Abstract
Background: In areas with high clarithromycin resistance, bismuth quadruple therapy (BQT) is recommended instead of clarithromycin triple therapy (CTT) as the first-line treatment for Helicobacter pylori eradication. Methods: Randomized clinical trials (RCTs) comparing BQT to CTT were identified through electronic and manual searches. A meta-analysis was performed to compare the efficacy and tolerability of these two regimens as first-line treatments for H. pylori infection. The effect of antibiotic resistance on treatment efficacy was also analyzed. Results: Twelve RCTs were included. BQT achieved eradication in 77.6% of patients, whereas CTT achieved an eradication rate of 68.9% [risk difference (RD) = 0.06, 95% CI: -0.01/0.13]. A high heterogeneity among the trials was found (χ2 = 50.16, p < 0.00001; I2 = 78%). In the subgroup analysis for treatment duration, the 10-day BQT was more effective than the 7-day CTT (RD = 0.25, 95% CI: 0.18/0.32), whereas no differences were observed between CTT and BQT given for 7 or 10 days. There were no statistical differences in side effects and compliance between both therapies (RD = 0.92, 95% CI: 0.76/1.12, and RD = -0.03, 95% CI: -0.05/0.00, respectively). The effect of antibiotic resistance on eradication rates was reported in 4 of the 12 RCTs. Clarithromycin resistance significantly affected the efficacy of CTT (RD = 0.75, 95% CI: 0.63/0.87), whereas BQT efficacy was not affected by metronidazole resistance (RD = 0.09, 95% CI: -0.06/0.25). Conclusions: The 10-day BQT was more effective than the 7-day CTT as a first-line therapy for H. pylori infection, whereas BQT and CTT for 7 or 10 days yielded similar eradication rates. Compliance and side effect rates were similar for both therapies. BQT overcomes clarithromycin resistance and its efficacy is not affected by metronidazole resistance.
Introduction
Helicobacter pyloriinfection is still one of the world's most frequent infections and accounts for high morbidity and mortality. About 20% of subjects infected with the bacterium will develop complications of the infection including peptic ulcer disease and gastric cancer, which accounts for at least 738,000 deaths annually [1]. Given the high prevalence and serious health burden of H. pylori infection, it is crucial to use a highly effective and well-tolerated eradication regimen.
The recent Maastricht IV consensus conference recommends for first-line empirical treatment of H. pylori infection a combination of a proton pump inhibitor (PPI), clarithromycin, and amoxicillin or metronidazole [clarithromycin triple therapy (CTT)] for 7-14 days in areas of low clarithromycin resistance and a combination of a PPI, bismuth, metronidazole, and tetracycline [bismuth quadruple therapy (BQT)] for 10-14 days in areas of clarithromycin resistance >15-20% [2]. The rationale for these recommendations is the increasing clarithromycin resistance rate in Europe and other geographical regions [3, 4]. Two previous meta-analyses on the effect of CTT and BQT found similar eradication rates for both regimens when used as first-line therapies for H. pylori infection. Compliance and side effects were similar for both therapy regimens as well [5, 6]. In the present meta-analysis, we further evaluated and compared the efficacy and tolerability of BQT and CTT by including recent trials. We included the analysis on the effect of resistance to clarithromycin and metronidazole on the efficacy of treatments.
Methods
Data Sources and Search Strategy
To find relevant articles for this review, a search of MEDLINE, EMBASE, and the Cochrane Library for randomized clinical trials (RCTs) published between 1995 and November 2011 was carried out. Only the following search terms were used: ‘Helicobacter pylori', ‘triple therapy', ‘quadruple therapy', ‘amoxicillin', ‘clarithromycin', ‘bismuth', ‘metronidazole', ‘tetracycline', and ‘eradication'. No limits for language were entered for the search. Boolean operators (‘AND', ‘OR', ‘NOT') were used to narrow and widen the search results. The titles from the search results were examined closely and determined to be suitable for potential inclusion into the study. In addition, the references from selected articles were examined as a further search tool.
