Accuracy of multichannel intraluminal impedance and pH testing for detection of gastro-oesophageal reflux (GOR) in anaesthetised dogs

Abstract

Objective To compare the sensitivity and specificity of pH with multiple intraluminal impedance (pH-MII), pH-metry (pH) alone and multiple intraluminal impedance (MII) alone to direct observation of GOR by endoscopy in anaesthetised dogs. We hypothesized that pH-MII is more sensitive and specific in detecting GOR in anaesthetised dogs compared to pH or MII alone. Study Design A prospective comparative trial in a live canine model Animals Thirty-five dogs (22 females, 13 males) of various breeds undergoing elective pelvic limb orthopaedic procedures. The mean (range) mass and age were 31.9 (14.0 to 40.0) kg and 5.6 (0.8 to 12.0) years, respectively. Methods All dogs were premedicated with medetomidine and morphine, anaesthesia was induced with propofol and maintained on isoflurane in oxygen. A monitoring assembly consisting of an endoscopy camera, endotracheal tube and a disposable flexible pH-MII catheter, was utilized to measure oesophageal pH, MII and directly visualise reflux. Visual reflux was assigned a score (0: none; 3: severe) and pH was recorded on a data capture sheet. Reflux was considered to have occurred whenever oesophageal pH was below 4.0 or above 7.5, device software analyzing MII data detected fluid refluxate or a visual reflux score of 2 or 3 were assigned. ROC analysis was used to determine sensitivity and specificity for each monitoring method to detect GOR. Area under the curve (AUC) was used to discern between an accurate method and non-accurate method (AUC ≤ 0.5), a method with poor accuracy (AUC 0.5-0.6), low accuracy (AUC 0.6-0.7), fair accuracy (0.7-0.8), good accuracy (AUC 0.8-0.9) and excellent accuracy (AUC ≥ 0.9). Results Endoscopy identified GOR in 20 dogs (57%), pH-MII in 19 dogs (54%), pH alone in 13 dogs (37%), and MII alone in 12 dogs (24%). As planned, the AUC for the ROC of endoscopy was 1.0 and demonstrated 100% sensitivity and specificity, respectively. AUC analysis for the ROC curve showed fair accuracy for pH-MII and pH alone. pH-MII and pH alone demonstrated a sensitivity and specificity of 69% and 76%, and 71% and 75%, respectively. While MII demonstrated low accuracy with a sensitivity and specificity of 98% and 24%, respectively. Prevalence for detecting GOR events per measured data point was greatest in endoscopy (35%), followed by pH-MII (25%), then pH (21%) with the least detected in MII (7%). pH-MII and pH alone exhibited almost perfect agreement. Conclusions and clinical relevance pH-MII is a reliable method for detecting GOR and emerges as a promising tool for future research. Endoscopy is reliable and provides the ability to subjectively quantify the volume of reflux, however, lacks the ability to discern the pH of the refluxate. pH alone misses reflux episodes with intermediate pH (4.1-7.4). Incorporation of impedance addresses some of the limitations associated with pH alone and enhances diagnostic accuracy.