Skip to main content

Emergency Medicine Training

← Back to blog

Paracentesis Procedure: Indications, Site Selection and Step-by-Step Technique

Jason Hine 14 min read

Quick Answer Paracentesis is the placement of a needle or catheter into the peritoneal cavity to sample ascitic fluid (diagnostic paracentesis) or drain it (therapeutic paracentesis). Tap every patient with new-onset ascites that is accessible for safe sampling, and every patient with cirrhosis and ascites who is admitted emergently, to rule out spontaneous bacterial peritonitis. Enter lateral to the rectus sheath, usually in the left lower quadrant, at a site you’ve confirmed with ultrasound. In patients with cirrhosis, an elevated INR or a low platelet count is not, by itself, a reason to delay.

Most emergency clinicians have done a paracentesis. Fewer have done one recently on the patient who makes you hesitate: febrile, encephalopathic, INR of 3, no interventional radiology until morning. This guide covers the abdominal paracentesis procedure for that patient, with the evidence behind each decision.

When to Do a Paracentesis

The indication splits into two questions. Do you need fluid to answer a diagnostic question, or does the patient need fluid removed for comfort and breathing? Often the answer is both, and one needle does the job.

Diagnostic paracentesis

The 2021 AASLD practice guidance is direct: “A diagnostic paracentesis should be performed in all patients with new-onset ascites that is accessible for sampling.” It also says patients with cirrhosis and ascites who are admitted emergently should have a diagnostic tap to rule out spontaneous bacterial peritonitis (SBP), whether or not they have symptoms. A patient with ascites who develops signs or lab findings that suggest infection needs one too.

StatPearls lists the clinical picture that should prompt a tap in a patient with known ascites: abdominal pain, fever, GI bleeding, worsening encephalopathy, new or worsening renal or hepatic dysfunction, hypotension, or signs of sepsis. The AASLD recommendation for emergently admitted patients applies even when none of these are present.

Therapeutic and large-volume paracentesis

Therapeutic paracentesis is for tense ascites causing discomfort or respiratory compromise, especially when diuretics have failed. StatPearls defines large-volume paracentesis (LVP) as removal of more than 5 L. It also lists temporary decompression in abdominal compartment syndrome as an indication, with surgery remaining the definitive treatment.

Why the timing matters in the ED

Delay has a measurable cost. In a two-center study of 239 patients with SBP, Kim and colleagues found in-hospital mortality of 27% when paracentesis was delayed 12–72 hours after admission, compared with 13% when it was done within 12 hours. After adjusting for MELD score and creatinine, each hour of delay was associated with a 3.3% increase in in-hospital mortality.

The tap is also skipped outright. In a national cohort of 17,711 admissions for cirrhosis with ascites, only 61% of patients had a paracentesis. Those who did had lower in-hospital mortality (6.5% vs 8.5%, adjusted odds ratio 0.55). Both studies are observational, so they show association, not proof that the tap itself saved lives. They are still the strongest argument for doing the diagnostic tap in the ED instead of leaving it for the admitting team.

Contraindications and the Coagulopathy Question

StatPearls lists two absolute contraindications: disseminated intravascular coagulation and an acute abdomen that needs surgery. Relative contraindications include pregnancy, organomegaly, ileus, bowel obstruction, a distended bladder and surgical scars, since bowel can adhere to the abdominal wall under a scar.

Abnormal coagulation studies are where most hesitation comes from, and in patients with cirrhosis the evidence says they shouldn’t stop you. Grabau and colleagues at Mayo Clinic reported 1,100 large-volume paracenteses in 628 patients with INR values up to 8.7 and platelet counts as low as 19 × 10³/µL. They found no significant procedure-related complications, including in patients with marked thrombocytopenia or a prolonged prothrombin time.

