The value of serum bile acid concentrations for predicting prognosis in cirrhotics was compared with the prognostic significance of clinical and laboratory findings in a prospective 1-year study of 76 patients with cirrhosis. A commercial radioimmunoassay for total serum-conjugated primary bile acids was used. Of 76 patients, 16 died within the follow-up period. The concentration of bile acids in serum more closely correlated with mortality in cirrhosis than the commonly used clinical and laboratory parameters such as the Number Connection Test, ascites, albumin, pseudocholinesterase, bilirubin, prothrombin time and nutritional state. Serum bile acids alone yielded a prediction of mortality comparable to the Child classification. When logistic regression analysis was performed, optimal prediction of prognosis was achieved with the combination of serum bile acids and the Number Connection Test. Serum bile acid levels alone or in combination with the Number Connection Test may be a clinically useful prognostic index in cirrhosis.
The value of serum bile acid concentrations for detection of liver disease has been extensively studied (1). Measurements of serum bile acids appear t o be more sensitive for detection of cirrhosis than commonly used tests (2). In contrast, the value of serum bile acid concentrations for prediction of prognosis in cirrhoisis has not been investigated. Therefore, we compared t h e prognostic significance of serum bile acids with clinical and laboratory findings and Childs' classification (3) in a prospective 1 -year study of patients with cirrhosis.
PATIENTS AND METHODS
There were 77 consecutive patients with cirrhosis which was proven by biopsy and/or laparoscopy. A successful follow-up for 1 year was possible in 76 patients (17 females, 59 males), and data were obtained from a prospective analysis of these patients. Ages ranged from 17 to 76 years (mean = 52.4 years). Twenty-one patients suffered from posthepatitic cirrhosis and 55 from alcoholic cirrhosis. Twelve patients had cholecystectomy and 14 had cholelithiasis as documented by ultrasonography. The following clinical findings were noted and graded to determine prognostic significance with regard to mortality during follow-up.
Ascites was graded as absent, slight, moderate or severe. The