Full Text 1

Aliment Pharmacol Ther 2003; 17: 1125–1135.

doi: 10.1046/j.0269-2813.2003.01560.x

Helicobacter pylori eradication and gastric ulcer healing — comparison of three pantoprazole-based triple therapies P. MA LFERTHEINER *, T. KIR CHNER , M. KISTà, A. LEOD OLTER *, U. PEITZ*, S. STR OBELà, M. BOH USCHK E§, G. GA TZ§ & THE BY K AD VANCED GASTR IC ULCER STUDY (BAGUS) G ROUP *Otto-von-Guericke-University, Magdeburg, Germany;  University of Erlangen-Nu¨rnberg, Germany; àDepartment of Medical Microbiology and Hygiene, University Hospital of Freiburg, Germany; §ALTANA Pharma AG, Konstanz, Germany Accepted for publication 26 February 2003

SUMMARY

Aim: To study the efficacy of three pantoprazole-based triple therapy regimens for the eradication of Helicobacter pylori infection and gastric ulcer healing. Methods: In an open, multi-centre, randomized study, 519 H. pylori-positive patients with active gastric ulcer were randomized to receive pantoprazole (40 mg) (P) and two of three antibiotics: clarithromycin (500 mg) (C), metronidazole (500 mg) (M) or amoxicillin (1000 mg) (A). Triple therapy (PAC, PCM, PAM) was administered twice daily for 7 days, followed by pantoprazole until the ulcer had healed. Antrum and corpus biopsies were taken to determine the pattern of gastritis, to assess the H. pylori status and to determine the strain susceptibility to antibiotics, and from the ulcer margins and base to exclude malignancy. Scores based on the Sydney system were used to categorize the gastritis phenotypically.

INTRODUCTION

Helicobacter pylori infection is the pathogenic key to the development of the majority of peptic ulcers, supported by strong biological plausibility. Several bacterial factors and inflammatory mechanisms, resulting from the infection, damage the gastric mucosal integrity.1 The

Correspondence to: Professor P. Malfertheiner, Department of Gastroenterology, Hepatology and Infectious Diseases, Otto-von-Guericke-University Magdeburg, Leipziger Strasse 44, D-39120 Magdeburg, Germany. E-mail: [email protected]
2003 Blackwell Publishing Ltd

Results: The H. pylori eradication rates for the per protocol (intention-to-treat) analysis were 89% (67%) for PAC, 83% (68%) for PCM and 76% (60%) for PAM, with a significant difference between PAC and PAM. Healing rates after 4 weeks were 91% for PAM, 90% for PCM and 88% for PAC (per protocol analysis). The eradication rates were lower in patients in whom strains resistant to any antibiotic used in the triple therapies were detected. Successful eradication [odds ratio, 5.2 (3.3; 8.3)] and the ulcer size (< 15 mm) were significant predictors for healing after 4 weeks. The regimens showed a comparable safety profile and compliance. Conclusions: Pantoprazole-based triple therapies are effective in the eradication of H. pylori infection in gastric ulcer patients, as reported in previous similar sized studies in duodenal ulcer patients. Successful eradication and an ulcer size of < 15 mm are the best predictors of gastric ulcer healing after 4 weeks.

final proof of the role of H. pylori, however, was the therapeutic achievement of a definitive cure of peptic ulcers through eradication therapy. Relapse rates after successful eradication are almost nonexistent in gastric ulcer2, 3 and duodenal ulcer.4–6 If there are relapses, they are rarely due to re-infection and are normally caused by non-steroidal anti-inflammatory drugs (NSAIDs) or acetylsalicylic acid.7 Despite the similarities between gastric ulcer and duodenal ulcer, the role of H. pylori infection in gastric ulcer pathogenesis appears to be less well established than in duodenal ulcer. There are differences between 1125

1126

P. MALFERTHEINER et al.

gastric ulcer and duodenal ulcer, such as a lower prevalence of H. pylori infection in gastric ulcer patients of 70% vs. 90% in duodenal ulcer patients.7, 8 Gastric acid secretion is often reduced in gastric ulcer, whereas it is normal or increased in duodenal ulcer.9 The difference in acid secretion is related to the different phenotypic expression of gastritis in gastric ulcer vs. duodenal ulcer.10, 11 One of the most important differences, however, is the time required for healing of duodenal ulcer and gastric ulcer. The reasons for the delayed healing of gastric ulcer compared to duodenal ulcer are not clear, but an influence exerted by the different pattern and activity of gastritis in duodenal ulcer and gastric ulcer is a possible explanation.12, 13 One-week proton pump inhibitor-based triple therapy is accepted as the standard treatment for H. pylori infection in patients with peptic ulcer.14–17 With proper use, eradication rates of around 90% have been achieved. For patients with uncomplicated duodenal ulcer, the 7-day eradication treatment is sufficient for duodenal ulcer healing if the infection has been successfully eradicated. However, this has not been shown for patients with gastric ulcer, in whom the healing process is usually slower with less gastric ulcers healed after 4 weeks. Therefore, the continued administration of proton pump inhibitor until healing is complete is currently advised. The aim of this study was to investigate the efficacy of three proton pump inhibitor-based regimens [pantoprazole in combination with amoxicillin and metronidazole (PAM), clarithromycin and metronidazole (PCM) or amoxicillin and clarithromycin (PAC)] in the eradication of H. pylori infection, and to determine whether successful H. pylori eradication resulted in the better healing of gastric ulcer. In addition, we studied the efficacy of the eradication regimen as a function of primary resistance and, in the case of treatment failure, the development of resistance of H. pylori to metronidazole, amoxicillin and clarithromycin. MATERIALS AND METHODS

