Cumulative rates of healing at 4 and 8 weeks during treatment with ranitidine hydrochloride, 150 mg twice daily; lansoprazole, 15 mg once daily; or lansoprazole, 30 mg once daily, among intent-to-treat patients.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Agrawal NM, Campbell DR, Safdi MA, et al. Superiority of Lansoprazole vs Ranitidine in Healing Nonsteroidal Anti-inflammatory Drug–Associated Gastric Ulcers: Results of a Double-blind, Randomized, Multicenter Study. Arch Intern Med. 2000;160(10):1455–1461. doi:10.1001/archinte.160.10.1455
The usefulness of nonsteroidal anti-inflammatory drugs (NSAIDs) is limited by adverse gastrointestinal tract events.
To identify the optimal antisecretory therapy for healing of gastric ulcer in patients using NSAIDs and the impact of concurrent Helicobacter pylori infection on ulcer healing.
Prospective, double-blind, multicenter, parallel-group study.
Gastroenterology practices in ambulatory and referral center settings.
Three hundred fifty-three patients with an active, nonmalignant gastric ulcer at least 5 mm in diameter confirmed by endoscopy and biopsy and who continued to receive stable doses of NSAIDs.
Patients were randomized to receive ranitidine hydrochloride, 150 mg twice daily, or lansoprazole, 15 mg or 30 mg once daily, for 8 weeks.
Healing was assessed by endoscopy at 4 and 8 weeks in an intent-to-treat population. Helicobacter pylori status was assessed by histological examination.
After 8 weeks of treatment, healing was observed in 61 (53%) of 115, 81 (69%) of 118, and 85 (73%) of 117 patients receiving ranitidine lansoprazole, 15 mg, and lansoprazole, 30 mg, respectively (P<.05 for ranitidine vs both lansoprazole doses; 95% confidence interval, 3.2-28.0 for ranitidine vs lansoprazole, 15 mg, and 7.4-31.8 for ranitidine vs lansoprazole, 30 mg). The gastric ulcer healing rates were similar between H pylori–infected and –noninfected patients, with a statistically significant increase with the use of lansoprazole vs ranitidine.
In patients who require continuous treatment with NSAIDs, lansoprazole is superior to ranitidine for healing of NSAID-associated gastric ulcers. Healing is not delayed by the presence of H pylori infection.
NONSTEROIDAL anti-inflammatory drugs (NSAIDs) are considered standard therapy for pain and inflammatory conditions associated with arthritis and other musculoskeletal disorders, with more than 1% of Americans using them on a regular basis.1 However, the usefulness of NSAIDs is limited by adverse events, which are reported more frequently than those of any other medication class.2 These agents cause a spectrum of gastrointestinal tract disorders, from dyspepsia and heartburn to damage of the gastroduodenal mucosa, including erosion, ulceration, bleeding, and perforation. The point prevalence for peptic ulcer disease among patients who use NSAIDs on a long-term basis is estimated at 10% to 30%, which is 10 to 30 times higher than that of the general population.1 In a recent meta-analysis of cohort and case-control studies, a risk ratio of 3.7 was calculated for adverse gastrointestinal tract outcomes among patients receiving long-term NSAID therapy.3
Most gastroduodenal ulcers are associated with chronic Helicobacter pylori infection of the gastric mucosa or with NSAID use,2-4 both of which impair the normal mucosal defense against damage (by different mechanisms), thus allowing gastric acid or other noxious agents to cause injury. Therefore, gastric acid is still considered central to the pathogenesis of gastroduodenal ulceration,2 and acid suppression is considered pivotal to treatment of NSAID-associated mucosal damage.
