|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Persistent infection with Helicobacter pylori is associated with the recurrence of duodenal ulcer. Whether the efficacy of bismuth therapy in reducing the rate of recurrence of duodenal ulcer is due to its antimicrobial effects on H. pylori or to a direct protective action on the mucosa is still a matter of debate.
Methods To study the effect of the eradication of H. pylori on the recurrence of duodenal ulcer, we treated 104 patients with H. pylori infection and recurrent duodenal ulcer with either amoxicillin (750 mg three times daily) plus metronidazole (500 mg three times daily) or identical-appearing placebos, given orally for 12 days. All patients also received ranitidine (300 mg each night) for 6 or 10 weeks. Endoscopy was performed before treatment and periodically during follow-up for up to 12 months after healing.
Results Among the 52 patients given antibiotics, H. pylori was eradicated in 46, as compared with 1 of the 52 given placebo (89 percent vs. 2 percent, P<0.001). After six weeks, the ulcers were healed in 48 patients given antibiotics and 39 given placebo (92 percent vs. 75 percent, P = 0.011). Side effects, mainly diarrhea, occurred in 15 percent of the patients given antibiotics. Among the patients followed up for 12 months, duodenal ulcers recurred in 4 of 50 patients given antibiotics and 42 of 49 given placebo (8 percent vs. 86 percent, P<0.001). Ulcers recurred in 1 of 46 patients in whom H. pylori had been eradicated, as compared with 45 of 53 in whom H. pylori persisted (2 percent vs. 85 percent, P<0.001).
Conclusions In patients with recurrent duodenal ulcer, eradication of H. pylori by a regimen that does not have any direct action on the mucosa is followed by a marked reduction in the rate of recurrence, suggesting a causal role for H. pylori in recurrent duodenal ulcer.
To investigate the effect of a regimen without bismuth on the eradication of H. pylori and the rate of recurrence of duodenal ulcer, we conducted a prospective, one-year, double-blind, randomized trial comparing the combination of amoxicillin and metronidazole with placebo. Our primary objective was to test the hypothesis that the eradication of H. pylori can prevent the recurrence of duodenal ulcer. A secondary objective was to evaluate the efficacy and safety of the combination of antibiotics when given with ranitidine.
Methods
The study protocol was approved by the appropriate institutional review committees, and written informed consent was obtained from all patients.
Patients
We studied 104 patients who had had at least two symptomatic recurrences of duodenal ulcer, diagnosed by endoscopy or radiography, that were separated by a symptom-free period of at least three months. To be eligible for the study, the patients had to have an ulcer at least 5 mm in diameter and H. pylori infection as documented by the presence of the organism in biopsy specimens of the gastric antral mucosa.
The patients were selected from a cross-section of social groups; 15 percent were immigrants from southeastern Europe. They were referred by their general practitioners, with the exception of 27 who had taken part at least six months earlier in trials involving short-term and maintenance treatment of duodenal ulcer with pirenzepine, various histamine-receptor antagonists, sucralfate, and omeprazole. Of these 27 patients, 12 were assigned to the antibiotic group and 15 to the placebo group. Patients who had a bleeding ulcer, had previously had gastric surgery, or had severe disease of any other kind or who had taken corticosteroids, antiinflammatory drugs, or antibiotics in the month before entry were excluded, as were women who were pregnant or lactating or who were not using contraceptives.
Study Design
After endoscopic diagnosis of an active duodenal ulcer and confirmation of the presence of H. pylori in antral specimens, the patients were randomly assigned to a 12-day course of either amoxicillin (750 mg given orally three times daily) combined with metronidazole (500 mg given orally three times daily) or identical-appearing placebos. In addition, all patients simultaneously received ranitidine (300 mg orally at bedtime) for six weeks and for another four weeks if ulcer healing was not apparent at six weeks. Gastroduodenal endoscopy was repeated after 6 weeks and, if the ulcer was not healed, after 10 weeks. Patients entered the 1-year follow-up period either 6 or 10 weeks after the start of treatment, depending on when their ulcers healed. If healing had not occurred after 10 weeks, the patient was dropped from the study. Patients whose ulcers healed were examined every two months for one year. Follow-up gastroduodenal endoscopy was performed 2, 6, and 12 months after the start of follow-up or whenever clinical findings suggested the recurrence of duodenal ulcer. The randomization code was not broken until the last patient had completed the 12-month visit (December 1991).