Study Selection
For inclusion in the meta-analysis, a study had to meet the following criteria: (1) randomized clinical trial; (2) treatment with BQT (metronidazole, bismuth-containing compound, tetracycline, and PPI) versus CTT (amoxicillin, clarithromycin, and PPI) as the first-line therapy for eradication of H. pyloriinfection; (3) method of H. pyloridiagnosis by urea breath test, rapid urease test, histology, and/or fecal antigen testing; (4) main outcome measure as an intention-to-treat (ITT) eradication rate, and (5) eradication testing with urea breath test and/or histology at least 4 weeks after completion of therapy.
Data Extraction
Only RCTs were included in our analysis. Using a predefined meta-analysis form, two independent reviewers (M.V. and T.K.) extracted data from each study. This process resulted in high interobserver agreement (ĸ = 0.91). Information contained the names of the authors, title of the study, journal in which the study was published, country and year of the study, treatment regimen, dosage, length of the therapy, method by which H. pyloriinfection was diagnosed and method of eradication, testing sample size (with sex differentiation if applicable), the number of patients receiving each regimen and within the group the number of patients who succeeded and the number of patients who failed to eradicate the H. pylori infection, the number of patients who were compliant and the number of patients who were not compliant in each group, and the number of patients reporting side effects and the number of patients with no side effects in each arm. In studies where susceptibility tests for clarithromycin and metronidazole in H. pyloristrains were performed, the number of eradicated subjects with resistant and sensitive strains was collected. After completing the data extraction, the two independent reviewers discussed the results and, if discrepancies were present, a consensus was reached.
Data Synthesis and Analysis
Study outcomes for the meta-analysis included the following: (1) eradication rate of BQT compared with CTT reported as ITT, (2) compliance rate of BQT versus CTT, (3) incidence of side effects associated with BQT versus CTT, and (4) the effect of resistance to clarithromycin and metronidazole on the efficacy of treatments. Risk difference (RD), which represents the difference between the frequency of the events in the two groups (quadruple therapy and triple therapy), was used to calculate the pooled effect of BQT versus CTT. A 95% confidence interval (95% CI) was calculated using both a fixed-effects model and a random-effects model. The test for funnel plot asymmetry was carried out. To assess whether the variation in the effects of treatment across trials was greater than might be expected, a statistical evaluation of heterogeneity by χ2 test was used. Heterogeneity was considered to be present if the χ2 test delivered a p < 0.05. An I2 statistic was used to quantify the proportion of total variation in the study estimate due to heterogeneity. Any heterogeneity identified would prompt subgroup analysis on the basis of random-effects models in attempt to explain these findings. For subgroups with considerable heterogeneity, meta-regression for the year of publication was performed. The regression coefficients (r) for the meta-regression analysis were reported. An r value >0.7 was considered relevant. Furthermore, the cumulative effect over time was analyzed using statistical software [7]. This same method was applied to assess compliance rates and the incidence of adverse events between the two study groups. All computations and plots were carried out with meta-analysis software [7].
Results
Description of Included Studies
The initial search strategy identified a total of 89 potential articles for inclusion. After detailed review of selected articles, 12 RCTs [8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19] with 2,753 patients fulfilled the inclusion criteria of the meta-analysis (fig. 1). The most common reasons for exclusion from the meta-analysis included treatment regimens offered as second-line treatment and regimens composed of medications inconsistent with traditional CTT or BQT. Tables 1, 2 and Appendix table 1 summarize the characteristics of the included studies. Antibiotic dosing regimens for triple therapy were quite consistent among trials (clarithromycin 1,000 mg/day, amoxicillin 2,000 mg/day), but varied considerably for quadruple therapy (bismuth 240-1,680 mg/day, metronidazole 400-1,500 mg/day, tetracycline 1,500-2,000 mg/day).
Meta-Analysis of ITT Eradication Rates
In the meta-analysis of ITT eradication rates of the 12 included studies (Appendix fig. 1) the quadruple therapy achieved an eradication rate of 77.6%, whereas the eradication rate with triple therapy was 68.9% (RD = 0.06, 95% CI: -0.01/0.13). There was no evidence to suggest significant publication bias according to the test for asymmetry of the funnel plot (Appendix fig. 2). Our analysis did have high heterogeneity (χ2 = 50.16, p < 0.00001; I2 = 78%). In an attempt to explain heterogeneity, we conducted subgroup analyses (table 3).
