Chapter 104: Paracentesis

Abdominal paracentesis is a procedure in which fluid is removed from the peritoneal cavity with a needle or cannula in patients with ascites. It is a relatively quick and safe procedure and can be done with a minimal amount of equipment as an outpatient procedure or at the bedside for inpatients. Now routinely used as an adjunct to paracentesis, point-of-care ultrasound is a simple procedure that confirms the presence of ascites. Proper analysis of fluid is invaluable in determining the etiology of ascites. The procedure can be done by any trained physician, surgeon, or a midlevel provider. At our institution we have a dedicated team of proceduralists and intensivists experienced in performing paracentesis using exclusively ultrasound guidance.

The procedure can be done for therapeutic, diagnostic, or both purposes.

Diagnostic: The procedure helps diagnose.

• The cause of a new-onset ascites or the status of preexisting ascites in patients who are admitted to the hospital for any reason. This is particularly important if there is evidence of infection, hepatic encephalopathy, fever, leukocytosis hypotension, and acute kidney injury.¹

• Spontaneous bacterial peritonitis where detection at an early stage and expedient initiation of antibiotics can lower mortality. Therefore, the procedure must be performed promptly in virtually all circumstances where a patient with known or newly discovered ascites sustains any change in clinical status. Delays that may occur due to lack of an experienced operator, unfounded concerns for the presence of coagulopathy, or unnecessary administration of blood products prior to paracentesis should be avoided. Analysis of the ascitic fluid may allow not only assessment of the likelihood of spontaneous bacterial peritonitis (SBP) but, with proper collection, identification of a specific microorganism and susceptibility testing to antibiotics guiding treatment in 90% of cases.

• Secondary peritonitis when free fluid is present in peritoneal cavity from rupture or perforation of an abdominal organ. The total protein may be a key differentiating factor in this situation as it is generally low in SBP but may be normal or elevated in secondary causes.

• Malignancy-related ascites (not to be confused with peritoneal carcinomatosis) may be seen with several tumors, including malignancies of the ovary, pancreas, colon, breast, lung, and liver. In addition, the presence of chylous ascites may indicate lymphoma as the cause of the ascites.

• Hemoperitoneum related to trauma, postprocedural or spontaneous intraperitoneal hemorrhage.

Therapeutic: The procedure is usually performed to relieve intra-abdominal pressure causing symptoms of dyspnea, fatigue, early satiety, and/or abdominal discomfort. The amount of fluid a patient can tolerate and the rate of accumulation is highly variable between patients, but will generally establish a predictable pattern in any given patient. Symptoms can be relieved by removal of as little as 1 to 2 L of fluid in a small individual.

In general, the diagnostic value of paracentesis outweighs the risks in virtually all circumstances. One of the most frequent concerns in the cirrhotic patient is the potential for bleeding due to coagulopathy and the risks this may pose, yet this concern is largely unfounded. When tested the majority of cirrhotic patients with ascites will be found to have some degree of coagulopathy, but this should preclude paracentesis only in severe cases of disseminated intravascular coagulopathy with overt bleeding or clinically evident primary fibrinolysis. Bleeding from a paracentesis puncture is rare and when it occurs is generally due to injury to a vascular structure which is consistent with the fact that these patients usually have normal coagulation function despite abnormal test results such as the prothrombin time (PT)/international normalized ratio (INR), activated partial thromboplastin time (aPTT), or platelet count. This is partly due to the fact that there is deficiency of both anticoagulants and procoagulants. In fact liver disease can lead to either a hypocoagulable state or a hypercoagulable state. The relative balance or imbalance of these factors is not reflected in conventional tests of coagulation. Clinical evaluation for overt bleeding tends to be a more valuable determinant of bleeding risk than coagulation testing. The inappropriate practice of transfusion of fresh frozen plasma (FFP) and/or platelets’ blood product to reverse the coagulopathy before paracentesis is therefore unnecessary and must be discouraged. Routine administration of blood products exposes the patient to a risk of infection and fluid overload and is costly; furthermore, there are no available data to support a threshold or a cutoff value for coagulation parameters beyond which paracentesis should be avoided. It has been shown that approximately 100 to 200 units of FFP would need to be administered prior to paracentesis to prevent transfusion of 2 units of red cells making this practice impractical at best. Grabau and colleagues reported no bleeding complications in series of 1100 patients who underwent large-volume paracenteses without pre- or postprocedure transfusions required despite INRs as high as 8.7 and platelet counts as low as 19,000/mL highlighting the lack of benefit of “corrective” blood product administration in virtually all instances.²

