Journal of Gastroenterology

, Volume 47, Issue 11, pp 1186–1197

Rabeprazole reduces the recurrence risk of peptic ulcers associated with low-dose aspirin in patients with cardiovascular or cerebrovascular disease: a prospective randomized active-controlled trial

Authors

  • Tsuyoshi Sanuki
    • Department of GastroenterologyKobe University Graduate School of Medicine
  • Tsuyoshi Fujita
    • Department of GastroenterologyKobe University Graduate School of Medicine
  • Hiromu Kutsumi
    • Department of GastroenterologyKobe University Graduate School of Medicine
  • Takanobu Hayakumo
    • Department of GastroenterologyKobe University Graduate School of Medicine
  • Shun-ichi Yoshida
    • Tabata Gastrointestinal Hospital
  • Hideto Inokuchi
    • Hyogo Cancer Center
  • Manabu Murakami
    • Department of GastroenterologyKobe University Graduate School of Medicine
  • Yoshihiro Matsubara
    • Department of Data Management and AnalysesTranslational Research Informatics Center
  • Hajime Kuwayama
    • Center for Bioethics and HumanitiesState University of New York
    • Department of GastroenterologyDokkyo Medical University Koshigaya Hospital
  • Takashi Kawai
    • Endoscopy CenterTokyo Medical University Hospital
  • Hideki Miyaji
    • Miyaji Medical Clinic
  • Takashi Fujisawa
    • Department of GastroenterologyKakogawa West City Hospital
  • Shuichi Terao
    • Department of GastroenterologyKakogawa East City Hospital
  • Yukinao Yamazaki
    • Department of Endoscopic MedicineUniversity of Fukui Hospital
    • Department of GastroenterologyKobe University Graduate School of Medicine
  • Care Study Group
Original Article—Alimentary Tract

DOI: 10.1007/s00535-012-0588-x

Cite this article as:
Sanuki, T., Fujita, T., Kutsumi, H. et al. J Gastroenterol (2012) 47: 1186. doi:10.1007/s00535-012-0588-x

Abstract

Background

Patients using low-dose aspirin (LDA) have an increased risk of gastroduodenal mucosal lesions and upper gastrointestinal symptoms. We aimed to clarify the efficacy of rabeprazole for preventing peptic ulcer, esophagitis, and gastrointestinal symptoms associated with LDA.

Methods

Patients with a history of peptic ulcers who were receiving LDA for cardiovascular or cerebrovascular disease were randomly assigned to receive rabeprazole at 10 mg daily, rabeprazole at 20 mg daily, or gefarnate (a cytoprotective anti-ulcer agent) at 50 mg twice daily. The primary endpoint was the development of gastric and/or duodenal ulcer at 12 weeks. The modified Lanza score (MLS) and gastrointestinal symptoms were evaluated at baseline and at 12 weeks.

Results

The full analysis set comprised 261 patients (rabeprazole 10 mg: n = 87, rabeprazole 20 mg: n = 89, gefarnate 100 mg: n = 85). The cumulative incidences of gastroduodenal ulcers at 12 weeks in the 10 mg rabeprazole group, 20 mg rabeprazole group, and gefarnate group were 7.4, 3.7, and 26.7 %, respectively (rabeprazole group 5.5 % vs. gefarnate group 26.7 %, hazard ratio [HR] 0.179; 95 % confidence interval [CI] 0.082–0.394; p < 0.0001). The proportions of patients with an MLS of ≥1 and erosive esophagitis were significantly lower in the rabeprazole group than in the gefarnate group at 12 weeks (gastric lesions 33.5 vs. 62.4 %, p < 0.0001; duodenal lesions 5.7 vs. 24.7 %, p < 0.0001; erosive esophagitis 5.8 vs. 19.4 %, p < 0.0001). Rabeprazole was significantly more effective than gefarnate for the resolution and prevention of gastrointestinal symptoms (resolution 53.6 vs. 25.0 %, p = 0.017; occurrence 9.2 vs. 28.3 %, p = 0.0026).

Conclusions

Rabeprazole is more effective than gefarnate for reducing the risk of recurrence of peptic ulcer, esophagitis, and gastrointestinal symptoms in LDA users.

Keywords

Randomized clinical trialRabeprazoleLow-dose aspirinPeptic ulcer

Introduction

Low-dose aspirin (LDA) therapy (75–325 mg once daily) is widely used for the primary and secondary prevention of cardiovascular and cerebrovascular disorders. However, aspirin causes gastrointestinal damage (including gastroduodenal ulcers), which is potentially associated with major complications such as gastrointestinal bleeding and perforation. We previously reported that LDA-induced gastroduodenal ulcers were endoscopically identified in 41 (10 %) of 400 Japanese patients [1]. In an observational study in Japan, it was reported that LDA was a prominent cause of bleeding ulcers in patients who underwent emergency endoscopy [2].

LDA users who develop gastrointestinal bleeding usually require discontinuation of aspirin, but this may raise the risk of thrombosis [3, 4]. Therefore, the prevention of LDA-induced gastroduodenal damage is important in LDA users having risk factors. The factors associated with LDA-induced gastroduodenal damage include aspirin dose, history of gastroduodenal ulcer or ulcer complication, concomitant use of non-steroidal anti-inflammatory drugs (NSAIDs), advanced age, concurrent use of anticoagulants, and Helicobacter pylori infection [5].