Serious bleeding does happen, but it tracks with how sick the patient is rather than with the INR. In a 10-year series of 4,729 paracenteses, severe hemorrhage occurred in 9 procedures (0.19%), with a death rate of 0.016%. Bleeding was not associated with elevated INR, low platelets or operator experience. It occurred in patients with high MELD and Child-Pugh scores, and all but one had some degree of renal failure.

In patients with cirrhosis, the AASLD guidance does not recommend routine transfusion of plasma or platelets before paracentesis. It notes that desmopressin may be considered in uremic patients with thrombocytopenia and a history of bleeding. StatPearls adds that platelet transfusion may be considered when the count is below 20,000/µL. Those are the edges of the evidence, and all of it comes from patients with cirrhosis.

Equipment

A prepackaged paracentesis kit is the easiest option. If you’re assembling one, the list below comes from StatPearls and the Merck Manual Professional Edition.

ItemSize or detailSource
Skin anesthesia needle25- or 27-gaugeStatPearls; Merck
Deeper infiltration needle20- or 22-gaugeMerck
Local anesthetic1% lidocaine, 5 mLStatPearls
Diagnostic needle18- to 22-gauge, 3.8 cm or 8.9 cmMerck
Therapeutic needle or catheter14- to 18-gauge, or a 15-gauge Caldwell needle with outer catheter; kit catheters are often 8 French over a 15- or 18-gauge needleMerck; StatPearls
Scalpel#11 blade, for a skin nick with larger needlesStatPearls; Merck
Flow control3-way stopcock, 30- to 60-mL syringes, tubingStatPearls; Merck
DrainageVacuum bottles, plastic canister or collection bagStatPearls
Specimen containersPurple-top tube, red-top tube, aerobic and anaerobic blood culture bottlesStatPearls
ImagingUltrasound machineStatPearls

Smaller needles leak less but drain more slowly, so match the needle to the job. A diagnostic tap needs a small needle. A 6 L drain needs a catheter.

Choosing the Site and Positioning the Patient

The goal is a deep pocket of fluid, away from bowel, bladder and the inferior epigastric vessels, which run inside the rectus sheath. That leaves two standard sites, and the sources give slightly different distances for the lateral one.

SiteLandmark as publishedNotesSource
Left lower quadrantLateral to the rectus abdominis, 2–4 cm superomedial to the anterior superior iliac spine (StatPearls); 3–5 cm superior and medial to the ASIS (Merck)Described by StatPearls as the safest site without imaging, because the abdominal wall is thinner and the fluid pocket deeperStatPearls; Merck
Right lower quadrantMirror of the left-sided landmarkWatch for the cecumStatPearls
MidlineAbout 2 cm below the umbilicus, through the linea albaHave the patient empty the bladder firstStatPearls; Merck

Treat the published distances as a starting point for the probe, not the endpoint. The site you actually use is the one ultrasound shows is deep, clear of bowel and free of vessels. Avoid surgical scars, visible abdominal wall veins and infected skin.

Positioning moves the fluid to you. StatPearls describes a supine patient with a wedge under one side to tilt them slightly toward the target quadrant. Merck describes elevating the head 45–90° and rolling the patient partly onto the left side for a left lower quadrant approach, so gas-filled bowel floats away from the entry point. Either way, have the patient empty their bladder first.

Using Ultrasound

Current guidance favors ultrasound, and the bleeding data support it. The Society of Hospital Medicine position statement recommends ultrasound guidance to reduce serious complications, mainly bleeding, and to avoid attempting a tap when there is too little fluid to drain.

Its practical recommendations translate directly to the bedside:

  • Assess the volume and location of fluid, then choose the site based on pocket size, abdominal wall thickness and distance from organs.
  • Use color flow Doppler at the insertion site to find abdominal wall vessels along the planned path.
  • Look in more than one plane to confirm the path is clear.
  • Mark the site immediately before the procedure, and don’t let the patient move between marking and needle entry.
  • Consider real-time guidance when the pocket is small or hard to reach.

