The authors tried to differentiate malignancy-related from nonmalignant ascites with a sequence of sensitive followed by specific ascitic-fluid parameters. There were four results of this study. First, of nine parameters investigated in a first series of 48 patients, 28 with nonmalignant and 20 with malignancy-related ascites, ascitic-fluid cholesterol and fibronectin yielded the best negative predictive value of 92% each. Carcinoembryonic antigen (CEA) and cytologic examination both showed a positive predictive value of 100%. Second, combining cytologic examination (sensitivity, 70%) and CEA determination (sensitivity, 45%) increased the sensitivity to 80%. Third, cytologic findings were negative in all ascitic-fluid samples with a cholesterol concentration below the cutoff value of 45 mg/100 ml. Fourth, based on the results of the first series of 48 patients, the diagnostic sequence with cholesterol as a sensitive parameter, followed by the combination of cytologic examination and CEA determination as specific parameters, was tested in a second series of 71 patients, 37 with nonmalignant and 34 with malignancy-related ascites. Again cytologic examination was negative in all samples with cholesterol levels below 45 mg/100 ml. In the total of 119 patients, this diagnostic sequence did not identify 9% of patients with malignancy-related ascites, and 82% of samples classified as malignancy related by cholesterol levels above 45 mg/100 ml were confirmed by positive cytologic examination and/or CEA level above 2.5 ng/ml. Thus, a diagnostic sequence with ascitic-fluid cholesterol determination, followed by cytologic examination and CEA determination, in samples with cholesterol levels above 45 mg/100 ml should permit a cost-efficient routine differentiation of malignancy-related from nonmalignant ascites.