J. P. Gisbert Dr Playa de Mojácar 29, Urb. Bonanza, 28669 Boadilla del Monte, Madrid, Spain. E-mail: email@example.com
To perform a meta-analysis of studies comparing twice daily, one-week triple therapy with a proton pump inhibitor, clarithromycin (C) and amoxycillin (A) (PCA) vs. those using proton pump inhibitor, clarithromycin and a nitroimidazole (N) (PCN) for H. pylori eradication.
Selection criteria: Comparative randomized trials of PCA vs. PCN were included. Data sources: PubMed database and abstracts from congresses until September 1999. Statistics: Meta-analysis was performed combining the Odds Ratios (OR) of the individual studies in a global OR (Peto method) both on an intention-to-treat (ITT) and on a per protocol (PP) basis.
Twenty-two studies fulfilled the inclusion criteria. Eighteen studies reported ITT and 20 PP analysis. Mean H. pylori eradication rates were 81% (95% CI: 79–83%) ITT, and 84% (82–86%) PP with PCA, and 81% (78–83%) ITT and 84% (82–86%) PP with PCN; the odds ratio for the effect of PCA vs. PCN was 1 (0.83–1.22) on an ITT, and 0.98 (0.8–1.2) on a PP basis. Subanalysis showed that mean H. pylori eradication efficacy with PC(250 b.d.)A was 81% (78–85%) ITT, vs. 86% (83–89%) with PC(250 b.d.)N. The odds ratio for this comparison was 0.68 (0.48–0.98). Finally, when comparing PC(500 b.d.)A against PC(250 b.d.)N ITT cure rates were 77% (74–80%), and 75% (72–78%) with an odds ratio of 1.18 (0.93–1.5).
Overall, one-week combination regimens of PCA and PCN present similar H. pylori eradication efficacy. Nevertheless, the PCN regimen obtains significantly better results when using low doses of C (250 mg b.d.).
At present, a 7-day quadruple therapy with a proton pump inhibitor, bismuth, tetracycline and metronidazole, and triple therapies with proton pump inhibitor, clarithromycin and amoxycillin or a nitroimidazole are considered the most effective combinations for the treatment of Helicobacter pylori infection. Because of their safety, cost-effectiveness and simplicity, proton pump inhibitor-based triple therapies are the most frequently recommended treatments world-wide.1–6 The two most widely used are a combination of a proton pump inhibitor, clarithromycin and amoxycillin, and a combination of a proton pump inhibitor, clarithromycin and a nitroimidazole. It has not been established whether one regimen is superior to the other. Although several meta-analyses of H. pylori eradication regimens have been published recently, none has been specifically designed to compare these two regimens. The aim of the present study was to perform a meta-analysis of the studies comparing these two 1-week proton pump inhibitor-based triple therapies for H. pylori infection.
PATIENTS AND METHODS
Bibliographical searches were performed in the PubMed (Internet) database, including studies available until September 1999, looking for the following words: Helicobacter or pylori (all fields) and each of the following terms: clarithromycin, amoxycillin, amoxycillin, metronidazole, tinidazole, imidazole, nitroimidazole, omeprazole, lansoprazole, pantoprazole, proton pump inhibitor, ‘proton pump’, 7, seven, triple, week, weeks (titles). We also conducted a manual search of abstracts from 1995 to 1999 from the International Workshop on Gastroduodenal Pathology and Helicobacter pylori, and American Digestive Disease Week. We included abstracts from congresses on the grounds that many negative or redundant studies are never published as a full paper, and the inclusion of abstracts thus prevents, or at least reduces, publication bias. Abstracts of the articles selected in each of these multiple searches were reviewed and those meeting the inclusion criteria were recorded. References of reviews on H. pylori treatment with proton pump inhibitor plus two antibiotics, and from the articles selected for the study, were also examined in searches of articles meeting the inclusion criteria. Articles published in any language were included. The bibliographic searches were performed independently by two different reviewers.
