|
|
| |||||||||||||||||||||||||||||||||||||||
Background Hemorrhage is the most common complication of duodenal ulcer disease, but there is little information about the effectiveness and safety of long-term maintenance therapy with histamine H2-receptor blockers.
Methods We conducted a double-blind study in patients with endoscopically documented hemorrhage from duodenal ulcers. Patients were randomly assigned to maintenance therapy with ranitidine (150 mg at night) or placebo and were followed for up to three years. Endoscopy was performed at base line (to document that the ulcers had healed), at exit from the study, and when a patient had persistent ulcer symptoms unrelieved by antacids or had gastrointestinal bleeding. Symptomatic relapses without bleeding were treated with ranitidine; if the ulcer healed within eight weeks, the patient resumed taking the assigned study medication.
Results The two groups were similar at entry, which usually occurred about three months after the index hemorrhage. After a mean follow-up of 61 weeks, 3 of the 32 patients treated with ranitidine had recurrent hemorrhage, as compared with 12 of the 33 given placebo (P<0.05). Half the episodes of recurrent bleeding were asymptomatic. One patient in the ranitidine group withdrew from the study because of asymptomatic thrombocytopenia during the first month.
Conclusions For patients whose duodenal ulcers heal after severe hemorrhage, long-term maintenance therapy with ranitidine is safe and reduces the risk of recurrent bleeding.
We compared the efficacy and safety of maintenance treatment with ranitidine with the effects of placebo for the prevention of recurrent bleeding from duodenal ulcers. We studied patients with a recent severe hemorrhage, the group most likely to benefit from maintenance treatment9,12,13.
Methods
In this multicenter, double-blind study, 65 patients with endoscopically documented hemorrhage from an ulcer and subsequent healing were studied as outpatients from May 1986 through December 1989. The study centers were the UCLA Center for the Health Sciences, the Center for Ulcer Research and Education, and Sunset Kaiser Permanente in Los Angeles; Valley Medical Center in Van Nuys, California; Highland Hospital in Oakland, California; and the Ochsner Medical Institutions in New Orleans. All the patients gave written informed consent before they entered the study. The study was approved by the research and human-studies committees at each center, all of which agreed that the study was ethically and scientifically justified and had substantial clinical relevance.
Study Population
The criteria for entry were an age of at least 18 years, a life expectancy of at least 3 years, upper gastrointestinal hemorrhage severe enough to lead to hospitalization, and a decline in the hematocrit of at least 5 percentage points (or the transfusion of red cells). The hemorrhage, from a duodenal or pyloric-channel ulcer, had to have been documented by endoscopy within one year before randomization. In addition, the healing of all ulcers was confirmed endoscopically within five days before randomization, and the use of ulcer medications (except antacids) was suspended for five days before randomization.
We excluded patients who had ingested ulcerogenic drugs (aspirin or other nonsteroidal antiinflammatory drugs) before the hemorrhage or who had pyloric obstruction, previous ulcer surgery except simple oversewing of a perforation, uncompensated or unstable medical conditions, hepatic disease, cancer (except of the skin), Zollinger-Ellison syndrome, hypersensitivity to any H2-receptor antagonist, or a history of drug or alcohol abuse that would interfere with participation in the study. Women of reproductive age who were pregnant, lactating, or not using birth control were excluded.
Protocol
Within five days before randomization, the patients underwent outpatient endoscopy to document healing (re-epithelialization) of the ulcer that had bled and the absence of other ulcers or major abnormalities of the upper gastrointestinal tract. All the patients underwent medical examination, medical history taking, and testing for fecal occult blood. Blood and urine samples were obtained. The patients received tablets to be taken nightly as maintenance therapy that were prepared, coded, and randomly distributed by Glaxo. The tablets, which were identical in appearance, contained 150 mg of ranitidine (Zantac) or placebo. Antacid tablets (Maalox) were also given to the patients for the relief of ulcer pain. Each patient kept a diary recording the occurrence of ulcer symptoms, evidence of gastrointestinal bleeding, the ingestion of medications, and the occurrence of non-ulcer illness.
Follow-up visits were scheduled for 1, 2, and 3 months after randomization and every 3 months thereafter for as long as 36 months. At each visit, the patient's diary was reviewed and a history was taken, with emphasis on symptoms of ulcer or evidence of gastrointestinal bleeding; laboratory tests were performed; and tests for fecal occult blood were evaluated. The study tablets and antacid tablets taken were counted, and a new supply of medications was issued.
There were unscheduled follow-up visits in the event of persistent symptoms of ulcer unrelieved by study medication and antacids, evidence of gastrointestinal bleeding, major non-ulcer illness, or hospitalization for any reason. The patients underwent endoscopy only if they reported persistent symptoms for at least seven days that suggested recurrent ulceration, if they had an upper gastrointestinal hemorrhage (hematemesis, melena, or hematochezia), or if they completed or withdrew from the study.