Ulcers heal readily with acid suppressant therapy if patients discontinue their NSAID therapy.5-7 In contrast, a high-dose histamine2 (H2)–receptor antagonist or a proton pump inhibitor is needed to heal ulcers if the NSAID therapy must be continued,7-10 although NSAID use may delay gastric ulcer healing with H2-receptor antagonists.6,7
Our large, double-blind, multicenter, randomized, parallel-group study was designed to identify the optimal antisecretory therapy for healing of gastric ulcer in patients receiving NSAIDs by comparing the healing rates of lansoprazole with those of the H2-receptor antagonist ranitidine hydrochloride while patients continued using their NSAIDs. Given the controversial role of H pylori in NSAID-associated ulcers,4 the impact of concurrent H pylori infection on ulcer healing rate was evaluated also.
Patients who met the entry criteria, including age 18 years or older, active, with nonmalignant gastric ulcer (primary lesion of ≥5 mm in diameter) documented by endoscopic and biopsy findings within 7 days of initiating treatment (at which time H pylori status was determined), and who were receiving stable doses of an NSAID for at least the previous month were eligible for enrollment into the study. Most patients (>90%) were taking a full therapeutic dose of an NSAID for their underlying indication. Patients with multiple ulcers (n=117), coexisting duodenal ulcer (n=48), or reflux esophagitis (n=43) were also eligible to participate in the study. Patients with active bleeding or perforated ulcers at screening were excluded. Use of any proton pump inhibitor within 2 weeks before initiating study drug therapy or use of an H2-receptor antagonist within 24 hours before study entry excluded patients. Patients requiring anticoagulants or corticosteroid therapy (at dosages greater than the equivalent of prednisone, 10 mg/d) were also excluded. Approval for the study was obtained from the institutional review board of each of the 43 participating centers in North America, and written informed consent was obtained before a patient's enrollment in the study.
Patients were assigned randomly in a 1:1:1 ratio to receive ranitidine hydrochloride, 150 mg twice daily, or lansoprazole, 15 or 30 mg once daily. A double-dummy design was used to maintain the double-blind nature of the study. Patients self-administered 1 capsule before breakfast and dinner for 8 weeks. Patients also received antacid tablets (a combination of 0.4 g each of aluminum hydroxide and magnesium hydroxide and 25 mg of simethicone [Gelusil; Parke-Davis, Morris Plains, NJ]) and were instructed to use them only as needed for symptom relief. Patients were instructed to avoid antiulcer (H2-receptor antagonist, misoprostol, or sucralfate) or ulcerogenic medication (except NSAIDs as noted or low-dose aspirin) other than study drug or Gelusil, from study enrollment until completion of the study.
Compliance and adverse events were assessed by returned capsule count and direct questioning at each treatment visit. Symptoms were assessed by patient diary on a daily basis. Patients recorded episodes of daytime and nighttime abdominal pain (defined as none, mild, moderate, or severe), study drug and NSAID dosing information, and frequency of antacid consumption.
Endoscopy with biopsy was repeated at the week-4 and -8 visits to document the healing status of the gastric ulcers and to assess gastritis and H pylori status (week-8 visit only). Patients with H pylori infection did not receive H pylori eradication treatment, so that the effect of H pylori status on gastric ulcer healing rates could be determined. Healing of a gastric ulcer was defined as the complete reepithelialization of gastric mucosa, with no evidence of an ulcer crater or erosion at the ulcer site.
Six mucosal biopsy specimens, 2 each from the greater and lesser curvatures of the antrum and greater curvature of the corpus, were obtained at each endoscopy. Formalin-fixed, paraffin-embedded tissue sections were stained separately with hematoxylin-eosin for the evaluation of gastritis, and with Warthin-Starry silver stain to detect H pylori. All histological preparations were evaluated by the same pathologist (M.M.H.), who was unaware of the patients' clinical status, treatment, and the results of other tests. Results were reported as positive for H pylori infection if the organism was detected in any of the examined tissue samples. Biopsy specimens were also examined to assess the degree of gastritis, including acute and chronic inflammation, atrophy, and intestinal metaplasia, and were graded semiquantitatively according to the Sydney system.11
A sample size of at least 300 patients (100 patients in each treatment group) was chosen, as that number would allow for the detection of significant differences between the lansoprazole and ranitidine treatment groups in gastric ulcer healing rates at the week-8 visit at the level of P = .05 (2 tailed) with approximately 89% power, assuming healing rates of 85% and 65%, respectively. Statistical analyses were conducted using SAS, version 6.12 (SAS Institute Inc, Cary, NC).
