When it is not possible to perform a liver biopsy in cirrhosis, one has to rely on suggestive clinical and biochemical parameters and imaging procedures to arrive at a presumptive diagnosis. Based on the statistical method described by Spiegelhalter and Knill-Jones, we have devised a scoring system to reliably differentiate cirrhotic from non-cirrhotic portal hypertension without a liver biopsy. Age, presence of ascites, liver scan result and serum albumin, taken together, could confirm or rule out the diagnosis of cirrhosis in two-thirds of patients with portal hypertension. A score of 5 or more suggested cirrhosis (sensitivity 78%), and of 5 or more suggested cirrhosis (sensitivity 78%), and a score of -6 or less suggested a non-cirrhotic cause (sensitivity 64%) for portal hypertension, both with 100% specificity. Eliminating liver scan result reduced both sensitivity and specificity, suggesting that liver scan is an important component of the score. This score may be useful not only in the management of individual patients but also to classify them in clinical trials.