Context: Abdominal paracenteses are performed in patients with ascites, most commonly to assess for infection or portal hypertension and to manage refractory ascites.
Objectives: To systematically review evidence for paracentesis methods that may decrease risk of adverse events or improve diagnostic yield and to determine the accuracy of ascitic fluid analysis for spontaneous bacterial peritonitis or portal hypertension.
Data sources: Relevant English-language studies from Medline (1966-April 2007) and EMBASE (1980-April 2007).
Study selection: Paracentesis studies evaluating interventions (use of preprocedure coagulation parameters, needle type, insertion location, ultrasound guidance, bedside inoculation into blood culture bottles, and use of plasma expanders in therapeutic taps) for reducing adverse events or improving the diagnostic yield, and studies assessing the accuracy of ascitic fluid biochemical analyses for spontaneous bacterial peritonitis or portal hypertension.
Data extraction: For technique studies, data on intervention and outcome; and for diagnostic studies, data on parameters for diagnosing spontaneous bacterial peritonitis and portal hypertension (ie, ascitic fluid white blood cell and polymorphonuclear leukocyte [PMN] count, ascitic fluid pH, blood-ascitic fluid pH gradient, and serum-ascites albumin gradient).
Data synthesis: Thirty-seven studies met inclusion criteria: 2 showed that obtaining preprocedure coagulation was likely unnecessary prior to paracentesis; 1 showed the 15-gauge, 3.25-inch needle-cannula results in less multiple peritoneal punctures [P = .05] and termination due to poor fluid return [P = .02] vs a 14-gauge needle in therapeutic paracentesis; 1 showed immediate inoculation of culture bottles improves diagnostic yield vs delayed (from 77% to 100% [95% CI for the difference, 5.3%-40.0%]); 9 evaluated therapeutic paracentesis, performed with or without albumin or nonalbumin plasma expanders, and found no consistent effect on morbidity or mortality; 16 showed the accuracy of biochemical analysis of ascitic fluid in patients suspected of having spontaneous bacterial peritonitis to increase the likelihood of spontaneous bacterial peritonitis (PMN count >250 cells/microL [summary likelihood ratio {LR}, 6.4] 95% CI, 4.6-8.8; ascitic fluid leukocyte count >1000 cells/microL [summary LR, 9.1] 95% CI, 5.5-15.1; pH < 7.35 [summary LR, 9.0] 95% CI, 2.0-40.6; or a blood-ascitic fluid pH gradient > or = 0.10 [LR, 11.3] 95% CI, 4.3-29.9) and other levels lowered the likelihood (PMN count < or = 250 cells/microL [summary LR, 0.2] 95% CI, 0.11-0.37; or a blood-ascitic fluid pH gradient < 0.10 [summary LR, 0.12] 95% CI, 0.02-0.77); and 4 showed the diagnostic accuracy of the serum-ascites albumin gradient lowers the likelihood of portal hypertension (< 1.1 g/dL [summary LR, 0.06] 95% CI, 0.02-0.20).
Conclusions: Ascitic fluid should be inoculated into blood culture bottles at the bedside. Spontaneous bacterial peritonitis is more likely at predescribed parameters of ascitic PMN count or blood-ascitic fluid pH, and portal hypertension is less likely below a predescribed serum-ascites albumin gradient.