Recently, randomized controlled trials for the prevention of LDA-induced peptic ulcer, using proton pump inhibitors (PPIs) (lansoprazole, esomeprazole, pantoprazole) or H2-receptor antagonists (famotidine) have been reported [69]. However, the efficacy of rabeprazole for the prevention of the gastroduodenal damage associated with LDA has not yet been examined. Because the metabolism of rabeprazole is thought to be primarily non-enzymatic, in contrast to the metabolism of omeprazole and lansoprazole, and to be little affected by CYP2C19 genotypes [10, 11], the effect of rabeprazole is considered to have little difference between individuals.

The aim of this cardio-cerebrovascular disease and aspirin-ulcer recurrence and evaluation (CARE) study was to clarify the efficacy of rabeprazole for preventing gastroduodenal and esophageal mucosal lesions associated with LDA, and to investigate the effect of rabeprazole on upper gastrointestinal symptoms, in patients with a history of gastroduodenal ulcers who were receiving LDA for cardiovascular or cerebrovascular disease. In this study, we used two doses of rabeprazole (10 and 20 mg/day) to verify the protective efficacy of 10 mg/day rabeprazole for LDA-induced gastroduodenal ulcers. Because it is unethical to conduct a placebo-controlled trial in patients with risk factors for LDA-induced gastroduodenal damage, we used 100 mg/day gefarnate, a cytoprotective anti-ulcer agent for the treatment of gastroduodenal ulcers, as a control.

Methods

This phase-III, randomized, open-label, parallel-group, active-controlled, multi-center study [Foundation for Biomedical Research and Innovation, Translational Research Informatics Center (TRI) study code: TRIGID0801, ClinicalTrials.gov identifier: NCT01051388, UMIN-CTR: UMIN000002901] was conducted between August 2008 and July 2010 at 28 centers in Japan, in accordance with the Declaration of Helsinki and its amendments. This study protocol was approved by the ethics committee of each center, and all patients provided written informed consent before enrollment.

Participants

Adult patients, aged 20 years or over, were eligible for enrollment if they: required long-term LDA therapy for the prevention of recurrence of cardiovascular or cerebrovascular attack; had a peptic ulcer history confirmed by endoscopic findings of ulcer scar; and had no evidence of active peptic ulcers at baseline endoscopy. Patients were excluded if they: were in an acute or unstable phase of cardiovascular or cerebrovascular disease; had undergone cardiovascular stenting treatment within 6 months of the beginning of the study; had suffered the initial attack of a cerebrovascular event within 3 months of the beginning of the study; had uncontrollable disease, such as liver or renal dysfunction and thrombocytopenia; were receiving corticosteroid; were pregnant or lactating; were alcoholic; had a history of hypersensitivity to the medicine in this clinical trial; or had already participated in other clinical trials. Although NSAID users were also excluded, the use of NSAIDs for a period of less than seven continuous days in the investigation period was permitted. Enrolled patients were allowed to continue to take other antiplatelet agents and anticoagulants without changing the dose. Enrollment for this study was carried out on the website of the data center (TRI) within 2 weeks after baseline endoscopy.

Randomization

Eligible patients were randomly assigned on the website of the data center (TRI) at a ratio of 1:1:1 to receive one 20 mg tablet of rabeprazole (Eisai, Tokyo, Japan) once daily, one 10 mg tablet of rabeprazole once daily, or a gefarnate (Dainippon-Sumitomo, Osaka, Japan) 50 mg capsule twice daily for 12 weeks.

Procedures

The participants’ clinical baseline characteristics, including gender, age, height, weight, concomitant use of anticoagulant, underlying diseases (cardiovascular disease, cerebrovascular disease, hypertension, diabetes mellitus, hyperlipidemia, articular rheumatism), smoking, alcohol consumption, and status of Helicobacter pylori infection, were investigated upon enrollment in this study. Helicobacter pylori infection was diagnosed by 13C-urea breath test. Endoscopic assessments were carried out at baseline and at 12 weeks after randomization, and clinical assessments, including subjective symptoms, blood tests, and questionnaire on health-related quality of life, were carried out at baseline and at 4, 8, and 12 weeks after randomization. The upper gastrointestinal symptoms scale (GSRS), the frequency scale for the symptoms of gastroesophageal reflux disease (FSSG) [12], and the Medical Outcomes Study Short Form 8 questionnaire (SF-8) were used to assess the health-related quality of life. Endoscopic and clinical assessments were also performed whenever judged clinically necessary by the investigator.

The primary endpoint was the recurrence of gastric and/or duodenal ulcers confirmed by endoscopy during the 12-week treatment period. An ulcer was defined as a mucosal defect measuring 3 mm or more. The Endoscopic Evaluation Committee organized by three experienced endoscopists evaluated the presence or absence of gastroduodenal ulcers at baseline and 12 weeks by reviewing the endoscopic photographs of all enrolled patients in a blinded manner.

The secondary endpoints were: (1) the modified Lanza score (MLS) for gastroduodenal mucosal damage; (2) esophageal mucosal lesions described according to the Los Angeles (LA) classification; (3) treatment discontinuation due to lack of efficacy; (4) gastrointestinal symptoms; and (5) the GSRS, FSSG, and SF-8 scores. The MLS and the LA classification were assessed by each investigator who performed endoscopy.

Safety during the 12 weeks was assessed using adverse events. Adverse events were defined as all events including gastrointestinal symptoms that occurred newly during the investigation period. When a serious adverse event occurred, the Effect and Safety Assessment Committee organized by three experts evaluated that event, and the Steering Committee organized by five experts, including the principal investigator (T.A.), decided on the advisability of study continuation if needed.