Paracentesis is a safe procedure with a major complication rate of less than 1 in 1000 procedures. Deaths caused by paracentesis are rare but do occur and require meticulous attention to technique. Major complications include procedurally related intraperitoneal bleeding, bowel perforation, and injury to internal organs. The last 2 complications have been virtually eliminated by the adjunctive use of ultrasound. However, until recently hemoperitonium remained a potentially tragic and unpredictable complication. Recently, this occurrence has largely been attributed to the presence of both normal and abnormal vascular structures buried in the abdominal wall that are generally not palpable, visible, or seen by routine fluid localization with a low-frequency abdominal ultrasound probe. By using in addition, a high-frequency linear probe to specifically evaluate the abdominal wall for the presence and location of these vessels, they may be identified and subsequently avoided (Figures 104–1A, B, and C). In our experience ultrasound assessment both for the optimal location of the fluid and the presence of abdominal wall vessels has nearly eliminated the risk of postprocedural hemoperitonium and is now considered standard of practice at our institution.³

Minor complications are relatively infrequent occurring in 2% to 5% of cases and may include abdominal wall hematoma, leakage from the paracentesis site, the need for more than one attempt to obtain fluid, excessive pain during the procedure, and the inability to completely drain the abdominal cavity. While abdominal wall or rectus sheath hematoma may be extensive, in general these complications can be managed with local measures. Of note, we also track operator injury in the form of needle stick or fluid contact contamination as a complication to emphasize the need for sharp safety and procedural hygiene to our operators.

Aseptic technique is required when performing this procedure though the need for wide sterile barriers is unnecessary. Caution should be taken especially in patients with massive bowel distension or ileus and in patients with scars indicative of prior abdominal surgeries. Ultrasound should be used in all cases and is invaluable in avoiding these potential anatomic pitfalls. Ultrasound, while essential to the overall safety of the procedure carries its own potential hazards if not applied appropriately. While the learning curve is relatively steep for this application, appropriate training is required to avoid interpretive errors.

Patient Positioning

The procedure is usually done with the patient in supine position and the head of the bed flat or slightly elevated. It may be helpful to have the patient tilt slightly to the site of where the entry is planned to allow for pooling and fluid shift.

Sites of Needle Entry and Anatomy

The patient is tilted slightly to the side to allow for fluid shift. The entire abdominal cavity must be scanned with ultrasound using a low-frequency phase array or curvilinear probe to determine the area of “maximum echogenicity” representing the most favorable pocket of fluid. The potential site of entry is then evaluated using a high-frequency linear probe to identify a site within this region devoid of abnormal vasculature. This optimal site of entry is marked on the abdominal wall with an indelible marker (Figures 104–2A, B, and C).

An informed consent must be obtained and a formal time-out performed to verify procedure and patient identifier. The following supplies and equipment are needed (Table 104–1):

• Spring-tipped paracentesis drainage kit (Figures 104–3A and B)

• Sterile gloves

• Collection canisters

• Chaux

The use of sterile gown, hair cover, or face mask is not required. But the overlying skin is to be prepped in the usual sterile fashion and sterile gloves are used with sterile barriers to create the appropriate sterile field.

For local anesthesia the preferred needle is 1.5-in 22- or 25-gauge. In obese patients, a 3.5-in 22-gauge “spinal” needle can be used for diagnostic paracentesis.

