1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Large-volume paracentesis, the preferred treatment for patients with symptomatic tense ascites due to cirrhosis, has traditionally been performed by physicians as an inpatient procedure. Our objectives were to determine (1) whether large-volume paracentesis could be performed safely and effectively by gastrointestinal endoscopy assistants and as an outpatient procedure, (2) whether the risk of bleeding was associated with either thrombocytopenia or prolongation of the prothrombin time, and (3) the resources used for large-volume paracentesis. Gastrointestinal endoscopy assistants performed 1,100 large-volume paracenteses in 628 patients, 513 of whom had cirrhosis of the liver. The preprocedure mean international normalized ratio for prothrombin time was 1.7 ± 0.46 (range, 0.9-8.7; interquartile range, 1.4-2.2), and the mean platelet count was 50.4 × 103/μL, (range, 19 × 103/μL − 341 × 103/μL; interquartile range, 42-56 × 103/μL). Performance of 3 to 7 supervised paracenteses was required before competence was achieved. There were no significant procedure-related complications, even in patients with marked thrombocytopenia or prolongation in the prothrombin time. The mean duration of large-volume paracentesis was 97 ± 24 minutes, and the mean volume of ascitic fluid removed was 8.7 ± 2.8 L. In conclusion, large-volume paracentesis can be performed safely as an outpatient procedure by trained gastrointestinal endoscopy assistants. Ten supervised paracenteses would be optimal for training the operators carrying out the procedure. The practice guideline of the American Association for the Study of Liver Diseases which states that routine correction of prolonged prothrombin time or thrombocytopenia is not required is appropriate when experienced personnel carry out paracentesis. (HEPATOLOGY 2004;40:484–488.)

Large-volume paracentesis involves removal of as much ascitic fluid as possible to relieve symptoms of a tense abdomen and dyspnea in patients who have cirrhosis of the liver, ascites of cardiac failure, or malignant ascites.1–4 Randomized controlled trials have confirmed that paracentesis is more effective than therapy with diuretics for the elimination of ascitic fluid.5 Hospitalization is shortened, systemic hemodynamics and renal function are better preserved, and hepatic encephalopathy is less frequent in patients treated with large-volume paracentesis supported by plasma volume expansion than in patients treated with diuretics.

Plasma volume expansion is recommended with therapeutic paracentesis; albumin (6-8 g/L of ascitic fluid removed) is the preferred volume expander.6 Removal of more than 5 L of ascitic fluid without plasma volume expansion is associated with impairment of renal function because of increased activity in the renin-angiotensin-aldosterone system.7 Additional benefits of paracentesis include improvement in hepatic hemodynamics and a decreased risk of developing spontaneous bacterial peritonitis.8 Thus, considerable evidence favors the use of paracentesis with albumin replacement as the preferred treatment of tense ascites.

However, large-volume paracentesis may have drawbacks. Retrospective studies have reported complications, including infection,9 bleeding,10 and bowel perforation, which can rarely result in mortality related to the procedure.11 Generally, it has been recommended that paracentesis be performed as an inpatient procedure and that a physician remain at the bedside throughout the treatment.1 In common practice, the physician may insert the drainage catheter into the abdomen, but a nurse may supervise the procedure after this. In addition, it is not certain whether measurement of the prothrombin time or platelet count is necessary as a routine measure before therapeutic paracentesis. The practice guidelines from the American Association for the Study of Liver Diseases, which have been recently updated, do not consider paracentesis to be unsafe in the presence of marked thrombocytopenia or prolongations in the prothrombin time.3, 4 A recent consensus statement from the International Ascites Club states that “there are no data to support the correction of mild coagulopathy with blood products prior to therapeutic paracentesis, but caution is needed when severe thrombocytopenia is present.”12 Thus, there is still uncertainty as to the optimal platelet count and prothrombin time for the safe conduct of paracentesis. Guidelines regarding paracentesis are not available from the publications of American Society for Gastrointestinal Endoscopy among their recommendations for carrying out procedures in patients with disorders of clotting and bleeding.13, 14

The ideal treatment of ascites would avoid complications and use few resources. This could be facilitated by nonphysician health care providers' performing paracentesis as an outpatient procedure and safely increasing the rate of fluid aspiration, shortening the procedure. Toward this ideal, we trained gastrointestinal endoscopy assistants, all of whom were licensed practical nurses whose education was comparable to that of licensed practical nurses elsewhere in the country. The purposes of the present study were (1) to determine how many paracentesis procedures gastrointestinal endoscopy assistants needed to perform to develop competence, (2) to retrospectively analyze the safe cut-off for prothrombin time and platelet count prior to paracentesis, and (3) to determine how much time was needed to perform paracentesis.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References


The decision to perform large-volume paracentesis was made by the primary physician, usually a hepatologist, who then referred the patient for the procedure. A total of 1,100 large-volume paracenteses were performed as outpatient procedures by gastrointestinal endoscopy assistants in 628 patients between June 1994 and June 2001. Of these patients, 513 had cirrhosis of the liver; the second-most common diagnosis was malignant ascites (n = 91).


