Background/Aim: To evaluate the optimal dosage of rabeprazole for proton-pump inhibitor (PPI) testing of gastroesophageal reflux disease (GERD) and to test the influence of cytochrome P (CYP) 2C19 polymorphism in a population with a high prevalence of people who poorly metabolize PPIs.

Patients and Methods: In this randomized, open-label trial, patients with symptoms suggestive of GERD were randomized to receive a two-week test with 20-mg or 40-mg rabeprazole after diagnostic endoscopy. Symptom response was assessed with a four-grade daily record; in addition, DNA from peripheral blood leukocytes was genotyped for CYP2C19 polymorphism with polymerase chain reaction-restrict fragment length polymorphism (PCR-RFLP) technique.

INTRODUCTION A broad spectrum of symptoms are commonly associated with gastroesophageal reflux disease (GERD), which has an incidence of roughly 20% in the general population.[1] Despite recent, substantial advances in understanding GERD's pathogenesis, diagnosis still presents many challenges. A rapid symptomatic response to proton pump inhibitors (PPIs) in patients with a presumptive diagnosis of GERD is useful to validate diagnosis, which is known as the PPI test. Studies addressing diagnostic PPI testing have produced valuable estimates for Western populations regarding prediction of GERD in individual patients.[2][3][4] Some population-specific features distinguish Asian patients from their Western counterparts. First, the majority (75-90%) of Asian reflux patients have endoscopy-negative reflux disease (ENRD).[1][5] Not all of them demonstrate a favorable response to PPI treatment because the pathogenesis of ENRD is in part associated with psychosomatic pathways.[6][7] Second, PPIs are eliminated by a hepatic route, with a polymorphically expressed cytochrome P (CYP) 2C19 primarily responsible for the rate of metabolism.[8] Compared with people who have a homozygous wild-type genotype, people with a variant CYP2C19 allele exhibit lower rates of PPI degradation, higher plasma PPI concentrations (3-to-13 times higher), and a lesser degree of gastric acid secretion.[9] The prevalence of people who are poor metabolizers (i.e., have a homozygous variant genotype) has been cited as 1.2% to 3.8% for Caucasian-European populations, but 12.6% to 22.5% for Asian populations.[10] Lower gastric parietal cell mass is also prevalent in Asians. These characteristics may augment the therapeutic effect of PPIs in Asian patients.

When considering the pros and cons of PPI testing in an Asian population, the major pro is that there will be a high prevalence of poor or intermediate metabolizers who may have an increased serum level of PPI, resulting in a higher sensitivity, lower required dose of PPI, and less cost in testing. The con is mainly related to the increased proportion of ENRD patients. Up to 50% of ENRD patients, whose condition is termed functional heartburn, report sufficient heartburn relief with PPI treatment.[6][11] Therefore, a higher proportion of false positive cases may occur, which can lead to a decrease in test specificity.

The prediction of intra-esophageal damage is of paramount importance because patients with erosive and non-erosive disease have distinctive manifestations and prognosis.[12] Therefore, in this study, we tested the hypothesis that the PPI test can be used as a valid tool for diagnosis of esophagitis in a Chinese population. Validation of the accuracy of PPI testing would provide important information for comparison of diagnosis with the traditional endoscopy-first approach. We selected rabeprazole as the PPI test agent because it is well tolerated and can effectively prevent pathological and symptomatic GERD relapse.[13][14] Another potential benefit is that a variable proportion of rabeprazole degradation proceeds through a non-enzymatic pathway.[15][16] Thus, a decrease in inter-individual variability in serum PPI level may prove the rabeprazole-based regimen to be a diagnostic test with stable accuracy.

PATIENTS & METHODS Patients A consecutive series of patients with symptoms suggestive of GERD were enrolled from the Gastroenterology outpatient clinic in our institution. The typical GERD symptom was defined as heartburn and/or acid regurgitation of at least three episodes per week for a minimum of three months. Patients who received concurrent PPI treatment, had a medical contraindication to rabeprazole therapy, reported a history of peptic ulcer disease or gastrointestinal surgery, peptic ulcer disease or malignancy proven by endoscopy, the presence of alarm features (e.g., dysphagia, weight loss, bleeding, abdominal mass, and/or anemia), or who were unwilling or unable to provide informed consent were excluded from the study. Participants provided informed consent, and the Ethics Committee of National Taiwan University Hospital approved the study protocol prior to implementation (no. 940711).