The selection criteria were as follows: (i) Articles had to report comparative randomized trials. (ii) They had to include at least two branches of treatment consisting of (a) a proton pump inhibitor (omeprazole, lansoprazole or pantoprazole, which were pooled together in the present meta-analysis), clarithromycin and amoxycillin, and (b) a proton pump inhibitor, clarithromycin and nitroimidazole (metronidazole or tinidazole, also pooled together). (iii) All drugs were given twice a day and for a period of 7 days. (iv) Only studies prescribing a twice a day proton pump inhibitor standard dose (omeprazole 20 mg b.d., lansoprazole 30 mg b.d., or pantoprazole 40 mg b.d.) were included; the extended use of a proton pump inhibitor was allowed for ulcer healing. (v) Assessment of eradication had to be performed by at least one reliable method (histology or 13C-urea breath test) at least 4-weeks after therapy. (vi) Studies evaluating these therapies as second-line treatment for previous eradication failures were not included. (vii) Only studies that had clearly stated information about the number of treated patients in each therapeutic group, drug dosage and schedule and H. pylori eradication rates were included. Publications identified as duplicates were excluded.
Subanalyses were planned to compare studies in subgroups using different doses of clarithromycin (e.g. 250 mg b.d. or 500 mg b.d.), different doses of nitroimidazole (e.g. 250 mg b.d. or 400–500 mg b.i.d) or different nitroimidazoles (metronidazole or tinidazole). A meta-analysis included proton pump inhibitor plus clarithromycin at high doses (500 mg b.d.) and amoxycillin vs. proton pump inhibitor plus clarithromycin at low doses (250 mg b.d.) and nitroimidazole.
The main outcome considered was eradication of H. pylori. Eradication was analysed both on an ‘intention-to-treat’ (ITT) and on a ‘per protocol’ (PP) basis. Articles that did not specify the type of analysis were assumed to report per protocol data. Separate comparisons depending on the type of antibiotics prescribed (clarithromycin plus amoxycillin on the one hand, and clarithromycin plus nitroimidazole on the other) were performed. Before combining the size effects of the individual studies, the homogeneity of effects throughout studies was appraised using a homogeneity test based on the χ2 test. Due to the low power of this test, a minimum cut-off P-value of 0.20 was established as a threshold for homogeneity: lower values indicated heterogeneity and prevented us from relying on the combination of the study results. The meta-analysis was performed combining the Peto Odds Ratios (OR) of the individual studies in a global OR, under the assumption-free model (or fixed effects model). Significance and 95% confidence intervals (95% CI) are provided for the combined OR. All calculations were performed with the freeware program REVIEW MANAGER 4.0. The statistical methods and formulae are described in the Handbook of the Cochrane Collaboration and the RevMan User Guide.7
Description of studies
Twenty-two studies fulfilled the inclusion criteria and contained data for at least one of the planned comparisons. For the main comparison, 18 studies reported ITT analysis and 20 PP analysis.8–29 Detailed characteristics of the studies are shown in Table 11.
Table 1. . Characteristics of studies comparing H. pylori eradication efficacy with one-week combination regimens of proton pump inhibitor plus clarithromycin and amoxycillin vs. proton pump inhibitor plus clarithromycin and nitroimidazole
Table 2. Table 1 . Continued
Efficacy on H. pylori eradication (overall results)
The main results of the meta-analysis comparing the two regimens on an ITT basis are summarized in Figure 1. Mean H. pylori eradication efficacy (pooled data) with proton pump inhibitor, clarithromycin and amoxycillin was 81% (95% CI: 79–83%) ITT, and 84% (82–86%) PP (Table 1). H. pylori eradication rates (pooled data) with proton pump inhibitor, clarithromycin and nitroimidazole were 81% (78–83%) and 84% (82–86%), respectively, by ITT and PP analysis (Table 1). The OR for the effect of proton pump inhibitor, clarithromycin and amoxycillin vs. nitroimidazole on H. pylori eradication was 1 (95% CI: 0.83–1.22) for ITT, and 0.98 (0.8–1.2) for PP. Chi-square homogeneity test was 24 (P: 0.14) by ITT, and 39 (P: 0.01) by PP; the P-value was lower than 0.2, indicating heterogeneity.