The outcome data are observational but consistent. In a claims analysis of 69,859 paracenteses, bleeding complications occurred in 0.8% overall, and ultrasound guidance was associated with a 68% lower risk of bleeding (odds ratio 0.32). StatPearls cites an overall complication rate of 1.4% with ultrasound vs 4.7% without.

Emergency physicians do this well. In an ED series of 131 patients, success was 97.7% with real-time ultrasound guidance and 95.6% with ultrasound-assisted site marking. All four failures were by first- or second-year residents using a curvilinear transducer, and every procedure done with a linear transducer succeeded. That is a small, single-center sample, so read it as a signal rather than a rule. If the needle keeps getting lost on a curvilinear image, a linear transducer is worth trying.

If you’ve placed lines under real-time guidance, the needle-tracking skill is the same one you use for ultrasound-guided central line placement. The decision logic (find the biggest pocket, map what’s in the way, commit) is close to our pericardiocentesis guide.

Paracentesis Steps

These paracentesis procedure steps follow StatPearls and the Merck Manual. Confirm your institution’s consent and time-out requirements before you start.

  1. Explain the procedure, its risks and alternatives, and document informed consent.
  2. Ask the patient to empty their bladder.
  3. Position the patient to pool fluid at your chosen site.
  4. Scan to confirm the pocket, check for bowel, bladder and vessels with color Doppler, and mark the site.
  5. Clean the skin with chlorhexidine or povidone-iodine, let it dry, and drape sterilely.
  6. Raise a skin wheal with 1% lidocaine through a 25- or 27-gauge needle.
  7. Infiltrate deeper along the planned tract, aspirating intermittently to check for blood, until you aspirate ascitic fluid.
  8. Pull the skin caudally (Z-track) or insert at about 45°.
  9. Advance the paracentesis needle or catheter slowly, using intermittent rather than continuous suction.
  10. Stop advancing when fluid flows freely, usually after a palpable give as you enter the peritoneum.
  11. Release the skin.
  12. Withdraw 25–50 mL for diagnostic studies and fill the culture bottles at the bedside.
  13. For a therapeutic drain, advance the catheter over the needle, remove the needle, and connect the stopcock and tubing to the drainage container.
  14. Remove the catheter when drainage is complete and hold pressure over the site.

The Z-track matters more than it looks. Displacing the skin before entry leaves a misaligned tract when it’s released, which StatPearls and Merck both describe as the main defense against persistent leakage. Intermittent suction keeps omentum or bowel from being pulled onto the tip and lets you see blood before you’ve gone further.

When the flow stops

Bowel or omentum usually occludes the catheter tip. Merck suggests rotating the catheter or advancing it in 1–2 mm increments, then briefly releasing suction at the stopcock before resuming. You can also withdraw in 1–2 mm steps, but a catheter that leaves the peritoneum can’t be pushed back in, so go slowly.

Sending the Fluid

The AASLD guidance lists the initial ascitic fluid workup as a neutrophil count, total protein and albumin, with a same-day serum albumin to calculate the serum-ascites albumin gradient (SAAG). Culture belongs at the bedside, in aerobic and anaerobic blood culture bottles, before antibiotics are given.

TestWhat it answersSource
Cell count with differentialSBP is diagnosed with a fluid polymorphonuclear count of 250 cells/mm³ or higherAASLD 2021
Culture in blood culture bottles, at the bedsideIdentifies the organism; inoculate before antibioticsAASLD 2021; StatPearls
Ascitic albumin with same-day serum albumin (SAAG)SAAG of 1.1 g/dL or higher points to portal hypertension; below 1.1 g/dL points elsewhereAASLD 2021; StatPearls
Total proteinPart of the initial workup; interpret it alongside the SAAG, not alone. With a SAAG of 1.1 g/dL or higher, StatPearls notes protein above 2.5 g/dL suggests heart failure and below 2.5 g/dL suggests cirrhosisAASLD 2021; StatPearls
TriglyceridesMilky fluid suggests chylous ascitesStatPearls
Cytology, LDH, adenosine deaminase, tumor markersBloody or unexplained fluid raises concern for malignancy or tuberculous peritonitisStatPearls

Don’t wait for the cell count to treat a patient who looks septic. StatPearls advises starting empiric antibiotics when SBP is strongly suspected, before results return.

