We read with great interest the article of Grabau et al.1 regarding performance of therapeutic abdominal paracentesis, because the authors highlight no significant procedure-related complications even in patients with marked thrombocytopenia or prolongation in the prothrombin time (PT) in an outpatient setting. We have launched a comparable study in an emergency setting and would like to share our results.
For a 1-year period starting in August 2003 in an emergency department of a tertiary center, a total of 186 abdominal paracenteses were carried out in 60 patients. The number of procedures carried out in a single patient ranged between 1 and 17. All patients underwent complete blood cell counts, biochemistries, and PT before the procedure. In the absence of a cutoff for coagulation parameters that would restrict paracentesis,2 all patients were eligible. The emergency physicians performed the procedures using ultrasonography to define the puncture site in the outer-left lower abdomen and an 18-gauge aspirating catheter (Surflo, Terumo Corporation, Tokyo, Japan) using sterile technique. Their age (mean ± SD) was 59.0 ± 14.0 years. The underlying diseases were hepatocellular carcinoma in 46, viral-related liver cirrhosis in 107, alcoholic cirrhosis in 7, and other malignancies in 26. The status of Child classification of patients was A in 8, B in 81, and C in 97. The preprocedure mean international normalized ratio (INR) for PT was 1.6 ± 0.5 (range, 0.9-4.7), and the mean platelet count was 124 ± 103 ×103/μL (range, 6-641 ×103/μL). Details of the data are given in Table 1.
|Diagnostic Paracentesis*, n||Therapeutic Paracentesis†, n||Complications, n|
|Platelet count (×103/μL)|
There were no procedure-related complications that required hospitalization, transfusions, or plasma volume expansion. Only two of 186 procedures (incidence, 1.1%; 95% CI, 0.3%-3.8%) were associated with minor complications in the same patient (incidence, 1.7%; 95% CI, 0.3%-8.9%) at different visits. One minor complication with removal of 1,200 mL of ascitic fluid for this 45-year-old male, a patient with hepatitis B virus-related cirrhosis, was local ecchymosis at the puncture site, with a platelet count of 81 ×103/μL and an INR of 2.6. The diameter of ecchymosis was 3.5 cm. The other episode of cutaneous bleeding (estimated 10 mL) occurred with removal of 4,000 mL of ascitic fluid when he had a platelet count of 51 ×103/μL and an INR of 2.9. It was promptly controlled within 10 minutes with local compression.
From our data, it appears that bleeding complications of abdominal paracentesis in an emergency department are rare, and even if present, appear to be very mild, regardless of preprocedure INR or platelet count. Considering the results of Grabau et al. in an outpatient setting, we propose such tests are unnecessary before abdominal paracentesis in an emergency setting. A limitation to the procedure is clinically evident fibrinolysis or disseminated intravascular coagulation.3 Otherwise, our data should translate into the avoidance of unnecessary transfusion and related complications, cost savings, and shortening of the length of stay for patients in emergency departments.