The indications, risks, and alternatives were discussed with patients before the procedure. Although recent complete blood cell counts, as well as prothrombin time, were available before the procedure, in keeping with the prevailing practice guidelines of the American Association for the Study of Liver Diseases,3 we considered no degree of thrombocytopenia or prolongation of prothrombin time to be unsafe for paracentesis. Details of the preprocedure prothrombin time and platelet count are given in Table 1. Procedures were performed in a dedicated paracentesis room in the endoscopy suite without ultrasonographic guidance. With the patient supine, the presence of ascitic fluid was confirmed on physical examination by the demonstration of shifting dullness or a fluid wave. Usually a site in the outer-left lower-abdominal quadrant, where dullness could be demonstrated on percussion, was chosen for paracentesis. The presence of fluid at the selected puncture site was reconfirmed by percussion before local anesthetic was administered.

Table 1. Preprocedure Prothrombin Times and Platelet Counts
Prothrombin Time (INR)Number of Patients*
  • *

    The number of procedures carried out in a single patient ranged between 1 and 16. A total of 1,100 procedures were carried out in 628 patients. Prothrombin time and platelet count are within 72 hours in most patients; in 27 patients, values are within 7–14 days of the procedure.

  • Patients with noncirrhotic causes of ascites are mainly in this group.

Platelet count (×103/μL)Number of Patients*

The procedure was performed with a sterile technique. The aspirating catheter consisted of an 18-gauge metal introducer needle with an outer plastic sheath (diameter, 8F; length, 20 cm) with multiple side holes (Allegiance Healthcare Corporation, McGaw Park, IL). Intravenous plastic tubing (Interlink System, Baxter Healthcare Corporation, Deerfield, IL) was used to connect the suction device to the aspirating catheter before it was inserted into the abdomen. The site for paracentesis was prepared, extending at least 7.5 cm in each direction, with povidone-iodine (Betadine, The Purdue Frederick Company, Stamford, CT). Thus, the prepared area had a diameter of at least 15 cm. The skin was anesthetized by intradermal injection of 1% lidocaine (Xylocaine, AstraZeneca, Wilmington, DE). A longer, 3.75-cm, 21-gauge needle was used to anesthetize deeper tissue down to the peritoneal cavity. The distance to the peritoneal cavity to be traversed by the aspirating catheter was judged by the distance the needle traversed before ascitic fluid was obtained. Before the aspirating catheter was inserted, the presence of peritoneal fluid was confirmed by aspiration via the needle used for local anesthesia. A stab wound was made to the depth of the beveled edge of the blade of the scalpel provided in the paracentesis kit. Bleeding was controlled by pressure with cotton gauze. After hemostasis was achieved, the aspirating catheter was inserted to the depth previously determined. Final entry into the peritoneal cavity was facilitated by asking the patient to protrude the abdomen. The aspirating catheter was inserted an additional 1 cm, then the needle was withdrawn while the outer sheath slid approximately 5 cm into the peritoneal cavity.

The suction device was an “octopus,” consisting of a central stand with 4 arms to which 4 3-L-capacity canisters could be connected in series (Baxter Healthcare Corporation). When each canister became full, the fluid automatically flowed into the next canister. After initial collection of ascitic fluid samples for testing, the three-way stopcock was opened to allow drainage of the ascitic fluid into the canisters. If more than 12 L of fluid was aspirated, the first full canister was replaced with an empty one. With initiation of paracentesis, an infusion of albumin (Buminate 25%, Baxter Healthcare Corporation) was started (6-8 g/L of ascitic fluid removed—that is, approximately 25 g of albumin for every 3-L cannister filled) and was usually completed by the end of paracentesis.