Large-Volume Paracentesis and Albumin

Albumin is the one part of a therapeutic drain that is easy to forget at 3 a.m. The AASLD guidance recommends albumin infusion at the time of LVP when more than 5 L is removed, to reduce the risk of post-paracentesis circulatory dysfunction (PPCD). The dose is 6–8 g for every liter removed, calculated on the total volume drained and not only the liters above 5, which the guidance labels as expert opinion. Some higher-risk patients may need albumin even when less than 5 L is removed.

The trial evidence is stronger than the dosing. A meta-analysis of 17 randomized trials with 1,225 patients found that albumin, compared with alternative treatments, reduced PPCD (odds ratio 0.39), hyponatremia (odds ratio 0.58) and mortality (odds ratio 0.64).

On volume, Merck states that removal of 5–6 L is generally well tolerated, and up to 8 L in some patients. The AASLD guidance notes that PPCD risk may rise above 8 L in a single session. After a large-volume drain, Merck recommends monitoring blood pressure for 2–4 hours.

Complications

Paracentesis complications are uncommon, and most are minor. The list from StatPearls and Merck:

  • Persistent leakage of ascitic fluid from the site, which StatPearls notes is often managed with a single skin suture.
  • Abdominal wall hematoma or bleeding, rarely severe and, as above, concentrated in the sickest patients.
  • Infection at the site.
  • Perforation of bowel or a vessel, which StatPearls describes as extremely rare.
  • Hypotension after large-volume removal, along with transient hyponatremia or a rise in creatinine.

StatPearls lists hypotension, tachycardia and hemorrhage among the things to watch for after the procedure. Keep Pache and Bilodeau’s finding in mind here: the patients who bled severely were the ones with high MELD scores and renal failure.

Keeping the Skill Between Taps

Paracentesis has a short learning curve but a real one. Grabau’s Mayo Clinic team found that 3–7 supervised procedures were needed before competence. Independent practice depends on demonstrated competence and institutional credentialing requirements, not a fixed procedure count. If you want the broader picture of how procedural competence builds, our podcast episode on learning curves walks through the intubation literature.

The harder problem comes later. A clinician who drains ascites weekly on a hepatology service and then moves to a community ED may go months between taps, and the skill decay problem applies here as it does to rarer procedures. The SHM position statement recommends simulation-based practice, when available, to build the knowledge and skills for paracentesis, and it recommends demonstrated competence before anyone performs the procedure independently.

Simulation supports confidence and retention. It doesn’t replace supervised experience or your institution’s credentialing, and the evidence linking it to patient outcomes is still thin, as we covered in what the simulation literature does and doesn’t show. SimKit doesn’t ship a paracentesis model. What carries over from practicing other needle procedures is the sequence itself: map the target, then commit at the decision point.

FAQ

How long does a paracentesis take? Large-volume drains take the longest. In Grabau’s series of 1,100 large-volume paracenteses, the mean procedure time was 97 ± 24 minutes for a mean of 8.7 L removed.

How much fluid can you remove at once? Merck states that 5–6 L is generally well tolerated and up to 8 L in some patients, and the AASLD guidance notes PPCD risk may rise above 8 L in one session. Give albumin at 6–8 g per liter of total fluid removed whenever more than 5 L comes off, and some higher-risk patients may need it with smaller volumes.

Do you need to correct the INR or platelets before a paracentesis? In patients with cirrhosis, usually not. The AASLD guidance doesn’t recommend routine plasma or platelet transfusion, and Grabau’s series reported no significant complications with INR values up to 8.7 and platelets as low as 19 × 10³/µL.