Statistical analysis

Efficacy analyses were performed for the full analysis set (FAS) and the per protocol set (PPS). The FAS included all randomized patients who received at least one dose of study medication, except for the patients who did not meet the inclusion criteria. The patients who did not have the final endoscopic examination were assumed to have had normal findings. The PPS included all those in the FAS who took more than 75 % of the study medication, did not take NSAIDs (the use of NSAIDs for less than seven continuous days in the investigation period was permitted), and who underwent final endoscopy within the permitted period. The safety analysis population comprised all patients who took at least one dose of study medication.

The Kaplan–Meier method was used to estimate the cumulative incidence of the primary endpoint, and the log-rank test was used to assess the differences between the rabeprazole treatment group and the gefarnate treatment group, and between the rabeprazole 10 mg group and the rabeprazole 20 mg group. Differences in ulcer sizes between the treatment groups were compared using Mann–Whitney’s U-test. Differences in the baseline characteristics of the patients between the rabeprazole treatment group and the gefarnate treatment group were compared using the χ2 test and Student’s t-test. Differences between the treatment groups in the frequencies of patients with an MLS of ≥1 were analyzed using the χ2 test. The Wilcoxon rank-sum test, stratified by the baseline LA classification, was used for the analysis of differences between treatment groups in the incidence of erosive esophagitis. Differences between the treatment groups in the frequencies of the resolution and occurrence of gastrointestinal symptoms at 12 weeks were analyzed using the χ2 test. To investigate the factors associated with the development of new gastroduodenal ulcers, univariate and multivariate Cox regression analyses were performed. These factors included treatment group, sex, age, obesity, underlying disease, smoking habit, alcohol consumption habit, Helicobacter pylori infection, and the SF-8 score at baseline.

We assumed that the 12-week cumulative recurrence rate of gastroduodenal ulcers in patients receiving LDA for cardiovascular or cerebrovascular disease was 15 %. If the risk reduction effects of rabeprazole at 10 mg, rabeprazole at 20 mg, and gefarnate at 100 mg were assumed to be 80, 85, and 10 %, the recurrence rates of gastroduodenal ulcers were estimated to be 3, 2, and 13 %, respectively. Although the patients were assigned to three groups, the principal objective was a comparison between the rabeprazole treatment group and the gefarnate treatment group, and the comparison between the rabeprazole 10 mg group and the rabeprazole 20 mg group was a secondary objective. Therefore, the difference between the rabeprazole treatment group and the gefarnate treatment group was defined as significant at p < 0.025, and the difference between the rabeprazole 10 mg group and the rabeprazole 20 mg group was defined as significant at p < 0.05. A sample size of 222 patients (74 per group) was required to provide 80 % power using a log-rank test with a two-sided alpha of 2.5 %. We calculated that a sample size of 285 patients would be needed, if the drop-out rate was 20 %. p < 0.05 was defined as significant in the secondary endpoints. Analysis was performed using SAS software (ver. 9.1) (SAS Institute, Cary, NC, USA) and JMP 8 (SAS Institute, Cary, NC, USA).

Results

Trial flow

Figure 1 shows the trial profile. All 280 patients enrolled were randomized to receive 10 mg of rabeprazole (95 patients), 20 mg of rabeprazole (94 patients), and 100 mg of gefarnate (91 patients). Nineteen of these patients were excluded from the analysis because 11 had gastroduodenal ulcers at baseline endoscopy, 6 did not take the study medication, and 2 did not meet the inclusion criteria. Thus, the FAS comprised 261 patients (10 mg of rabeprazole, 87 patients; 20 mg of rabeprazole, 89 patients; 100 mg of gefarnate, 85 patients). The PPS comprised 223 patients (10 mg of rabeprazole, 75 patients; 20 mg of rabeprazole, 77 patients; 100 mg of gefarnate, 71 patients). Common reasons for exclusion from the FAS included discontinuation due to adverse events (12 patients), withdrawal of consent (10 patients), and deviation from the permission period for endoscopy (7 patients).
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0588-x/MediaObjects/535_2012_588_Fig1_HTML.gif
Fig. 1

Trial profile. FAS full analysis set, PPS per protocol set

Baseline characteristics

Table 1 shows the baseline characteristics of the FAS. The median age was 73.0 years and 185 (70.9 %) patients were men. As for the underlying disease that required aspirin therapy, 149 (57.1 %) patients had cardiovascular disease and 119 (45.6 %) patients had cerebrovascular disease. The status of Helicobacter pylori infection was negative in 119 patients (including 7 patients who had received eradication therapy for Helicobacter pylori infection 1 month or more before enrollment in this study). There was no significant difference in the baseline characteristics between the rabeprazole treatment group and the gefarnate treatment group.
Table 1

Baseline characteristics of study participants (FAS population)

 

Rabeprazole 10 mg (n = 87)

Rabeprazole 20 mg (n = 89)

Rabeprazole group (n = 176)

Gefarnate 100 mg (n = 85)

Age, years (range)

72 (54–95)

75 (43–87)

73.5 (43–95)

73 (48–86)

 <65

18 (20.7 %)

19 (21.3 %)

37 (21.0 %)

16 (18.8 %)

 ≥65

69 (79.3 %)

70 (78.7 %)

139 (79.0 %)

69 (81.2 %)

Male sex

63 (72.4 %)

59 (66.3 %)

122 (69.3 %)

