Background: Endoscopic ultrasonography (EUS) has been suggested to investigate portal hypertension (PHT). We compared EUS and endoscopy in the exploration of PHT in cirrhosis.
Methods: In 58 patients with cirrhosis and 16 controls, the presence of esophageal varices and gastric varices was recorded by both endoscopy and EUS, peridigestive varices and perforating veins by EUS. Relationships of patterns with each other and with clinical and biologic data were evaluated.
Results: In patients with cirrhosis: esophageal varices were found by endoscopy and EUS in 88% and 55% of cirrhotic cases (p < 0.01), and gastric varices in 17% and 41%, respectively (p < 0.01). Perforating veins, below the gastroesophageal junction, were diagnosed by EUS in 40% of cases and never in controls. A significant relationship was found between perforating veins and the following patterns: peri-digestive varices, gastric varices at EUS, and esophageal varices at endoscopy. In patients with and without cirrhosis: after multivariate analysis, only esophageal varices at endoscopy were independently related to cirrhosis. However, in Child-Pugh class A patients, both esophageal varices and perforating veins were independently related to cirrhosis (diagnostic accuracy, 92%).
Conclusions: EUS is of limited value in the diagnosis of cirrhosis because it gives little additional information to upper gastrointestinal endoscopy. The relationships between perforating veins, varices, and peridigestive varices support the hypothesis that perforating veins are important in the pathogenesis of esophageal and gastric varices.