This open, multi-centre, randomized, parallel-group comparison between three treatment groups was conducted in 60 centres in Germany in accordance with Good Clinical Practice and the Declaration of Helsinki. Fifteen independent local ethics committees approved

the study and written informed consent was obtained from all patients. Patients Five hundred and nineteen patients of at least 18 years of age, with one or two endoscopically diagnosed active gastric ulcers with a diameter of 5–20 mm and with a positive rapid urease test, were randomized into the study. H. pylori infection was confirmed by at least one other positive test result (13C-urea breath test, histology, culture testing). Major exclusion criteria were: ulcer complications, duodenal and/or intrapyloric ulcers, a medical history of Zollinger–Ellison syndrome, reflux oesophagitis (grades II–IV), pyloric stenosis, subtotal gastrectomy or vagotomy. Patients with known allergy, especially to any study drug, malignant diseases, other severe concurrent diseases or malfunctions or a history of drug or alcohol abuse were also excluded. The intake of systemic glucocorticoids or NSAIDs on more than two consecutive days per week during the last 28 days before the start of the study was forbidden, except for a daily dose of up to 150 mg acetylsalicylic acid. Treatment H. pylori-infected patients were randomly assigned to one of the three treatment groups: PAC, PCM or PAM for 7 days, with a twice daily dose of pantoprazole (40 mg), amoxicillin (1000 mg), clarithromycin (500 mg) or metronidazole (500 mg), followed by pantoprazole (40 mg) for another 3 weeks. If the ulcer had not healed at the control endoscopy after 4 weeks, this antisecretory treatment was continued for another 4 weeks. Conduct of the study All patients were treated for a period of 28 days (or 56 days), followed by a 6-week treatment-free phase. For each patient, a maximum of five visits was scheduled, with two to three endoscopies carried out in order to determine the ulcer diameter, the grade of gastro-oesophageal reflux disease (GERD), H. pylori status and the histological pattern of gastritis and malignancy. Standard laboratory investigations were performed, with demographic data, medical history, gastrointestinal symptoms and concomitant medication also being documented. At baseline and after the


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

H. PYLORI ERADICATION AND GASTRIC ULCER HEALING

treatment-free phase, the H. pylori status was determined. Compliance and clinical symptoms were checked at each follow-up visit, and clinical symptoms, changes in concomitant medication and possible adverse events were documented. Endoscopy Endoscopies were performed at baseline and at every follow-up visit after starting monotherapy. The ulcer size and location were reported exactly. Additional abnormalities, such as reflux oesophagitis (graded according to Savary–Miller), were also carefully described. For the histological assessment of gastritis and the determination of H. pylori infection, a number of biopsies were carried out as follows: two antrum biopsies were taken 2–4 cm pre-pylorically at the greater and lesser curvature, two from the corpus at the lesser curvature, one approximately 4 cm proximal to the angulus, and one from the greater curvature approximately 8 cm from the cardia. For the rapid urease and culture tests, four biopsies were taken, two each from the antrum and corpus. Malignancy was determined from additional biopsies of the gastric ulcer. At the final examination, three biopsies each were taken from the antrum and corpus for the determination of the histological parameters and H. pylori status. Histology and determination of the H. pylori status The activity of gastritis and the density of H. pylori colonization were evaluated by scoring the parameters according to the updated Sydney classification on a four-point scale: none (0), mild (1), moderate (2), severe (3).18 The gastritis was classified using haematoxylin and eosin staining. The presence of H. pylori was determined using Warthin–Starry staining. The score for lymphoid follicles/aggregates was defined as 0 (none), 1 (aggregates), 2 (follicles), 3 (aggregates and follicles) and 4 (not investigated). The grade of intestinal metaplasia was categorized as none, incomplete or complete and analysed quantitatively. Culture and susceptibility testing of H. pylori isolates were performed at the Institut fu¨r Medizinische Mikrobiologie und Hygiene (Universita¨tsklinikum Freiburg, Germany). H. pylori was determined by selective cultivation, biochemical characteristics and microscopic examination after Gram staining, as described previously.19 Susceptibility testing to given antibiotics was performed using the

1127

E-test (AB Biodisk, Solna, Sweden), according to Glupczynski et al.20 modified after Heep et al.21 The 13C-urea breath test was performed according to a standard validated protocol and assayed at IMTM GmbH (Magdeburg, Germany).22 A patient was considered to be H. pylori positive if the D13C value exceeded 4&. For the evaluation of the gastrointestinal symptoms of ulcer pain (day and night), nausea, vomiting, retching, heartburn, acid eructation and dysphagia, a score was assigned (none, 0; mild, 1; moderate, 2; severe, 3) to quantify the severity of each symptom at each visit. Statistical methods The number of patients required for the study was calculated such that a difference of 7% in the eradication rates between the three study groups could be detected. Based on an expected mean eradication rate of 85%, a significance level of 5% and a power of 80%, at least 125 patients per group were required to detect this difference for the per protocol analysis. Assuming a drop-out rate of 25%, at least 500 patients were needed to achieve a total of at least 375 patients. The primary criterion of this study was the H. pylori eradication rate. The point estimate was determined according to the binomial distribution, and the 95% confidence interval (CI) was determined according to the approximation to the F-distribution, restricted to 0,100. For the comparison of eradication rates between the three groups, the chi-squared test was used, with a significance level of 5%. Ulcer healing rates after 4 and 8 weeks were evaluated in a similar manner to the eradication rate. Logistic regression models were used for the sub-group analyses of healing rates as a function of the intake of NSAIDs and ulcer diameter. In addition, for the healing rates, the odds ratios were calculated with regard to gender, age (older vs. younger than 57 years), GERD grade I at study entry (present vs. absent), ulcer size (5–10 mm vs. 10–20 mm) and eradication status (eradicated vs. not eradicated). Dichotomous parameters (gender, smoking behaviour and alcohol habits) were compared using Fisher’s exact two-tailed test. Ordinal variables (age, height, weight and body mass index) were compared using the Kruskal–Wallis test. The median and 68% range were calculated. Analyses of the results were performed in three populations. The safety population comprised all


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

1128

P. MALFERTHEINER et al.

screened patients with a positive rapid urease test result who took the study medication at least once. Patients for whom a carcinoma of the stomach was diagnosed were included in this population. The intention-to-treat population comprised all patients fulfilling the inclusion criteria who had taken the study medication at least once. Patients who completed the study according to the protocol and whose H. pylori status could be determined at the end of the study were included in the per protocol analysis. Patients who took antacids and/or NSAIDs during a time interval of 28 days before the start of the study were included in the per protocol population, but considered separately.

