Inter- and intra-rater reproducibility of automated and integrated pressure-flow analysis of esophageal pressure-impedance recordings

Abstract

<h4>Background</h4>Automated impedance manometry pressure-flow analysis (AIM analysis) determines pressure measurements relative to bolus flow and has to date shown subtle variations in esophageal motility in relation to dysphagia. In this study, we assessed intra- and inter-rater reproducibility of AIM metrics derived using purpose designed software.<h4>Methods</h4>Fifty patients referred for evaluation of gastro-esophageal reflux symptoms (33 men, age 52 ± 1.9 years) underwent combined high-resolution impedance manometry and completed a dysphagia questionnaire. From 10 liquid and 10 viscous swallows, a subset of four swallows (two saline and two viscous) was systematically selected from each patient for manual and AIMplot analysis, which was performed twice by five observers (two experts, three non-experts). Intra- and inter-rater agreement were determined using intraclass correlation coefficients.<h4>Key results</h4>AIMplot-based analysis showed high intra-rater and inter-rater reproducibility for all metrics (mean ICCs of 0.95 and 0.94, respectively). Reproducibility of metrics derived for liquid and viscous did not differ (ICCs of 0.96 and 0.91 for liquid and viscous, respectively). In addition, metrics derived by experts had an equivalent level of reproducibility compared to non-experts (ICCs of 0.96 and 0.94, respectively). Variables that could be derived with commercial software (ManoView™) correlated highly with variables from AIMplot-based analysis, such as 4-s integrated relaxation pressure (r = 0.85) and the 20-mmHg isobaric contour defect (r = 0.92).<h4>Conclusions & inferences</h4>Esophageal AIM analysis is highly reproducible, independent of an observer's level of experience in esophageal motility. Therefore, AIM analysis produces data that are reliable for clinical and research purposes.