Portal hypertension-related ascites (SAAG ≥11 g/L)

The possibilities of intrinsic liver disease, right heart failure and venous obstruction must be addressed in patients with portal hypertension-related ascites. Standard liver function tests are important but may be entirely normal. Abdominal imaging, generally with ultrasound in the first instance, may demonstrate features of cirrhosis or extensive infiltration, along with other features of associated portal hypertension, such as splenomegaly and venous collaterals. Evidence of cytopenia on full blood count may point to hypersplenism. A clinical association between cirrhotic ascites and the development of hepatorenal syndrome, especially in hyponatraemic patients, is well recognised and, accordingly, markers of renal function and the serum sodium concentration can be important clues pointing to otherwise clinically inapparent (silent) liver disease with portal hypertension as the cause of ascites. Liver biopsy may be required. To minimise the risk of haemoperitoneum, this should be performed only after ascites has been effectively treated. The transjugular biopsy route is an alternative to the percutaneous approach if this is problematic. Coagulopathy must be reversed with clotting factor support and/or platelet transfusion prior to biopsy. If right heart failure is suspected, investigations such as echocardiography and lung function tests may be needed to establish the cause.

Patency of the portal and hepatic veins is usually adequately assessed by Doppler-ultrasound or dynamic computed tomography (CT). A high index of clinical suspicion is important for diagnosing hepatic veno-occlusive disease and Budd-Chiari syndrome, the latter resulting from hepatic venous outflow block. Tender, non-pulsatile hepatomegaly is a clinical hallmark of these disorders, although can be difficult to detect when ascites is tense. An underlying hypercoagulable state, most commonly related to a myeloproliferative disorder, is present in approximately 75% of patients with Budd-Chiari syndrome and should be sought. The possibility of an underlying malignancy should also be excluded. Associated thromboses in the inferior vena cava and/or portal vein are present in up to 20% of cases.

Ascites due to causes other than portal hypertension (SAAG <11 g/L)

Cytological examination of ascitic fluid remains the gold standard test for ascites due to peritoneal malignancy, although sensitivity is as low as 40%–60% in some series, even when up to 500 mL of fluid is analysed. Ascitic fluid fibronectin levels have been reported to have high diagnostic accuracy in discriminating between malignant and non-malignant causes of ascites. Laparoscopy-guided peritoneal biopsy may be required. The latter is the most reliable test for tuberculous peritonitis. A raised adenosine deaminase level in ascitic fluid has a specificity >90%, although reported sensitivities are widely variable. Mycobacteria are identified on culture of ascitic fluid in fewer than 20% of cases. Molecular diagnosis by polymerase chain reaction techniques is becoming more routinely available. Over 70% of patients with tuberculous peritonitis have no evidence of pulmonary disease. A raised ascitic amylase level, generally in the order of 2,000 IU/L, is typical of pancreatic ascites, while an ascitic bilirubin value in excess of that in serum is compatible with a biliary aetiology. Further imaging of the pancreas and biliary tree, respectively, should be undertaken in these latter circumstances.

Spontaneous bacterial peritonitis

The spectrum of clinical features of spontaneous bacterial peritonitis (SBP) extends from a fulminant presentation with peritonism and shock to an asymptomatic state manifest only by deterioration in laboratory variables. Abdominal pain with fever or hypothermia, often with an increased volume of ascites which may become refractory to diuretic therapy, are characteristic features. However, the majority of patients present with such non-specific symptoms as general malaise, anorexia, nausea or vomiting. Presentation with deterioration in mental status due to precipitation or exacerbation of hepatic encephalopathy is well recognised. Consequently, a high index of suspicion is mandatory if patients with SBP are to be identified and this diagnosis should be sought in any patient with ascites whose clinical status deteriorates.

An ascitic fluid neutrophil count ≥250/μL is the single most reliable test for SBP, with sensitivity approaching 90%. Other causes of neutrocytic ascites (as listed in Figure 31.1) must be excluded but, due to the relative frequencies with which these occur in cirrhotic patients, most instances of a raised ascitic neutrophil count in patients with cirrhosis are due to SBP. The sensitivity of traditional ascitic fluid culture is only 33%. This is improved to around 75% by the bedside inoculation of ascitic fluid into blood culture bottles. Gram stain of ascitic fluid is usually negative, as the concentration of infecting bacteria is generally only very low. Since neither the ascitic fluid neutrophil count nor ascitic culture has 100% sensitivity for SBP, even in combination, it is appropriate to make a clinical diagnosis of this disorder in patients with a suggestive clinical picture, even when these standard tests are negative. Assessment for the presence of bacterial DNA in ascites may be helpful in this latter circumstance, but is not yet widely available in clinical practice. The SAAG has no value in identifying patients with SBP. This gradient remains ≥11 g/L in patients with underlying portal hypertension-related ascites and <11 g/L in those with SBP complicating ascites of another aetiology, such as nephrotic syndrome.

