Splenomegaly is obvious in portal hypertension, but controversy still exists over the relationship between splenic size or size of esophageal varices. Previous methods to assess spleen size are less accurate than ultrasonic estimation of spleen size by splenic volumetric index (SVI). In a prospective study, we evaluated 101 consecutive patients with portal hypertension for spleen size measured ultrasonically by SVI, presence and size of esophageal varices, and etiology of portal hypertension. A total of 219 age-matched controls were evaluated ultrasonically to define a normal SVI. Splenomegaly defined by 1 or 2 standard deviations of normal SVI had high accuracy in predicting portal hypertension, presence of esophageal varices, and provided a clue to the etiology of portal hypertension. However, there was no correlation between spleen size and size of esophageal varices.