To determine whether liver damage correlates with typical changes in alpha 1-acid glycoprotein (alpha 1-AGP) carbohydrate branching, we selected patients with liver disease determined by histological liver findings. The severity of their illness was assessed by a clinical classification depending on the presence singly or together of four clinical complications (jaundice, ascites, hepatic encephalopathy and weight loss). An alpha 1-AGP crossed immunoelectrophoresis with Concanavalin A, an easy-to-perform method, revealed 3 or 4 subpopulations, the areas of which were calculated. A ratio R was determined as the most anodic peak area relative to the other ones. In our experimental conditions a ratio R value exceeding 1 correlated with the presence of one or more clinical complications. These results, evidencing fluctuations in the proportion of the carbohydrate variants of alpha 1-AGP, lead us to propose such a ratio as an index for grading liver damage. The sensitivity of this test was 86% and its specificity was 83%.