Using nonsterile gloves first the skin is sterilized with chlorhexidine. Sterile gloves are then used. If the original X mark has been erased during site preparation, a new mark can be placed with a sterile pen. A sterile drape with a round hole in the middle is placed over the prepped skin and another sterile drape is placed on the side next to the patient to create a wider area. The needle with the syringe to administer lidocaine is inserted intradermally to create a skin wheel. Then the same needle is advanced ideally to the peritoneum providing anesthesia along the needle track. Once the needle enters the peritoneal cavity, fluid is aspirated but this may not be accomplished in obese patients and should not preclude proceeding with the procedure if adequate ultrasound landmarks have been established. At this point the needle is withdrawn and a small skin nick is to be made with a #11 size scalpel at the exact site of needle entry. Attention should be paid to avoid an unnecessarily large skin nick as this will lead to postprocedural leakage and issues with hemostasis. A 6- or 8-Fr catheter over a blunt spring-tipped trocar is then introduced through the skin incision into the peritoneal space with a return of fluid aspirate. When the trocar is at the level of the peritoneum it is helpful to ask the patient to take a deep breath. This tightens the peritoneum and provides countertraction to the trocar making entry into the peritoneal space more comfortable with less rebound. A click of the spring-tip catheter will be appreciated indicating passage of the needle tip past the peritoneum membrane into the peritoneal space. The metal trocar is then held in place and the soft catheter is advanced into the cavity (Figure 104–4A-D). Samples are collected sterilely with the specimens for cell count and chemistries are placed in the clear tubes provided. Cultures are obtained by inoculating culture bottles directly and if cytology is needed the entire canister is sent to the laboratory to be concentrated. The catheter is then connected to a negative suction bottle or vacuum suction system (Figure 104–5A and B).

As drainage ensues, the bowel and the surrounding omentum may block the flow of the ascitic fluid. Residual ascites may be verified and occasionally the catheter tip itself can then be visualized by placing the ultrasound probe directly adjacent to the catheter insertion site. The stopcock can be opened to release the suction that may have occurred at the catheter tip and the catheter can then be pulled back a few centimeters until flow is restored. In persistent cases it may be worth trying to reposition the patient by further tilting him/her slightly toward the draining site in an attempt to shift the fluid toward the catheter before withdrawing the catheter completely. When ultrasound examination shows no further fluid to be drained, the 3-way stopcock is turned toward the patient and the catheter is gently pulled out. A sterile gauze and adhesive dressing is applied at the site.

Occasionally, we have noticed a leak from the insertion site after removal of the draining catheter. This, if unrecognized is not serious, but invariably is distressing to the patient and may cause local chemical cellulitis. To address this potential problem, we have been applying a high-viscosity tissue adhesive (Octylseal) or a skin adhesive (Dermabond) to help prevent leakage. This has effectively replaced our prior practice of suturing the puncture site with less skin irritation and no need for suture removal. Multiple references describe the practice of “Z-tracking” to prevent leakage. We have not found this to be an effective option in most patients.

Removal of 5 L of ascitic fluid in a single session is defined as large-volume paracentesis (LVP) and is considered safe without the need to replace colloids⁴ and it used to be considered the safe practical limit to relieve symptoms while preventing hepatorenal syndrome. More recent studies however document the safety of “total paracentesis” where all of the fluid is removed and albumin replacement is given. Standard albumin replacement is performed with 25% albumin administered as 6 to 8 g/L of ascites removed. Occasionally a larger dose of albumin (10 g/L removed) may be given in the setting of baseline renal insufficiency, hypoalbuminemia, or significant edema/anasarca. Additional albumin may also be given postprocedure in the setting of hypotension related to volume shifts. The largest-volume removal reported is 42 L. We have removed 38.8 L without incident.⁵

In our experience, we routinely schedule patients with recurrent ascites refractory to diuretics to return to the clinic at regular intervals that will vary per patient but in general a pattern will be established after 1 to 2 visits. We aim to see them before their ascites interferes with their ability to eat or to exercise in order to prevent muscle breakdown. This has the additional benefit of preventing unwanted emergency department visits for symptom relief and an opportunity for dietary reinforcement.

