We consecutively enrolled a total of 203 H. pylori-positive patients with chronic gastritis from September 2014 to August 2015 at Kyung Hee University Hospital, Seoul, Korea. The participants were excluded from the study if they had a PUD, gastric cancer, previous gastrectomy, severe underlying diseases, or received antibiotics or PPI within 4 weeks [15]. At the first time of enrollment, we tried to include patients with PUD and chronic gastritis, but the number of H. pylori-related PUD cases was much smaller relative to chronic gastritis. Therefore, we included only chronic atrophic or nonatrophic gastritis. All the participants received a diagnosis of H. pylori infection by a rapid urease test or detection of H. pylori genomic DNA from gastric mucosal tissues.

We collected data including age, gender, cigarette smoking, alcohol intake, salt in the diet, family history of gastric cancer, BMI, hypertension, and diabetes. Although we tried to acquire prior history of antibiotic use, most subjects did not remember the name or the amount used of antibiotics. All the patients were prescribed the triple regimen consisting of a standard dose of PPI (pantoprazole), 1 g of amoxicillin, and 0.5 g of CAM twice daily for 7 days. Four weeks after the end of eradication therapy, the ¹³C-urea breath test was performed to confirm successful eradication of H. pylori infection. A week after completion of treatment, information about adverse events and compliance was collected from the outpatient department by means of telephone interviews. When patients had consumed more than 85% of the study medications, compliance was considered satisfactory. The adverse events were metallic taste, diarrhea, nausea, epigastric discomfort, urticaria, and headache. All the patients provided written informed consent prior to enrollment. The Institutional Review Board of the Kyung Hee University Hospital approved the study protocol (KMC IRB 1427-01).

It is known that the mutation of 23S rRNA results in increased antibiotic resistance [16]. CAM resistance of H. pylori was determined by analysis of the A2142G or A2143G mutation in 23S rRNA. H. pylori genomic DNA was extracted from two biopsy specimens of the antrum and body collected during endoscopy. DNA was amplified by Seeplex Home-brew primer mix of the Seeplex ClaR-H. pylori polymerase chain reaction (PCR) kit (Seegene Inc., Seoul, Korea) that was developed using the dual-priming oligonucleotide system to detect the A2142G or A2143G mutation.

The genotyping of CYP2C19 was conducted using PCR with restriction fragment length polymorphism [17]. To detect the polymorphism of CYP2C19 in exon 5, we designed the sense (5′-AATTACAACCAGAGCTTGGC-3′) and antisense primers (5′-TATCACTTTCCATAAAAGCAAG-3′). In exon 4, we used the sense (5′-GCTTCACCCTGTGATCCCAC-3′) and antisense primer (5′-TCAGGGCTTGGTCAATATAG-3′). The PCR product containing G681A in exon 5 was digested by incubation with Sma I at 37°C for 3 hours, resulting in two bands (120 and 49 bp) for the G/G genotype (wild), three bands (169, 120, and 49 bp) for the G/A genotype (heterozygote), and one band of 169 bp for the A/A genotype (mutant type). The PCR product containing G636A in exon 4 was digested by incubation with BamH I at 37°C for 3 hours, resulting in two bands (155 and 93 bp) for the G/G genotype (wild type), three bands (248, 155, and 93 bp) for the G/A genotype (heterozygote), and one band of 248 bp for the A/A genotype (mutant type). The patients were classified into three groups by CYP2C19 genotype: RM had exon 5, wild type/exon 4, wild type; intermediate metabolizers (IMs) had exon 5, wild type/exon 4, heterozygote or exon 5, heterozygote/exon 4, wild type; and poor metabolizers (PMs) had both mutated alleles.

Eradication was evaluated by the ITT and PP analyses with 95% confidence intervals (CIs). Continuous variables are expressed as mean ± standard deviation, and categorical variables as percentages. Student t test for parametric data and the Mann-Whitney U test for nonparametric data were used for comparisons of the means of continuous variables, whereas differences in categorical variables were assessed using chi-square or Fisher exact test for demographic and clinical data. Logistic regression analyses were conducted to determine the independent contribution of each possible predictor to eradication failure. PASW Statistics for Windows version 22.0 (IBM Co., Armonk, NY, USA) was used for statistical analysis. All statistical tests were two-tailed, and p values < 0.05 were assumed to indicate statistical significance.

Thirteen cases among the total of 203 patients did not participate. After receiving CAM-based triple therapy, 10 cases were lost to follow-up, one case was non-compliant, and two cases discontinued treatment owing to severe adverse events. Eventually, 190 patients were analyzed (Fig. 1). The baseline characteristics of all the participants are summarized in Table 1.

The overall eradication rates of CAM-based standard triple therapy according to ITT and PP analyses were 64.0% (130/203 patients) and 68.4% (130/190 patients), respectively (Table 1). The frequency of adverse events is shown in Table 2. The most frequent adverse event was metallic taste (65/190, 34.2%). Diarrhea, nausea, epigastric discomfort, urticaria, and headache followed. The adverse events were not different between the success group and failure group. Among 60 patients with first eradication failure, 53 patients received 10 days of bismuth-containing quadruple therapy. In 48 patients, the H. pylori infection was eradicated with a quadruple therapy (48/53, 90.6%) (Fig. 1).

The rate of CAM resistance was 17.4% (33/190 cases) (Table 1). The A2143G mutation was predominantly observed (30/33 cases, 90.9%). Two cases (6.1%) had A2142G mutation, and one case had a dual point mutation at A2142G and A2143G. All patients with CAM resistance revealed eradication failure (Table 3). The frequency of CAM resistance was different between females (20/total 101 patients, 19.8%) and males (13/total 89 patients, 14.6%), but the difference did not reach statistical significance (p = 0.087).

Analysis of the CYP2C19 genotypes showed that the RM genotype was present in 40/190 cases (21.0%), IM genotype in 75/190 cases (39.5%), and PM genotype in 75/190 cases (39.5%). There were no statistically significant differences among the three genotypes in H. pylori eradication (p = 0.682) (Table 3).

The clinical characteristics of the eradication success group and failure group are shown in Table 3. In univariate analysis, there were no statistically significant differences in clinical parameters except for female gender. CAM resistance was detected in 33/190 (17.4%) of cases and their eradication rate was zero (0/33 cases). In multivariate logistic regression analysis, female gender (odds ratio [OR], 1.73; 95% CI, 1.15 to 4.251; p = 0.016) and CAM resistance (OR, 19.13; 95% CI, 9.354 to 35.094; p < 0.001) were identified as powerful predictive factors of eradication failure (Table 4).