxen group (median follow-up, 6 months; range, 10 days to 6 months) and 180 in the aspirin group (median follow-up, 6 months; range, 2 to 6 months). The results showed that in the aspirin group, the probability of recurrent bleeding was similar for patients receiving either treatment (P=1.00). In the naproxen group, however, there was a marked difference in the probability of recurrent bleeding: it was 18.8 percent among patients treated with eradication therapy and 4.4 percent among patients treated with omeprazole (P= 0.005). Because this difference reached the criterion for early stopping, the assignment of patients to the naproxen group was terminated after the second interim analysis. The final analysis was performed in July 2000, after 250 patients in the aspirin group had completed the study. Data analyses were carried out exclusively by a data-review committee at these specified dates.
The Kaplan–Meier method was used to estimate the likelihood of reaching the end point of recurrent upper gastrointestinal bleeding within six months according to the intention-to-treat principle. The log-rank test was used for comparisons between treatment groups. The intention-to-treat analysis included all patients who had taken at least one dose of a study medication. Failure to take at least 70 percent of the study drugs or use of nonstudy drugs, including antiulcer agents, corticosteroids, anticoagulants, NSAIDs other than aspirin or naproxen, or antiplatelet agents, was considered a protocol violation. In the naproxen group, a Cox proportional-hazards model was used to adjust for possibly confounding covariates, including age, the presence or absence of other illnesses, location of the ulcers (stomach, duodenum, or both), diameter of the ulcers (greater than 2 cm, or 2 cm or less), and the presence or absence of previous ulcer disease (symptomatic ulcer or bleeding ulcers).14 The patients' base-line characteristics were compared with use of Student's t-test for parametric data, the Mann–Whitney U test for nonparametric data, and Pearson's chi-square test for proportions (SPSS software, version 10.0, SPSS, Chicago). All P values are two-tailed.
Results
Study Groups
We screened 896 users of low-dose aspirin (from May 1995 through January 2000) or other NSAIDs (from May 1995 through July 1999) who presented with upper gastrointestinal bleeding. A total of 58 percent of aspirin users and 49 percent of users of other NSAIDs were positive for H. pylori infection. We enrolled 400 patients with H. pylori infection: 250 in the aspirin group and 150 in the naproxen group (Figure 1). The study groups were similar with respect to demographic characteristics, the prevalence of coexisting conditions, the location and size of the ulcers, the severity of bleeding at presentation, and the prevalence of previous ulcer disease (Table 1 and Table 2).
|
|
|
We enrolled 250 patients in the aspirin group (median follow-up, 6 months; range, 2 to 6 months). A predefined level of drug compliance was achieved in 98 percent of the patients who received eradication therapy or omeprazole. Eradication therapy was effective in 93 percent of the patients who received it. Seven percent of the patients in the omeprazole group had a negative urea breath test.
Five patients met the criteria for repeated endoscopy: three had hematemesis or melena, and two had only a drop in the hemoglobin level of at least 2 g per deciliter. All five patients underwent endoscopy. Three of them were confirmed to have recurrent bleeding: two had been assigned to eradication therapy, and one to omeprazole (Table 3). Of the two patients who had recurrent bleeding after the eradication of H. pylori, one took a concomitant nonaspirin NSAID for musculoskeletal pain. The two patients with a drop in hemoglobin of at least 2 g per deciliter had normal findings on endoscopy (one received eradication therapy, and one received omeprazole). The estimated probability of recurrent bleeding during the six-month study was 1.9 percent for patients who received eradication therapy and 0.9 percent for patients who received omeprazole (absolute difference, 1.0 percent; 95 percent confidence interval for the difference, –1.9 to 3.9 percent) (Table 4).
|
|
Naproxen Group
We enrolled 150 patients in the naproxen group (median follow-up, 6 months; range, 10 days to 6 months). Predefined drug compliance was achieved in 88 percent of the patients who received eradication therapy and 87 percent of those who received omeprazole. Eradication therapy was effective in 91 percent of the patients who received it. Ten percent of the patients in the omeprazole group had a negative urea breath test.