Should you use the left or right lower quadrant? StatPearls describes the left lower quadrant as the safest site without imaging because the abdominal wall is thinner and the pocket is often deeper. With ultrasound, use whichever side shows the largest pocket clear of bowel and vessels.

What’s the difference between paracentesis and thoracentesis? Paracentesis removes ascitic fluid from the peritoneal cavity, while thoracentesis removes pleural fluid from the chest. Both are needle drainage procedures that benefit from ultrasound guidance.

Conclusion

A paracentesis is a low-risk procedure that gets delayed for high-risk reasons. Tap early when SBP is on the list, don’t let an INR hold you up in a patient with cirrhosis, find your site with ultrasound, and give albumin when more than 5 L comes off.

If you want structured practice on the procedural skills you use less often, the Procedural Training Course Year 2 covers central venous catheterization, pericardiocentesis, transvenous pacing and cranial burr hole.

References

  1. Biggins SW, Angeli P, Garcia-Tsao G, et al. Diagnosis, evaluation, and management of ascites, spontaneous bacterial peritonitis and hepatorenal syndrome: 2021 practice guidance by the American Association for the Study of Liver Diseases. Hepatology. 2021;74(2):1014-1048. doi:10.1002/hep.31884
  2. Kurapati R, Katta S, O’Rourke MC. Paracentesis. In: StatPearls. StatPearls Publishing; updated September 16, 2025. https://www.ncbi.nlm.nih.gov/books/NBK435998/
  3. Malik Z. How to do paracentesis. Merck Manual Professional Edition. Reviewed April 2025. https://www.merckmanuals.com/professional/gastrointestinal-disorders/how-to-do-gastrointestinal-procedures/how-to-do-paracentesis
  4. Cho J, Jensen TP, Reierson K, et al. Recommendations on the use of ultrasound guidance for adult abdominal paracentesis: a position statement of the Society of Hospital Medicine. J Hosp Med. 2019;14(1). doi:10.12788/jhm.3095
  5. Kim JJ, Tsukamoto MM, Mathur AK, et al. Delayed paracentesis is associated with increased in-hospital mortality in patients with spontaneous bacterial peritonitis. Am J Gastroenterol. 2014;109(9):1436-1442. doi:10.1038/ajg.2014.212
  6. Orman ES, Hayashi PH, Bataller R, Barritt AS IV. Paracentesis is associated with reduced mortality in patients hospitalized with cirrhosis and ascites. Clin Gastroenterol Hepatol. 2014;12(3):496-503.e1. doi:10.1016/j.cgh.2013.08.025
  7. Grabau CM, Crago SF, Hoff LK, et al. Performance standards for therapeutic abdominal paracentesis. Hepatology. 2004;40(2):484-488. doi:10.1002/hep.20317
  8. Pache I, Bilodeau M. Severe haemorrhage following abdominal paracentesis for ascites in patients with liver disease. Aliment Pharmacol Ther. 2005;21(5):525-529. doi:10.1111/j.1365-2036.2005.02387.x
  9. Mercaldi CJ, Lanes SF. Ultrasound guidance decreases complications and improves the cost of care among patients undergoing thoracentesis and paracentesis. Chest. 2013;143(2):532-538. doi:10.1378/chest.12-0447
  10. Bernardi M, Caraceni P, Navickis RJ, Wilkes MM. Albumin infusion in patients undergoing large-volume paracentesis: a meta-analysis of randomized trials. Hepatology. 2012;55(4):1172-1181. doi:10.1002/hep.24786
  11. Wubben BM, Dandashi J, Rizvi O, Adhikari S. Emergency physician performed ultrasound-guided abdominal paracentesis: a retrospective analysis. POCUS J. 2024;9(1):67-71. doi:10.24908/pocus.v9i1.16668

Need help? We're here to assist you!

Join Our Newsletter

* indicates required