RESULTS

Table 1. Baseline characteristics

Characteristic

PCM (n ¼ 181)

Age (years ± s.d.) 57 ± 13 Height (cm ± s.d.) 170 ± 9 Weight (kg ± s.d.) 75 ± 13 26 ± 4 Body mass index (kg/m2 ± s.d.) Race, n (%) Caucasian 178 (98) Other 3 (2) Gender, n (%) Male 110 (61) Female 71 (39) Smoking status, n (%) Yes 83 (46)

PAC (n ¼ 169)

PAM (n ¼ 167)

57 169 75 26

56 169 75 26

± ± ± ±

13 9 14 5

± ± ± ±

13 8 14 4

168 (99) 1 (1)

166 (99) 1 (1)

93 (55) 76 (45)

91 (54) 76 (46)

82 (48)

86 (52)

PAC, pantoprazole–amoxicillin–clarithromycin; PAM, pantoprazole– amoxicillin–metronidazole; PCM, pantoprazole–clarithromycin–metronidazole; s.d., standard deviation.

Patients Of the 519 patients enrolled in the study, two did not take any study medication and, in 53, the H. pylori status could not be confirmed by another test or a carcinoma of the stomach was diagnosed. Thus, 517 patients were included in the safety population and 464 patients in the intention-to-treat population. All patients who took the medication at least once and completed the study without major protocol violations were included in the per protocol population, which comprised 313 patients. The most prominent protocol violation was absence from the final visit, and therefore information on the H. pylori status at the end of the study was missed. These patients were categorized as ‘not eradicated’ in the intention-to-treat analysis. No substantial differences with regard to demographic characteristics were found between the treatment groups (Table 1). Efficacy The eradication rates were 89% (67%) in the PAC group, 83% (68%) in the PCM group and 76% (60%) in the PAM group after the treatment-free phase for per protocol (intention-to-treat) analysis. The numerical difference of about 5% in the hierarchy PAC > PCM > PAM was significant between the PAC and PAM groups (P < 0.05). Determination of resistance Different primary resistance rates to the antibiotics in the treatment groups might be the reason for the

different efficacy in eradication. Therefore, susceptibility tests for the three antibiotics were performed at baseline and again at the final visit. These tests could only be performed in patients who were positively tested for H. pylori, resulting in a low number of tests at the end of the study. At study entry, tests were assessable for a total of 354 patients. At this time, the majority of H. pylori-positive patients were sensitive to clarithromycin and metronidazole (295, 83.3%). Resistance at baseline (primary resistance) was observed in 14.1% of the examined samples to metronidazole, in 1.4% to clarithromycin and in 1.1% to both antibiotics. No resistance to amoxicillin was observed (Table 2). Resistance in the three treatment groups was similar: 19 patients in the PCM group, 18 in the PAC group and 17 in the PAM group were resistant to metronidazole, and three patients in each group were resistant to clarithromycin. At the end of the study, susceptibility tests were performed in all patients (n ¼ 34) who remained H. pylori positive after therapy. Resistance to clarithromycin was observed in 2.9% (1/34), to metronidazole in 50% (17/34) and to both in 23.5% (8/34) (Table 2). Resistance to amoxicillin was not found at baseline or at any time point during the study. A statistically relevant difference in the metronidazole resistance rates was observed in the 54 cases between males (40.7%) and females (59.3%). This was reflected by significantly lower eradication rates in females (74%) compared with males (82%) in the PAM group.


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

H. PYLORI ERADICATION AND GASTRIC ULCER HEALING Table 2. Determination of resistance to clarithromycin (C), metronidazole (M) or both (C + M) at baseline and at the end of the study. Susceptibility testing could only be performed in Helicobacter pylori-positive patients (n ¼ number of patients)

At baseline (Assessable in 354 patients) n At study end (Assessable in 34 patients) n

C

M

C+M

1.4% (5/354) 2.9% (1/34)

14.1% (50/354) 50% (17/34)

1.1% (4/354) 23.5% (8/34)

1129

19 patients were not included in the following intention-to-treat and per protocol analyses (Table 3). In patients with strains resistant to at least one given antibiotic relevant for the regimen, eradication rates were 54% and 36% in the PCM and PAM groups, respectively (per protocol). In the PAC group, only two patients were infected with H. pylori strains resistant to clarithromycin, and could not be cured. Ulcer healing

It was of interest to determine whether the observed resistant strains were present at baseline or were newly developed (secondary resistance). Secondary resistance to metronidazole or clarithromycin was observed in strains of detected in 14 patients who had fully sensitive strains at baseline: in five patients in the PCM group (two with new double resistance, two with additional clarithromycin resistance and one with additional metronidazole resistance), seven patients in the PAM group (all to metronidazole) and one patient in the PAC group (to metronidazole). One patient in the PAM group with no assessable baseline tests showed a strain which was resistant to metronidazole. It was of particular interest in this study to determine whether eradication rates were lower in patients with resistant strains than in those with sensitive strains. In order to obtain useful evidence on this question, analyses were performed only in those patients with an antibiotic resistance relevant for their treatment group. From the 59 patients with resistant strains at baseline, 17 patients were resistant to metronidazole but belonged to the PAC group, and two patients were not included in the intention-to-treat population. These