Criteria for Completing the Trial
The patients were considered to have completed the trial if the ulcers had not healed after 10 weeks of treatment or if duodenal ulcers recurred during follow-up; recurrence was defined as the reappearance of a fibrin-coated niche. Patients with healing and no recurrence during 12 months of follow-up were considered to have completed the study. Data on patients who were not followed for 12 months, up to the last information available, were included in the analysis.
Gastroduodenal Endoscopy
Five biopsy specimens (two for culture, two for histologic examination, and one for a rapid urease test) were taken from the gastric antrum at each endoscopic examination. The endoscopes (Olympus GIF Q10, Q20, and XQ10 [Olympus Optical, Hamburg, Germany] and Pentax FG29H [Asahi Optical, Tokyo, Japan]) were disinfected with an automatic washing machine (Endo Thermo Disinfector, Olympus Optical) or manually by immersing them for 15 to 30 minutes in a solution of 10 percent succinic acid aldehyde and dimethoxytetrahydrofuran (Gigasept FF, Schulke and Mayr, Hamburg). Special attention was paid to rinsing all channels of each endoscope with this solution. Biopsy forceps were sterilized with ethylene oxide or disinfected as described above. The biopsy samples obtained for histologic examination were immediately placed in buffered formalin, and those obtained for microbiologic examination were transported in the top portion of a tube that contained 0.2 ml of sterile saline to maintain humidity19. For the rapid detection of urease production, the CU (campylobacter urease) test (Temmler, Marburg, Germany) was used.
Microbiologic Examination
The biopsy samples were processed within three hours after collection. The culture medium used was Mueller-Hinton agar (Oxoid, Basingstoke, United Kingdom) supplemented with 5 percent heat-lysed sheep's blood and antibiotic, as described by Dent and McNulty20. The plates were incubated for seven days in a microaerobic atmosphere (85 percent nitrogen, 10 percent carbon dioxide, and 5 percent oxygen) at 37 °C. H. pylori was identified by its typical microscopical appearance and a positive urease test. The cultured organisms were stored at -70 °C in horse serum with 17 percent glycerol for later testing of their susceptibility to metronidazole and amoxicillin, as described elsewhere19. Minimal inhibitory concentrations of 4 mg per liter for metronidazole and 16 mg per liter for amoxicillin were considered to indicate resistance.
Histologic Examination
Paraffin-embedded sections of tissue fixed with 8 percent phosphate-buffered formalin (pH 7.4) were stained with hematoxylin and eosin. For this study, they were evaluated only for the presence or absence of H. pylori.
The identification of H. pylori was based on the results of culture, histologic examination, and the rapid urease test of biopsy specimens. The organism was considered present if two of the three tests were positive; it was considered eradicated if all three tests were negative four weeks after antibiotic treatment ended. The efficacy or failure of antibiotic therapy with regard to the eradication of H. pylori was therefore assessed by evaluation of the biopsy specimens after six weeks. During the one-year course of the study, 19 of 164 specimens that were negative for H. pylori on culture were equivocal or weakly positive on either histologic examination (11 specimens) or urease testing (8 specimens). Because both the culture and one of the other two tests of these specimens were clearly negative, and because each weakly positive result was followed by a clearly negative result at the next scheduled appointment, these specimens were classified as negative for H. pylori.
Statistical Analysis
Differences between the two study groups in the incidence of ulcer healing and eradication of H. pylori, as well as the cumulative rate of recurrence of duodenal ulcer after follow-up of 12 months, were assessed with the chi-square test. The probability of remaining free of recurrence during the one-year follow-up period was analyzed by means of Kaplan-Meier survival estimates, and the event curves were compared by the log-rank test. All tests performed were two-tailed.
