Noninvasive Prediction of Erosive Esophagitis Using a Controlled Attenuation Parameter (CAP)-Based Risk Estimation Model.

Abstract

Erosive esophagitis and fatty liver share obesity and visceral fat as common critical pathogenesis. However, the relationship between the amount of hepatic fat and the severity of erosive esophagitis was not well investigated, and there is no risk estimation model for erosive esophagitis. To evaluate the relationship between the amount of hepatic fat and the severity of erosive esophagitis and then develop a risk estimation model for erosive esophagitis. We enrolled 1045 consecutive participants (training cohort, n = 705; validation cohort, n = 340) who underwent esophagogastroduodenoscopy and CAP. The relationship between severity of fatty liver and erosive esophagitis was investigated, and independent predictors for erosive esophagitis that have been investigated through logistic regression analyses were used as components for establishing a risk estimation model. The prevalence of erosive gastritis was 10.7 %, and the severity of erosive esophagitis was positively correlated with the degree of hepatic fatty accumulation (P < 0.05). A CAP-based risk estimation model for erosive esophagitis using CAP, Body mass index, and significant alcohol Drinking as constituent variables was established and was dubbed the CBD score (AUROC = 0.819, range 0-11). The high-risk group (CBD score ≥3) showed significantly higher risk of having erosive esophagitis than the low-risk group (CBD score <3) (24.1 vs. 2.7 %, respectively; P < 0.001). The diagnostic accuracy of CBD score was maintained in the validation cohort (AUROC = 0.848). The severity of erosive esophagitis was positively correlated with the degree of hepatic fatty accumulation, and the CBD score might be a simple CAP-based risk model for predicting erosive esophagitis.