No significant differences in the ulcer healing rates were found between the three treatment groups (Figure 1), either in the intention-to-treat (not shown) or per protocol analysis. A healing rate of 91% in the PAM group was reached after 4 weeks. The rates increased to over 97% in all groups after a further 4 weeks, with PAC reaching the highest rate with 100%. In the subgroup of patients in whom susceptibility testing was performed, the influence of H. pylori eradication on ulcer healing was analysed. It was found that, in patients in the intention-to-treat population in whom H. pylori had been eradicated, ulcer healing rates were clearly higher than in those with eradication failure. However, in the per protocol population, the healing rates were similar (Figure 2). Furthermore, in order to identify additional factors influencing ulcer healing, odds ratios were determined with regard to gender, age, GERD status at baseline, ulcer size and H. pylori status (Table 4). Patients with successful eradication and an ulcer size of < 15 mm had a significantly higher chance of ulcer healing compared with those with unsuccessful eradication. In 48% of patients, ulcers had a size of 5–10 mm at baseline. Medium and large ulcers of 10–15 mm and

Table 3. Comparison of the eradication rates in patients with sensitive or resistant Helicobacter pylori strains to at least one given antibiotic relevant to the treatment group at baseline [intention-to-treat (ITT) and per protocol (PP) populations] Eradication rates in patients with sensitive strains (%) [95% CI]

Eradication rates in patients with resistant strains (%) [95% CI]

Population

PCM

PAC

PAM

PCM

PAC

PAM

ITT

69/93 (74%) [64–83%] 60/70 (86%) [75–93%]

82/116 (71%) [62–79%] 68/76 (90%) [80–95%]

58/90 (64%) [54–74%] 50/61 (82%) [70–91%]

8/21 (38%) [18–62%] 7/13 (54%) [25–81%]

0/2 (0%) — * —

5/17 (29%) [10–56%] 5/14 (36%) [13–65%]

PP

CI, confidence interval; PAC, pantoprazole–amoxicillin–clarithromycin; PAM, pantoprazole–amoxicillin–metronidazole; PCM, pantoprazole– clarithromycin–metronidazole. * No patients with resistant strains at baseline.
2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

P. MALFERTHEINER et al.

1130

90%

Ulcer healing rates (%)

100%

98%

100%

88%

Relapse

97% 91%

80% 60% 4 weeks 8 weeks

40% 20% 0%

PCM

PAC

PAM

Figure 1. Ulcer healing rates after 4 and 8 weeks (per protocol population). PAC, pantoprazole–amoxicillin–clarithromycin; PAM, pantoprazole–amoxicillin–metronidazole; PCM, pantoprazole–clarithromycin–metronidazole. 100%

98·2%

100·0%

98·9%

Ulcer healing rates (%)

Eradicated Not eradicated

80% 68·6%

60%

70·7%

57·7%

40% 20% 0%

PCM

PAC

PAM

Figure 2. Ulcer healing rates in patients with successful and failed eradication after 4 weeks (per protocol population). PAC, pantoprazole–amoxicillin–clarithromycin; PAM, pantoprazole– amoxicillin–metronidazole; PCM, pantoprazole–clarithromycin– metronidazole.

15–20 mm were observed in 25% and 27% of intentionto-treat patients, respectively. The distribution of these ulcer sizes was very similar in all three treatment groups (data not shown). High healing rates were reached in all ulcer categories after 8 weeks, clearly demonstrating that the speed of healing was dependent on the ulcer size. After 4 weeks, 84% of patients with small ulcers (5–10 mm) were healed, compared with only 65% of patients with large ulcers (15–20 mm) (intentionto-treat) (Table 5). In addition, the ulcers (5–10 and 10–15 mm) of H. pylori-eradicated patients healed faster than those of patients who were still H. pylori positive (intention-to-treat: > 90% vs. < 70% after 4 and 8 weeks). This difference was even more prominent in patients with large ulcers (intention-to-treat: 41% in H. pylori-positive vs. 74% in H. pylori-negative patients).

Ulcer relapse after stopping antisecretory treatment was observed in 16 of 464 patients (3.4%) in the intentionto-treat population (12/313 in the per protocol population). With regard to the intention-to-treat analysis, eradication had a slight influence on the relapse rates, as 3.96% of patients (12/303) in whom H. pylori had been eradicated experienced ulcer relapse, compared with a relapse rate of 2.5% in patients with unsuccessful eradication. With regard to the per protocol analysis, the eradication of H. pylori had no obvious influence on the relapse rates, i.e. 3.9% of patients (10/259) in whom H. pylori had been eradicated, in contrast with 3.7% (2/54) in whom H. pylori had not been eradicated, experienced an ulcer relapse within a short observation period of up to 14 weeks. Relapse rates were slightly higher in the PAM group (4.9%) than in the PAC (4.0%) or PCM (2.7%) groups. Long-term investigations are ongoing to observe the patients in this study with remission. Although the intake of NSAIDs was not permitted during the study, 18 intention-to-treat patients who took NSAIDs during a time interval of 28 days before the start of the study were considered separately. These patients were not statistically significantly different, but lower healing rates (78%) were seen in comparison with patients who did not take NSAIDs (88%) at the end of the study. The intake of NSAIDs also had no influence on the ulcer relapse rate or the development of new ulcers, i.e. only one patient with ulcer relapse or a new ulcer took NSAIDs before and during the study. Histological examination The H. pylori density was equal in the antrum and the corpus and the number of colonies decreased strongly in both sites after treatment. It was conspicuous that the improvement of gastritis was better in patients in whom H. pylori had been eradicated. At baseline, signs of chronic and active inflammation, determined by the mean density of lymphoplasmacellular infiltrates and the density of granulocyte infiltrates, respectively, were more prominent in the antrum than in the corpus. The scores for lymphatic follicles/aggregates and the atrophy of the gland body were also higher in the antrum than in the corpus. In addition, the biopsies from the antrum revealed a higher proportion of intestinal metaplasia than did biopsies from the corpus. At the


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

H. PYLORI ERADICATION AND GASTRIC ULCER HEALING Table 4. Odds ratios and 95% confidence intervals (CI) for ulcer healing after 4 weeks as a function of several parameters

Gender Male/female Age of 57 years Younger/older GERD I at baseline No GERD/GERD I Ulcer size 5–10 mm/10–15 mm 5–10 mm/15–20 mm 10–15 mm/15–20 mm Eradication status Eradicated/not eradicated

1131

ITT [95% CI]

PP [95% CI]

0.8635 [0.5554; 1.3425] N.S.