Intent-to-treat analyses were conducted for ulcer healing and symptom relief based on diary data. For all efficacy and safety end points, comparisons were made between the results obtained with either lansoprazole dose and those obtained with ranitidine. For the efficacy end points, a comparison between both lansoprazole dose groups was also conducted.
The comparability of the treatment groups at baseline was assessed with respect to demographic variables, medical and social histories, gastric ulcer size, and H pylori status using the χ2 test for categorical variables and 1-way analysis of variance for continuous variables. Baseline severity of symptoms, based on the investigator interview, was compared among the treatment groups using the Cochran-Mantel-Haenszel method for ordered response variables.
Pairwise treatment group differences were assessed at week-4 and -8 visits with respect to ulcer healing using the Cochran-Mantel-Haenszel test with study sites as a stratification factor. The treatment groups were also compared after adjusting for concomitant factors (ie, baseline H pylori status, baseline gastric ulcer size, presence of hiatal hernia, sex, age, race, tobacco use, alcohol use, and caffeine use) using the Cochran-Mantel-Haenszel test with the concomitant factor as a stratification factor. Pairwise treatment group comparisons of the mean severity of daytime and nighttime abdominal pain, the percentage of days and nights with abdominal pain, the mean number of antacid tablets used per day, and the percentage of days that antacid was used during the first 4 weeks of treatment and during the entire 8-week treatment period were performed using the Wilcoxon 2-sample test for patient diary data. Resolution or improvement of gastritis findings from baseline to the week-8 visit was assessed using the Cochran-Mantel-Haenszel test with baseline severity as a stratification factor. The Fisher exact test was used to compare the incidence of treatment-related adverse events (defined as possibly or probably related) between the treatment groups.
Endoscopy was performed on 669 patients. Of these, 353 met the study criteria for presence of ulcer disease and were randomized to receive ranitidine, 150 mg twice daily (n=117), lansoprazole, 15 mg once daily (n=119), or lansoprazole, 30 mg once daily (n=117), for 8 weeks. Three (1%) of the 353 enrolled patients were excluded from all efficacy analyses: 2 patients with no gastric ulcer and 1 patient who did not receive an NSAID before or during treatment. In addition, 9 patients (3%) were not included in the analyses of patient diary data because they did not record data. The primary indication for NSAID use was arthritis, and the most commonly used NSAIDs in the study population were ibuprofen (33%), naproxen sodium (24%), aspirin and aspirin combinations (21%), diclofenac sodium (18%), and oxaprozin (8%). The treatment groups were well matched at baseline, including demographic characteristics, gastric ulcer size, H pylori status, severity of symptoms, history of gastrointestinal tract disorders, and histological gastritis (Table 1). Most patients (58%, 61%, and 62% in the rantidine, lansoprazole, 15 mg, and lansoprazole, 30 mg, treatment groups, respectively) reported mild or moderate daytime abdominal pain, and approximately half reported mild or moderate nighttime abdominal pain. One hundred (29%) of 350 patients were positive for H pylori infection at baseline. The prevalence of H pylori positivity among the treatment groups at week 8 was similar to the baseline values (74/310 [24%]).
