Managing Ascites: Hazards of Fluid Removal
A 50-year-old man with longstanding alcoholic cirrhosis presented to the emergency department (ED) with several days of progressive abdominal pain and worsening distension. He had no fevers or other infectious symptoms, bleeding, or recent changes to his medications. He also reported that his last drink was more than a year ago, reflecting a behavior change motivated by his desire for liver transplant consideration. In the ED, he was afebrile with normal vital signs and no evidence of jaundice. His abdominal exam was notable for marked distension with a fluid wave and a bedside ultrasound that confirmed significant ascites. The patient last required a paracentesis several months earlier for similar symptom relief. His laboratory results were notable for a stable anemia, a low platelet count of 50, an elevated INR of 2.5, and a mild transaminitis and leukocytosis. In the ED, his working diagnosis was acute decompensation of cirrhosis without a clear etiology.
Eager to relieve his symptoms, the providers performed a large volume paracentesis. While there was discussion about "how much to take off," the patient became acutely hypotensive as the 10th liter of fluid was being removed. The hypotension failed to respond to initial fluid resuscitation, and the patient required a short duration of vasopressors in the intensive care unit before being weaned off. He was ultimately discharged after a 3-day hospitalization with an increase in his outpatient diuretic regimen and close follow-up in the liver clinic. The case prompted a formal review since the entire hospitalization was deemed preventable given the belief that his hypotension was a direct result of the aggressive fluid removal attempts.
by Kevin Moore, MBBS, PhD
Approach to managing patients with ascites
Ascites is the most common complication of cirrhosis. Approximately 60% of patients with compensated cirrhosis develop ascites, accompanied by portal hypertension, within 10 years.(1,2) The development of ascites is caused by impaired ability to excrete sodium into urine, leading to a positive sodium balance, and is associated with a poor prognosis in patients with cirrhosis. The probability of survival at 2 and 5 years after the appearance of ascites in patients with cirrhosis is estimated to be 60%–70% and 30%–40%, respectively.(1) Hyponatremia is present in approximately 50% of patients presenting with ascites and significant hyponatremia is associated with poor prognosis. Finally, and relevant to this case presentation, patients with alcoholic liver disease who have stopped drinking have continued improvement of liver function for up to 12 months following cessation of alcohol intake. Therefore, the re-occurrence of ascites in someone who is supposedly abstinent raises the possibility of renewed alcohol intake, or hepatic decompensation secondary to a bacterial infection. The mainstay to managing patients with ascites is modest salt restriction, treatment with diuretic therapy, and cessation of alcohol in patients with alcoholic liver disease.
Common pitfalls in managing ascites
Diagnosing spontaneous bacterial peritonitis. Many providers assume that a patient without fever or abdominal pain cannot have spontaneous bacterial peritonitis (SBP). However, fever is only present in approximately 50% of patients with SBP. Therefore, the threshold for performing a diagnostic paracentesis and sending the fluid for culture should be very low, as the consequences of missing the diagnosis can be life-threatening. Furthermore, every patient presenting with decompensated cirrhosis and ascites warrants a diagnostic paracentesis. An ascitic neutrophil count greater than 250 cells/mm3 is considered diagnostic of presumed SBP and is present in about 15% of patients with ascites. It is unclear in this case whether a diagnosis of SBP was present or excluded, but the evaluation for it should have occurred based on the presentation described. Patients with decompensated cirrhosis should receive a thorough evaluation for other evidence of systemic infection, including a urinalysis, blood cultures, and a chest radiograph.
Look for impairment of kidney function and electrolyte abnormalities. It is critical to evaluate kidney function and electrolyte abnormalities in all patients presenting with ascites. This is because the risk of developing acute kidney injury and/or hepatorenal syndrome is highest in those with a significant deterioration of their liver function. Failure to recognize this important relationship between liver and kidney function can impact treatment and prognosis. The patient in this case has significant liver dysfunction, evidenced by an elevated INR at 2.5. It is important to determine whether the INR has increased significantly compared to baseline as this may indicate an acute decompensation in the presentation.
Re-evaluate alcohol intake. It is critical to carefully question patients about recent alcohol intake, whether they have a longstanding history, report abstinence, or recently began drinking. In patients being evaluated for liver transplantation, blood alcohol levels should also be measured upon hospital admission since the presence of alcohol confirms continued drinking and precludes liver transplantation. Given the importance of alcohol use in patients being considered for transplantation, providers should also consider measuring urinary ethyl glucuronide, which remains elevated up to 90 hours after last alcohol intake.(3) The importance of uncovering alcohol use in all patients with ascites goes beyond simply precluding liver transplantation. If patients can stop drinking, their liver function will improve and their ascites may resolve; they will also have a better prognosis.
Take a dietary and drug history. Has the patient's diet changed recently? Is the patient taking antacids rich in salt? Has the patient stopped a prescribed diuretic? For many patients with compensated cirrhosis, even seemingly subtle changes in their salt intake or medication adherence can impact their liver function. Perhaps the patient in this case was eating more pre-prepared meals, which are high in salt content, or he relaxed his strict no-salt regimen because he hadn't needed paracentesis for a while.
Measure urine sodium concentration. Measurement of a urine sodium concentration in a random urine sample is an easy test and can help predict response to subsequent diuretic therapy. If the urine sodium is high, then patients are salt excreters who are ingesting too much salt. If the level is very low (below 10 mmol/L or 10 mEq/L) then they are salt retainers. This simple test helps gauge future therapy. For example, when treating with a diuretic (e.g., spironolactone), if a patient's urine sodium has increased from 20 mEq/L to 90 mEq/L, it raises confidence that the patient will respond to therapy; whereas, if urine sodium remains very low or is virtually undetectable, then the patient is less likely to respond.
