Use of a variety of biological parameters in distinguishing cirrhotic frommalignant ascites

Twenty-two different protein measurements were taken in the serum and ascit ic fluid of fifty consecutive patients in an attempt to investigate which t ests are the most reliable for the differential diagnosis of ascites. Serum and ascitic fluid total proteins (TPR), albumin (ALB), lactate (LAC), ferr itin (FER), C3 and C4 complement factors, C-reactive protein (CRP), cerulop lasmin (CER), alpha2-macroglobulin (alpha 2MG), haptoglobin (HAP), alpha1-a ntitrypsin (alpha 1AT), alpha1-acid glycoprotein (alpha 1AG), transferrin ( TRF), immunoglobulins IgG, IgA, IgM and cytokines such as interleukin-1 alp ha (IL-1 alpha), interleukin-1 beta (IL-1 beta), interleukin-2 (IL-2), inte rleukin-6 (IL-6), interleukin-8 (IL-8) and tumor necrosis factor-alpha (TNF -alpha) were measured to distinguish between malignant and cirrhotic ascite s. Correlations and non-parametric Mann-Whitney tests-were used for ascitic fluid:serum ratio comparisons between the two groups. Multivariate analyse s were used to determine the most significant biochemical ratio predictors for the differential diagnosis and a recursive partitioning model was const ructed. Highly positive correlations (r>0.50) were found between the ratios IgA, IgG, IgM, CER, alpha2 MG, HAP, alpha 1AT, alpha 1AG and TRF. There wa s evidence that TPR, ALB, LAC, FER, IgG, CER, alpha 2MG, alpha 1AT, alpha 1 AG, TRF and IL-8 ascitic fluid:serum ratios are significnatly higher in pat ients with malignant neoplasms than in cirrhotics. In the recursive partiti oning model the most significant parameters were found to be the ratios of albumin and IL-1 alpha. The model fitted allowed for 100% correct classific ation of ascites. In conclusion, we have shown that a simple and very accur ate model based on two ascitic fluid:serum measurements is able to differen tiate between malignant and non-malignant ascites.