Among patients who continued their NSAID therapy, lansoprazole treatment resulted in a statistically significant increase in healing rate for NSAID-associated gastric ulcer at the week-4 and -8 visits, compared with ranitidine (Figure 1). The gastric ulcer healing rate was numerically higher, although the difference was not statistically significant, among patients receiving lansoprazole, 30 mg. Healing after 4 weeks of treatment was observed among 34 (30%) of 115 patients treated with ranitidine, compared with 56 (47%) of 118 treated with lansoprazole, 15 mg (P=.009 vs ranitidine; 95% confidence interval [CI], 5.6-30.2), and 67 (57%) of 117 patients treated with lansoprazole, 30 mg (P<.001 vs ranitidine; 95% CI, 15.5-39.9). The healing rates after 8 weeks of treatment were 53% (61/115 patients) for ranitidine, compared with 69% (81/118 patients; P=.01 vs ranitidine; 95% CI, 3.2-28.0) for lansoprazole, 15 mg, and 73% (85/117 patients; P=.009 vs ranitidine; 95% CI, 7.4-31.8) for lansoprazole, 30 mg (Table 2). These observations were unaffected after adjustment for potentially influential factors, including age, sex, race, baseline gastric ulcer size, presence of hiatal hernia, baseline H pylori status, and use of alcohol, tobacco, or caffeine.
After 4 weeks of antisecretory treatment, healing was observed in 37% of H pylori–infected and 26% of non–H pylori–infected patients treated with ranitidine, in 47% and 46%, respectively, of patients treated with lansoprazole, 15 mg (P<.01 vs ranitidine), and in 56% and 58%, respectively, of patients treated with lansoprazole, 30 mg (P<.001 vs ranitidine), after controlling for H pylori status. After 8 weeks of treatment, the gastric ulcer healing rates in H pylori–infected patients remained similar to the healing rates in non–H pylori–infected patients, whereas healing rates were higher (the difference was statistically significant) in those treated with lansoprazole compared with ranitidine (Table 2). Although 4 of the 6 observations across treatment groups demonstrated higher healing rates for patients infected with H pylori, this numerical difference was not statistically significant.
Among patients with gastric and duodenal ulcers who continued NSAID use, healing of the duodenal ulcers was documented after 8 weeks of treatment in 15 (88%) of 17, 14 (93%) of 15, and 13 (81%) of 16 patients treated with ranitidine, lansoprazole, 15 mg, and lansoprazole, 30 mg, respectively. Among patients with gastric ulcers and erosive esophagitis (erosions and/or ulcerations) and who required long-term NSAID use, healing of the esophagitis, defined as normal-appearing mucosa observed at endoscopy, was documented after 8 weeks of treatment in 12 (86%) of 14, 17 (85%) of 20, and 9 (100%) of 9 patients treated with ranitidine, lansoprazole, 15 mg, and lansoprazole, 30 mg, respectively. There were no statistically significant differences between the groups after 8 weeks of treatment based on healing rate for duodenal ulcer or erosive reflux esophagitis; however, only a small number of patients were in these categories.
During the initial 4 weeks of treatment and during the entire 8 weeks of treatment, a numerical, although not a statistically significant, increase in symptom relief and a statistically significant decrease in antacid use were observed with lansoprazole compared with ranitidine based on analyses of patient diary data (Table 3). The early improvements noted during the first 4 weeks of treatment included a statistically significant smaller percentage of days of antacid use and fewer antacid tablets taken per day in the lansoprazole, 30 mg, group compared with the ranitidine group. Antacid use was significantly reduced among patients in the lansoprazole groups compared with patients in the ranitidine group during the entire 8-week treatment period, based on fewer antacid tablets taken per day (both lansoprazole groups) and a smaller percentage of days of antacid use (lansoprazole, 30 mg, group).
In general, patients did not experience a net change in histological markers of chronic gastritis at the week-8 visit compared with the pretreatment visit. Among the 328 patients with findings of acute inflammation at baseline, the grade of inflammation remained unchanged at the final treatment visit for 55%, whereas 25% demonstrated resolution or improvement and 19% demonstrated worsening. Among the 327 patients with findings of chronic inflammation at baseline, the grade of inflammation remained unchanged at the final treatment visit for 56%, whereas 25% demonstrated resolution or improvement and 19% demonstrated worsening. No statistically significant differences were observed between the treatment groups in the percentage of patients with resolution or improvement of acute or chronic inflammation. Among the patients with atrophy at baseline, these atrophic changes were not observed at the final visit. Focal atrophic changes developed by the final visit in 3 (3%) of 101 ranitidine-treated and 7 (3%) of 212 lansoprazole-treated patients with negative findings at baseline. No patient had intestinal metaplasia at baseline or at the final visit.