63 (74.1 %)

Obesity (body mass index ≥25)

 Positive

22 (25.3 %)

19 (21.3 %)

41 (23.3 %)

27 (31.8 %)

 Negative

59 (67.8 %)

67 (75.3 %)

126 (71.5 %)

53 (62.4 %)

 Not determined

6 (6.9 %)

3 (3.4 %)

9 (5.1 %)

5 (5.9 %)

Type of aspirin

 Aspirin 100 mg

84 (96.6 %)

88 (98.9 %)

172 (97.7 %)

85 (100 %)

 Aspirin dialminate 81 mg

3 (3.4 %)

1 (1.1 %)

4 (2.3 %)

0 (0 %)

Current smoker

13 (14.9 %)

18 (20.2 %)

31 (17.6 %)

12 (14.1 %)

Alcohol consumption

34 (39.1 %)

24 (39.1 %)

58 (33.0 %)

35 (41.2 %)

Underlying disease

 Cardiovascular disease

47 (54.0 %)

50 (56.2 %)

97 (55.1 %)

52 (61.2 %)

 Cerebrovascular disease

41 (47.1 %)

41 (46.1 %)

82 (46.6 %)

37 (43.5 %)

 Hypertension

57 (65.5 %)

69 (75.3 %)

126 (71.6 %)

64 (75.3 %)

 Diabetes mellitus

32 (36.8 %)

30 (33.7 %)

62 (35.2 %)

35 (41.2 %)

 Hyperlipidemia

46 (52.9 %)

41 (46.1 %)

87 (49.4 %)

39 (45.9 %)

 Articular rheumatism

4 (4.6 %)

2 (2.2 %)

6 (3.4 %)

2 (2.4 %)

Concomitant anticoagulant

8 (9.2 %)

11 (12.4 %)

19 (10.8 %)

4 (4.7 %)

Helicobacter pylori statusa

 Positive

34 (39.1 %)

36 (40.4 %)

70 (39.8 %)

40 (47.1 %)

 Negative

41 (47.1 %)

40 (44.9 %)

81 (46.0 %)

38 (44.7 %)

 Not determined

12 (13.7 %)

13 (14.6 %)

25 (14.2 %)

7 (8.2 %)

Gastrointestinal symptoms

28 (32.2 %)

28 (31.5 %)

56 (31.8 %)

32 (37.6 %)

 Heartburn

14

12

26

17

 Epigastric pain

1

6

7

10

 Bloating

10

12

22

14

 Nausea

4

5

9

8

 Others

11

6

17

5

Peptic ulcer scarb

 Stomach

64 (73.6 %)

67 (75.3 %)

131 (74.4 %)

65 (76.5 %)

 Duodenum

26 (29.9 %)

25 (28.1 %)

51 (29.0 %)

33 (38.8 %)

 Unconfirmed

4 (4.6 %)

4 (4.5 %)

8 (4.5 %)

2 (2.4 %)

Gastroduodenal mucosal lesions

 Gastric lesions (MLS ≥1)

51 (58.6 %)

45 (50.6 %)

96 (54.5 %)

48 (56.5 %)

 Duodenal lesions (MLS ≥1)

16 (18.4 %)

7 (7.9 %)

23 (13.1 %)

13 (15.3 %)

FAS full analysis set, MLS modified Lanza score

aApproximately 10 % of participants were diagnosed by serologic test, rapid urease test, bacterial culture, or histology

bUnconfirmed scar was defined as a scar not confirmed by the Endoscope Evaluation Committee

The recurrence rates of gastric and/or duodenal ulcers

The recurrence rates of gastric and/or duodenal ulcers are shown in Table 2. The cumulative rates of ulcer recurrence by 12 weeks were estimated as 7.4 % (6 cases) in the rabeprazole 10 mg group, 3.7 % (3 cases) in the rabeprazole 20 mg group, and 26.7 % (20 cases) in the gefarnate 100 mg group. A significant difference between the rabeprazole treatment group and the gefarnate treatment group was observed [5.5 vs. 26.7 %; hazard ratio (95 % confidence interval [CI]) 0.179 (0.082–0.394); p < 0.0001]. Similar results were obtained from the PPS analysis [6.0 vs. 28.5 %; hazard ratio (95 % CI) 0.200 (0.081–0.495); p < 0.0005]. Although the cumulative rate of ulcer recurrence in the rabeprazole 20 mg group was lower than that in the rabeprazole 10 mg group, there was no significant difference between the two rabeprazole groups. Gastric ulcers occurred more frequently than duodenal ulcers, and the hazard ratios for gastric ulcer and duodenal ulcer were estimated as 0.197 and 0.147, respectively, in the rabeprazole treatment group compared with the gefarnate treatment group. Ulcer size according to location and treatment is shown in Table 3. The size of ulcer in the rabeprazole treatment group was significantly smaller than that in the gefarnate treatment group [median ulcer size (range) 3 mm (3–10 mm) vs. 5 mm (3–30 mm), p = 0.0165]. Seven (35 %) of 20 patients in the gefarnate treatment group had gastroduodenal ulcers of 10 mm or greater in size, whereas only one of nine patients in the rabeprazole treatment group (11.1 %) had gastroduodenal ulcers of 10 mm or greater in size. In addition, there were no patients with an ulcer of 5 mm or greater in size in the rabeprazole 20 mg group. In regard to the sites of the recurrent ulcers, 18 (90 %) of 20 gastric ulcers and 8 (89 %) of 9 duodenal ulcers developed in sites similar to those of the scars observed at baseline endoscopy.
Table 2