Eighteen patients met the criteria for repeated endoscopy: 16 had hematemesis or melena, and 2 had only a drop in hemoglobin of at least 2 g per deciliter. All 18 patients underwent endoscopy, and 16 were confirmed to have recurrent bleeding (13 had received eradication therapy and 3 had received omeprazole) (Table 3). The two patients with a drop in hemoglobin of at least 2 g per deciliter had normal findings on endoscopy (one had received eradication therapy and one had received omeprazole). The estimated probability of recurrent bleeding during the six-month study was 18.8 percent for patients who received eradication therapy and 4.4 percent for patients who received omeprazole (absolute difference, 14.4 percent; 95 percent confidence interval for the difference, 4.4 to 24.4 percent; P=0.005) (Table 4 and Figure 2). In the Cox proportional-hazards model, only a history of ulcer disease was significantly associated with an increased risk of recurrent bleeding (hazard ratio, 5.4; 95 percent confidence interval, 1.9 to 15.5). After adjustment for the covariates of age, the presence or absence of other illnesses, the size and location of the ulcers, and the presence or absence of a history of ulcer disease, the adjusted hazard ratio was 7.1 (95 percent confidence interval, 1.9 to 27.6).
|
Discussion
We set out to test the hypothesis that eradication of H. pylori is equivalent to maintenance treatment with omeprazole for the secondary prevention of upper gastrointestinal bleeding in patients who are taking low-dose aspirin or other NSAIDs. The patients enrolled in this study were at risk for recurrent bleeding because they had a recent history of upper gastrointestinal bleeding.3,4 Our results show that in patients infected with H. pylori who are taking low-dose aspirin, the eradication of H. pylori alone is as effective as maintenance treatment with omeprazole in preventing recurrent upper gastrointestinal bleeding. The equivalence was demonstrated by the absence of a clinically important difference between the results of the two treatments. In contrast, we found that omeprazole is superior to the eradication of H. pylori in preventing recurrent bleeding in patients with H. pylori infection who are taking naproxen.
Among persons who are infected with H. pylori and who take aspirin, bleeding ulcers could be attributed to H. pylori, aspirin, or both. It is impossible to distinguish ulcers related to H. pylori from ulcers related to aspirin.15 One might argue that the eradication of H. pylori may not prevent ulcers related to aspirin use. However, our findings suggest that the eradication of H. pylori prevents recurrent bleeding in patients who are taking low-dose aspirin irrespective of the location or cause of the ulcers. Thus, H. pylori and low-dose aspirin may have a synergistic effect that increases the risk of bleeding from ulcers, so that curing the infection would substantially lower the risk of recurrent bleeding.
In contrast to the case with recurrent bleeding associated with low-dose aspirin use, our results indicate that the eradication of H. pylori is inferior to omeprazole in preventing recurrent bleeding associated with nonaspirin NSAIDs. Whether the eradication of H. pylori can reduce the risk of ulcers appears to vary according to the group of patients using nonaspirin NSAIDs. For people who have never been treated with nonaspirin NSAIDs, we showed in an earlier study that eradicating H. pylori before NSAID treatment was initiated reduces the risk of ulcers.16 For long-term users of nonaspirin NSAIDs, however, the eradication of H. pylori has not been shown to prevent gastroduodenal injury.17 Our current findings indicate that the eradication of H. pylori alone is not sufficient to prevent recurrent bleeding in susceptible long-term users of nonaspirin NSAIDs.
The divergent outcomes in patients taking aspirin and those taking other NSAIDs suggest that H. pylori may have a more important role in ulcer bleeding associated with low-dose aspirin than in bleeding associated with other NSAIDs. One explanation is that low-dose aspirin may not be as ulcerogenic as other NSAIDs. Other factors, such as H. pylori infection, may need to be present for aspirin to induce substantial gastroduodenal bleeding. Infection with H. pylori has been shown to impair gastric adaptation to aspirin. The eradication of H. pylori restores this ability and increases the mucosal resistance to aspirin.18 Alternatively, low-dose aspirin may provoke bleeding from preexisting H. pylori–induced ulcers through its antiplatelet effect. Eradication of H. pylori restores the mucosal barrier and thus the ability to withstand the damaging effects of aspirin. In contrast, other NSAIDs can induce peptic ulcers in the absence of H. pylori.19 It is possible that in H. pylori–infected long-term users of nonaspirin NSAIDs, many ulcers are induced by NSAIDs rather than by H. pylori.
