About 50% of patients with cirrhosis develop ascites within 10 years of diagnosis [1,2]. The main drivers are sodium retention due to splanchnic arterial vasodilatation and perceived hypovolaemia (acting via the renin–angiotensin–aldosterone and sympathetic nervous systems) and portal hypertension [1–3]. Refractory ascites is classified either as resistant (ie, no response to diuretic treatment with spironolactone 400mg/day, furosemide 160mg/day and sodium restriction for >7 days) or intractable (ie, the patient does not tolerate diuretics). Lack of response is also characterised by weight loss of less than 0.8 kg over 4 days and urinary sodium concentrations being lower than sodium intake. Early recurrence is indicated by the reappearance of ascites with a moderate to large volume within 4 weeks of fluid mobilisation. Patients might experience diuretic-induced complications, such as encephalopathy, renal impairment, hyponatraemia and hypo- or hyperkalaemia. About half of patients with refractory ascites die within 1 year of onset [5,6] and, therefore, liver transplantation should be considered if feasible (Figure).

In about 20% of patients with cirrhosis, ascites has an alternative cause [3]. These include thrombosis in the portal or hepatic vein, pancreatitis, cardiac disease (constrictive pericarditis might be present in patients with recurrent ascites) and abdominal tuberculosis or malignancy. Initial or repeat investigations to consider are full fluid analysis and imaging of the heart, liver and other abdominal and pelvic organs. Testing blood for tumour markers might be misleading when ascites is present [4]



Large-volume paracentesis

Repeated large-volume paracentesis (LVP) with albumin replacement is a safe and effective first-line treatment for refractory ascites [1–5,7,8]. Albumin infusion prevents post-paracentesis circulatory dysfunction. This complication is caused by fluid shift, hypovolaemia and further activation of the renin–angiotensin system [9], and can lead to renal-failure hyponatraemia, rapid re-accumulation of ascites, and shortened survival.

In a study of 171 LVP procedures, two patients (1.2%) had significant haemorrhages and one (0.6%) had a bowel perforation [10]. In another study of 4729 patients, significant haemorrhage occurred in nine patients (0.2%), among whom eight had renal impairment [11].

Severity of coagulopathy does not predict bleeding in patients with cirrhosis [2–4, 12, 13] and no data support routine use of fresh-frozen plasma or platelets before LVP. Guidelines acknowledge the absence of evidence-based cut-off values for coagulation parameters contraindicating paracentesis [2–4]. Nevertheless, platelet count might be a useful predictor of bleeding [4,12,13]. Contraindications to LVP include disseminated intravascular coagulation, bowel distension, pregnancy, infection at the proposed puncture site and an uncooperative patient [3]. Specialist vascular guidelines suggest continuing aspirin, but, if feasible, stopping clopidogrel or ticagrelor 5 days before the procedure, which is graded as having a low risk of bleeding [14], and seeking cardiological advice when needed [14,15].



Diuretics and non-selective beta-blockers

Diuretics are usually ineffective, but, if tolerated, they might be tried in patients who have urinary sodium excretion greater than 30 mmol per day [3].

Use of non-selective beta-blockers to treat advanced liver disease remains controversial [16,18]. These drugs may be continued if the patients demonstrates cardiovascular stability [16], but should be stopped or the dose reduced if  systolic blood pressure falls below 90 mmHg or if renal impairment or hyponatraemia (serum sodium <130 mEq/L) develop [3,18].

Non-steroidal anti-inflammatory drugs, angiotensin-converting-enzyme inhibitors, and angiotensin-II-receptor or alpha-1 adrenergic-receptor blockers should be avoided due to the risk of renal impairment [3,4].


Transjugular intrahepatic portosystemic shunts

Transjugular intrahepatic portosystemic shunts (TIPS) dramatically reduce portal pressure in patients with cirrhosis [2–4]. Meta-analyses of randomised controlled trials of TIPS versus continued LVP have indicated that use of TIPS effectively controls ascites and reduces the need for LVP [19,20]. TIPS also improve renal function and possibly (although not conclusively) transplant-free survival [19,20]. The main complication is hepatic encephalopathy, for which incidence is about 25% over 6 months and is higher than that in patients managed by LVP [21].