Cumulative incidence of gastroduodenal ulcers in the study period (FAS population)

 

Rabeprazole 10 mg (n = 87)

Rabeprazole 20 mg (n = 89)

Rabeprazole group (n = 176)

Gefarnate 100 mg (n = 85)

Hazard ratio (95 % CI)

p value

Gastric or duodenal ulcer

6 (7.4 %)

3 (3.7 %)

9 (5.5 %)

20 (26.7 %)

0.179 (0.082–0.394)

<0.0001

Gastric ulcer

6 (7.4 %)

1 (1.2 %)

7 (4.3 %)

14 (20.0 %)

0.197 (0.080–0.489)

0.0005

Duodenal ulcer

0 (0 %)

2 (2.4 %)

2 (1.2 %)

6 (7.8 %)

0.147 (0.030–0.727)

0.0187

CI confidence interval

p values: comparison between the rabeprazole group and the gefarnate group

Table 3

Numbers of patients with either gastric ulcer or duodenal ulcer according to ulcer size and treatment

Location/size (mm) of largest ulcer

Rabeprazole 10 mg (n = 6)

Rabeprazole 20 mg (n = 1)

Rabeprazole group (n = 7)

Gefarnate 100 mg (n = 14)

Gastric ulcer

 3–4

4

1

5

5

 5–9

1

0

1

3

 ≥10

1

0

1

6

Location/size (mm) of largest ulcer

Rabeprazole 10 mg (n = 0)

Rabeprazole 20 mg (n = 2)

Rabeprazole group (n = 2)

Gefarnate 100 mg (n = 6)

Duodenal ulcer

 3–4

0

2

2

1

 5–9

0

0

0

4

 ≥10

0

0

0

1

Gastroduodenal mucosal lesions

The frequency of patients with gastroduodenal mucosal lesions (MLS ≥1) in each treatment group is shown in Table 4. Consistent with the results of gastroduodenal ulcer recurrence, the LDA-induced gastroduodenal mucosal lesions at 12 weeks were observed significantly less frequently in the rabeprazole treatment group than in the gefarnate treatment group (gastric lesion 33.5 vs. 62.4 %, p < 0.0001; duodenal lesion 5.7 vs. 24.7 %, p < 0.0001; FAS population). Although the rabeprazole 20 mg group had a lower incidence of LDA-induced gastroduodenal mucosal lesions at 12 weeks than the rabeprazole 10 mg group (gastric lesion 27.0 vs. 40.2 %; duodenal lesion 3.4 vs. 8.0 %), there was no significant difference between the two dosage groups.
Table 4

Proportions of patients with gastric or duodenal mucosal damage evaluated by the modified Lanza score at 12 weeks (FAS population)

 

Rabeprazole 10 mg (n = 87)

Rabeprazole 20 mg (n = 89)

Rabeprazole group (n = 176)

Gefarnate 100 mg (n = 85)

p value

Gastric mucosal lesions (modified Lanza score ≥1)

35 (40.2 %)

24 (27.0 %)

59 (33.5 %)

53 (62.4 %)

<0.0001

Duodenal mucosal lesions (modified Lanza score ≥1)

7 (8.0 %)

3 (3.4 %)

10 (5.7 %)

21 (24.7 %)

<0.0001

p values: comparison between the rabeprazole group and the gefarnate group

Erosive esophagitis

Figure 2 shows the prevalence of erosive esophagitis described according to the LA classification in both the rabeprazole treatment group and the gefarnate treatment group at baseline and at 12 weeks. The proportion of patients having erosive esophagitis at 12 weeks was significantly lower in the rabeprazole treatment group than that in the gefarnate treatment group (5.8 vs. 19.4 %, p < 0.0001).
https://static-content.springer.com/image/art%3A10.1007%2Fs00535-012-0588-x/MediaObjects/535_2012_588_Fig2_HTML.gif
Fig. 2

The proportions of patients with erosive esophagitis at baseline and at 12 weeks. The grade of erosive esophagitis is described according to the Los Angeles (LA) classification. The Wilcoxon rank-sum test, stratified by the baseline LA classification, was used for the analysis of differences between treatment groups in the incidence of erosive esophagitis. The data at baseline are shown as the frequency in the FAS population, and the data at 12 weeks are shown as the frequency in the patients who underwent endoscopy at 12 weeks

Gastrointestinal symptoms and quality of life

Table 5 shows the proportions of patients with resolution and occurrence of gastrointestinal symptoms at 12 weeks; there were significant differences between the rabeprazole treatment group and the gefarnate treatment group in both the resolution of gastrointestinal symptoms (53.6 vs. 25.0 %, p = 0.017) and the occurrence of gastrointestinal symptoms (9.2 vs. 28.3 %, p = 0.0026). Rabeprazole was significantly more effective than gefarnate for the resolution of heartburn (80.8 vs. 43.8 %, p = 0.02) among patients with the symptom at baseline, and it was also significantly more effective than gefarnate for preventing the development of heartburn (3.3 vs. 14.7 %, p = 0.0054) among patients without the symptom at baseline. As for health-related quality of life assessed using the GSRS, FSSG, and SF-8, there were no significant differences between the two treatment groups at 12 weeks.
Table 5

Proportions of patients with resolution and occurrence of gastrointestinal symptoms at 12 weeks

 

Rabeprazole 10 mg group

n/N (%)

Rabeprazole 20 mg group

n/N (%)

