In a two-year period, ascitic fluid concentrations of complement 3c and complement 4 were measured in 110 patients with sterile cirrhotic ascites, 31 patients with spontaneous bacterial peritonitis, 65 patients with hepatocellular carcinoma, 36 patients with peritoneal carcinomatosis and 12 patients with miscellaneous diseases (nephrotic syndrome 4, systemic lupus erythematosus 3, secondary peritonitis 2, cardiac ascites 1, eosinophilic peritonitis 1 and tuberculosis peritonitis 1) to assess the clinical utility of ascitic fluid complements. The ascitic fluid level of complements 3c or C4 was significantly higher in patients with peritoneal carcinomatosis (32.8 +/- 10.2, 13.4 +/- 7.4 mg/dL) than in patients with sterile cirrhotic ascites (9.2 +/- 5.2, 4.5 +/- 3.9 mg/dL, p < 0.001), spontaneous bacterial peritonitis (8.2 +/- 4.1, 3.8 +/- 2.4 mg/dL, p < 0.001) or hepatocellular carcinoma (12.8 +/- 8.3, 5.6 +/- 4.4 mg/dL, p < 0.001). However, it was not significantly different from the miscellaneous disease group. To verify that ascites formation is not related to liver disease origin, diagnostic sensitivity, specificity and accuracy were 83.3%, 92.7% and 90.9%, respectively, by the ascitic fluid level of complement 3c higher than the cut-off value (20 mg/dl); or 60.4%, 89.8% and 84.3%, respectively, by the ascitic fluid level of complement 4 higher than the cut-off value (10 mg/dL). A direct correlation was found between the ascitic fluid protein level and the ascitic fluid complement 3c (r = 0.70) or complement 4 (r = 0.57) level. Based on results in this study, we can conclude that measuring ascitic fluid complements is clinically useful in disapproving the liver disease origin of ascites formation. However, it is of little value in diagnosing spontaneous bacterial peritonitis or hepatocellular carcinoma.