Recurrence, Retreatment, and Treatment Failure
In cases of recurrent symptomatic ulcers without bleeding, coded treatment was discontinued and the patient entered a relapse phase of the protocol, receiving open-label ranitidine (300 mg twice a day) and antacid pills as needed for the relief of pain. If there was endoscopic healing within eight weeks, the patient returned to the maintenance study and received the same coded treatment as before the symptomatic relapse. The relapse phase was counted as part of the three years of the study. If the ulcer did not heal within eight weeks, the coded treatment was considered a failure and the patient was withdrawn from the study.
Study treatment was considered to have failed in patients who had clinical evidence of upper gastrointestinal bleeding and endoscopic evidence of a bleeding ulcer or the stigmata of recent hemorrhage from ulcer,14,15,16 and these patients were withdrawn from the study. Patients were also withdrawn from the study if they had a third symptomatic recurrence of ulcer (without recurrent bleeding), a marked adverse reaction to a study medication that required its discontinuation, or a life-threatening illness that required hospitalization or the discontinuation of the coded treatment.
Data Collection and Statistical Analysis
Data were collected on standard forms by a research study nurse, checked for accuracy, and sent to the study data center, where they were prospectively entered into computer files and checked. An interim analysis was planned according to the methods of O'Brien and Fleming17 and was performed during the fourth year of study. Because chi-square analysis found a significant difference between the treatment groups in the incidence of recurrent hemorrhage from ulcer, the study was terminated. A comprehensive biostatistical analysis was then performed. All the patients who entered the study were included in the final analysis, whether they dropped out, had a failure of treatment, or had recurrent bleeding during the study.
The base-line characteristics of the study groups were compared by t-test and chi-square analysis. The primary outcome was the recurrence of hemorrhage from ulcer. This was analyzed in two ways, according to treatment group: by chi-square analysis and survival analysis. The chi-square test was used for comparisons of dichotomous outcomes, such as the presence or absence of recurrent bleeding. The time to recurrence was analyzed by product-limit survival analysis with the BMDP software program18. The equality of event curves between study groups was analyzed with the log-rank test (Mantel-Cox generalized Savage test) and the Breslow generalized Wilcoxon test18. For all analyses, a two-sided P value of less than 0.05 was considered to indicate statistical significance.
Results
The study groups were comparable at the time of randomization (Table 1). There were no significant differences in base-line characteristics. Most patients were men, and their mean age was 49 years. Two risk factors, cigarette smoking and alcohol consumption, were more frequent in the ranitidine group, but the difference from the placebo group was not significant. Although smoking may influence the recurrence of ulcer,11 no patient stopped smoking during the study.
|
Table 2 shows the outcomes in each study group. The mean follow-up time was 61 weeks. Twelve patients in the placebo group and three in the ranitidine group had the primary outcome, recurrent bleeding ulcers (P<0.05 by chi-square test). Because bleeding ulcers were significantly less frequent in the ranitidine group, the study was stopped before all patients had completed 36 months of follow-up.
|
There was a higher dropout rate in the placebo group (27 percent) than in the ranitidine group (19 percent) (Table 2). In the first three months 15 percent of the placebo group (five patients) and 16 percent of the ranitidine group (five patients) dropped out of the study. An additional 12 percent (four patients) and 3 percent (one patient), respectively, dropped out during the remainder of the study. We considered all patients who discontinued follow-up because of medical reasons, adverse events, noncompliance or a refusal to return to the study center, or a move to have dropped out (Table 2).
Details of Ulcer and Bleeding Recurrences
Table 3 gives details of recurrent painful ulcers and bleeding ulcers. The median time to recurrences of bleeding was similar in the two groups -- 8.5 months in the placebo group and 6.2 months in the ranitidine group. In the placebo group, 50 percent of the recurrent bleeding ulcers (6 of 12 patients) were "silent" -- i.e., not associated with antecedent or concurrent pain or other symptoms of ulcer. Of the three patients in the ranitidine group with recurrent bleeding, two also had dyspepsia. In both groups bleeding ulcers frequently recurred in the same location, as indicated by endoscopy and their occurrence at the site of previously documented scars.
|
Figure 1 shows the time to recurrent bleeding from ulcer. In the survival analysis, the fraction of patients without recurrent bleeding was significantly greater in the ranitidine group than in the placebo group. After the first year there was no recurrent bleeding in the ranitidine group. In contrast, recurrences continued in the placebo group throughout the study. Patients in the placebo group appeared to remain at comparable levels of risk during the entire study. According to survival analysis, approximately 50 percent of the patients in the placebo group had recurrent hemorrhage within two years, as compared with 10 percent of the patients given ranitidine (P<0.02) (Figure 1).
|
Of the 65 patients who entered the study with a recent upper gastrointestinal hemorrhage from a duodenal ulcer, 15 had a recurrent hemorrhage during the study period. One other patient in whom a nonbleeding visible vessel was found on endoscopy was withdrawn for symptoms of ulcer recurrence without bleeding. No other complications of ulcer disease were found.