Rabeprazole group

n/N (%)

Gefarnate 100 mg

n/N (%)

p value

Resolution of gastrointestinal symptoms

14/28 (50.0)

17/28 (60.7)

31/56 (55.4)

8/32 (25.0)

0.011

 Heartburn

10/14 (71.4)

11/12 (91.7)

21/26 (80.8)

7/16 (43.8)

0.02

 Epigastric pain

1/1 (100)

4/6 (66.7)

5/7 (71.4)

8/10 (80.0)

1

 Bloating

7/10 (70.0)

7/12 (58.3)

14/22 (63.6)

5/14 (35.7)

0.17

 Nausea

4/4 (100)

5/5 (100)

9/9 (100)

6/8 (75.0)

0.21

 Others

6/11 (54.5)

4/6 (66.7)

10/17 (58.8)

2/5 (40.0)

0.18

Occurrence of gastrointestinal symptoms

4/59 (6.8)

7/61 (11.5)

11/120 (9.2)

15/53 (28.3)

0.0026

 Heartburn

2/73 (2.7)

3/77 (3.9)

5/150 (3.3)

10/68 (14.7)

0.0054

 Epigastric pain

1/86 (1.2)

1/83 (1.2)

2/169 (1.2)

5/75 (6.7)

0.051

 Bloating

5/77 (6.5)

1/77 (1.3)

6/154 (3.9)

5/71 (7.0)

0.49

 Nausea

1/83 (1.2)

2/84 (2.4)

3/167 (1.8)

6/77 (7.8)

0.052

 Others

2/76 (2.6)

6/83 (7.2)

8/159 (5.0)

5/80 (6.3)

0.93

Resolution N: The number of patients who had the symptom at baseline

Resolution n: The number of patients who experienced resolution of the symptom at 12 weeks

Occurrence N: The number of patients who did not have the symptom at baseline

Occurrence n: The number of patients who developed the symptom at 12 weeks

Univariate and multivariate Cox regression analyses

To investigate the effects of baseline factors on the risk of development of new gastroduodenal ulcers, univariate and multivariate Cox regression analyses were performed. Table 6 shows the results of the univariate Cox regression analysis. No significant effect caused by Helicobacter pylori infection was observed [hazard ratio (95 % CI) 0.860 (0.390–1.894), p = 0.708]. Table 7 shows the results of the multivariate Cox regression analysis using the variables for which p values in the univariate Cox regression analysis were <0.2. Only the treatment group was a significant factor in the multivariate Cox regression analysis. The hazard ratio was estimated as 0.173 (95 % CI 0.075–0.401, p < 0.0001) in the rabeprazole treatment group compared with the gefarnate treatment group after adjustment for the baseline factors.
Table 6

Results of univariate Cox regression analysis using baseline variables

Baseline characteristics

Hazard ratio

95 % CI

p value

Treatment group (rabeprazole/gefarnate)

0.179

0.082–0.394

<0.0001

Gender (male/female)

0.960

0.437–2.107

0.9183

Age (10-year increment)

1.306

0.832–2.051

0.2456

Obesity (yes/no)

1.611

0.738–3.519

0.2314

Concomitant anticoagulant (yes/no)

0.352

0.048–2.582

0.3042

Cardiovascular disease (yes/no)

1.651

0.752–3.626

0.2115

Cerebrovascular disease (yes/no)

0.741

0.350–1.570

0.4344

Hypertension (yes/no)

0.552

0.263–1.155

0.1148

Diabetes mellitus (yes/no)

1.682

0.812–3.485

0.1618

Hyperlipidemia (yes/no)

0.823

0.396–1.712

0.6027

Current smoking (yes/no)

1.157

0.556–2.405

0.6966

Alcohol consumption (yes/no)

1.741

0.822–3.687

0.1473

Helicobacter pylori status (positive/negative)

0.860

0.390–1.894

0.7080

SF-8 score

 Physical functioning (10-point increment)

0.666

0.458–0.969

0.0335

 Role physical (10-point increment)

0.996

0.601–1.650

0.9876

 Bodily pain (10-point increment)

0.711

0.483–1.048

0.0850

 General health (10-point increment)

1.409

0.771–2.577

0.2654

 Vitality (10-point increment)

1.027

0.593–1.778

0.9244

 Social functioning (10-point increment)

0.828

0.535–1.282

0.3980

 Role emotional (10-point increment)

0.910

0.583–1.423

0.6804

 Mental health (10-point increment)

0.763

0.450–1.292

0.3135

CI confidence interval, SF-8 Medical Outcomes Study Short Form 8 questionnaire

Table 7

Results of multivariate Cox regression analysis using baseline variables

Baseline characteristics

Hazard Ratio

95 % CI

p value

Treatment group (rabeprazole/gefarnate)

0.173

0.075–0.401

<0.0001

Hypertension (yes/no)

0.535

0.237–1.208

0.1322

Diabetes mellitus (yes/no)

1.143

0.518–2.523

0.7411

Alcohol consumption (yes/no)

1.962

0.812–4.741

0.1343

SF-8 score (physical functioning) (10-point increment)

0.734

0.459–1.175

0.1978

SF-8 score (bodily pain) (10-point increment)

