Background: Endoscopic ultrasound (EUS) was used to evaluate cirrhotic patients with and without prior variceal hemorrhage. The findings were compared with those of EUS in noncirrhotic control subjects to determine EUS features indicative of cirrhosis and of a risk for variceal hemorrhage.
Methods: Patients with cirrhosis undergoing indicated endoscopic screening for varices or surveillance after endoscopic therapy for variceal hemorrhage were studied and compared with healthy noncirrhotic control patients undergoing EUS for benign conditions.
Results: Sixty-six cirrhotic patients (31 with prior hemorrhage) and 32 control patients were studied. Nonhemorrhage cirrhotic patients had more severe liver disease by Child's class (p = 0.02) and less beta-adrenergic blocker usage (p < 0.0001). Paraesophageal varices were detected in 97% of cirrhotic patients and 3% of control patients (p < 0.001) and were a more sensitive predictor of cirrhosis than varices at endoscopy (74%, p < 0.0001). Azygos vein and thoracic duct diameters, and gastric mucosa and submucosa thickness were greater for cirrhotic than control patients (p < 0.001). More hemorrhage patients had large (5 mm or greater) paraesophageal varices (odds ratio 3.1: 95% CI [1.1, 8.3]; p < 0.05) and paragastric varices (odds ratio 3.7: 95% CI [1.3, 10.2]; p = 0.01). Paraesophageal varix size correlated with ascites (p = 0.03) and, for nonhemorrhage patients, with Child's class (p < 0.01).
Conclusions: Paraesophageal and paragastric varices correlate with the presence and severity of liver disease and portal hypertension. These data support the hypothesis that large paraesophageal and paragastric varices (5 mm or greater) may be risk factors for variceal hemorrhage, an observation that merits further prospective study. Cirrhosis causes dilation of the azygos vein and thoracic duct and thickening of gastric mucosa and submucosa.