If interruption of the flow of ascitic fluid could not be corrected by changing the position of the patient, or if it stopped completely, the stopcock was closed and the aspirating catheter withdrawn while pressure was applied over the insertion site. The position of the aspirating catheter was not changed if the flow of ascitic fluid was interrupted, because repositioining of the plastic aspirating catheter required reinsertion of the metallic stilette that could be associated with a risk of shearing off the catheter. After the site was cleaned with an alcohol wipe, a 3 × 3–cm cotton-gauze sponge was placed and secured with an elastic adhesive bandage (Microfoam; 3M, St. Paul, MN). After the procedure, the patient was observed in a recovery room for 60 minutes. When the patient was asymptomatic and had no postural hypotension, the patient was dismissed.

After dismissal, the patients returned to their physicians, either at Mayo Clinic or at home. Patients who were not seeing their primary physician at Mayo Clinic within 24 hours were contacted by telephone to confirm that they had no postprocedure complications.


When training began, the supervising physician (P.S.K.) sat with the trainee before the procedure and discussed the indications, risks, and benefits of paracentesis. The physician described the relevant surface anatomical landmarks and explained how to use the contents of the paracentesis tray. The connections between the aspirating cannula and the suction device were demonstrated.

For each trainee, the first training paracentesis was performed by the physician with the gastrointestinal endoscopy assistant assisting. The second and third paracenteses were performed by the gastrointestinal endoscopy assistant with the physician assisting. Paracenteses 4 through 10 were performed independently by the gastrointestinal endoscopy assistant with the physician observing but not assisting unless necessary. The number of paracenteses completed by the assistant before the supervising physician felt comfortable with the assistant's performance was noted. A checklist was developed (Table 2), and the various skills required for the procedure were checked off during the training period. Competence was considered attained when the gastrointestinal endoscopy assistant performed 3 consecutive procedures without any help from the supervising physician and when the gastrointestinal endoscopy assistant had completed the steps of the entire checklist. After the 10th procedure, the gastrointestinal endoscopy assistant performed the procedure independently; a physician, who was almost always the supervising training physician (P.S.K.), was available for consultation as necessary.

Table 2. Skills Checklist for Abdominal Paracentesis
  1. NOTE. In practice, this checklist can be incorporated into a form with 10 blank columns (for 10 paracenteses) in which the supervising physician initials and dates the demonstration of each skill by a trainee.

Consulted with physician before the procedure
Checked history and laboratory results
Recorded vital signs before the procedure
Discussed indications and risks with patient
Identified correct site
Connected “octopus”
Started albumin infusion
Put on sterile gloves correctly
Cleaned site adequately
Withdrew local anesthetic from ampule
Infiltrated anesthetic intradermally
Injected local anesthetic to peritoneal lining
Checked aspiring catheter connections
Made stab wound of appropriate length and depth
Correctly inserted catheter through abdominal wall
Withdrew needle with minimal leakage of ascitic fluid
Collected ascitic fluid samples for tests
Replaced albumin adequately (6–8 g/L of fluid removed)
Withdrew catheter at end of procedure
Cleaned and covered site adequately
Recorded vital signs after the procedure
Looked for complications
Dismissed patient in stable condition

Use of Resources.

The Division of Gastroenterology and Hepatology at Mayo Clinic uses a database to record all patient data. The database contains information on all gastroenterologic procedures, including various types of endoscopy, motility studies, and paracentesis. The information includes patient demographics, pertinent history, procedure details, and nursing issues, as well as the appointment time, arrival time, check-in time, procedure start and finish times, and dismissal times. The record includes the name of the person providing each segment of care, as well as details such as the room and equipment used.

The duration of paracentesis was defined as the time interval between when the patient entered the procedure suite and when the patient left the suite. The total duration of the procedure, however, was longer and included the check-in time at the procedure area, preparation time, procedure time, and recovery time. The total duration of the procedure was the time between when the patient checked into and when the patient checked out of the procedure area.


  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References


The number of independent procedures performed by the gastrointestinal endoscopy assistants before the supervising physician was comfortable with their procedural skill ranged from 3 to 7 (mean, 4.4 ± 1.6). The number of large-volume paracenteses performed by gastrointestinal endoscopy assistants during the study period ranged from 37 to 318 (mean, 170 ± 12; median, 158).

