Influence of Cytochrome P450 2C19 Genotype on Helicobacter pylori Proton Pump Inhibitor-Amoxicillin-Clarithromycin Eradication Therapy: A Meta-Analysis

Background: Proton pump inhibitors (PPIs) are the first-line treatment for acid-related diseases. The pharmacokinetics and therapeutic efficacy of PPIs, however, are influenced by genetic factors such as variants in genes encoding drug-metabolizing enzymes (e.g., cytochrome P450 2C19 [CYP2C19]) and drug transporters. We performed a meta-analysis to evaluate the influence of CYP2C19 genotype and PPI class, PPI dose, treatment duration and clarithromycin dose on the cure rate of PPI-containing Helicobacter pylori eradication therapy. Methods: Randomized control trials (RCTs) investigating cure rates using a PPI-amoxicillin-clarithromycin regimen among different CYP2C19 genotypes through May 2021 were included. Results: A total of 25 studies (5,318 patients) were included. The overall eradication rate in the intention-to-treat analysis was 79.0% (3,689/4,669, 95% confidence interval [CI]: 77.8–80.2%), and that in CYP2C19 extensive metabolizers (EMs), intermediate metabolizer (IMs) and poor metabolizers (PMs) was 77.7% (1,137/1,464, 95% CI: 75.3–79.6%), 81.2% (1,498/1,844, 95% CI: 79.3–83.0%) and 86.8% (644/742, 95% CI: 83.9–88.9%), respectively. Meta-analysis showed that the relaTakashitive risk of failed eradication in CYP2C19 EMs compared with IMs and PMs was 1.21 (95% CI: 1.06–1.39, P = 0.006) and 1.57 (95% CI: 1.27–1.94, P < 0.001), respectively, in the fixed-effects model. The cure rate of omeprazole and lansoprazole-containing eradication regimens differed among CYP2C19 genotypes (P < 0.05), while that of rabeprazole and esomeprazole-containing regimens was similar. Conclusion: The cure rates of PPI-amoxicillin-clarithromycin H. pylori eradication regimen, especially those containing omeprazole and lansoprazole, differ among CYP2C19 genotypes. Therefore, selection of a second-generation PPI or tailored treatment may achieve higher eradication rates than first-generation PPI-amoxicillin-clarithromycin triple regimen.


INTRODUCTION
The Maastricht V/Florence Consensus Report issued by the European Helicobacter Study Group in 2017 provides a guideline on how to manage Helicobacter pylori (H. pylori) infection (Malfertheiner et al., 2017). This guideline indicates that, despite increasing H. pylori resistance rates to antibiotics such as clarithromycin (CAM), metronidazole (MNZ) and levofloxacin (LVFX) in most parts of the world, clinicians worldwide continue to use general first-line H. pylori triple eradication therapy consisting of a proton pump inhibitor (PPI) and two kinds of antimicrobial agents [i.e., CAM, amoxicillin (AMPC), MNZ, or LVFX], especially in areas with lower rates of the clarithromycin-resistant strain (Malfertheiner et al., 2017). In Japan, H. pylori eradication therapy is limited to regimens comprising an acid-inhibitory drug such as a PPI (e.g., esomeprazole, rabeprazole, lansoprazole and omeprazole) or vonoprazan at a standard dose at twice-daily dosing (bid), AMPC 750 mg bid, and CAM 200 mg or 400 mg bid for 7 days as a first-line regimen (Kato et al., 2019). This is despite the fact that eradication therapy for all H. pylori-positive patients with gastritis confirmed by endoscopy is currently covered by the Japanese National Health Insurance system. From 1990 to 2000, eradication rates achieved in Japan using PPI-AMPC-CAM triple therapy ranged from approximately 85-91% (Asaka et al., 2001). However, because this rate has subsequently decreased to 60-75%, the factors affecting cure rates of PPI-AMPC-CAM therapy need to be identified.
Here, we performed a meta-analysis to re-evaluate the efficacy and safety of PPI-AMPC-CAM therapy alone among different CYP2C19 genotypes in RCTs written in English.

Search Strategy and Inclusion Criteria
This meta-analysis was conducted using data from RCTs identified by searching the medical literature on PubMed and Cochrane Library databases. We compared H. pylori eradication rates of PPI-AMPC-CAM triple therapy (each given over 7-14 days) as first-line therapy among different CYP2C19 genotypes (EMs, IMs and PMs). Two researchers (YM and MS) independently searched both the PubMed and Cochrane Library databases using the terms "Helicobacter pylori," "eradication," and "CYP2C19" and reviewed the titles and abstracts of all potential studies (Supplementary Figure S1). The inclusion criteria were 1) RCTs published through June 2021; 2) studies that used PPI-AMPC-CAM as first-line treatment; 3) studies comparing cure rates of PPI-AMPC-CAM regimen for H. pylori infection; and 4) studies written in English. Exclusion criteria were 1) non-RCT studies, 2) studies performed on non-PPI-AMPC-CAM regimen, 3) studies written in non-English languages, and 4) studies with abstracts alone. The author's names, year of publication, country where the study was conducted, number of patients, eradication rate of each regimen, patient characteristics (sex and age), CYP2C19 genotype, susceptibility to antimicrobial agents, and incidence of adverse events were extracted from each study.