Is the diuretic therapy optimal? The aim of diuretic therapy is to enhance salt and water excretion. Unless this is done in conjunction with a modest reduction of salt intake, then ascites is unlikely to resolve in the short term. Thus, diuretics are given to initially enable resolution of the ascites, and then to prevent recurrence. What is the best diuretic to use? Many providers believe that furosemide is better than spironolactone based on the experience that the former medication causes patients to produce a lot more urine than the latter. However, controlled clinical studies have shown that spironolactone is far superior—its use leads to resolution of ascites in almost 90% of patients, compared to about 50% of those given furosemide alone. This finding is not surprising. Many of these patients have elevated circulating aldosterone concentrations, which leads to avid salt retention that overcomes the loop diuretic effect of furosemide; aldosterone acts on a different part of the kidney (i.e., the distal tubule) to achieve its beneficial impact.
Managing patients with large volume ascites
In general, large volume paracentesis is recommended for patients presenting with tense and large volume, such as the patient in this case. We typically characterize more than 10 liters as large volume, and this patient had at least 10 liters removed. For patients requiring intermittent paracentesis for symptom relief, it would also be important to ensure they are taking diuretic therapy or not requiring an increase in that therapy. Measurement of a post urine sodium concentration can help in this assessment. Otherwise, for patients with large ascites, the recommendation is to simply drain it, and drain it as rapidly as possible.
There are several reasons for this approach. First, let's consider the hemodynamic consequences of paracentesis. As you drain fluid, the pressure in the abdomen falls exponentially and very rapidly, leading to a rapid reduction in right atrial pressure. Since the right atrium tends to incompletely fill in the setting of tense ascites, and venous return increases to the right atrium as ascites is drained, there is a resulting increase in cardiac output due to splanchnic hyperemia. The vasodilatation that occurs in the splanchnic bed leads to a lowering of mean arterial pressure and a compensatory an increase in cardiac output. Thus, most if not all patients undergo some hemodynamic changes during paracentesis. If a patient develops a large fall in blood pressure, as occasionally occurs, I would give vasopressor therapy as the clinicians did in this particular case. I would not regard this patient's fluid removal as too much or too fast. I would regard it as potentially predictable and just treat the complication as needed.
Does the patient have SBP? And if so, does that affect your management? There are no data on the safety of large volume paracentesis in patients with SBP. My personal practice is that if I had a patient with SBP and large volume ascites, I would drain it and treat with antibiotics. That said, it would not be considered wrong to treat the infection and leave the ascitic fluid in place, since there are no data to support one action over another.
When should the patient be given albumin? It is now well established that failure to give albumin after paracentesis can result in post paracentesis circulatory dysfunction. However, it is often not clear when to give albumin post paracentesis to prevent this complication. All paracenteses should be completed as rapidly as possible, within about 4 hours of commencing the paracentesis (i.e., do not leave a draining cannula in situ overnight or for more than 6 hours). This recommendation is based on our early experience of a high incidence of infection when the draining cannula was left in situ (unpublished). My personal recommendation is that albumin should be given at the end of paracentesis, rather than during paracentesis. This recommendation is based on hemodynamic studies that we undertook 25 years ago (4), which observed that the reduction of intravascular volume does not begin until about 3 hours after the paracentesis has commenced, and further at the end of paracentesis, you will know how much albumin to prescribe.
In reflecting on the case presented, some information that was not provided may have improved the care for this patient's decompensated cirrhosis, including a more comprehensive evaluation of precipitating factors and evaluation for SBP. However, it does appear there may have been an error in managing his large volume paracentesis if albumin was not provided, as it could have prevented the need for vasopressors.
- Spontaneous bacterial peritonitis should be considered in all patients presenting with ascites. A diagnostic paracentesis should be performed in all patients presenting with ascites since fever and abdominal pain are not always present.
- Common pitfalls in managing patients with ascites require a systematic approach to understanding the factors that lead to acute decompensation (e.g., infection, dietary indiscretion, alcohol intake, poor medication adherence, etc.).
- When performing a large volume paracentesis, it should be completed as rapidly as possible once started but at least within 6 hours, with a goal to remove as much fluid as possible.
- Large volume paracentesis is generally safe, but blood pressure may fall and require treatment.
- Albumin should be used to prevent hemodynamic consequences from large volume paracentesis. My practice is to give albumin at the end of paracentesis, not during it.
Kevin Moore, MBBS, PhD Professor of Hepatology UCL Institute for Liver and Digestive Health Royal Free Campus University College London
1. European Association for the Study of the Liver. EASL clinical practice guidelines on the management of ascites, spontaneous bacterial peritonitis, and hepatorenal syndrome in cirrhosis. J Hepatol. 2010;53:397-417. [go to PubMed]
2. Angeli P, Wong F, Watson H, Ginès P; CAPPS Investigators. Hyponatremia in cirrhosis: results of a patient population survey. Hepatology. 2006;44:1535-1542. [go to PubMed]
3. Walsham NE, Sherwood RA. Ethyl glucuronide and ethyl sulfate. Adv Clin Chem. 2014;67:47-71. [go to PubMed]
4. Panos MZ, Moore K, Vlavianos P, et al. Single, total paracentesis for tense ascites: sequential hemodynamic changes and right atrial size. Hepatology. 1990;11:662-667. [go to PubMed]