The gastrointestinal endoscopy assistants determined that paracentesis was not feasible in 23 of 1100 patients. The supervising physician agreed with that assessment for all 23. Thus, there was complete agreement between the gastrointestinal endoscopy assistants and the supervising physician for the appropriateness of large-volume paracentesis (κ statistic, 1.0). Two patients had significant gaseous abdominal distention that made blind, large-volume paracentesis unsafe; paracentesis was deferred in these patients. In 9 patients referred for therapeutic paracentesis, physical examination by the gastrointestinal endoscopy assistant before the procedure did not demonstrate sufficient ascitic fluid for paracentesis; in each instance, the supervising physician confirmed that finding. These patients were then referred to the Radiology Department for ultrasonography-directed paracentesis. At ultrasonography, there was no ascitic fluid in 3 of the 9 patients and only a small volume of ascitic fluid in the other 6. Twelve patients with several abdominal surgical scars had paracentesis carried out safely under ultrasound guidance.

Use of Resources.

The total duration of the procedure from the time the patient checked in to the time the patient was dismissed was less than 4 hours. The mean duration of the paracentesis procedure (the time between when the patient entered and exited the procedure room) was 97 ± 24 minutes. The mean volume of ascitic fluid removed was 8.7 ± 2.8 L, at a rate of approximately 1 L every 6 minutes. The largest volume of ascitic fluid removed in a procedure was 31 L.

Thrombocytopenia, Prothrombin Time Prolongation, and Procedure Risk.

The lowest platelet count recorded was 19 × 103/μL (reference range, 150-450 × 103/μL), and the most prolonged prothrombin time had an international normalized ratio of 8.7. Both of those procedures were carried out in the presence of the supervising physician, but they were not carried out during the training period of the gastrointestinal endoscopy assistants.

The range for prothrombin time international normalized ratio was 0.9 –8.7, with the interquartile range of 1.4–2.2. The range for platelet counts was 19 × 103/μμL to 341 × 103/μL, with the interquartile range 42 ×103/μL to 56 × 103/μL. There were 612 paracenteses carried out safely with platelet counts below 50 × 103/μL. There were no procedure-related complications that required hospitalization. Postural hypotension (decrease in systolic pressure >20 mm Hg) at 1 hour after the procedure was noted in 3 patients. These patients were observed in a supine position for a longer period; all were dismissed within the next hour without postural hypotension. In one patient, the ascitic fluid tap was initially hemorrhagic but subsequently clear. The platelet count in this patient was 64 × 103/μL. Four patients had leakage of ascitic fluid, which lasted less than 48 hours and was treated with additional dressings over the wound. All of the patients with leakage of ascitic fluid had a chronic cough, with fluid leakage tending to occur during bouts of coughing. There was no significant bleeding in any patient, even in the presence of significant abnormalities of prothrombin time and platelet count.


  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

This study demonstrates that paracentesis performed in an outpatient setting by our gastrointestinal endoscopy assistants is safe and efficient, even in the presence of thrombocytopenia and prolongation of the prothrombin time. The major benefit of our approach is that it decreases the requirement for physician resources, as physician participation may be limited to training of personnel. Finally, ultrasonography-guided paracentesis is rarely required in patients who have not undergone previous abdominal operations and only increases the cost of the procedure.

Several aspects of the study need to be addressed further. Although large-volume paracentesis has traditionally been considered a procedure for an inpatient setting, our experience clearly demonstrates that the procedure can be performed safely in an outpatient setting, even with marked abnormalities of the prothrombin time and platelet count. In fact, the majority of patients who underwent paracentesis had platelet counts below (50 × 103/μL). However, the procedure may be safe only in the hands of experienced operators, since significant bleeding following paracentesis has been reported.10 Therefore, individual physicians may have different thresholds for correcting prothrombin time prolongation or thrombocytopenia, in spite of the data from this study.

Adequate training does not require many procedures. Clearly, 10 supervised procedures should be sufficient for an operator to learn the procedure. The lack of complications associated with the procedure confirms that this is an appropriate training period for achieving competence. Moreover, each time the gastrointestinal endoscopy assistants deemed that paracentesis was not feasible, the supervising physician agreed, and the decisions were confirmed to be correct upon ultrasonographic examination of the abdomen. Therefore, accreditation bodies such as the Residency Review Commission may recommend a minimum of 10 supervised paracentesis as part of training programs.

The complications of paracentesis, such as bleeding, bowel perforation, and death, occur infrequently. Our study demonstrates that with adequate training, these complications usually can be avoided. Also, in our hands, hemorrhagic complications were not seen, even in the presence of thrombocytopenia and a prolonged prothrombin time. Paracentesis was complicated by postprocedure leakage of ascitic fluid in only 4 patients with chronic cough, suggesting that the best safeguard against leakage of ascitic fluid postprocedure is aspiration of ascitic fluid as completely as possible. If only a small amount of ascitic fluid is left, leakage is unlikely, and the abdominal wall seals off before ascitic fluid reaccumulates.