Subanalysis on H. pylori eradication
The initial intention was to perform separate comparisons for the two types of nitroimidazole used, metronidazole or tinidazole; however, almost all studies prescribed metronidazole, and only two protocols (ITT) used tinidazole, so we were unable to combine and compare the study results. Again, when trying to perform separate comparisons depending on the nitroimidazole dose, we found that almost all studies gave 400–500 mg b.d., and only three protocols (ITT) prescribed 250 mg b.d., precluding an adequate comparison of the study results.
A range of comparisons of clarithromycin doses were performed. Firstly, both regimens (with amoxycillin and with nitroimidazole) using clarithromycin at low doses (e.g. 250 mg b.d.) were compared. Results of the meta-analysis based on this comparison on an ITT basis are summarized in Figure 2. Mean H. pylori eradication efficacy (pooled data) with proton pump inhibitor, clarithromycin 250 mg b.d. and amoxycillin was 81% (78–85%) ITT, while mean eradication rate with proton pump inhibitor, clarithromycin 250 mg b.d. and nitroimidazole was 86% (83–89%). The OR for this comparison was 0.68 (95% CI: 0.48–0.98).
Finally, proton pump inhibitor plus clarithromycin at high doses (500 mg b.d.) and amoxycillin was compared with proton pump inhibitor plus clarithromycin at low doses (250 mg b.d.) and nitroimidazole. Meta-analyses on an ITT basis are summarized in Figure 3. Mean H. pylori eradication efficacy (pooled data) with proton pump inhibitor, clarithromycin 500 mg b.d. and amoxycillin was 77% (74–80%) ITT, and mean eradication rate with proton pump inhibitor, clarithromycin 250 mg b.d. and nitroimidazole was 75% (72–78%). The OR for this comparison was 1.18 (0.93–1.5).
This meta-analysis studied 22 comparisons of H. pylori eradication efficacy with 1-week combination regimens of proton pump inhibitor plus clarithromycin and amoxycillin vs. nitroimidazole (Table 1).8–29 The overall results show that the effectiveness of the two therapies is similar (Figure 1) with an OR for these two regimens of 1 (0.83–1.22) on an ITT basis, and 0.98 (0.8–1.2) on a PP basis. Nevertheless, results were heterogeneous, thus preventing us from relying on the combination of the overall study results, and advising dividing the treatments by subanalysis (e.g. depending on clarithromycin doses). The efficacy of both regimens seemed to be relatively low (at least, lower than that previously reported in initial studies), with eradication rates of only 80–85%. These results coincide with previously published systematic reviews. For example, Laheij et al.30 reported an ITT cure rate of 80% and 83%, respectively, with proton pump inhibitor, clarithromycin and amoxycillin or nitroimidazole; Pipkin et al.6 calculated a pooled ITT eradication rate of 83% with omeprazole-clarithromycin-amoxycillin, and 83% with omeprazole-clarithromycin-metronidazole regimens. Huang & Hunt4 reported slightly higher ITT rates, 85%, with the amoxycillin regimen. Finally, Unge3 calculated a mean efficacy of only 83% (ITT) with proton pump inhibitor plus amoxycillin and a macrolide, while higher eradication rates, 90% (ITT), were obtained with nitroimidazole regimens.