Most patients in each of the treatment groups were compliant with study medication (ie, they took at least 67% of study medication). Fifteen patients (4%) of 353 stopped taking study medication prematurely at least in part because of adverse events (6, 5, and 4 patients in the ranitidine, lansoprazole, 15 mg, and lansoprazole, 30 mg, groups, respectively).
The lansoprazole and ranitidine treatment groups were comparable based on the percentage of patients who reported a treatment-related adverse event: 13 (11%) of 117 patients in the ranitidine group, 10 (8%) of 119 in the lansoprazole, 15 mg, group, and 10 (9%) of 117 in the lansoprazole, 30 mg, group. The most commonly reported treatment-related event was diarrhea (4, 5, and 3 patients in the ranitidine, lansoprazole, 15 mg, and lansoprazole, 30 mg, groups, respectively), and there were no statistically significant differences between treatment groups for any specific drug-related adverse event. Most treatment-related adverse events experienced by study patients were mild or moderate in severity. There were no reports of gastric ulcer–related complications in any group during the trial.
Ulcers heal readily with H2-receptor antagonist therapy if patients discontinue their NSAID (90%-100% at 8 weeks).6,8,10 For those who require continuous NSAID therapy, acid suppression with high-dose H2-receptor antagonist leads to 8-week healing of NSAID-associated gastric ulcer in 50% to 80% of patients.6-8 The respective healing rates for patients using proton pump inhibitors who require continuous NSAID therapy are 80% to 87%.12,13 To determine the most appropriate strategy for those who require continuous NSAID therapy, our study was designed to compare lansoprazole with ranitidine for healing of gastric ulcers.
The study population consisted of patients with endoscopic documentation of gastric ulcer(s) at least 5 mm in diameter; treatment success was defined as complete healing of index ulcer, and failure to heal was defined as the presence of an ulcer or erosion at the original site. Increase in healing rates and decrease in antacid use (as a measure of symptom relief) were statistically significant with lansoprazole, compared with ranitidine. There was a trend toward greater symptom relief observed in both lansoprazole groups compared with the ranitidine group. These results are consistent with the findings that proton pump inhibitors are more effective than H2-receptor antagonists in healing gastric ulcers among patients who used NSAIDs regularly.12
Data from epidemiological studies show that H pylori is not a necessary cofactor for NSAID-associated ulcers.14-18 Among patients receiving long-term NSAID therapy, the prevalence of H pylori infection was similar among those with and without ulcers (50% and 51%, respectively).14 Laine et al18 showed that the prevalence of H pylori infection was lower among patients with gastric ulcers receiving NSAIDs compared with those who were not using the drugs (53% vs 83%, respectively; P=.01). Data from this study demonstrate a low (29%) prevalence of H pylori infection among patients with NSAID-associated gastric ulcers.
The results of studies evaluating the impact of H pylori infection on NSAID-associated mucosal injury are conflicting, with some suggesting increased damage,19-22 one suggesting a lower incidence of erosions in the setting of H pylori infection,17 and others suggesting that there is no association between NSAID-associated damage and H pylori status.23-25 In a study comparing the mucosal eicosanoid synthesis in NSAID users with that in control subjects, Hudson and colleagues26 showed that H pylori– infected patients were protected from the prostaglandin-lowering effects of the drugs.
