Study of the NDO Endoscopic Plication System For the Treatment of Symptomatic Gastroesophageal Reflux Disease

The primary objective for this study was to measure the reduction in GERD symptoms as evidenced by analysis of the GERD-HRQL questionnaire at 3-months post-procedure with an objective of achieving a 50% or greater improvement. The trial was powered to detect a 50% reduction in mean GERD-HRQL at 3-months using a one-sided t-test with an α of .05 and a β level of .10, testing versus the equality of means. The calculation referenced above includes the added assumption that the standard deviation will be no more than 20 percent. The null hypothesis stated that the mean percent reduction in GERD symptoms was less than or equal to 50 percent at 3-months versus the alternate hypothesis that it was greater. The device treatment was considered a success if the statistical test rejected the null hypothesis at a one-sided p-value of 0.05 or less. Primary endpoint success was thus related to the statistical conclusion that the mean percent reduction was greater than 50%. A one-sided 95 percent confidence interval was constructed for the percent reduction in GERD symptoms. In order to assess the data with an "Intent to Treat" spirit, the number of patients who achieved a 50% reduction was analyzed as a fraction of the total number of patients who received the treatment. For secondary endpoint measures, statistical tests for medians were based on a Wilcoxon sign rank test of the percent improvement in a given study measure. This was based on the paired patient data of the pre-treatment scores versus the 6-month scores. The issue of multiple statistical tests of hypothesis being performed on data arising from individual patients was addressed in the following way. The comparison of GERD-HRQL scores was taken as the main results for which no correction of significance level was necessary. To recognize multiple testing using the method of Bonferroni, statistical significance was claimed for the secondary results only if, for a single test, the nominal p-value was \<.01. Given that some patients, during the course of the clinical study, were lost to follow-up, all outcomes were examined using the method of last visit carried forward, provided that at least one follow-up visit had been completed. It should be noted that using this method had little impact on the results; as compared to an analysis of the data of just those patients who completed follow-up, excluding those who missed the visit or were lost to follow up. However, this method was employed to account for those patients who were lost to follow-up, with specific consideration to those who had been lost to follow-up due to unsatisfactory treatment effect. Means and standard deviations are reported using the mean (SD) format, medians and interquartile ranges are reported using the median (IQR) format.