What’s the Story, H. pylori?  New Regimens in the Toolbox?

The list of regimens for Helicobacter pylori eradication is longer than ever.  In recent years, American and European guidelines have recommended clarithromycin-based triple therapy or bismuth-containing quadruple therapy for primary treatment of H. pylori infection (Table 1).^1,2  However, increasing resistance to these regimens has forced us to consider alternative treatments, including but not limited to: the addition of probiotics, use of sequential regimens, and the inclusion of levofloxacin in the antibiotic cocktail. With all these choices, the clinician is left wondering — is there a way to compare this wide array of treatment options?³

^{PPI = proton pump inhibitor; H₂RA = H₂ receptor antagonist}

^{a. Bismuth quadruple therapy preferred in patients with recent macrolide or metronidazole exposure}

^{b. Extension of clarithromycin-based triple therapy from 7 to 10-14 days may be considered}

Last year, Li et al. published a network meta-analysis that compared the efficacy and tolerability of H. pylori eradication regimens (Table 2).⁴  Unlike a traditional pairwise meta-analysis, a network meta-analysis allows the researchers to compare interventions that have not been evaluated in a head-to-head trial.  Rather, it compares interventions through a common comparator.  For instance, if interventions A and C were compared in a trial, and interventions B and C in another, then interventions A and B may be compared on the basis of their performance against the common comparator.  In many cases, the network meta-analysis can provide more robust results than a pairwise meta-analysis.⁵  When conducting a network meta-analyses, a comparison network should be mapped (Figure 1).  The map is then used to visualize the relative strength of both direct and indirect comparisons.  Any incoherence between treatment effects from direct (pairwise meta-analysis) and indirect (network meta-analysis) comparisons should be investigated.  These differences suggest a problem with study heterogeneity.⁵

Li et al. only considered randomized controlled trials (RCTs) in fully published form.  Studies were included in the meta-analysis if they reported an intention to treat (ITT) analysis; used regimens that were 7, 10, or 14 days in duration; enrolled patients who were treatment-naïve; and the eradication of H. pylori was assessed at least four weeks after treatment.  The primary outcome of the meta-analysis was the eradication rate per ITT population. The secondary outcome was incidence of adverse effects.  Treatment regimens were compared to the standard 7-day triple therapy by way of relative risk (RR) with 95% confidence intervals (or the analogous 95% credible interval for indirect comparisons).  A random effects model was utilized, which accounts for variation across studies and provides more conservative estimates of treatment effect differences.

For the efficacy analysis, 143 RCTs were included, with over 32,056 patients enrolled in these studies.  Given the 14 eradication regimens represented (Table 2), 91 possible pairwise comparisons were possible — 34 of which were directly compared by at least one RCT.  Table 3 summarizes the relative efficacy of each treatment regimen relative to 7-day triple therapy (PPI + clarithromycin + amoxicillin/metronidazole).  With the exception of 7-day regimens of levofloxacin- and bismuth-based therapy, all regimens were statistically more effective than 7-day triple therapy via indirect comparisons.  For regimens with available direct comparisons, traditional pairwise meta-analysis corroborated superiority, except for the 10 and 14-day probiotic-containing regimens.  The 14 treatments were ranked by efficacy within the network meta-analysis model, with 7-day concomitant quadruple therapy (PPI + three antibiotics) being the most effective and 7-day triple therapy being the least.  The number needed to treat (NNT) with 7-day concomitant quadruple therapy to successfully eradicate H. pylori when compared to the 7-day triple regimen is five.

Figure 1: Network map for efficacy analysis.  (click link to view image)  Each node represents an intervention, and each line represents direct comparisons between each intervention.  The weight of each line represents the number of separate studies comparing two adjacent treatments, and the size of each node represents the size of the pooled population in which that intervention was studied.