The results of our study suggest that H pylori infection does not delay healing of gastric ulcers in NSAID users, although the choice of antisecretory agent has a significant impact on the healing rate achieved. This is likely based on the differences among antisecretory agents in their ability to increase intragastric pH during a 24-hour period, with lansoprazole being superior to ranitidine.27 After 4 and 8 weeks of antisecretory treatment, the gastric ulcer healing rate in H pylori–infected patients was similar to the healing rate in non–H pylori–infected patients within each treatment group; a statistically significant increase was seen with lansoprazole compared with ranitidine. Combined across all treatment groups, H pylori-infected patients (70/100 [70%]) had a numerically higher healing rate compared with non–H pylori–infected patients (153/243 [63%]). Others have also shown healing of NSAID-associated ulcers to be unaffected by H pylori infection. Healing rates at 4 and 8 weeks were the same in patients without H pylori infection treated with omeprazole sodium, 20 mg twice daily, and H pylori–infected patients treated with omeprazole sodium alone (20 mg twice daily) or omeprazole and amoxicillin, 1 g twice daily.28 In contrast, H pylori status was found to be a statistically significant positive prognostic factor for treatment success in 2 recently conducted studies.12,13 Additional well-designed studies are needed to resolve this important issue, but intriguing trends are being recognized.
Our study definitively demonstrates that lansoprazole, 15 mg or 30 mg once daily, is clearly superior to ranitidine hydrochloride, 150 mg twice daily, for healing of NSAID-associated gastric ulcers, and that H pylori infection does not have an impact on healing rates. Patients with gastric or duodenal ulcers who require continuous NSAID treatment can be treated safely with the proton pump inhibitor lansoprazole.
Accepted for publication October 4, 1999.
This study was supported by a grant from TAP Holdings Inc, Deerfield, Ill.
NSAID-Associated Gastric Ulcer Study Group: Robert J. Bailey, MD, Alberta, Ontario; Charles F. Barish, MD, Raleigh, NC; Thomas Bianci, MD, Tallassee, Ala; Charles Allen Birbara, MD, Worchester, Mass; Phillip C. Bird, MD, Norman, Okla; Jeffrey R. Breiter, MD, Manchester, Conn; Edward Cheng, MD, Northport, NY; Charles Collip, MD, Portland, Ore; Carleton Davis, MD, Monroe, Wis; Michael DeMicco, MD, Anaheim, Calif; James Doyle, MD, Spokane, Wash; Roy Fleischmann, MD, Dallas, Tex; Syam P. Gaddam, MD, Anaheim; William Harford, MD, Dallas; Samuel Ho, MD, Minneapolis, Minn; Keith P. Hussey, MD, Naples, Fla; James V. Jones, MD, Ruston, La; Mukul Khandelwal, MD, Hershey, Pa; David G. Kogut, MD, Statesville, NC; Richard Krause, MD, Chattanooga, Tenn; Steven Krumholz, Palm Beach, Fla; Paul N. Maton, MD, Oklahoma City, Okla; Aubrey McElroy, MD, Johnson City, Tenn; Morry Moskovitz, MD, Beaver, Pa; John Ondrejicka Jr, MD, Jacksonville Beach, Fla; Daniel Pambianco, MD, Charlottesville, Va; Terry Ponich, MD, Ontario; Ronald E. Pruitt, MD, Nashville, Tenn; Malcom Robinson, MD, Oklahoma City; Seymour Sabesin, MD, Chicago, Ill; Bruce Sahba, MD, San Diego, Calif; Howard I. Schwartz, MD, Miami, Fla; Jerrold Schwartz, MD, Arlington Heights, Ill; Nayan R. Shah, Leonardtown, Md; David Silvers, MD, Metairie, La; Stephen Sontag, MD, Hines, Ill; Lewis Strong, MD, Loveland, Colo; Peter Winkle, MD, Orange, Calif; Barry Winston, MD, Houston, Tex; and James Wolosin, MD, Escondido Calif.
Reprints: Naurang M. Agrawal, MD, Division of Gastroenterology, Duke University Medical Center, DUMC 3662, Durham, NC 27710 (e-mail: firstname.lastname@example.org)