Statistical Analysis
All meta-analyses were conducted using open-source statistical software (Review Manager Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). First, a meta-analysis of RCTs comparing the cure rates of PPI-AMPC-CAM therapy in all patients and among CYP2C19 genotypes was performed. For each comparison, intention-to-treat (ITT) and per-protocol (PP) analyses of cure rates were conducted. Relative risk (RR) and the corresponding 95% CI were used to summarize the effect of each comparison in fixed-effects models and random-effects models (46)(47)(48). Potential bias in each study was evaluated using funnel plots. Heterogeneity was evaluated using the I 2 value and Cochran's Q. The I 2 values used to define heterogeneity were as follows: 0-39%, low heterogeneity; and 40-74%, moderate heterogeneity. All p values were two-sided, and p < 0.05 was considered statistically significant. Calculations were performed using commercial software (SPSS version 27, IBM Inc.; Armonk NY, United States).

DISCUSSION
This meta-analysis included 25 RCTs from 24 Asian and 1 South American country to re-evaluate the association between the cure rate of first-line PPI-AMPC-CAM triple H. pylori eradication therapy and CYP2C19 genotype. Of a total of 5,318 patients with resistance rates of 13.1% for CAM and 8.9% for AMPC, the

Acid Inhibition and H. pylori Eradication Therapy
An optimal intragastric pH during prolonged potent acid inhibition in the stomach increases the stability and bioavailability of acid-sensitive antibacterial agents (e.g., CAM and AMPC) (Grayson et al., 1989;Hunt, 1993). This in turn leads to an increase in the gastric mucosal concentration of antimicrobial agents and exhibits antibacterial effects (Grayson et al., 1989;Goddard et al., 1996;Scott et al., 1998). In addition, acid inhibition increases the susceptibility of H. pylori to antimicrobial agents (Scott et al., 1998). Therefore, rapid, potent and prolonged neutralization of pH after optimal PPI treatment is required to cure H. pylori infection and increases eradication rate. We previously showed, using a PPI/AMPC/CAM regimen, that the median 24-h pH during eradication therapy was higher (6.4) and the median pH < 4 holding time ratio (HTR) (0.5%) was shorter in patients who experienced successful eradication than those who experienced failed eradication (pH 5.2 and pH < 4 HTR 26.7%) . Because the degree and duration of acid inhibition correlates with the cure rate, pH > 4 should be maintained for 24 h, and a target pH higher than 6.0 is needed for acid inhibition to cure patients using PPI/AMPC/ CAM therapy, irrespective with infection of CAM-resistant strain.
A critical subsequent question then becomes: how do we plan treatment using acid inhibitory drugs to achieve this target pH for acid inhibition? Controlling pH with PPIs depends on the dosing schedule, dose, and a combination of acid inhibitors (Shirai et al., 2001;Shirai et al., 2002;Sugimoto et al., 2004;Sugimoto et al., 2005;Sugimoto et al., 2012). Further, polymorphisms in genes encoding drug-metabolizing enzymes and drug transporters such as CYP2C19 and multidrug resistance protein-1 (ABCB1) affect pH during treatment (Furuta et al., 1999;Shirai et al., 2001;Shirai et al., 2002;Sugimoto et al., 2004;Sugimoto et al., 2005;Kodaira et al., 2009;Sugimoto et al., 2012). In general, the pharmacokinetics and pharmacodynamics of PPIs significantly differ among CYP2C19 genotypes. A previous meta-analysis reported significantly different cure rates of PPI therapy in patients with gastroesophageal reflux disease, a major acid-related disease, with different CYP2C19 genotypes (ITT analysis: EMs, 52.2%; IMs, 56.7%; PMs, 61.3%; p 0.047), and that CYP2C19 EMs had an increased risk of being refractory to PPI therapy compared to PMs (OR: 1.661, 95% CI 1.023-2.659, p 0.040) . Therefore, it is necessary to note that some patients who become resistant to PPI treatment may be CYP2C19 EMs. It may be possible to overcome this disadvantage of genetic variations affecting response to PPIs by using frequent dosing of PPIs, which has been shown to be more efficacious for acid inhibition through 24 h than a onetime increase in dose (Furuta et al., 2001c;Sugimoto