0.7717 [0.3661; 1.6264] N.S.

1.2429 [0.8030; 1.9239] N.S.

1.6142 [0.7604; 3.4268] N.S.

1.4621 [0.6199; 3.4487] N.S.

1.3178 [0.2831; 6.1337] N.S.

1.5042 [0.8571; 2.6398] N.S. 2.7136 [1.6276; 4.5244] P < 0.05 1.8229 [1.0308; 3.2236] P < 0.05

1.0301 [0.3019; 3.5140] N.S. 5.2965 [2.2218; 12.6262] P < 0.05 5.1415 [1.6863; 15.6764] P < 0.05

5.2228 [3.2797; 8.3170] P < 0.05

1.4794 [0.5974; 3.6638] N.S.

GERD, gastro-oesophageal reflux disease; ITT, intention-to-treat; PP, per protocol; N.S., not significant.

Table 5. Ulcer healing rates and ulcer diameter [and 95% confidence intervals (CI)] in patients with successful Helicobacter pylori eradication or eradication failure H. pylori status

Population

5–10 mm (%) [95% CI]

10–15 mm (%) [95% CI]

15–20 mm (%) [95% CI]

Not eradicated

ITT 4 weeks 8 weeks

n ¼ 84 56/84 (67%) [56–77%] 58/84 (69%) [58–77%]

n ¼ 43 23/43 (54%) [38–69%] 27/43 (63%) [47–77%]

n ¼ 34 14/34 (41%) [25–59%] 21/34 (62%) [44–78%]

PP 4 weeks 8 weeks ITT 4 weeks 8 weeks

n ¼ 31 28/31 (90%) [74–98%] 28/31 (90%) [74–98%] n ¼ 141 132/141 (94%) [88–97%] 140/141 (99%) [96–100%]

n ¼ 12 11/12 (92%) 11/12 (92%) n ¼ 73 67/73 (92%) 72/73 (99%)

[83–97%] [93–100%]

n ¼ 11 8/11 (73%) [39–94%] 11/11 (100%) [71–100%] n ¼ 89 66/89 (74%) [64–83%] 88/89 (99%) [94–100%]

PP 4 weeks 8 weeks

n ¼ 119 114/119 (96%) [90–99%] 119/119 (100%) [97–100%]

n ¼ 65 62/65 (95%) [87–99%] 64/65 (98%) [92–100%]

n ¼ 75 58/75 (77%) [66–86%] 75/75 (100%) [95–100%]

Eradicated

[61–100%] [61–100%]

ITT, intention-to-treat; PP, per protocol.

end of the study, the gastritis scores had decreased in both the antrum and the corpus due to the suppression of inflammatory processes caused by H. pylori infection. The degree of both atrophy and metaplasia remained unaffected by treatment. However, most of the gastritis parameters improved during the trial period (Table 6), and it can be concluded that treatment success was similar for both antrum and corpus gastritis. In the absence of endoscopic criteria for suspected neoplasia, gastric malignancy was diagnosed by histol-

ogy at the initial visit in a total of 25 patients (4.8% of the safety population). Gastro-oesophageal reflux disease At baseline, no oesophagitis was observed in 294 patients (per protocol: 94%) and oesophagitis of grade I in 16 patients (per protocol: 5.1%). During the first 4 weeks of therapy, improvement of GERD grade I to normal was observed in 11 patients. Only in two


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

1132

P. MALFERTHEINER et al.

Parameter

Phase

Antrum Mean scores ± s.d.

H. pylori colonization

Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post Pre Post

2.09 0.20 2.12 1.44 1.86 0.17 0.38 0.39 0.87 0.52 7.5 5.0 25.6 25.4

Chronic inflammation Active inflammation Atrophy of gland body Lymphatic follicles/aggregates Incomplete metaplasia  (%) Complete metaplasia  (%)

± ± ± ± ± ± ± ± ± ±

0.91 0.62 0.51 0.57 0.75 0.49 0.78 0.77 1.03 0.80

Corpus Mean scores ± s.d. 2.04 0.19 1.60 1.08 1.43 0.15 0.12 0.11 0.35 0.24 1.4 1.0 8.1 8.8

± ± ± ± ± ± ± ± ± ±

Table 6. Histological evaluation of gastritis in the 310 patients in the per protocol population*

0.70 0.55 0.59 0.56 0.77 0.48 0.47 0.43 0.69 0.56

s.d., standard deviation. * Drop-outs were excluded from the analysis.   Intestinal metaplasia was assessed only qualitatively, and therefore was not scored.

patients, who initially showed no signs of GERD, was grade I oesophagitis diagnosed after 4 weeks. None of the patients showed a deterioration of GERD to higher grades, i.e. grades II–IV. Six weeks after the end of treatment (final visit), six patients who initially had no oesophagitis showed GERD grade I, one patient deteriorated to GERD grade III and one patient changed from grade I to II. These eight patients who showed a deterioration of GERD grade between the initial and final visit were all H. pylori negative at the end of the study. In total, 297 patients (per protocol: 95%) showed no signs of GERD and 11 patients (per protocol: 3.5%) showed signs of grade I at the final visit. Thus, the results of this trial indicate that successful H. pylori eradication therapy does not have a promoting effect on GERD. Patients were asked about gastrointestinal symptoms at each visit. These were evaluated by assigning a score for each symptom (data not shown). The symptom with the highest score in all treatment groups was ulcer pain during the daytime, followed by ulcer pain at night, nausea and heartburn. The scores of all investigated symptoms decreased to nearly zero during the study course, and there were only small differences between the treatment regimens. There was no association between the deterioration of GERD and specific symptoms, as the eight patients with deterioration of GERD experienced all kinds of symptoms.

