Diagnostic paracentesis is usually considered the first test to be performed in the assessment of the ascitic patient and a large number of investigations on ascitic fluid have been proposed. To assess the value of a simplified procedure, serum to ascites albumin gradient and ascitic white blood cell counts were employed as a first step. One hundred and fifty-three paired serum and ascitic fluid samples were analysed and allowed patients to be divided into three groups: 1) serum to ascites albumin gradient > = 11 g/L and white blood cells < 0.5 x 10(9)/L predicted cirrhosis (or liver carcinoma) without peritonitis with 83% efficacy, 96% positive predictive value and 65% negative predictive value; 2) serum to ascites albumin gradient > = 11 g/L and white blood cells > = 0.5 x 10(9)/L predicted cirrhosis (or liver carcinoma) with peritonitis with 86% efficacy, 45% positive predictive value and 99% negative predictive value; 3) serum to ascites albumin gradient < 11 g/L predicted the other diagnoses with 92% efficacy, 77% positive predictive value and 95% negative predictive value. As serum to ascites albumin gradient > = 11 g/L and white blood cells < 0.5 x 10(9)/L predicted cirrhosis (or liver carcinoma) without peritonitis in 96% of the cases and excluded peritonitis in 99% of the cases, further fluid ascitic analyses could be considered as a second step only in patients with serum to ascites albumin gradient < 11 g/L and/or white blood cells > = 0.5 x 10(9)/L. In a group of ascitic patients where the prevailing diagnosis is cirrhosis (or liver carcinoma) without peritonitis, this simplified approach could provide a favourable cost/benefit ratio.