S0414 A Predictive Model for LES Location Independent of Manometry

Abstract

INTRODUCTION: In current practice, multichannel intraluminal impedance (MII)/pH testing must be preceded by manometry to determine LES location for accurate placement of the impedance catheter. However, when solely used to localize LES for MII placement, manometry adds increased cost, time, and patient dissatisfaction without clinical value. A non-invasive predictive model for LES location based on clinical demographics could obviate the need for manometry. We aimed to create a pragmatic, predictive model for LES location based on routine clinical parameters. METHODS: A retrospective chart review was performed on patients who underwent esophageal manometry from 2007–2018 at a single, tertiary medical center. Pre-selected variables of baseline demographics were evaluated in correlation with manometric parameters of proximal LES location, LES length, esophageal length, presence and size of hiatal hernia. Multivariable linear regression was used to estimate the association between predictors and proximal LES location. Spearman's rank correlation was used to determine potential predictive power to prevent overfitting (Figure 1). Missing data on predefined variables was minimal; for missing covariates multiple imputations was used. RESULTS: 465 patient records were analyzed. Proximal LES location ranged from 29–51 cm. Height and gender were the most important predictors of LES location (P < 0.001). Height increased distance to LES location by 2 cm for every 15 cm increase in height. Male gender predicted an increased distance to LES location by 2cm. Age and weight had no significant impact on LES location. The predictive equations for LES location were 18.7 + 0.136*height (males) and 17.2 + 0.154*height (females) (Figure 2). CONCLUSION: This is the first predictive model for LES location using readily-accessible, non-invasive patient characteristics of height and gender. This simple, pragmatic model prevents the need for costly, invasive manometry and allows accurate location of the LES. Future studies are ongoing for prospective validation.Figure 1.: Prediction coefficients and association of clinical parameters on LES location as calculated by multivariable linear regression. (Significance at P < 0.01).Figure 2.: Prediction Model: Predicted LES location as function of height and gender. (LES, Lower Esophageal Sphincter).