Subgroup Analysis for Duration of Treatment Regimens
The different durations of the treatment regimens (7, 10, or 14 days for each regimen in different combinations) best explains the high heterogeneity among the studies. Therefore, we stratified trials in 3 subgroups according to the duration of treatment regimens (A = 7 days for both groups; B = 10-14 days for both groups; C = 7 days for triple therapy and 10 days for quadruple therapy; fig. 1, 2). No differences were observed between the two therapy regimens in subgroup A. Heterogeneity within group B was still high. When only trials with 10 days of treatment for both regimens were considered [11, 12, 13], summation of individual studies was possible. There were also no significant differences observed between the two therapy regimens in this subanalysis. In the sensitivity analysis of group C, BQT achieved eradication in 82.5% of patients, whereas triple therapy achieved an eradication rate of 57.7% (RD = 0.25, 95% CI: 0.18/0.32).
Subgroup Analysis for Year of Publication
The year of publication of the single studies may also represent a source of heterogeneity. The meta-regression analysis for year of publication showed a significant relationship between year of publication and efficacy of therapy regimen (r = 0.88, data not shown). In the subanalysis per year of publication comparing studies published between 2000 and 2005 versus studies published between 2006 and 2011, BQT was more effective than CTT within studies conducted in the period 2006-2011, but comparable to CTT within studies conducted in 2000-2005 (table 3). Heterogeneity was negligible within the group of studies published between 2000 and 2005, and moderate within the groups with a more recent year of publication (2006-2011).
Subgroup Analysis for Study Location
Study location may also represent a source of heterogeneity and different antimicrobial resistance rates between the Western and Eastern hemispheres may have an impact on treatment outcomes. Our analysis showed no significant differences in eradication rates between BQT and CTT in studies conducted in the Western hemisphere (Appendix fig. 3, table 3). In the subanalysis of ITT eradication rates of the 7 studies conducted in the Eastern hemisphere, the BQT achieved an eradication rate of 72.5%, whereas the eradication rate with CTT was 65.6% (RD = 0.052, 95% CI: -0.06/0.16). Heterogeneity within the Eastern hemisphere group exceeded our baseline results. In an attempt to explain heterogeneity within this subgroup, a meta-regression analysis for year of publication and frequency of eradication was performed, and showed a significant correlation for both variables (r = 0.88 and -0.75, respectively, data not shown). Thus, again the heterogeneity in this subgroup can be partially explained by the decreasing efficacy of CTT over the years. In studies conducted in 2006-2011 in the Eastern hemisphere, CTT achieved an eradication rate of 51.5% compared with 78.6% in studies conducted in 2000-2005 (data not shown). Furthermore, in the Eastern hemisphere, BQT achieved an ITT eradication rate of 70.3% in studies conducted in 2006-2011, whereas in studies conducted in 2000-2005, an ITT eradication rate of 74.7% was achieved (data not shown).
Influence of Local Clarithromycin Resistance on Efficacy of Therapy Regimens
Local clarithromycin resistance rates may influence the efficacy of both therapy regimens. Therefore, available data on local clarithromycin resistance in the countries of the included trials were collected and a meta-regression analysis for local clarithromycin resistance was performed. The available data on local clarithromycin resistance are as follows: Spain 14% [4], Australia/New Zealand 8% [10], Greece 24.7% [4], India 4.7% [20], United States/Canada 11% [13], Korea 10.8% [21], Turkey 48.2% [22], United Kingdom 9% [4], and China 20.8% [18]. The meta-regression analysis for local clarithromycin resistance showed no influence of local clarithromycin resistance rates on the efficacy of either therapy regimen (r = 0.17).
Subgroup Analysis according to Treatment Populations
The failure rate of H. pylori eradication has been reported to be significantly lower in patients with duodenal ulcer than in patients with nonulcer dyspepsia [18]. Therefore, we further stratified studies according to treatment populations, i.e. studies investigating H. pylori eradication rates in patients with dyspeptic symptoms, nonulcer dyspepsia, and active peptic ulcer (table 3, Appendix figure 4). BQT was more effective than CTT for H. pylori eradication in patients with dyspeptic symptoms and nonulcer dyspepsia, yet within both groups we found significant heterogeneity. Our analysis did not find any significant difference between BQT and CTT in patients with peptic ulcer disease, yet within this group no heterogeneity was found.
