O0088 COMPONENTS OF GASTROESOPHAGEAL REFLUX (GER) THAT CONTRIBUTE TO RESPIRATORY SYMPTOMS: A STUDY USING MULTICHANNEL INTRALUMINAL IMPEDANCE (MII) WITH PH

Abstract

Introduction: We have previously shown that the rate of symptom index positivity increases with the use of MII compared to pH probe in children on antacid medications. It is the aim of this study to determine which components of a GER episode contribute to symptoms. Methods: 28 children with atypical reflux symptoms (cough, stridor, wheezing, desaturations and tachypnea) on antacid medications were evaluated using the MII/pH for 24 hours. Impedance tracings were manually reviewed. Reflux episodes were correlated with a symptom if the reflux event occurred two minutes before or two minutes after a recorded symptom. Logistic regression was performed to determine which components of a reflux episode (acid content, bolus clearance time, height of the refluxate, and composition of the refluxate) predicted symptom association. Results: The mean age of the patients was 6.5±5.7 years. There were 10 females and 18 males. A total of 1822 reflux episodes were detected by either by pH probe or MII. 45% were non-acid reflux and 55% acid reflux. 75.4% of the episodes were liquid, 18.8% were mixed and 7.7% were gas reflux. 33.8% of the episodes reached the proximal esophagus, 16.2% reached the mid-esophagus and 50% reached the distal esophagus. 467 atypical symptoms were recorded by 18 patients; 10 patients had no symptoms during the study period. 141 of the symptoms were correlated with reflux episodes. In the univariate analysis, non-acid reflux (OR: 4.33, p<0.0001), liquid reflux (OR: 1.81, p=0.006) and proximal esophageal reflux (OR: 1.31, p=0.02) were the variables significantly associated with the outcome, reflux-associated symptoms. Controlling for all of the reflux characteristics in the final multivariate model, non-acid reflux (OR: 6.14, p<0.0001) and liquid reflux (OR: 1.84, p=0.01) were the only variables significantly correlated with symptoms; bolus clearance time and proximal reflux were not predictors of symptoms. Conclusion: Non-acid reflux is a significant predictor of symptom association even though more than 50% of the reflux episodes were acidic. Because non-acid reflux is not detected by pH probe, the use of MII may be important in the evalaution of children with atypical symptoms and may change the clinical management of patients since antacids do not reduce the amount of non-acid reflux.