Adverse events Safety data were available for 517 patients and 37% (n ¼ 193) reported adverse events. Most (21%, n ¼ 111) occurred during triple therapy, with 13% (n ¼ 67) during monotherapy and 9% (n ¼ 49) during the drug-free period. The majority of adverse events were considered to be unrelated or unlikely to be related to the administration of the study medication. Only 19 symptoms (6%) were considered as likely to be related to pantoprazole and no symptom as definitely related to pantoprazole. On the other hand, 105 symptoms (31%) were assessed as likely to be related to the intake of antibiotics and two symptoms as definitely related to the intake of antibiotics. Twentyfive serious adverse events in 21 patients were observed during the trial, but only four of these were assessed as likely to be related to the medication. No relevant differences in the safety profile of the three treatment groups with regard to the overall incidence of adverse events were apparent. The most frequently observed adverse event in the PCM group was influenza-like symptoms (4%), and diarrhoea in the PAC and PAM groups (12% and 5%, respectively). Compliance Compliance was assessed by tablet count. During the 7-day triple therapy, 92% of patients were compliant,


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

H. PYLORI ERADICATION AND GASTRIC ULCER HEALING

with 89% taking the following pantoprazole monotherapy, i.e. more than 80% of the study medication. DISCUSSION

In this study, the H. pylori eradication rates in patients with gastric ulcer were 76% with PAM, 83% with PCM and 89% with PAC. These results from our large trial confirm previous findings observed in patients with gastric ulcer from smaller studies.2, 3 The eradication efficacy found in patients with gastric ulcer is no different from that reported in large trials conducted in patients with duodenal ulcer.4–6 The combination of a proton pump inhibitor with clarithromycin and either amoxicillin or metronidazole is superior to the combination of a proton pump inhibitor with amoxicillin and metronidazole.2, 6 Primary resistance was most frequent to metronidazole (14%) and rare to clarithromycin (1.4%), with the expected impact on the eradication rates and a clear therapeutic superiority of PCM vs. PAM. In contrast with other European countries, primary clarithromycin resistance is still low in Germany and amoxicillin resistance has not been observed.24 Following treatment failure, resistance to metronidazole was detected in 73.5% (25/34) and secondary clarithromycin resistance occurred in 26.5% (9/34). In eight of nine cases, clarithromycin resistance was accompanied by a simultaneous resistance to metronidazole. The proportion of secondary clarithromycin resistance was at the lower limit of post-treatment clarithromycin resistance rates, which often have been found to be of the order of 30– 70%.25, 26 In vitro exposure of metronidazole-resistant H. pylori strains to metronidazole has been reported to be accompanied by a dramatic increase in their mutation rate, thus favouring base exchanges, which are also involved in clarithromycin resistance.27 The low rate of secondary clarithromycin resistance observed in this study is likely to be due to the low proportion of H. pylori strains with primary resistance to metronidazole. The ulcer healing rates after 4 weeks in the three treatment groups were similar. However, patients with successful H. pylori eradication — independent of the administered treatment regimen — showed significantly higher healing rates after 4 weeks than those with persistent infection. This adds important information to the healing mechanism of gastric ulcers, as so far convincing evidence exists only for a tight correlation

1133

between the power of acid suppression and the speed of ulcer healing. Recently, based on a meta-analysis, Huang and Hunt have shown an additional effect of antibiotics on the acceleration of duodenal ulcer healing, but this has never been demonstrated to be the case in gastric ulcer healing.28 In this study, all patients continued to receive strong acid inhibition with pantoprazole (40 mg/day) following the 7-day eradication regimen. Therefore, the observed difference in ulcer healing between those with and without persistent infection at 4 weeks clearly indicates an additional positive effect of H. pylori eradication, which is probably mediated by healing of the mucosal inflammation. Indeed, active inflammation was completely reversed in those with successful eradication by week 4. The second crucial factor for ulcer healing identified in our study was the ulcer size. The critical ulcer size predicting delayed healing is a diameter of more than 15 mm. In terms of the clinical consequence of this observation, gastric ulcers larger than 15 mm should be considered for prolonged acid inhibitory therapy. Even in the case of successful eradication, an ulcer size larger than 15 mm suggests that treatment for a longer period should be employed. In clinical practice, ulcer size is easier to use as a predictor of the speed of ulcer healing than is the success of eradication therapy. Successful eradication can be determined only later, after having stopped eradication therapy, and, in practical terms, is not useful for deciding whether acid inhibitory therapy should be continued beyond the phase of eradication. A comment needs to be dedicated to the phenotypic expression of gastritis in conjunction with gastric ulcer. Although, in patients with duodenal ulcer disease, gastric inflammation is characteristically antrum predominant, the gastritis pattern in gastric ulcer varies from corpus predominant to pangastritis, and even a slightly greater involvement of the antrum has been reported.2, 7, 12, 13 In this large series of patients with gastric ulcer, the density of granulocyte infiltrates was slightly higher in the antrum than in the corpus mucosa at baseline. This finding indicates that, in H. pylori-induced gastric ulcer, global involvement of the gastric mucosa usually takes place, with slightly less histological damage to the acid-secreting gastric compartments. Although active inflammation improved dramatically or even disappeared in both the antrum and the corpus from the first control visit onwards, the elements of chronic gastritis (assessed by the density of