Finally, there is a significant saving in physician time without any increased patient risk. The procedure can usually be completed within a 2-hour period, and paracentesis schedules can be organized accordingly. However, our interpretation of the Joint Commission on Accreditation of Healthcare Organizations guidelines15 was that the practice of having gastrointestinal endoscopy assistants carry out large-volume paracentesis should be discontinued. Thus, as of June 2001, our gastrointestinal endoscopy assistants no longer carry out paracenteses.

In summary, our study demonstrates that large-volume paracentesis can be performed safely as an outpatient procedure and by personnel who are not physicians; specifically, gastrointestinal endoscopy assistants can be trained to perform the procedure. These personnel can be trained adequately by performing paracentesis about 10 times in a supervised setting. Most therapeutic paracenteses should be completed within a 2-hour period. If our results are confirmed in another large study, routine measurement of the prothrombin time and platelet count may not be required, provided that experienced operators carry out the paracentesis. Finally, the Joint Commission on Accreditation of Healthcare Organizations guidelines for carrying out large-volume paracentesis procedures on patients may need to be readdressed.


  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References
  • 1
    Arroyo V, Sort P, Ginès P, Planas R. Treatment of ascites by paracentesis. In: ArroyoV, GinèsP, RodésJ, SchrierRW, eds. Ascites and Renal Dysfunction in Liver Disease: Pathogenesis, Diagnosis, and Treatment. Malden, MA: Blackwell Science; 1999: 463479.
  • 2
    Garcia-Tsao G. Current management of the complications of cirrhosis and portal hypertension: variceal hemorrhage, ascites, and spontaneous bacterial peritonitis. Gastroenterology 2001; 120: 726748.
  • 3
    Runyon BA. Management of adult patients with ascites caused by cirrhosis. HEPATOLOGY 1998; 27: 264272.
  • 4
    Runyon BA. Management of adult patients with ascites caused by cirrhosis. HEPATOLOGY 2004; 39: 841856.
  • 5
    Ginès P, Arroyo V, Quintero E, Planas R, Bory F, Cabrera J, et al. Comparison of paracentesis and diuretics in the treatment of cirrhotics with tense ascites. Results of a randomized study. Gastroenterology 1987; 93: 234241.
  • 6
    Ginès A, Fernandez-Esparrach G, Monescillo A, Vila C, Domenech E, Abecasis R, et al. Randomized trial comparing albumin, dextran 70, and polygeline in cirrhotic patients with ascites treated by paracentesis. Gastroenterology 1996; 111: 10021010.
  • 7
    Tito L, Ginès P, Arroyo V, Planas R, Panes J, Rimola A, et al. Total paracentesis associated with intravenous albumin management of patients with cirrhosis and ascites. Gastroenterology 1990; 98: 146151.
  • 8
    Luca A, Feu F, Garcia-Pagan JC, Jimenez W, Arroyo V, Bosch J, et al. Favorable effects of total paracentesis on splanchnic hemodynamics in cirrhotic patients with tense ascites. HEPATOLOGY 1994; 20: 3033.
  • 9
    Mallory A. Complications of diagnostic paracentesis in patients with liver disease. JAMA 1978; 239: 628630.
  • 10
    Webster ST. Hemorrhagic complications of large volume paracentesis. Am J Gastroenterol 1996; 91: 366368.
  • 11
    Minocha A. A fatal case of paracentesis. Am J Gastroenterol 1999; 94: 856.
    Direct Link:
  • 12
    Moore KP, Wong F, Gines P, Bernardi M, Ochs A, Salerno F, et al. The management of ascites in cirrhosis: report on the consensus conference of the International Ascites Club. HEPATOLOGY 2003; 38: 258266.
  • 13
    Eisen GM, Baron TH, Dominitz JA, Feigel DO, Goldstein JL, Johnson JF, et al. American Society for Gastrointestinal Endoscopy. Guideline on the management of anticoagulation and antiplatelet therapy for endoscopic procedures. Gastrointestinal Endoscopy 2002; 55: 775779.
  • 14
    Van Os EC, Kamath PS, Gostout CJ, Heit JA, Gastrointestinal procedures among patient with disorders of hemostasis: evaluation and management recommendations. Gastrointestinal Endoscopy 1999; 50: 536543.
  • 15
    Joint Commission on Accreditation of Healthcare Organizations. JCAHO Manual. Available at: Accessed October 9, 2003.