The Model for End-Stage Liver disease (MELD) score (based on international normalised ratio of prothrombin time and serum bilirubin and creatinine concentrations) predicts 3-month survival following TIPS [1,2]. The risk of encephalopathy is also predicted by higher MELD scores and low mean arterial pressure.

Contraindications to TIPS include hepatic encephalopathy and a high MELD and/or Child-Pugh score (although agreement on optimum cut-off values is lacking), heart failure, pulmonary hypertension, multiple hepatic cysts, portal or hepatic vein thrombosis, large central hepatic tumours and severe thrombocytopenia [22].

Small-diameter polytetrafluroethylene-covered stents are now favoured over uncovered stents [3,23] as they reduce the risk of TIPS dysfunction (including stenosis, thrombosis or migration) and ascites re-accumulation. Use of these stents was associated with better survival than repeated LVP in one randomised controlled trial [24]. Furthermore, encephalopathy risk is not increased and might be reduced if smaller (8 mm) covered stents are used [25]


Other treatments

In a pilot randomised controlled trial in 40 patients with cirrhosis and refractory ascites, midodrine added to diuretics achieved better control of ascites and possibly survival than diuretics alone [4,26]. This and other vasoactive drugs, such as vasopressin, alpha-2 agonists and vaptans, require further evaluation before they can be recommended for routine use [26–29].

Regular infusions of albumin in patients with decompensated cirrhosis reduced the need for LVP and improved survival in one randomised controlled trial [30] but not in another in which lower doses were used and combined with midodrine [31]. In a non-randomised study involving patients with refractory ascites, albumin administration improved ascites control and reduced the 2-year hospitalisation and mortality rates [32].

The Alfapump (Sequana, Ghent, Belgium) is an automated low-flow pump system that is implanted subcutaneously and pumps ascitic fluid into the bladder at an adjustable rate [33]. It might be considered for patients unsuitable for TIPS or liver transplant [3]. A multicentre randomised controlled trial involving 58 patients showed improved ascites control but a serious complication rate of 85% compared with 45% in those who underwent LVP [34].

The PleurX drainage system (BD, Eysins, Switzerland) is a tunnelled indwelling peritoneal catheter that represents a potential community-based palliative treatment for refractory cirrhotic ascites to improve quality of life [35]. Typically, 2 L of ascitic fluid is drained three times per week under district nurse supervision. In a retrospective study, drains remained patent in 30 of the 33 patients over a median of 117 days [35]. Micro-organisms from the catheter were isolated in 11 (38%) of patients after a median of 105 days. All patients were treated successfully. In another study of 25 patients with cirrhotic refractory ascites, mortality did not differ between patients treated with PleurX and those treated with LVP [36]. Prophylactic use of antibiotics has been suggested in recipients of PleurX, but evidence of a benefit is lacking [37].



Figure 1: Management protocol for refractory ascites in patients with decompensated cirrhosis

LVP: large volume paracentesis; TIPS: transjugular intrahepatic portosystemic shunts.



Refractory ascites remains difficult to manage with detrimental impact on patient’s quality of life and significant increase in mortality. Early consideration for liver transplantation and feasibility of TIPS is important in the selected group of patients. Novel therapies are promising but require further validation. Palliative approach in end stage liver disease has an important role in symptom control.



Dr Thazin Min, Liver Unit, Sheffield Teaching Hospitals, Sheffield, UK

Professor Dermot Gleeson, Liver Unit, Sheffield Teaching Hospitals, Sheffield, UK

Learning points

  • Consider the possibility of alternative aetiology for ascites
  • Early consideration of the feasibility of transjugular intrahepatic portosystemic shunts (TIPS) and of liver transplantation in selected patients
  • Other alternative treatment options if conventional treatment fails or unsuitable