Comparison of the grading of esophageal varices by transnasal endoluminal ultrasound and esophagogastroduodenoscopy.

Background & Aims: Esophagogastroduodenoscopy (EGD) is the current standard for evaluating esophageal varices, yet there is no universally accepted system of grading varices endoscopically and several studies have shown interobserver variability of endoscopic grading. High-resolution endoluminal ultrasound 20 MHz (HRES) has been shown to detect varices accurately and can be performed without sedation. Our aim was to compare the detection of esophageal varices by transnasal HRES and EGD.

Methods: We prospectively studied 37 cirrhotic patients being evaluated for esophageal varices. HRES was used to measure the largest esophageal variceal diameter and then EGD was performed. Photographs were taken for both procedures. Three blinded investigators graded the EGD photographs at 2 periods. End points were as follows: (1) the correlation of HRES variceal diameter and EGD grading of varices, and (2) the assessment of interobserver and intraobserver variation in varix grading by EGD.

Results: The correlation between the transnasal HRES and EGD was .63, with a 95% confidence interval of .37-.80. The HRES detected early varices that were not seen by EGD. The interobserver correlation for EGD scoring was .87 or greater in all comparisons and the intraobserver correlation was .91 or greater.

Conclusions: In conclusion, there is a significant correlation between transnasal HRES size measurement of esophageal varices and EGD. HRES is much more sensitive in detecting early esophageal varices and may not require sedation, suggesting that it may be more tolerable to patients and is worth further study.