Results
Treatment Period
The demographic and personal characteristics of the patients in the two study groups were similar (Table 1). One patient from each group withdrew during treatment and did not return for the second endoscopic procedure at six weeks. The patients who received both ranitidine and antibiotic therapy had a significantly higher rate of ulcer healing at six weeks than did the patients who received ranitidine and placebo (P = 0.011). Endoscopic biopsy showed that H. pylori was eradicated in 46 patients in the antibiotic group (89 percent), as compared with 1 patient in the placebo group (2 percent) (Table 2).
|
|
32 mg per liter). Thus, among the five patients in the antibiotic group in whom H. pylori was not eradicated, the isolates of three patients showed primary resistance to metronidazole, and the isolates of two patients showed secondary resistance. A total of eight patients in the antibiotic group had side effects during treatment. Seven patients had diarrhea, with four to seven bowel movements per day; six of these patients completed the 12-day period of antibiotic treatment, after which the diarrhea subsided. The other patient who had diarrhea also had heartburn and discontinued antibiotic treatment after seven days; his diarrhea also subsided promptly. His ulcer was healed after six weeks, but he declined to continue in the study. Another patient in the antibiotic group had a rash and stopped taking ranitidine after five days, but he continued to take the antibiotics and participate in the study. He recalled that a similar event had occurred while he was taking ranitidine alone during previous treatment. Since his ulcer was healed after six weeks, he was invited to continue the study, in accordance with the intention-to-treat principle. One patient in the placebo group had diarrhea.
Follow-up Period
All 49 patients in the placebo group whose ulcers were healed at 10 weeks entered the follow-up period. Of the 51 patients in the antibiotic group whose ulcers were healed at this time, 1 patient withdrew from the study because of diarrhea and heartburn that developed during the first week of antibiotic treatment, as noted previously. Thus, 50 patients in this group entered the follow-up period (Table 3). In each group, four patients failed to return for one of the follow-up visits scheduled for every two months and were considered to have withdrawn from the study. Four patients in the antibiotic group (8 percent) and 42 in the placebo group (86 percent) had a recurrence of their duodenal ulcers during the one-year follow-up period (P<0.001). In a worst-case situation -- i.e., all four patients who withdrew from the antibiotic group had a recurrence, and all four who withdrew from the placebo group remained in remission -- the difference between the groups would continue to be significant.
|
Table 3 also shows the results of follow-up according to status for H. pylori as assessed by endoscopy at six weeks. One (2 percent) of the 46 patients who were negative for H. pylori at six weeks had a recurrence of duodenal ulcer (while still negative), as compared with 45 (85 percent) of the 53 patients positive for H. pylori (P<0.001). Of the 46 patients who were negative for H. pylori at six weeks (i.e., 46 in the antibiotic group plus 1 in the placebo group, minus 1 patient who declined to continue in the study), only 1 patient (2 percent) became positive for H. pylori during follow-up. This reinfection was detected at 12 months but was not accompanied by an ulcer recurrence. Of the 53 patients (5 in the antibiotic group and 48 in the placebo group) who were found to be positive for H. pylori at endoscopy at six weeks, all continued to be positive.
The probability of remaining in remission is shown in Figure 1. The marked declines in the curves 2, 6, and 12 months after the start of the follow-up period are due to the detection of some asymptomatic recurrences at the endoscopic examinations scheduled at these times, and also to the tendency of patients to wait for the next appointment despite having some symptoms. Asymptomatic ulcers could have formed and healed spontaneously during the intervals between endoscopic examinations, without being detected at these evaluations. However, since this possibility applies to both study groups, a bias influencing the results appears to be unlikely.
|
We found that healing of recurrent duodenal ulcers occurred in more of the patients treated with ranitidine and antibiotics than the patients treated with ranitidine and placebo. These results confirm those of earlier studies: acute duodenal ulcers heal more rapidly in patients who receive treatment for H. pylori infection than in those who do not2,21. These findings indicate that H. pylori infection probably delays the process of ulcer healing.
Fifteen percent of the patients treated with antibiotics had side effects, but only one patient discontinued treatment for this reason. This dropout rate compares favorably with a rate of 21 percent6 for side effects of bismuth subcitrate, amoxicillin, and metronidazole. In two studies of patients treated with bismuth subcitrate, tetracycline, and metronidazole, nausea occurred in 32 percent7 and 25 percent,8 and diarrhea in 7 percent7 and 11 percent8. Undoubtedly, side effects and the large number of tablets to be taken are major obstacles to the large-scale use of the currently available triple-drug regimens.