Study Protocol All enrolled patients underwent an initial diagnostic evaluation with upper endoscopy and were classified with erosive or non-erosive disease. After endoscopy, patients in the two groups were randomly assigned to receive either one tablet of rabeprazole 20mg before breakfast or two tablets of rabeprazole before breakfast and dinner for 2 weeks. The random allocation was performed by choosing cards in sealed envelopes. Patient response to PPI treatment was recorded in a diary (illustrated below). Patients were notified about endoscopy results only at study completion.

Endoscopy After an overnight fast, all subjects were placed in the left lateral decubitus position. Endoscopy was performed by experienced endoscopists at our academic institution. During the procedure, the stomach and duodenum were inspected to exclude possible lesions. The distal portion of the esophagus was evaluated carefully to determine the presence of any mucosal injury. The definition of reflux esophagitis rested on demonstration of diffuse or streaking erythema, mucosal friability, and the presence of erosions or ulcers, i.e., mucosal breaks. Severity was assessed according to the Los Angeles (LA) classification with standard comparator photos, a reference standard in diagnosis of GERD.[17]

Symptom Assessment Patients kept a daily diary record of the severity of their reflux symptoms for the 14 days of PPI therapy. The evaluation scale was adapted from the scale used in the Dickman trial [18] and was translated into Chinese. The scale was as follows: (1) mild symptoms that were easily tolerated and did not last long; (2) moderate symptoms that caused some discomfort but did not interfere with usual activities; (3) severe symptoms that caused much discomfort and interfered with usual activities; and (4) disabling symptoms that were unbearable and interfered considerably with usual activities.

Analyses of CYP2C19 Genotypes For genotype analysis, each patient's peripheral blood leukocytes were obtained and their DNA was extracted. Genotyping procedures that identified the CYP2C19 wild-type allele and the two known mutant alleles, CYP2C19m1 in exon 5 and CYP2C19m2 in exon 4, were carried out by polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) technique with allele-specific primers.[19] On the basis of their ability to metabolize PPI, individuals were classified as extensive metabolizers (homEM: homozygous for wild-type allele), heterozygous extensive metabolizers (hetEM: carrier of only one mutant allele), or poor metabolizers (PM: homozygous for variant allele). All laboratory procedures were performed blind to the patient's randomization status.

Statistical Analysis Quantitative data were summarized as mean ± standard deviation (SD), and categorical variables were summarized as percentages. Between study groups, categorical variables were compared with the 2 test, quantitative data with the independent t-test, and ordinal data with Wilcoxon's two-sample test. To identify the best cutoff value of symptom reduction for prediction of GERD, we compared three levels for a positive response (25%, 50%, and 75% symptom reduction from baseline) by calculating the areas under the receiver-operating curve (ROC) in a prediction model. In this model, the dependent variable was diagnosis of GERD, and a positive therapeutic response was treated as the main independent variable. Covariants included age, male gender, body mass index (BMI), and social habits (e.g., smoking and alcohol), all of which are known to be associated with risk of GERD.[1] According to the best definition for positive response, diagnostic parameters were calculated for sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy. To assess the effect of CYP2C19 polymorphic genotype, results were stratified and compared according to carrier status (e.g., homEM, hetEM, and PM). The Cochran-Armitage trend test assessed whether there was a therapeutic difference across the three genotype classes. All P values were two sided, and P < 0.05 indicated statistical significance. Analyses were performed with statistical software (SAS, version 8.0; SAS Institute Inc., Cary, NC).

Inclusion Criteria:

• A consecutive series of patients with symptoms suggestive of GERD were enrolled from the Gastroenterology outpatient clinic in our institution. The typical GERD symptom was defined as heartburn and/or acid regurgitation of at least three episodes per week for a minimum of three months.

Exclusion Criteria:

• Patients who received concurrent PPI treatment, had a medical contraindication to rabeprazole therapy, reported a history of peptic ulcer disease or gastrointestinal surgery, peptic ulcer disease or malignancy proven by endoscopy, the presence of alarm features (e.g., dysphagia, weight loss, bleeding, abdominal mass, and/or anemia), or who were unwilling or unable to provide informed consent were excluded from the study.