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

1134

P. MALFERTHEINER et al.

lymphoplasmacellular infiltrates), although reduced, persisted in all patients during the follow-up period. In agreement with other authors, our study confirms that mucosal atrophy and intestinal metaplasia are more distinctive in the antrum than in the corpus, and more than twice as frequent in the antrum than in the corpus mucosa.13 For both parameters, there were no significant changes after therapy, and similar results have also been found in duodenal ulcer studies after treatment with omeprazole-based triple therapy.23 This suggests that a complete regeneration of atrophic mucosa, if it occurs, may take a long time. NSAIDs are another important aetiological factor in the development of gastric ulcer, and therefore the use of NSAIDs was not permitted in this study. Some patients, however, took NSAIDs before or during the study; the separate analysis of these patients showed an expected lower ulcer healing rate than in patients with no NSAID intake. Patients with malignant ulcers are frequently reported to have a higher prevalence of incomplete and complete metaplasia.18 Importantly, 25 gastric neoplasms were found at the inclusion visit, emphasizing that the benign appearance of gastric ulcer should never be taken as an indication of its benign nature. The fact that, during the study, another four cases of gastric cancer were diagnosed further emphasizes the need to perform an additional endoscopy and histological evaluation in patients with gastric ulcer to ensure ulcer healing and confirm the benign nature of the lesion. With regard to symptomatology in the short term, ulcer pain and the typical symptoms of GERD, such as heartburn, improved following H. pylori eradication. This finding needs to be followed over time, as H. pylori eradication itself has been reported to provoke the development of GERD in duodenal ulcer patients.29 In conclusion, the lessons learned from this large trial on gastric ulcer patients are that H. pylori eradication is highly effective for gastric ulcer healing, and accelerates the gastric ulcer healing process in comparison with acid suppressive therapy alone. Treatment with the PAC regimen was more effective than treatment with PCM, and both of these regimens had a significantly higher eradication rate than PAM. Factors determining the accelerated healing of gastric ulcer are successful H. pylori eradication and an ulcer size of < 15 mm. The clinical impact of this experience may direct us to shorten or prolong the use of proton pump inhibitor therapy beyond the phase of eradication.

ACKNOWLEDGEMENTS

We are grateful to Altana, Konstanz, Germany, for support of the study. We thank the IFE Institut fu¨r Forschung und Entwicklung, Witten, Germany, for monitoring at the study sites, data collection, source data verification and preparation of the manuscript. This work was supported by ALTANA Pharma AG, Konstanz, Germany, who also provided the study medication. To all investigators collaborating in the study, we express our gratitude (alphabetical order): K.-D. Aßmus (Nu¨rnberg), P. Balig (Freiburg), J. Bauer (Leipzig), B. Birkner (Mu¨nchen), B. Bokemeyer (Minden), H. Bordel (Ostercappeln), J. Bott (Elmshorn), M. Braschke (Celle), W. Burmeister (Hamburg), R. Daikeler (Sinsheim), H. Daus (Ludwigslust), A. Dettmer (Mu¨nchen), E. Dombrowski (Berlin), U. Ehrle (Pfungstadt), T. Eisenbach (Duisburg), H. Eisold (Mo¨ssingen), H. Eschenburg (Gu¨strow), J. Fiedler (Magdeburg), R. Fink (Freising), O. Friedrichs (Duisburg), R. Goes (Ditzingen), M. Grothoff (Ahlen), A. Halbach (Hannover), J. Hein (Marburg), G. Heptner (Dresden), K. Hey (Paderborn), A. Hoffmann (Magdeburg), H. Ju¨rgens (Oelde), W. Karlas (Nu¨rnberg), H.-U. Kellner (Kassel), C. Klein (Ku¨nzing), B. Knapp (Siegen), A. Kocjan (Lu¨denscheid), S. Kolbe (Bad Berleburg), W. Kra¨mer (Pirmasens), H.-G. Krezdorn (Mu¨nchen), K. Lenz (Isny), L. Lerche (Lu¨beck), A. Mares (Frankfurt), B. Marowski (Berlin), A. Petrides (Bochum), K. Rupp (Ellwangen), A. Ryschka (Berlin), A. Sammler (Friedrichsthal), J. Schmitz (Fu¨rth), M. Schumacher (Wolmirstedt), D. Sehland (Rostock), R. Sievers (Bederkesa), B. Sto¨lzle (Lindau), W. Tenbieg (Darmstadt), C. Vo¨lker (Mu¨nchen), E. von Fritsch (Erlangen), R. Warncke (Elmshorn), J. Wehnert (Dresden), J. Weismu¨ller (Koblenz), R. Wendland (Bergneustadt), P. Witzany (Marktheidenfeld), K. Ziegler (Berlin), W. Zink (Nu¨rnberg). REFERENCES 1 Penston JG, McColl KE. Eradication of Helicobacter pylori: an objective assessment of current therapies. Br J Clin Pharmacol 1997; 43: 223–43. 2 Malfertheiner P, Bayerdo¨rffer E, Diete U, et al. The GU-MACH study: the effect of 1-week omeprazole triple therapy on Helicobacter pylori infection in patients with gastric ulcer. Aliment Pharmacol Ther 1999; 13: 703–12. 3 Axon ATR, O’Morain CA, Bardhan KD, et al. Randomised double blind controlled study of recurrence of gastric ulcer