In the present trial, 89 percent of the patients were free of H. pylori infection after treatment with amoxicillin and metronidazole for 12 days. Glupczynski and Burette22 reported that H. pylori was eradicated in as many as 95 percent of patients treated with 2.0 g of amoxicillin a day for 16 days in combination with 1.5 g of metronidazole a day for 10 days. These authors also found that regimens for H. pylori infection had to be followed for at least 10 days to avert the development of secondary resistance to metronidazole during therapy. They further reported that H. pylori could be eradicated in up to 70 percent of patients with primary resistance to metronidazole if bismuth compounds were added to the combination of metronidazole and amoxicillin22.
Of the 46 patients cured of H. pylori infection in our study, only 1 (2 percent) had a recurrent duodenal ulcer. After achieving an eradication rate of 89 percent with triple-drug therapy and ranitidine in a group of 47 patients with duodenal ulcer, Graham et al.23 noted three recurrences among patients negative for H. pylori, all occurring in association with the administration of nonsteroidal antiinflammatory drugs. As a cause of relapse, reinfection appears to be of minor importance, since the rate of reinfection in this and other studies was 2 percent or less23,24.
In a recent review of the literature, Rabeneck and Ransohoff25 concluded that there was not sufficient evidence of a causal relation between H. pylori infection and the recurrence of duodenal ulcer. These authors suggested that to clarify the putative role of H. pylori as a pathogenic agent in duodenal ulcer disease, clinical trials should be performed to evaluate agents with antibacterial properties but without direct beneficial protective effects on the gastroduodenal mucosa. The study reported here meets these requirements: amoxicillin and metronidazole are effective against H. pylori and are not known to have direct actions on the mucosa.
Our results support the concept of a cause-and-effect relation between H. pylori infection and recurrent duodenal ulcer. Although H. pylori is certainly not the only cause, its eradication favorably alters the natural course of the disease. With respect to an eventual practical application of the regimen used in this study, we consider an eradication rate of more than 80 percent to be satisfactory. The benefit of a lasting remission in 84 percent of patients outweighs the 15 percent rate of side effects. We now offer this combination of antibiotics and ranitidine to patients with H. pylori-associated chronic duodenal ulcer disease, as an alternative to long-term drug-maintenance therapy or elective surgery.
Supported by a research grant from Biochemie Ltd. (Vienna).
We are indebted to Biochemie Ltd. for general support, to Dr. Jasmina Preinreich for carefully coordinating and monitoring the study, to Dr. Guenther Nirnberger and Miss Johanna Vierheilig for performing the statistical analysis, to Mrs. Sari Lindley for assistance in the preparation of the manuscript, and to Mrs. Maria Blaha for technical assistance.
Source Information
From the Medical Department I and the Department of Pathology, Hanusch Hospital, Vienna (E.H., H.N., K.S.); the Medical Department II, General Hospital, Graz (G.B., H.W.); the Outpatient Department South, Regional Public Medical Insurance, Vienna (B.D.); the Department of Clinical Microbiology, Hygiene Institute, University of Vienna School of Medicine, Vienna (A.M.H.); and the Department of Pathology, University of Graz School of Medicine, Graz (M.T.) -- all in Austria.
Address reprint requests to Dr. Hentschel at Hanusch Hospital, Medical Department I, Heinrich-Collinstr. 30, A-1140 Vienna, Austria.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Eradication of Helicobacter pylori in Recurrent Duodenal Ulcer
Vigneri S., Savarino V., Termini R., Mela G. S., Hentschel E.
Extract |
Full Text
N Engl J Med 1993;
329:59-60, Jul 1, 1993.
Correspondence
More on the Eradication of Helicobacter pylori and the Recurrence of Duodenal Ulcer
Schutze K., Hentschel E.
Extract |
Full Text
N Engl J Med 1993;
329:1356, Oct 28, 1993.
Correspondence
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
