The clinical value of the tumor marker human chorionic gonadotropin-beta (hCG-beta) in ascitic fluid for the differentiation of malignancy-related and non-malignant ascites was evaluated. Ascitic fluid protein, cholesterol and cytological examination were determined for comparison. Thirty-six patients with malignancy-related ascites (27 peritoneal carcinomatosis, 9 miscellaneous malignant causes without peritoneal carcinomatosis) and 69 patients with nonmalignant ascites (55 with liver cirrhosis, 14 with miscellaneous nonmalignant causes) were investigated. hCG-beta concentrations were elevated in malignant samples and with a cut-off value of 10 mIU/ml hCG-beta yielded a sensitivity of 61%, specificity of 94% and efficiency of 83%. Ascitic fluid protein (cut-off value 3.0 g/100 ml) and cholesterol (cut-off value 45 mg/100 ml) concentrations showed a sensitivity of 64%/83%, specificity of 77%/81% and efficiency of 72%/82%. The combination of hCG-beta and cytological examination yielded 89.5% differential diagnostic efficiency, superior to the combinations of protein and cytology or protein and hCG-beta. hCG-beta tended to be superior to protein/cholesterol determination regarding sensitivity (44% vs. 11%/33%) and specificity (79% vs. 50%/57%) in the subgroups of patients with miscellaneous causes of ascites. In conclusion, hCG-beta is frequently elevated in malignancy-related ascites and seems to be as useful a parameter as total protein for the differentiation of malignancy-related from nonmalignant ascites.