et al., 2004;Lou et al., 2009). This phenomenon is due to the fact that frequent dosing sustains plasma levels of PPIs through a 24-h period and continues to inactivate H + , K + -ATPase consistently for 24 h (Furuta et al., 2001c;Sugimoto et al., 2004;Lou et al., 2009). However, because acid inhibition by omeprazole, lansoprazole, rabeprazole and esomeprazole bid in CYP2C19 EMs is 5.0 (2.4-5.9), 4.7 (3.7-5.5), 4.8 (2.5-6.4) and 5.4 (3.5-6.8) (Sahara et al., 2013), respectively, maintaining acid secretion for 24 h in all patients may be difficult. In fact, a previous metaanalysis conducted in 2013 showed that CYP2C19 EM is a risk factor for eradication (Tang et al., 2013). In the present study, we re-evaluated the effect of CYP2C19 genotype on the cure rate of first-line PPI-AMPC-CAM eradication therapy by analyzing RCTs written in English and demonstrated that the cure rate was 77.7% in CYP2C19 EMs, 81.2% in IMs and 86.8% in PMs, and that the RR of failed eradication in CYP2C19 EMs compared with IMs and PMs was 1.21 and 1.57, respectively. Because the incidence rates of antimicrobial resistance change year-by-year, determining the antibiotic susceptibility of H. pylori using either culture or genetic testing or both is of great importance, particularly in populations with a high rate of drug-resistant strains. Therefore, we think our finding that CYP2C19 genotype has a major effect on patients' response to first-line PPI-AMPC-CAM therapy has great clinical relevance.
In 2015, the potassium-competitive acid blocker, vonoprazan, became clinically available in Japan. Vonoprazan competitively inhibits H + /K + -ATPase activity more firstly and potently than PPIs (Kagami et al., 2016). Therefore, a vonoprazan-containing eradication regimen is expected to increase the eradication rate compared with PPI-containing conventional regimens. In fact, a recent meta-analysis that investigated the efficacy of first-line eradication therapy showed that a vonoprazan-containing regimen achieved a Frontiers in Pharmacology | www.frontiersin.org October 2021 | Volume 12 | Article 759249 10 higher eradication rate than a PPI-containing regimen (88.1% (95%CI: 86.1-89.9%) in the vonoprazan therapy and 72.8% (95%CI: 71.0-75.4%) in PPI therapy) (Jung et al., 2017;Sugimoto and Yamaoka, 2018). Because eradication rate for patients infected with CAM-sensitive strain is similar between vonoprazan and PPI, a benefit by vonoprazan-induced potent acid inhibition may play in patients infected with CAMresistance strain. Therefore, as the first-line standard treatment for H. pylori eradication in Japan, a vonoprazancontaining regimen is selected. However, there is no study to investigate the efficacy and safety of vonoprazan-AMPC-CAM therapy among CYP2C19 genotypes by the RCT. Further trials should be planned to clarify association with vonoprazan and CYP2C19 genotypes and association with PPI/vonoprazan and other genetic variation including CYP3A4 (Sugimoto et al., 2020).

Class of PPI and H. pylori Eradication Therapy
PPIs can be divided into first generation (e.g., omeprazole and lansoprazole) and second generation (e.g., rabeprazole and esomeprazole), which differ in the degree to which they are metabolized by CYP2C19. Rabeprazole is metabolized to thioether-rabeprazole mainly via a non-enzymatic pathway, with minor contribution from CYP2C19 (Yasuda et al., 1995). Esomeprazole is a pure S-isomer of omeprazole and interindividual variations in plasma concentrations of the S-isomer are lower than those in the R-isomer. In this metaanalysis, we found that the cure rates of omeprazole-, lansoprazole-, rabeprazole-and esomeprazole-AMPC-CAM therapy were similar. However, for first-generation PPIs, the RR of failed eradication in CYP2C19 EMs compared with PMs was 1.66 (95% CI 1.12-2.46) for omeprazole and 2.47  (1.44-4.23) for lansoprazole. In contrast, there were no significant differences in cure rates of regimen using second-generation PPIs among genotypes. In addition, because second generation PPIs are more effective for eradication therapy than first generation PPIs, second-generation PPIs or vonoprazan should be selected to ensure high efficacy in all patients, irrespective of CYP2C19 genotype.