One patient receiving ranitidine and none receiving placebo had adverse events leading to withdrawal from the study. The episode of reversible thrombocytopenia in the patient who withdrew appeared to be related to ranitidine, although the patient refused rechallenge with it. There were no significant differences between the two groups on other laboratory tests. There were no deaths during the study.
Discussion
Our study was designed to test the hypothesis that the high risk of recurrent hemorrhage from ulcer in patients with chronic duodenal ulcer disease could be significantly reduced by the long-term administration of an H2-receptor antagonist. We chose to study patients with prior documented duodenal hemorrhage because there were no data from randomized, controlled trials indicating that medical treatment prevents this complication and because there is a high risk of recurrent hemorrhage in this group13.
Several risk factors significantly increase the likelihood of symptomatic recurrence of ulcer in studies of patients with uncomplicated disease (i.e., without hemorrhage) who are given placebo1,5,6,8,20,21,22. In a recent two-year maintenance study of patients with symptomatic ulcers, these risk factors were alcohol consumption, smoking, early onset of ulcers or repeated ulcers, and evidence of duodenal scarring or erosion on index endoscopy11. The incidence of these risk factors was similar in the two treatment groups in our study. Long-term maintenance therapy with an H2-receptor blocker has often been recommended in such patients and, by extrapolation, in patients with previous complications such as hemorrhage4,5,12,22,23. The rationale is that patients with recurrent, symptomatic duodenal ulceration are at increased risk for recurrence and future complications such as hemorrhage4,5,12,22,23. The problems with continuous maintenance therapy are the cost, patient compliance, and uncertainty about the safety of very-long-term acid suppression24,25. Unfortunately, patients with previous complications of ulcers have not been the primary focus of most maintenance studies, and they have been excluded from most randomized, controlled trials of maintenance therapy3,6,8.
Murray et al. recently reported on a randomized, double-blind, controlled trial of medical maintenance therapy for patients with hemorrhage from duodenal ulcers8. Their study did not find any difference in the incidence of recurrent hemorrhage between the placebo and the ranitidine groups. During that two-year trial, ulcerations or severe erosions (symptomatic or asymptomatic) recurred in 24 of the 40 patients studied, but hemorrhage recurred in only 2. In the light of our data, this was a surprisingly low frequency of recurrent bleeding in the patients who received placebo. A few other studies have examined the natural history of peptic ulcer disease and the possible efficacy of maintenance therapy with H2-receptor blockers in preventing complications4,5,6,7,12. However, these either focused on patients with uncomplicated disease4,12 or were nonrandomized or retrospective studies5,6,7.
Ulcer surgery is the only treatment besides maintenance medical therapy that has significantly reduced the risk of recurrent hemorrhage13,24,26. Ulcer surgery has risks, however, particularly in older patients. Our controlled study demonstrates that patients with a documented history of hemorrhage from duodenal ulcers have a high risk of recurrence that can be significantly reduced by continuous medical therapy. Conversely, according to survival analysis, with placebo the risk of bleeding was high over the 36 months of follow-up and did not decrease during the study. Intermittent medical treatment for symptomatic recurrences of ulcer in the placebo group apparently did not reduce the risk of subsequent hemorrhage. This may be because about 50 percent of the episodes of recurrent bleeding in both study groups were silent, unaccompanied by pain or other symptoms of ulcer.
These data have enormous public health implications. Patients with ulcer and prior hemorrhage have frequent recurrences when they are not receiving maintenance therapy, and they present with recurrent hemorrhage3,13. These patients account for a substantial portion of the hospital admissions for hemorrhage from ulcer each year. The subgroup of patients with documented hemorrhage is the most likely to benefit from maintenance medical therapy or elective surgery2,3,13. If these patients do receive long-term medical treatment, it could markedly reduce hospital admissions for hemorrhage from duodenal ulcer and reduce the cost of care24,25,27.
Supported in part by grants (AM 33273 and 41301 -- Human Studies CORE) from the National Institute of Diabetes and Digestive and Kidney Diseases and by Glaxo, Inc.
We are indebted to Assumpta Oturu for assistance in the preparation of the manuscript and to our colleagues who referred their patients for inclusion in this study.
Source Information
From the Center for Ulcer Research and Education, UCLA, and the West Los Angeles Veterans Affairs Medical Center, Los Angeles (D.M.J., S.C., T.O.G.K., G.R., M.E.J., T.R., G.V.D.); Kaiser Permanente, Los Angeles (H.F.); Valley Medical Center, Van Nuys, Calif. (G.M.); Ochsner Medical Institutions, New Orleans (J.S.); and Highland Hospital, Oakland, Calif. (M.S.).
Address reprint requests to Dr. Jensen at 44-133 CHS, GI Division, UCLA Center for the Health Sciences, Los Angeles, CA 90024-1684.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Ranitidine and Recurrent Hemorrhage from Duodenal Ulcer
Loffeld R.J.L.F., van der Putten A.B.M.M., Jensen D. M., Peterson W. L.
Extract |
Full Text
N Engl J Med 1994;
331:53-54, Jul 7, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