Clarity and color of the fluid should be noted. The clarity or opacity of the fluid depends largely on the presence and the amount of the neutrophil count or the presence of lipids. The amount of protein, bile, and bilirubin present will determine color. A “benign” fluid sample that has low protein and neutrophil count less than 250/mm³ is usually transparent and slightly yellow-tinged. The presence of bloody or serosanguinous fluid should be noted. A traumatic tap gives a bloody fluid that would clot easily in a nonanticoagulant-containing tube, while a blood-tinged nontraumatic ascitic fluid due to other reasons will not clot. Examples where bloody ascitic fluid may be expected are patients with hepatocellular carcinoma, trauma, and postsurgical and occasionally in peritoneal carcinomatosis. A milky or chylous fluid indicates a high triglyceride concentration and can be seen in advanced stage lymphomas. Bilirubin-stained ascitic fluid occurs in the setting of significant jaundice and appears tea-colored or dark brown. Bile-stained ascites appears greenish and is seen in patients with a bile leak and in patients with hemorrhagic or necrotic pancreatitis.

Only a limited number of diagnostic tests need to be ordered to “profile” the nature of the ascites. A cell count, total protein, and albumin are generally sufficient in most cases and need not be repeated with each paracentesis. The most important test in the symptomatic patient is the cell count and only a few milliliters are needed for evaluation. It is usually agreed that in uncomplicated cirrhosis, the white blood cell (WBC) count will be less than 500/mm³ and the absolute polymorphonuclear neutrophil (PMN) count should be less than 250/mm³. A PMN count of greater than 250/mm³ is considered presumptive evidence of SBP pending culture results but it is important to note that 10% of cases of culture-proven SBP may occur with a PMN count of less than this value. A bloody or serosanguinous ascitic fluid is more difficult to interpret however; using a correction factor subtracting one PMN for each 250 red blood cells (RBCs) may be useful in some scenarios.

Other basic diagnostic tests include Gram stain and culture (aerobic and anaerobic), albumin, total protein, and cytology. More specific tests for diagnosis are triglyceride level in chylous ascites, red blood cell count in trauma and malignancy, bilirubin concentration in bowel perforation, and amylase in pancreatitis (Table 104–2), but need not be ordered routinely.

The serum-to-ascites albumin gradient (SAAG) helps to differentiate etiologies of ascites. The SAAG can be calculated by subtracting the ascitic fluid albumin value from the serum albumin, both measured on the same day. SAAG above 1.1 g/dL suggests the presence of portal hypertension. Etiologies of ascites that are associated with portal hypertension include cirrhosis, congestive heart failure, portal vein thrombosis, and Budd-Chiari syndrome. SAAG less than 1.1 g/dL indicates absence of portal hypertension. Examples of causes of ascites without portal hypertension are nephrotic syndrome, peritoneal tuberculosis, and carcinomatosis (Table 104–3). In addition to the ascitic fluid albumin, a total ascites protein may be used to assess the risk of SBP with an increased risk associated with levels less than 1 g/dL. It is felt that this level predisposes to SBP because it correlates with complement levels and opsonic activity.

The use of colloids after a LVP has been debated, but it is generally agreed that paracenteses of 5 L or less would not routinely require albumin replacement. The usual formulation of albumin given in the United States is 25% solution. This provides less volume and less sodium load than the 5% solution. The term postparacentesis circulatory dysfunction (PCD) refers to a state of increased plasma renin activity as a marker of active hypovolemia following LVP and may be responsible for more avid fluid retention and more rapid reaccumulation of ascites. Whether albumin solution administration after LVP has value to correct hypovolemia and improve survival and whether this depends on the amount of fluid removed has been looked at in one study of 105 patients who underwent LVP. These patients were allocated to receive albumin (10 g/L of fluid removed) or no albumin. Those patients who did not receive albumin had an increase in plasma renin activity, developed hemodynamic instability, worsening renal function and hyponatremia.⁶ Additionally, albumin replacement has been posited to mitigate fluid extravasation leading to anasarca, and to improve pharmacokinetics of protein-bound medications though data are limited. As noted earlier it is our practice to routinely replace albumin based on volume removed and conditions related to hypovolemia and low oncotic pressure.

In conclusion, paracentesis is a relatively commonly performed procedure that is clearly in the purview of the internist in a variety of settings. It can provide invaluable diagnostic information and significant symptomatic relief. As a benchmark, individuals performing this procedure should maintain records of their procedural activity and be able to document major complication rates approaching 1/1000 procedures. Additionally, it is imperative that the operator be familiar with ultrasound characteristics of the patient with ascites and that ultrasound be used for guidance in all cases.