• Antibiotics—clinical trials
• gastric diseases
• Helicobacter pylori—treatment

We greatly appreciated Dr Gatta and colleagues'1 interest and valuable comments to our study published in this journal.2 Our response to their comments were as follows. First, 7 out of the 1173 patients were excluded because of positive culture alone in the screening for eligibility. Second, the successful rate of culture in naïve patients was approximately 74.5% in this study. Although the specificity of culture was high (near 100%), the sensitivity was usually lower (ranged from 55% to 90%) in the literature.3 Factors that might affect the sensitivity of culture for Helicobacter pylori included (1) a delay in transporting the specimen to the laboratory, (2) insufficient number of biopsied specimens, (3) exposure of the bacteria to an aerobic environment, (4) the presence of gastric atrophy and (5) recent use of antibiotics, proton pump inhibitor, H2-blocker, etc.4 Third, we agree with Gatta and colleagues that the potential bias could not be excluded in the interpretation of the influences of antibiotic resistance on the eradication rates because the susceptibility tests were available in only two-thirds of patients. However, the eradication rates were not significantly different according to the availability of positive culture in this study. In group A, the eradication rates were 75.3% (122/162) and 70.9% (39/55) among patients with and without positive culture (p=0.519), respectively. In group B, the eradication rates were 83.1% (133/160) and 85.5% (47/55) among patients with and without positive culture (p=0.686), respectively. The use of molecular methods to detect mutations in the gyrase A and 23S RNA might allow us to detect the antibiotic resistance using the gastric biopsy specimens without performing culture. Nevertheless, a recent study showed a relevant discordance between the phenotypic and genotypic analysis in the prediction of eradication rate after clarithromycin-based therapy.5 The impact of genotypic and phenotypic resistance to levofloxacin on the treatment responses of levofloxacin-based therapy deserves our further investigation. Fourth, the second endoscopy and gastric biopsy (before second-line therapy) were not required in this study. Because the intervals between the first and the second endoscopies were as short as 2 months, some patients might drop out from this study simply because they do not want to receive the second endoscopy. This might in turn lead to some bias because the reason of dropout in these patients was their unwillingness to receive the second endoscopy rather than unwillingness to receive the second-line therapy. Therefore, the second endoscopy was not required in our study protocol. However, patients were encouraged to receive the second endoscopy before the second-line therapy. In this study, 83.3% (55/66) of the patients received the second endoscopy, and the positive culture rate was 52.7% (29/55). Fifth, we assumed the eradication rate of levofloxacin-based triple therapy to be 90% in the sample size calculation. We also assumed a dropout rate to be 10% and estimated a sample size of 440 patients. However, the actual dropout rate in this study was only 5.8% and, therefore, a total of 432 patients were adequate to give a power of greater than 80%. Finally, the sample size in the second-line therapy should be 266 to give a power of 80%. This meant that approximately 1330 patients might be needed in this study (assumed the average eradication rate to be 80% in the first-line therapy). Because the 60% eradication rate is clinically unacceptable in the treatment of H pylori infection, we considered that clarithromycin-based triple therapy is not suitable as a rescue regimen for those who failed from levofloxacin-based therapy. However, we agree with their opinions that cautions should be taken in the interpretation of the results in the second-line therapy because the power to detect the difference between the two therapies in the second-line therapy was low.