Meta-Analysis of Compliance and Side Effects
Eleven out of the 12 studies [8, 9, 10, 11, 12, 13, 15, 16, 17, 18, 19] included data on compliance rates (table 2). Acceptable compliance was differently defined and varied from 75 to 100% of completed therapy. Overall, the compliance rate was 92 and 96% with BQT and CTT, respectively. There was a statistically significant difference between the two groups (RD = -0.03, 95% CI: -0.05/0.00). Heterogeneity was not present.
All but two of the included studies [10, 17] describe the incidence of side effects within each group. The overall incidence of side effects in patients receiving BQT and CTT was 46 and 46.3%, respectively (table 2). There were no statistically significant differences in side effects yielded by each group. Heterogeneity was not present.
Role of Antibiotic Resistance
Four out of 12 studies [10, 12, 18, 19] also assessed the effect of antimicrobial susceptibility on eradication rates. ITT eradication rates achieved by quadruple and triple therapies based on metronidazole and clarithromycin resistance are shown in figure 3. Clarithromycin resistance significantly affected the efficacy of triple therapy (RD = 0.75, 95% CI: 0.63/0.87). Indeed, CTT achieved eradication in 88% of clarithromycin-sensitive strains and in 14.3% of clarithromycin-resistant strains. However, metronidazole resistance did not affect the efficacy of quadruple therapy. Indeed, BQT achieved eradication in 92% of metronidazole-sensitive strains and 84.2% of metronidazole-resistant strains. The role of simultaneous susceptibility for metronidazole and clarithromycin was tested in 2 of the 12 studies [18, 19]. In the presence of simultaneous metronidazole and clarithromycin resistance, BQT was superior to CTT for eradication of H. pylori infection (RD = 26.52, 95% CI: 3.43/205.06, data not shown). Given the small number of patients with simultaneous metronidazole and clarithromycin resistance (14 in the group receiving BQT and 17 in the group receiving CTT) included, these results should be interpreted cautiously.
Discussion
In our meta-analysis, CTT and BQT given both for 7 or 10 days yielded similar eradication rates when used as a first-line therapy for eradication of H. pylori infection. Furthermore, no difference in compliance rate or incidence of treatment-associated side effects between BQT and CTT was found. The results of our updated meta-analysis at first glance are consistent with those from two previous meta-analyses that compared these treatment regimens [5, 6]. However, in the subgroup analysis of treatment duration, the 10-day BQT was shown to be more effective than the 7-day CTT. The comparison of two different treatment durations may raise criticism of our analysis. To rebut this point, a recent meta-analysis reported that a treatment with CTT beyond 7 days would confer only a slight benefit of about 5% in eradication efficacy if the treatment is extended to 14 days [23]. Furthermore, if this analysis is restricted to the four highest-quality trials as assessed by the Jadad scale [24], prolongation of CTT beyond 7 days (with extension up to 10 days) was not associated with an additional benefit. This has been confirmed in a recent randomized trial in patients infected with H. pylori without an increase in eradication efficacy if CTT was extended up to 14 days [25]. Clarithromycin resistance obviously cannot be overcome by increasing the dose or duration of therapy [26]. Regarding BQT, a 10- to 14-day treatment is recommended by international guidelines [27, 28, 29]. A recent survey in Europe reported that the H. pylori resistance rate for metronidazole is as high as 34.9% [4]. Unlike clarithromycin, metronidazole resistance can to a great extent be overcome by increasing the duration of therapy and the dose of metronidazole, which should not be lower than 1,500 mg [30]. Indeed, a meta-analysis of 91 treatment arms concluded that ‘if nitroimidazole resistance is present, a nitroimidazole-containing regimen should be avoided or a quadruple regimen should be given for >1 week' [31]. On the basis of these data, a first-line empirical H. pylori treatment with BQT given for 10 days or longer is expected to work.