Some consensus reports have advised against the use of nitroimidazoles in first line therapies,5 due to locally high rates of resistances to this antibiotic.31 In dual therapy and bismuth-based triple therapy, it has become accepted that in vitro measurement of metronidazole resistance accurately predicts eradication failure.32 However, the correlation between in vitro H. pylori resistance to nitroimidazoles and the clinical repercussion on proton pump inhibitor-based triple therapies is not as clear, and several studies have demonstrated that metronidazole resistance does not negatively influence treatment outcome when using proton pump inhibitor, clarithromycin and nitroimidazole in combination,14, 33–41 although other studies have reached the opposite conclusion.23,42–49 However, the negative effect of nitroimidazole resistance seems to be of limited importance,32, 50 according to the MACH 2 multicentre study: eradication rates in primary metronidazole-susceptible and resistant strains were 95% and 76%, respectively.21 In another multicentre study, Misiewicz et al.23 showed that amoxycillin and metronidazole-based regimens (both combined with proton pump inhibitor and clarithromycin) were equally effective in patients with pre-treatment nitroimidazole-resistant organisms. Finally, in a randomized study, Houben et al.14 eradicated the infection in 82% of the patients despite the fact that the strains were metronidazole-resistant (while efficacy in patients with metronidazole-susceptible strains was only slightly higher, at 86%). Consequently, the clinical value of measuring metronidazole resistance in vitro when a proton pump inhibitor-clarithromycin-nitroimidazole regimen is prescribed is still unclear.
Some authors have recommended a proton pump inhibitor-clarithromycin-amoxycillin regimen to eradicate H. pylori in geographical regions with high metronidazole resistance.5,42 As reviewed, previous reports have suggested that H. pylori eradication with proton pump inhibitor-clarithromycin-nitroimidazole in metronidazole-resistant strains is still relatively high. Bazzoli et al. calculated eradication rates to be about 75% in resistant strains. Those authors described a model which allowed an estimation of the cure rates according to the different prevalences of metronidazole resistance and the mean eradication rates in a given geographical area with a known frequency of metronidazole resistance.50 Thus, the expected eradication rate would still be greater than 85% in regions with a metronidazole resistance of up to 50%.50 Therefore, the recommendation that the proton pump inhibitor-clarithromycin-nitroimidazole regimen should not be prescribed in areas with high metronidazole resistance only seems to be valid for those rare geographical regions with very high resistance rates (for example > 50%).
Regardless of the influence of primary nitroimidazole resistance, one argument against the use of nitroimidazole-based regimens as first-line therapy is that, in cases of eradication failure, secondary (acquired) H. pylori resistance may appear. This would limit the usefulness of the widely used quadruple ‘rescue’ therapy, which includes metronidazole among its components.
It remains to be seen whether a low dose of clarithromycin (250 mg b.d.) is sufficient for H. pylori eradication when 1-week proton pump inhibitor-based triple therapies are used. The initial 1994 study by Bazzoli suggested that this low dose of clarithromycin was sufficient when the proton pump inhibitor-clarithromycin-nitroimidazole regimen was prescribed: an eradication rate of 95% was achieved.51 Furthermore, several other studies have confirmed favourable results using low doses of clarithromycin when prescribing this regimen, with eradication rates ranging from 86% to 95%.34,52–56 The present study evaluates both regimens (proton pump inhibitor-clarithromycin-amoxycillin and proton pump inhibitor-clarithromycin-nitroimidazole) using clarithromycin at low doses (250 mg b.d.; Figure 2), and the results of our meta-analysis indicate better results with the nitroimidazole regimen (86%) than with the amoxycillin regimen (81%). Nevertheless, although the differences were statistically significant, the upper 95% confidence interval limit of the OR was 0.98, very close to 1. Finally, a proton pump inhibitor plus clarithromycin at high doses (500 mg b.d.) and amoxycillin were compared with a proton pump inhibitor plus clarithromycin at low doses (250 mg b.d.) and nitroimidazole (Figure 3), and eradication rates of 77% and 75% were obtained (OR 1.18; 0.93–1.5), indicating that the effectiveness of the therapies hardly differed.