SBP must be differentiated from secondary bacterial peritonitis due to perforation of the gut or an intra-abdominal abscess. A polymicrobial ascitic culture, especially in association with a raised ascitic neutrophil count, is an important clue to these latter possibilities. Gram stain may be positive for multiple organisms, especially in the case of secondary bacterial peritonitis related to gut perforation. A polymicrobial culture in association with a neutrophil count <250/μL suggests needle perforation of the gut during diagnostic or therapeutic paracentesis.

Chylous ascites

When ascites appears milky, the triglyceride content should be measured in order to differentiate between true chylous and pseudochylous ascites, the latter due to scattering of light by aggregates of cholesterol, phospholipid and protein derived from degenerating malignant or inflammatory cells. Lymphatic obstruction in association with lymphoproliferative disorders or tuberculosis is a common cause of chylous ascites. This may also occur in patients with nephrotic syndrome and those with uncomplicated cirrhosis, due to rupture of overloaded lymphatics.

Portal hypertension-related ascites

As sodium retention is a key feature in the pathogenesis of ascites related to portal hypertension, the goal of medical therapy is to induce a negative sodium balance by use of diuretics in combination with dietary sodium restriction, aiming for loss of body weight in the order of 0.5–1.0 kg/24 h, without precipitating hyponatraemia or azotaemia (Table 31.3). Refractory ascites, defined by lack of response to a sodium-restricted diet and high-dose diuretic therapy, occurs in a minority of patients. Second-line options for this group are listed in Table 31.3. Specific interventions may be necessary in carefully selected patients with ascites due to Budd-Chiari syndrome or veno-occlusive disease. These include thrombolytic therapy, angioplasty with or without venous stenting, transjugular intrahepatic portosystemic shunt (TIPS), surgical portacaval or mesoatrial shunt and orthotopic liver transplantation (OLT). Treatment of any underlying hypercoagulable state is mandatory in those with Budd-Chiari syndrome.

TABLE 31.3 Treatment options for ascites related to portal hypertension

Ascites due to causes other than portal hypertension

The primary focus of treatment in patients with non-portal hypertension-related ascites is on the underlying cause, such as antituberculous chemotherapy in those with tuberculous peritonitis or endoscopic or percutaneous stenting of a biliary leak in cases of biliary peritonitis. Therapeutic paracentesis may be required. The intraperitoneal installation of cytotoxic agents may provide effective palliation in selected patients with chemotherapy-sensitive tumours.

Spontaneous bacterial peritonitis (SBP)

Empiric antibiotic treatment should be commenced immediately after diagnostic paracentesis in patients in whom SBP is clinically suspected or after the demonstration of neutrocytic ascites in other patients. It is inappropriate to wait for the result of ascitic culture in view not only of the suboptimal sensitivity of this test but also of the risk of rapid clinical deterioration. The antibiotic of choice is generally a third-generation cephalosporin. The quinolones, ofloxacin or ciprofloxacin, are also efficacious. The appropriateness of the chosen antibiotic regimen may be reviewed when ascitic fluid culture and sensitivity results eventually become available. Granulocyte-macrophage colony stimulating factor has been shown in vitro to reverse defects in neutrophil phagocytosis and chemotaxis in cirrhotic patients and may have a role in the management of SBP in selected patients, such as those with severe infection and an incomplete initial response to antibiotics.

Although resolution of SBP is usually achieved with the use of appropriate antibiotics and despite improvements in the general medical care of cirrhotic patients, the in-hospital mortality rate remains 20%–40%. This is largely related to underlying hepatic decompensation and a high prevalence of hepatorenal syndrome. The latter occurs in approximately 30% of patients with SBP, predominantly in those with pre-existing renal impairment, and is progressive despite cure of infection in half of these cases. The development of renal failure is the most important predictor of in-hospital mortality associated with SBP. Survivors of an episode of SBP should be evaluated for orthotopic liver transplanation in view of the high risk of recurrence and poor overall prognosis.