0.816

0.504–1.320

0.4077

CI confidence interval

Adverse events

Table 8 shows the details of adverse events reported during the investigation period. In the safety analysis population, 27.0 % (74/274) of the patients had adverse events. Most adverse events were related to gastrointestinal symptoms, and the incidence of adverse events in the rabeprazole treatment group was significantly lower than that in the gefarnate group (22.3 vs. 36.7 %, p = 0.018). Serious adverse events were reported for 10 patients, and three events had a fatal outcome (one cerebral infarction in the rabeprazole 20 mg group, one suicide in the rabeprazole 10 mg group, and one brainstem hemorrhage in the gefarnate 100 mg group). As for serious adverse events associated with gastrointestinal bleeding, one patient in the gefarnate 100 mg group experienced hemorrhagic duodenal ulcer and one patient in the rabeprazole 20 mg group had severe anemia. None of these serious adverse events was considered to be causally related to the study drug. In general, rabeprazole was well tolerated in combination with LDA.
Table 8

Frequency of adverse events (safety analysis population)

 

Rabeprazole 10 mg (n = 92)

Rabeprazole 20 mg (n = 92)

Rabeprazole group (n = 184)

Gefarnate (n = 90)

p value

All adverse events

20 (21.7 %)

21 (22.8 %)

41 (22.3 %)

33 (36.7 %)

0.018

Serious adverse events

4 (4.4 %)

3 (3.3 %)

7 (3.8 %)

3 (3.3 %)

0.88

 Duodenal ulcer hemorrhage

0 (0.0 %)

0 (0.0 %)

0 (0.0 %)

1 (1.1 %)

 

 Cerebral infarction

2 (2.2 %)

0 (0.0 %)

2 (1.1 %)

0 (0.0 %)

 

 Adhesion ileus

1 (1.1 %)

0 (0.0 %)

1 (0.5 %)

0 (0.0 %)

 

 Colonic diverticulitis

0 (0.0 %)

0 (0.0 %)

0 (0.0 %)

1 (1.1 %)

 

 Suicide

0 (0.0 %)

1 (1.1 %)

1 (0.5 %)

0 (0.0 %)

 

 Renal failure

0 (0.0 %)

1 (1.1 %)

1 (0.5 %)

0 (0.0 %)

 

 Brainstem hemorrhage

0 (0.0 %)

0 (0.0 %)

0 (0.0 %)

1 (1.1 %)

 

 Anemia

1 (1.1 %)

0 (0.0 %)

1 (0.5 %)

0 (0.0 %)

 

 Chronic subdural hematoma

0 (0.0 %)

1 (1.1 %)

1 (0.5 %)

0 (0.0 %)

 

  Death

1 (1.1 %)

1 (1.1 %)

2 (1.1 %)

1 (1.1 %)

 

Adverse events related to gastrointestinal symptoms

13 (14.1 %)

17 (18.5 %)

30 (16.3 %)

26 (28.9 %)

0.023

Adverse events except for serious events or gastrointestinal symptoms

5 (5.4 %)

3 (3.3 %)

8 (4.3 %)

5 (5.6 %)

0.89

p values: comparison between the rabeprazole group and the gefarnate group

Discussion

This study proved that rabeprazole reduced the risk of recurrence of gastroduodenal ulcer more efficiently than gefarnate in Japanese patients who were receiving LDA for cardiovascular or cerebrovascular disease. We also demonstrated that rabeprazole prevented or resolved erosive esophagitis and gastrointestinal symptoms more effectively than gefarnate in these patients. Although there are several reported studies that have shown the efficacy of PPIs for preventing LDA-induced peptic ulcer, rabeprazole has not yet been evaluated, and this is the first report on the effects of rabeprazole on LDA-induced gastroduodenal and esophageal lesions and gastrointestinal symptoms.

In the plan of this study, we assumed that the 12-week cumulative recurrence rate of gastroduodenal ulcers in patients receiving LDA and 100 mg of gefarnate would be 13 %. However, 26.7 % of the patients in the gefarnate treatment group had experienced the recurrence of gastroduodenal ulcers at 12 weeks in this study. This recurrence rate is higher than the rate in the report of a recent study (15.2 % on day 91 in the gefarnate group), which was performed by Sugano et al. [13] in LDA users with a history of peptic ulcer to compare the difference between 15 mg of lansoprazole and 100 mg of gefarnate. We consider that differences in patient backgrounds, such as age, sex, and Helicobacter pylori infection, were involved in the difference of the recurrence rate between the present study and Sugano’s study. On the other hand, Laine et al. [14] reported that the 12-week peptic ulcer rate was 26.7 % in aspirin users with osteoarthritis who had a previous history of ulcers or their complications, which is consistent with our results. In addition, in a trial of famotidine for the prevention of peptic ulcers and esophagitis in LDA users, which was recently reported by Taha et al. [8], the development of peptic ulcers at 12 weeks was observed in 38 (19.0 %) of 200 placebo group patients, of whom 47 % had scars of peptic ulcers at baseline. Considering the background of the patients, the recurrence rate of peptic ulcers in Taha’s study seems to be equivalent to the results in our study. As for the influence of interobserver variation in reporting the endoscopic findings, we consider that this influence was minimal in our study, because the Endoscopic Evaluation Committee organized by three experienced endoscopists reviewed the endoscopic photographs of all enrolled patients in a blinded manner.

In the present study, the risk reduction rate of rabeprazole for ulcer recurrence compared with gefarnate was found to be 82.1 %, which was similar to the results of previous studies that demonstrated that the risk reduction rate of PPIs was between 70 and 90 % [6, 7, 13].