Large multi-centre, international, randomized, double-blind trials seem to support these findings. The MACH I study,20 for example, concluded that the two most effective therapies were those combining omeprazole with either amoxycillin plus clarithromycin (at high doses, 500 mg b.d.), or metronidazole plus clarithromycin (at low doses, 250 mg b.d.). With this last regimen, the results were even better than with higher doses of clarithromycin (90% vs. 85%). In contrast, the results of omeprazole-clarithromycin-amoxycillin with low doses of clarithromycin (250 mg b.d.) were disappointing, with an eradication rate of only 79%. In the MACH 2 study21 the eradication rate with omeprazole and clarithromycin at low doses, and metronidazole, was as high as 87%. In another multicentre comparative study, Misiewicz et al.23 achieved H. pylori eradication in up to 87% of the patients with lansoprazole, clarithromycin 250 mg b.d. and metronidazole.
Furthermore, a recent meta-analysis of H. pylori eradication regimens concluded that the optimal dose of clarithromycin is uncertain when combined with proton pump inhibitor plus nitroimidazole, although data from Europe suggest that a low dose of 250 mg twice daily is sufficient.3 In a recent meta-analysis which evaluated the importance of clarithromycin dose in the management of H. pylori infection, Huang & Hunt4 showed that, when using the proton pump inhibitor-clarithromycin-amoxycillin regimen, significantly better results were obtained with high doses (500 mg b.d.) of clarithromycin (87%) than with low doses (78% with 250 mg b.d.). In contrast, these authors found similar results with high and low doses of clarithromycin when proton pump inhibitor-clarithromycin-nitroimidazole regimens were used (88% and 87%, respectively). Finally, Pipkin et al.6 showed that doubling the dose of clarithromycin from 250 mg b.d. to 500 mg b.d. in metronidazole-based regimens only marginally increased the eradication rate, from 83% to 87%.
One reason for the use of high-dose clarithromycin is to avoid the possibility of primary clarithromycin resistance in H. pylori.57 However, Cayla et al.58 have shown that, irrespective of the dose of clarithromycin used, the H. pylori cure rate is dependent on the pre-treatment strain sensitivity to this antibiotic. These data suggest that primary clarithromycin resistance is difficult to overcome, and a poor H. pylori cure rate of proton pump inhibitor-clarithromycin-nitroimidazole can be expected in patients carrying resistant strains, despite the higher doses of clarithromycin.57
If low doses of clarithromycin are sufficient when using a proton pump inhibitor-clarithromycin-nitroimidazole regimen, economic benefits could be inferred. A detailed review of the different H. pylori eradication regimens available at the present time concluded that the inclusion of a proton pump inhibitor, clarithromycin (250 mg b.d.) and nitroimidazole for 1 week was associated with the best cost–benefit ratio results.59 Duggan et al.60 used a decision analysis to examine the cost effectiveness of several eradication strategies and concluded that omeprazole, clarithromycin and metronidazole therapy was the most cost effective.
In addition, if low doses of clarithromycin are sufficient when using proton pump inhibitor-clarithromycin-nitroimidazole regimens, benefits in compliance and tolerance could also be inferred. A recent systematic review showed that significantly more patients on a proton pump inhibitor-clarithromycin-nitroimidazole regimen experienced adverse events in the sub-group taking clarithromycin 500 mg b.d. than in the clarithromycin 250 mg b.d. group.4
In summary, the results of the present meta-analysis indicate that, overall, 1-week combination regimens of proton pump inhibitor plus clarithromycin and amoxycillin present roughly the same efficacy for H. pylori eradication as proton pump inhibitor plus clarithromycin and nitroimidazole. Nevertheless, when using low doses of clarithromycin (250 mg b.d.), the nitroimidazole regimens seem to obtain better results.
We are indebted to Brenda Ashley and Michael Maudsley for assistance with the English.
This study was not funded by any pharmaceutical company.