InPATIENTS with liver disease one or more abnormalities of hemostasis can frequently be demonstrated in the laboratory.1 Although it is generally believed that these abnormalities predispose to hemorrhage, it has not been clearly established that hemostatic defects are more common or more severe in bleeding than in nonbleeding patients with liver diesase. The many comprehensive reports of hemostasis in liver disease2-7 have not systematically related laboratory tests to clinical hemorrhage. Similarly, previous studies of hemostatic defects in patients with bleeding8-11 do not present comparable data on nonbleeding cirrhotics. Correlation between laboratory tests and clinical bleeding is further complicated by the difficulty in classifying minor degrees of hemorrhage. In the present report hemostatic studies were performed on a consecutive series of patients with cirrhosis admitted to a medical ward and were correlated with the incidence of major hemorrhage, defined as bleeding sufficiently severe to require transfusion. Materials and Methods All patients with cirrhosis of the liver admit¬ ted to the George Washington University Med¬ ical Division of the District of Columbia General Hospital during August 1965 to April 1966 were studied. The diagnosis was confirmed histologically by biopsy or autopsy in 50% of the patients. Laennec's cirrhosis with variable de¬ grees of fatty metamorphosis was the most common condition. Liver function tests and Quick prothrombin time studies were done on all patients. Additional coagulation studies were done on patients with prothrombin times prolonged more than three seconds (<60%). In bleeding patients, blood samples were ob¬ tained prior to transfusion. To evaluate the effect of phytonadione (vitamin Kx) on pro¬ thrombin abnormalities, clotting studies were performed on 33 patients before and after treatment with 50 to 100 mg of phytonadione (Aqua Mephyton) administered parenterally. Platelet count, plasma recalcification time, partial thromboplastin time, thrombin time, Quick prothrombin time, assays for factors II, V and VII and X (combined assay), and deter¬ minations of fibrinogen and euglobulin lysis time were done by methods recently de¬ scribed.12 Factor IX (plasma thromboplastin component, PTC) was measured in a thrombo¬ plastin generation system, the incubation mix¬ ture of which consisted of 0.5 ml platelet substi¬ tute (Cephaloplastin), 0.5 ml barium sulfate adsorbed-oxalated plasma diluted 1:5, 0.25 ml PTC-deficient serum diluted 1:10, 0.25 ml test serum diluted 1:20, and 0.5 ml calcium chloride 0.025M. Six minutes after recalcification 0.2 ml of the incubation mixture was added to 0.1 ml of a standard substrate plasma and the subse¬ quent clotting time converted to percentage of PTC by reference to a dilution curve made from pooled normal serum. The normal range of values for the coagulation studies as per¬ formed in our laboratory is included in Table 1. Results