Patient factors predictive of 24-h pH normalization following endoluminal gastroplication for GERD.

摘要

BACKGROUND: Endoluminal full-thickness gastroplication has been documented to provide significant and long-lasting improvement of GERD symptoms and health-related quality of life (HRQL) with very little patient morbidity. These treatments, however, are criticized for normalizing esophageal acid exposure in only 30-40% of patients treated. We hypothesize that there are objective criteria that will identify those patients who will have a normal DeMeester score (DMS) following endoluminal treatment. METHODS: Data from a prospective multicenter trial using the NDO Plicator device to treat GERD were available for statistical analysis. All patients were treated with endoluminal full-thickness gastroplication. All patients had GERD symptoms and abnormal 24-h pH exposure preoperatively. Postoperative objective outcome was assessed by performing 24-h pH studies at 6 months. Univariate and multivariate regression analyses were performed to determine factors predictive of successful treatment (normalized 24-h pH). RESULTS: A total of 266 patients were included in the study. Mean preoperative DMS was 47.91 (+/-31.34). Postoperatively, mean DMS decreased significantly (37.11 +/- 24.63, p < 0.001), and 31.67% of patients had a DMS within normal range (DMS < 22). Results of multivariate regression analysis demonstrated that the following preoperative patient characteristics were predictive of postoperative success (normal DMS): DMS < 30 (odds ratio [OR] = 4.24, 95% confidence interval [CI] = 1.73, 10.36, p < 0.001), heartburn score < 2 (OR = 3.37, CI = 1.44, 7.89, p = 0.005), and BMI < 30 (OR = 4.93, CI = 1.55, 15.61, p = 0.007). CONCLUSION: Data analysis from this prospective study indicates that the odds of objective success would be significantly greater if the treatment was restricted to thinner patients with mild reflux disease. This may help define the optimal place for endoluminal therapy in a comprehensive GERD treatment algorithm.