Rebamipide does not protect against naproxen-induced gastric damage: a randomized double-blind controlled trial

Volunteers of both sexes aged 18 years old or older without any significant cardiac, hepatic, renal, pulmonary, neurological, gastrointestinal or hematological diseases, as determined by their medical history, physical examination, and routine laboratory tests (hematology, blood biochemistry, urine analysis and fecal occult blood test), were invited to participate in this double-blind, randomized, parallel placebo-controlled phase II single-center trial. All subjects tested negative for hepatitis B and C (except for serologic scar), as well as HIV I and II, and were instructed to abstain from taking any medication including over-the-counter medication for 2 weeks prior to and during the study period. Pregnancy was an exclusion criterion. All women enrolled had negative β-HCG. All volunteers were informed about the aim and risks of the study by the clinical investigator and they all signed a written informed consent statement before entering the study. The study was performed according to the 2008 revised Declaration of Helsinki for bio-medical research involving human subjects and the 1996 rules of Good Clinical Practices. The study protocol was approved by the Committee of Research Ethics of the University of Sao Paulo, Sao Paulo, Brazil, on 26 November 2014 (before the initiation of the study procedures), and the study is registered at ClinicalTrials.gov, ID NCT02632812.

After performing a preliminary gastroduodenal endoscopy at the Vera Cruz Hospital (Campinas, Brazil) to rule-out prior macroscopic upper GI damage in the subjects, twenty-four healthy volunteers (12 men) were enrolled in the study. Other exclusion criteria were: achlorhydria (pH > 6.5); positive fecal occult blood test; drug abuse, including alcohol or tobacco.

Volunteers were randomly divided into 2 groups: group A received 550 mg of sodium naproxen (Flanax® - Bayer) plus an effervescent placebo (Biolab Indústria Farmacêutica Ltda.) diluted in 200 mL of water b.i.d. for 7 consecutive days; group B received 550 mg of sodium naproxen (Flanax® - Bayer) plus 100 mg of effervescent rebamipide (Biolab Indústria Farmacêutica Ltda.) diluted in 200 mL of water b.i.d. for 7 consecutive days. The randomization sequence was determined using the randomization generator available at www.​randomization.​com, using 3 blocks of 8 subjects each. Volunteers were blind to whether they were receiving rebamipide or placebo. To assure adherence, the drugs were administered by a member of the research team (also blind to treatment allocation) early in the morning and late in the evening for the whole duration of the study. A pharmacist was responsible for labeling treatments as A or B, and for breaking the code at the end of the study. Volunteers were required to be fasted in the morning and were only allowed to eat at least 2 h after administration of medications. In the evening, volunteers were required to be at least for 2 h without ingestion of food, and were also asked to avoid eating for the following 2 h after the dose.

Adverse reactions were individually evaluated. Causality relationship to the treatment was assessed with the aid of the Karch & Lasagna algorithm [19]. The necessity of interruption of treatment or specific therapy was individually assessed by the Principal Investigator depending on the severity of the reaction and the causality relationship to the treatment.

Gastroduodenal endoscopy was performed before the start of the study and in the morning of the 8th day of treatment in all volunteers. Exams were performed by a single endoscopist. The primary outcome measure of the trial was the macroscopic damage observed in each treatment group at the end of treatment. The number of mucosal erosions after treatment was counted and macroscopic mucosal injury scored according to Cryer [20] and to the modified Lanza score (MLS; Table 1) [9, 21]. Gastroduodenal ulcers were classified according to the Sakita-Miwa classification. Additionally, a second endoscopist retrospectively evaluated the exams, blind to the scores given by the first endoscopist. If there was any difference between the score given by the two endoscopists, the final score would be the mean value of the two numbers. Both endoscopists were blind to treatment group allocation of volunteers. In each exam 3 biopsies were collected from the gastric antrum and 3 from the gastric corpus for PGE2 quantification. Samples were immediately washed in phosphate buffer saline at pH 7.4 and stored individually in a dry Eppendorf tube and kept at −20 °C until analysis. Another 2 biopsies from each stomach region (4 total) were collected at each endoscopic procedure for histological analysis. Volunteers were questioned about GI symptoms at every dose. Changes in gastric PGE2 concentration, histopathological characteristics and GI symptoms at the end of the study compared to baseline were evaluated as secondary outcome measures. A more detailed questionnaire about GI symptoms such as heartburn, abdominal pain, fullness, nausea, vomiting, blood in stools or others was completed before receiving the first dose and at the last day of treatment. In the questionnaires, when any symptom was reported, the volunteer was asked to describe the intensity and frequency of the symptom over the previous 7 days. Routine laboratory tests were repeated at the end of treatment, including fecal occult blood test.

Tissue PGE2 concentration was quantified by enzyme-linked immunosorbent assay (ELISA) using Cayman Chemical Monoclonal Prostaglandin E₂ EIA Kit (item number 514010). Briefly, biopsies from the gastric antrum and corpus collected in both endoscopic procedures were weighted and homogenized individually for 20 s by a Polytron in 1 mL of Na₃PO₄ buffer solution (pH = 7.4). Aliquots of 200 μL of the homogenate were placed in a dry bath at 37 °C for 20 min and then centrifuged at 20,800 g for 30 s. Supernatant (150 μL) was collected and diluted 1:100 (v/v) to be used in the ELISA. Each aliquot was assayed in duplicate and final PGE2 concentrations were adjusted for initial sample mass. Data are displayed as mean ± SEM. The person responsible for the quantification of PGE2 was blind to treatment group allocation of volunteers.

Biopsies from gastric antrum and corpus were immediately put in formaldehyde after collection. Samples were stained in hematoxylin and eosin for characterization and graded according to a score described in Table  2 . Giemsa stain was used to diagnose H. pylori infection. The pathologist was also blind to treatments.

For evaluation of statistical difference of tissue PGE2 between treatments, a t-test was performed. A paired t test was used to compare tissue PGE2 before and after treatments within each group. Endoscopic and histopathological scores between treatment groups were compared using the Mann–Whitney U-test. Gastrointestinal symptoms after treatments were compared by Fisher’s exact test. A p value <0.05 was considered significant. The sample size planned for the study was 24 volunteers of both sexes. This number was chosen based on the gastroprotective effects of rebamipide previously reported [9]. The calculation considered the Mann–Whitney U test analysis of the reported data on gastroduodenal damage scoring by MLS. The effect size observed (difference between treatments) was 67 % (median score after treatment of 3 and 1, in the placebo and rebamipide groups, respectively). Given a power of 80 % and a 0.05 chance of type 1 error, the original sample size estimation was 10 volunteers per group. The final number of 12 volunteers per group was chosen considering a 20 % rate of drop-out. Other studies with rebamipide in healthy volunteers also used similar sample sizes [10, 15].