CAM Resistance and H. pylori Eradication Therapy
There is growing evidence that the cure rate of PPI-AMPC-CAM regimen decreases with an increase in CAM-resistant strains, and recent rates of CAM-resistant strains in Japan and Europe have exceeded 35 and 20%, respectively (Furuta et al., 2001a;Asaka et al., 2001;Murakami et al., 2002;Megraud, 2007). Although previous studies have reported that cure rates in patients infected with CAM-resistant strains have markedly decreased to 20-40% (Fischbach and Evans, 2007;Megraud, 2007), this meta-analysis obtained cure rates of 89.7% (95% CI 88.3-91.0%) and 42.1% (95% CI 36.8-47.6%) in patients infected with CAM-sensitive and resistant strains, respectively. The Maastricht V/Florence Consensus Report recommends first-line therapy using a CAMcontaining regimen with PPI/AMPC or PPI/MNZ and an alternative treatment using bismuth-containing quadruple treatment in areas with low prevalence of CAM-resistant strains or bismuth or non-bismuth quadruple treatment and concomitant therapies in areas with high (>15%) CAM resistance (Malfertheiner et al., 2017). This indicates the need to carefully select antimicrobial agents and/or regimens based on individual's antibiotic resistance to H. pylori and/or known regional characteristics.
Recent studies have demonstrated the efficacy of tailored treatment based on sensitivity to CAM, with eradication rates exceeding 90% Kawai et al., 2008;Sugimoto et al., 2014). Although it is unfeasible to provide tailored treatment based on sensitivity to CAM to all patients with H. pylori infection or all areas/populations around the world due to limitations in culture testing and PCR in clinical practice, tailored treatment based on sensitivity to CAM should be considered in areas with a high incidence rate of CAM-resistant H. pylori.
Because eradication rate for patients infected with CAMsensitive strain is similar between vonoprazan and PPI (Jung et al., 2017;Sugimoto and Yamaoka, 2018), insufficient acid inhibition by PPI in CYP2C19 EMs may reduce eradication rate for CAM-resistance H. pylori strain. Therefore, different efficacy among CYP2C19 genotypes shown by our metaanalysis may depends on rate of CAM-resistance strain in each study. This observation will support hypothesis that potent acid inhibition (e.g., PPI high-dose, PPI qid with high adherence, vonoprazan and tailored PPI therapy based on CYP2C19 genotype) is important to be high eradication rate in patients infected with CAM-resistant strain Sugimoto et al., 2014). In fact, despite the dual therapy of vonoprazan/AMPC, it was reported to show a high eradication rate in patients infected with CAM-resistant strain (Suzuki et al., 2020). Although this observation suggests that when vonoprazan and AMPC are selected for eradication therapy, CAM may be unnecessary. Therefore, we think that culture test and susceptibility-based treatment provides best outcome for H. pylori eradication therapy. Although potency of PPI differs among kind of PPIs and resistant rate to CYP2C19 differs among kind of PPIs, acid inhibition by PPI also depends on the dosing schedule (oid, bid and qid) and dose. Therefore, I should be evaluated to determine optimal eradication regimen by combined analysis with various factors (dosing schedule, dose, genetic factor, potency, and generation) that affect intragastric pH at the time of PPI administration, as further study.

Limitations
This meta-analysis has a few limitations. First, there is a possibility of selection bias due to the exclusion of studies published in languages other than English, those with abstracts alone and studies that used non-RCT designs. Second, only one RCT has examined the association of CYP2C19 genotype with cure rate outside of East Asia. Third, the small sample size may have affected the statistical power of sub-group analyses. Because background (e.g., race and sex), treatment (e.g., class of PPI, regimen, dose of PPI, kinds of antimicrobial agents) and bacterial factors (e.g., resistance to antimicrobial agents and virulence factors) are associated with the outcome of eradication therapy, future meta-analyses of a larger number of RCTs should account for these factors to gain a clearer understanding of the association of cure rate with CYP2C19 genotype. Forth, because there were a few reports investigating effect of CYP2C19 genotype between clarithromycin-sensitive and resistant strain, we cannot meta-analyze associations in this time.

CONCLUSION
Cure rates of first-line PPI-AMPC-CAM eradication therapy differ among CYP2C19 genotypes, especially in Asian populations. This meta-analysis suggest that PPI-AMPC-CAM therapy has the potential to eradicate H. pylori infection in >85% of CYP2C19 PMs, but a low proportion of CYP2C19 EMs. Thus, the cure rate of this regimen in CYP2C19 EMs may be insufficient to eradicate H. pylori infection. Therefore, genotyping patients for CYP2C19 variants before eradication therapy may help to achieve higher eradication rates. To receive eradication rate >90% is important in clinical practice (Graham et al., 2007). However, most of regimen cannot reach to this goal, due to insufficient acid inhibition during treatment and susceptibility to antimicrobial agents. Therefore, second generation PPI-high-dose, highfrequent (qid) or vonoprazan-containing tailored treatment based on susceptibility to antimicrobial agents may optimal to ignore CYP2C19 genotype status and to receive high eradication rate.