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

H. PYLORI ERADICATION AND GASTRIC ULCER HEALING

4

5

6

7

8 9

10

11

12

13

14

15 16

17

18

after treatment for eradication of Helicobacter pylori infection. Br Med J 1997; 314: 565–8. Bayerdo¨rffer E, Miehlke S, Mannes GA, et al. Double-blind trial of omeprazole and amoxicillin to cure Helicobacter pylori infection in patients with duodenal ulcers. Gastroenterology 1995; 108: 1412–7. Labenz J, Tillenburg B, Peitz U, et al. Efficacy of omeprazole one year after cure of Helicobacter pylori infection in duodenal ulcer patients. Am J Gastroenterol 1997; 92: 576–81. Lind T, Veldhuyzen van Zanten S, Unge P, et al. Eradication of Helicobacter pylori using one-week triple therapies combining omeprazole with two antimicrobials: the MACH I study. Helicobacter 1996; 1: 138–44. Malfertheiner P, Leodolter A, Peitz U. Cure of Helicobacter pylori-associated ulcer disease through eradication. Baillieres Best Pract Res Clin Gastroenterol 2000; 14: 119–32. Marshall BJ. Helicobacter pylori. Am J Gastroenterol 1994; 89: 5116–28. Ruiz B, Correa P, Fontham ET, Ramakrishnan T. Antral atrophy, Helicobacter pylori colonization, and gastric pH. Am J Clin Pathol 1996; 105: 96–101. Dixon MF. Patterns of inflammation linked to ulcer disease. Baillieres Best Pract Res Clin Gastroenterol 2000; 14(1): 27–40. Sipponen P, Hyvarinen H, Siurala M. H. pylori corpus gastritis — relation to acid output. J Physiol Pharmacol 1996; 47(1): 151–9. ¨ a¨rynen M, Helske T, Kettunen P. Sipponen P, Seppa¨la¨ K, A Chronic gastritis and gastroduodenal ulcer: a case control study on risk of coexisting duodenal or gastric ulcer in patients with gastritis. Gut 1998; 30: 922–9. Schultze V, Hackelsberger A, Gu¨nther T, Miehlke S, Roessner A, Malfertheiner P. Differing patterns of Helicobacter pylori gastritis in patients with duodenal, prepyloric, and gastric ulcer disease. Scand J Gastroenterol 1998; 33: 137–42. Current European concepts in the management of Helicobacter pylori infection. The Maastricht Consensus Report. Gut 1997; 41: 8–13. Hunt RH, Thomson ABR. Canadian Helicobacter pylori Consensus Conference. Can J Gastroenterol 1998; 12: 31–41. Lam SK, Talley NJ. Report of the 1997 Asia Pacific Consensus Conference on the management of Helicobacter pylori infection. J Gastroenterol Hepatol 1998; 13: 1–12. Malfertheiner P. The Maastricht recommendations and their impact on general practice. Eur J Gastroenterol Hepatol 1999; 11: S63–7, discussion S73. Dixon MF, Genta RM, Yardley JH, Correa P. Classification and grading of gastritis. The updated Sydney System.

19

20

21

22

23

24

25

26

27

28

29


2003 Blackwell Publishing Ltd, Aliment Pharmacol Ther 17, 1125–1135

1135

International Workshop on the Histopathology of Gastritis, Houston, 1994. Am J Surg Pathol 1996; 20: 1161–81. Labenz J, Stolte M, Aygen S, Hennemann O, Bertrams J, Borsch G. Qualitative and semiquantitative invasive and noninvasive diagnosis of Helicobacter pylori colonization of gastric mucosa. Z Gastroenterol 1993; 31: 437–43. Glupczynski Y, Labme M, Hansen W, Crokaert F, Yourassowsky E. Evaluation of the E test for quantitative antimicrobial susceptibility testing of Helicobacter pylori. J Clin Microbiol 1991; 29: 2072–5. Heep M, Kist M, Strobel S, Beck D, Lehn N. Secondary resistance among 554 isolates of Helicobacter pylori after failure of therapy. Eur J Clin Microbiol Infect Dis 2000; 19: 538–41. Leodolter A, Dominguez-Mun˜os JA, von Arnim U, Kahl A, Peitz U, Malfertheiner P. Validity of a modified 13C-urea breath test for pre- and post-treatment diagnosis of Helicobacter pylori infection in the routine clinical setting. Am J Gastroenterol 1999; 94: 2100–4. Veldhuyzen van Zanten SJO, Bradette M, Farley A, et al. The DU-MACH study: eradication of Helicobacter pylori, ulcer healing and relapse in duodenal ulcer patients. Omeprazole and clarithromycin in combination with either amoxicillin or metronidazole. Gut 1997; 41: A103(Abstract). Megraud F, Lehn N, Lind T, et al. Antimicrobial susceptibility testing of Helicobacter pylori in a large multicenter trial: the MACH 2 study. Antimicrobial Agents Chemother 1999; 43: 2747–52. Lind T, Megraud F, Unge P, et al. The MACH2 study: role of omeprazole in eradication of Helicobacter pylori with 1-week triple therapies. Gastroenterology 1999; 116: 248–53. Kist M, Strobel S, Foelsch UR, et al. Prospective assessment of the impact of primary antimicrobial resistances on cure rates of Helicobacter pylori infections. Gut 1997; 41: A190 (Abstract). Sisson G, Jeong JY, Goodwin A, et al. Metronidazole activation is mutagenic and causes DNA fragmentation in Helicobacter pylori and in Escherichia coli containing a cloned H. pylori RdxA(+) (Nitroreductase) gene. J Bacteriol 2000; 182: 5091–5. Huang J, Hunt RH. Reply — the importance of clarithromycin dose in the management of Helicobacter pylori infection: a meta-analysis of triple therapies with a PPI, clarithromycin and amoxicillin or metronidazole. Aliment Pharmacol Ther 1999; 13(12): 1693–4. Labenz J, Blum AL, Bayerdo¨rffer E, Meining A, Stolte M, Bo¨rsch G. Curing Helicobacter pylori infection in patients with duodenal ulcer may provoke reflux esophagitis. Gastroenterology 1997; 112: 1442–7.