The strongest predictor of H. pylori treatment failure appears to be antimicrobial resistance [32]. Clarithromycin resistance is the most important clinically because resistance cannot be overcome by increasing the dose or duration of therapy. Amoxicillin and tetracycline resistance instead is negligible as resistance for these two antibiotics is reported to be very low (<1%) in most countries [4, 32]. In our analysis CTT achieved eradication in 88% of clarithromycin-sensitive strains but only in 14.3% of clarithromycin-resistant strains. Furthermore, the meta-regression analysis for year of publication showed a decreasing efficacy of triple therapy over years, mostly due to increasing clarithromycin resistance. With few exceptions worldwide, the prevalence of clarithromycin resistance has increased to a level where CTT is no longer effective [4, 32]. Thus, in the future, a clinical trial comparing a candidate effective first-line H. pylori therapy with CTT can no longer be considered ethical in countries with a high rate of clarithromycin resistance. On the contrary, BQT achieved an ITT-eradication rate as high as 92% in metronidazole-sensitive strains and 84.2% in metronidazole-resistant strains. Thus, BQT is effective even in cases with proven metronidazole resistance in vitro and represents a good treatment option for empirical first- and second-line H. pylori eradication therapy.
In patients with dyspeptic symptoms and nonulcer dyspepsia, BQT was more effective than CTT for H. pylori eradication. A higher resistance rate to clarithromycin reported in nonulcer dyspeptic patients in comparison to duodenal ulcer patients is likely to account for this [33].
In patients with peptic ulcer disease, BQT and CTT yielded similar eradication rates as a first-line treatment for H. pylori infection. The lower prevalence of H. pylori strains with clarithromycin resistance among patients with peptic ulcer disease accounts for the similar eradication rates between the CTT and the BQT in this specific group of patients. A lower prevalence of clarithromycin resistance among patients with peptic ulcer disease was described previously [4], but remains unexplained.
Strengths and Limitations
To minimize bias, two authors selected the studies and extracted the data; this process resulted in high interobserver agreement (ĸ = 0.91). The presence of publication bias was excluded by both graphical and statistical evaluation. All studies included in the analysis were RCTs. Our literature research was not restricted to English-language reports. We were able to extract data from each study, regardless of the language, with the exception of the exclusion criteria from the article by Jang et al. [14], which was published in Korean.
When considering the results of this meta-analysis, potential limitations should be considered. Heterogeneity was present and could not be accounted for by multiple subgroup analyses. Further, individual studies included in our analysis differed in several respects. Indeed, studies defined varying inclusion and exclusion criteria and used different PPIs. Relevant to BQT, antibiotic dosing regimens and bismuth formulations varied considerably among trials. The heterogeneity in the regimens included in this analysis could have affected eradication rates, as increasing dose and duration can improve efficacy [26]. Many efforts have been made to optimize the BQT. Possibly, the most effective BQT is the one employing the highest practical doses of the single drugs for the longest reasonable duration, such as 14 days. A one-arm, open-label pilot study of 47 H. pylori-infected, asymptomatic/mildly dyspeptic adult Hispanic residents of El Paso, Texas showed a per-protocol effectiveness of 97.7% for BQT given for 14 days [34]. In this study the combination capsule of bismuth, metronidazole, and tetracycline (daily doses: 1,680, 1,500, and 1,500 mg, respectively) plus omeprazole (daily dose: 40 mg) was used. Unfortunately, only one of the trials included in our meta-analysis evaluated the BQT for 14 days, without employing the highest practical doses of the single drugs. Thus, the heterogeneity in the treatment regimens included in the present analysis does not allow for the recommendation of an ‘optimal' BQT. Only 2 RCTs were available for the subgroup analysis comparing the 10-day BQT versus the 7-day CTT.
In conclusion, on the basis of our analysis, the 10-day BQT is more effective than the 7-day CTT as a first-line therapy for H. pylori infection, whereas BQT and CTT, both taken for 7 or 10 days, yielded similar eradication rates. BQT is independent of clarithromycin resistance and is at all only minimally affected by metronidazole resistance. Compliance rates and side effects were comparable for both treatment regimens.
Disclosure Statement
Peter Malfertheiner, MD is the guarantor of the article. No funding was received for the preparation of this manuscript. Marino Venerito has no conflicts of interest relevant to this study. Tina Krieger received honoraria from Aptalis Pharma. Thomas Ecker received honoraria from Aptalis Pharma. Gioacchino Leandro has no conflicts of interest relevant to this study. Peter Malfertheiner received honoraria from Aptalis Pharma, Abbott, Astra-Zeneca, and Takeda.