The size of a recurrent ulcer is a clinically important issue because larger ulcers have a greater need for treatment. In our rabeprazole treatment group, the recurrent ulcers were significantly smaller than those in the gefarnate treatment group. In addition, rabeprazole reduced the risk of LDA-induced gastroduodenal erosion more efficiently than gefarnate.

The relationship between Helicobacter pylori infection and LDA-induced gastroduodenal lesions is a controversial issue. In observational studies, Helicobacter pylori infection has been reported to be associated with an increased risk of LDA-induced gastroduodenal damage [15, 16]. In addition, Sugano et al. [13] reported that Helicobacter pylori infection was a significant risk factor for the recurrence of gastroduodenal ulcers in LDA users. On the other hand, Hart et al. [17] recently reported that Helicobacter pylori infection was a preventive factor for LDA-induced gastric lesions. In our study, Helicobacter pylori infection had no significant effect on the development of new gastroduodenal ulcers, which is consistent with the results of the study performed by Taha et al. [8]. In the development of LDA-induced gastric lesions, gastric acid has been reported to play an important role [1820]. The effects of Helicobacter pylori infection on LDA-induced gastroduodenal lesions might depend on the intragastric acidity, which is influenced by Helicobacter pylori infection through the development of gastric mucosal inflammation and/or the progression of gastric mucosal atrophy.

Erosive esophagitis associated with LDA was observed in 19.4 % of our patients in the gefarnate treatment group at 12 weeks, and rabeprazole significantly reduced the proportion of patients with erosive esophagitis. In addition, the proportion of patients with erosive esophagitis in the gefarnate treatment group increased slightly, from 15.3 at baseline to 19.4 % at 12 weeks. In our study, the incidence of erosive esophagitis at 12 weeks in the patients without erosive esophagitis at baseline was 8.3 % (6/72) in the gefarnate treatment group and 2.0 % (3/150) in the rabeprazole treatment group (p = 0.0606). These results suggest that erosive esophagitis is induced by LDA, and that rabeprazole reduces the development of LDA-induced erosive esophagitis. Consistent with our study, Taha et al. [8] reported that the incidence of LDA-induced erosive esophagitis at 12 weeks was 19.0 % in a placebo group that did not have erosive esophagitis at baseline. Prevention or treatment of erosive esophagitis is important because it is considered as one of the causes of gastrointestinal bleeding in LDA users.

In our study population, approximately one-third of LDA users had gastrointestinal symptoms, and heartburn and bloating were common. Such symptoms are likely to lead to discontinuation of or poor compliance with LDA therapy, resulting in an increased risk of recurrence of cardiovascular events [21]. Therefore, the efficacy of rabeprazole in both the resolution and prevention of gastrointestinal symptoms is important for patients who need to continue LDA therapy.

In the present study, the health-related quality of life of the patients was assessed using the GSRS, FSSG, and SF-8. Although rabeprazole prevented or resolved gastrointestinal symptoms more effectively than gefarnate, there were no significant differences in quality of life scores between the two treatment groups at 12 weeks. LDA-induced gastrointestinal mucosal injuries are not often associated with gastrointestinal symptoms, whereas gastrointestinal symptoms often develop without mucosal injury in LDA users. Generally, the correlation of gastrointestinal symptoms and mucosal injury is not good in LDA users. Therefore, it seems that efficacy of rabeprazole for health-related quality of life was not found in the present study. In previous studies, it was reported that the health-related quality of life assessed by the SF-36 or GSRS in LDA users was not improved by PPI treatment [22, 23], which is consistent with the results of our study.

Regarding the dose of rabeprazole, there was no significant difference in the cumulative rate of ulcer recurrence between the rabeprazole 20 mg group and the rabeprazole 10 mg group in the present study. However, the cumulative rate of ulcer recurrence in the rabeprazole 20 mg group (3.7 %) was lower than that in the rabeprazole 10 mg group (7.4 %), and LDA-induced gastroduodenal mucosal lesions at 12 weeks were less frequently observed in the rabeprazole 20 mg group than in the rabeprazole 10 mg group (gastric lesions 27.0 vs. 40.2 %, duodenal lesions 3.4 vs. 8.0 %). In addition, there were no patients with an ulcer of 5 mm or more in the rabeprazole 20 mg group, whereas two of six patients in the rabeprazole 10 mg group had ulcers of 5 mm or more. To determine whether there is a difference in efficacy for preventing the recurrence of gastroduodenal ulcers between 10 mg of rabeprazole and 20 mg of rabeprazole, a prospective controlled study of more patients is required.

This study has some limitations. First, this was an open-label and active-controlled study. Therefore, there might have been a bias for the evaluation of subjective symptoms not only by the patients but also by the investigators. Second, the investigation period of this study was relatively short. The degree of mucosal injury might change according to the investigation period.

In conclusion, rabeprazole was more effective than gefarnate in reducing the recurrence risk of peptic ulcer associated with LDA in Japanese patients with cardiovascular or cerebrovascular disease.

Acknowledgments

This study was funded by a grant provided by the NPO-Corporation, Gastrointestinal Medical Care Research and Education Center. We are very grateful to Mr. Tagaji Ikami and Miss Yuko Osawa for their dedicated secretarial assistance. We also thank Mr. Koichi Yamashiro, Ms. Kyoko Murata, Ms. Kikuko Oshima, and Ms. Tomomi Sakabayashi for their assistance with the web system and statistical analysis. We thank Mr. Hiroshi Furuichi for his assistance in the study.

Conflict of interest

The authors declare that they have no conflict of interest.

Copyright information

© Springer 2012