When compared to 7-day triple therapy, most regimens had similar incidences of adverse effects.  However, the 7-day probiotic-supplemented triple therapy and the 7-day levofloxacin-based regimen appear to be better tolerated and had lower incidences of adverse effects in the network meta-analysis (RR 0.65 [95% CrI 0.47-0.87] and RR 0.69 [95% CrI 0.49-0.94], respectively).  When ranked within the network meta-analysis model, 7-day regimens were generally ranked better for tolerability than 10 or 14-day regimens.

In their primary efficacy analysis, the authors utilized 7-day triple therapy as the comparator.  This regimen was likely chosen due to its large pooled sample population and its former place as first-line treatment.  However, the American College of Gastroenterology guidelines currently recommend 10 to 14 days of triple therapy — due to their higher eradication success rates.⁶  Therefore, comparisons to the 10 to 14-day regimen are of greater utility to clinicians practicing in North America.  When compared to 10 or 14-day triple therapy, the following regimens were associated with statistically significant improvements in eradication rates: 10 or 14-day bismuth quadruple therapy, 10 or 14-day concomitant therapy, 10 or 14-day probiotic-supplemented triple therapy, 10 or 14-day sequential therapy, and 10 or 14-day levofloxacin-based therapy.  With the exception of 10 or 14-day bismuth quadruple therapy, all of these regimens were also superior by pairwise meta-analysis, although the comparison with sequential therapy had significant heterogeneity (I² = 70%, P < 0.0001 for Cochran’s Q).

When considering the impact and application of a network meta-analysis, special attention should be paid to the network geometry (Figure 1).  The largest number of trials and largest pooled study populations featured the following therapies: 7-day triple therapy, 10 or 14-day triple therapy, 10 or 14-day sequential therapy, and 10 or 14-day bismuth quadruple therapy.  The comparisons between these regimens are the most statistically robust.  And thus we have greater confidence in the conclusion that 10 or 14-day bismuth quadruple therapy and 10 or 14-day sequential therapy are superior to 10 or 14-day triple therapy.

The safety and tolerability comparisons from this study are of minimal utility. The authors did report rates of individual adverse effects (abdominal pain, taste alterations, diarrhea, and headache), but rates were low and differences are difficult to interpret.  Most regimens had similar rates of adverse effects when compared to 7-day triple therapy.

The most significant limitation to this study is the relative underrepresentation of North Americans in the study populations and it is further compounded by the lack of antibiotic susceptibility data.  The majority of studies included in this meta-analysis were conducted in East Asia, Europe, and the Middle East.  Only six studies included patients from North America.  The success of H. pylori eradication is highly dependent on antibiotic resistance, which differs significantly by region.  For instance, clarithromycin resistance has been reported as high as 38% in Korea whereas a recent study in US veterans reported clarithromycin resistance to be 16.4%.^7,8  The authors performed sub-group analyses for the most common countries represented and found that eradication rates for 10 or 14-day triple therapy ranged from 58% in Turkey to 88% in Iran. 

Even if a study may not provide sufficient evidence to transform first-line therapy, it can still validate the expansion of treatment options.  In this case, where the primary literature is comprised of many different treatment options, heterogeneous study populations, and relatively small sample sizes, the network meta-analysis affords optimal statistical power to compare interventions.  Without H. pylori susceptibility data, it is difficult to conclude that any these regimens would outperform the standard of care in North America for treatment-naïve patients. However, this network meta-analysis adds to growing evidence that several newer regimens may be as or more effective than traditional triple or bismuth quadruple therapy.^9-11 From our perspective, the most promising regimens are 10 or 14-day levofloxacin-based triple treatment, 10 or 14-day probiotic-supplemented triple therapy, 10 or 14-day sequential treatment, and 10 or 14-day concomitant therapy.  While they may not supplant current first-line therapy, they are effective alternatives for patients with drug allergies or who experience treatment failure or intolerance.  Tell us what you think.  Are you routinely using some of the newer H. pylori eradication regimens in your practice?