Previous endoscopic studies found that among users of nonaspirin NSAIDs who were receiving omeprazole, those infected with H. pylori had less severe mucosal injury than those who were not infected with H. pylori.20,21 H. pylori augments the acid-suppressing effect of proton-pump inhibitors.22,23 It is unclear whether omeprazole can effectively prevent bleeding ulcers in users of nonaspirin NSAIDs who are not infected with H. pylori. Even so, augmentation of the acid-suppressing effect of proton-pump inhibitors by H. pylori cannot be used to justify the failure to treat H. pylori infection in users of nonaspirin NSAIDs who are receiving acid-suppressive therapy, since H. pylori is a risk factor both for peptic ulcer and for cancer of the stomach.
Our study had several limitations. First, the absolute reduction in the risk of recurrent bleeding that was attributable to the eradication of H. pylori cannot be determined, because a placebo group was not included. However, our objective was to investigate whether the eradication of H. pylori could substitute for omeprazole therapy in persons at high risk for upper gastrointestinal bleeding. It would be unethical to withhold treatment to prevent recurrence of ulcer bleeding from such patients. Second, therapy with bismuth may have a slight protective effect on mucosa, although it is very unlikely that one week of bismuth could prevent complications of ulcers, such as bleeding. Third, we studied only naproxen and thus did not address the differences in the ulcerogenicity of other NSAIDs. Since epidemiologic studies do not suggest that naproxen is associated with a higher risk of ulcer bleeding than are other commonly used nonaspirin NSAIDs,24 we believe it is reasonable to extend our findings to other nonaspirin NSAIDs.
In summary, we found that in patients who were infected with H. pylori and who were receiving low-dose aspirin, the eradication of H. pylori is as effective as maintenance therapy with omeprazole in preventing recurrent upper gastrointestinal bleeding. In the light of the increasing use of aspirin for cardiovascular prophylaxis, our findings suggest that patients who are at risk for bleeding from ulcers should be tested for H. pylori infection and treated for it if the infection is found. In contrast, we found that therapy with omeprazole is superior to the eradication of H. pylori for the secondary prevention of upper gastrointestinal bleeding in H. pylori–infected users of naproxen, and presumably other nonaspirin NSAIDs.
Supported by research grants from the Research Grant Council (CUHK 277/96M) and the Health Services Research Committee of Hong Kong (HSRC 711011 and HSRC 831013).
The interim results were presented at the plenary session of the annual meeting of the American Gastroenterology Association, San Diego, Calif., May 21–24, 2000.
We are indebted to Drs. Michael K.K. Li, Gabriel Yip, K.S. Wong, Raymond Wong, Michael Fu, K.M. Chow, and Thomas S.T. Li for their referrals; to Dr. K.F. To of the Department of Anatomical and Cellular Pathology at the Chinese University of Hong Kong for histologic assessment; to Mr. Albert Cheung of the Center of Clinical Trials and Epidemiological Research at the Chinese University of Hong Kong and Professor S.K. Lo at the Polytechnic University for statistical advice; to Dr. C.C. Szeto of the Department of Medicine and Therapeutics at the Chinese University of Hong Kong for his valuable comments; and to Ms. Jessica Ching, Ms. M.Y. Yung, Ms. Agnes Yu, Ms. Roxanna Liu, our surgical colleagues, and the nursing staff of the endoscopy center at the Prince of Wales Hospital for their generous support.
Source Information
From the Departments of Medicine and Therapeutics (F.K.L.C., Y.T.L., W.K.L., J.C.Y.W., Y.H., H.L.Y.C., J.J.Y.S.) and Surgery (S.C.S.C., B.Y.S., J.Y.W.L.), Prince of Wales Hospital, Chinese University of Hong Kong; and the Medical Unit, Alice Ho Miu Ling Nethersole Hospital (V.K.S.L., M.S.L.) — all in Hong Kong.
Address reprint requests to Dr. Francis K.L. Chan at the Department of Medicine and Therapeutics, Prince of Wales Hospital, 30–32 Ngan Shing St., Shatin, Hong Kong, China, or at fklchan{at}cuhk.edu.hk.
References
| |||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
Patients in this group received low-dose aspirin.