The development of ascites in patients with cirrhosis occurs via a complex mechanism and molecular pathway known as the “peripheral arterial vasodilation hypothesis.” The initial event appears to be the development of portal hypertension. This occurs in the setting of advanced fibrosis which leads to architectural distortion and increased hepatic vascular tone via the release of various vasoconstrictive agents such as angiotensin, endothelin, and thromboxane, as well as reduced levels of hepatic nitric oxide. This elevation of portal pressures leads to a reflexive systemic vasodilatation. This process is primarily driven by peripheral release of nitric oxide leading to a diffuse vasodilatory state, especially dilatation of the splanchnic circulation. This vasodilation leads to perceived underfilling and a drop in mean systemic arterial pressure. This triggers a baroreceptor-mediated response involving activation of the renin-angiotensin-aldosterone system, release of anti-diuretic hormone, and activation of the sympathetic nervous system. Sodium and water retention is promoted, and in the setting of splanchnic vasodilatation, permeability and leakage of fluid into the peritoneal space occurs. This cycle leads to more vasodilation, more underfilling, and more retention of sodium and water, forming essentially a vicious cycle. The mechanism and pathogenesis for development of ascites in the noncirrhotic patient is quite distinct from that which occurs in the setting of cirrhosis and will not be discussed in further detail.

In the United States ascites occurs due to underlying cirrhosis and portal hypertension in roughly 85 % of cases. The remaining cases of ascites include additional noncirrhotic causes of portal hypertension, cardiac disease, peritoneal carcinomatosis, and miscellaneous nonportal hypertension related disorders [2].

A thorough history including inquiry into specific risk factors for the development of cirrhosis including amount of alcohol consumption, history of intravenous drug use, tattoos, high-risk sexual behavior, blood transfusions received prior to 1992 when universal screening was implemented, presence of coexisting autoimmune disease, family history of iron or copper storage disorders, and country of origin. Assessment for underlying risk factors for nonalcoholic steatohepatitis (NASH) such as the presence of obesity, diabetes mellitus, and hyperlipidemia should also be made.

Presence of coexisting risk factors or medical comorbidities should guide the clinical evaluation in a patient with new onset ascites. A patient with known malignancy , especially of the colon, breast, pancreas, or lung, who suddenly develops ascites, should be suspected of having malignancy related ascites. Abdominal pain can be a useful distinguishing feature as patients with ascites due to malignancy are more likely than patients with ascites from cirrhosis to complain of abdominal pain. A patient with known cardiac disease or cardiac risk factors should be evaluated for heart failure or cardiomyopathy leading to ascites. Additional clinical clues suggesting pancreatitis, thyroid dysfunction, renal impairment, or risk factors for tuberculosis should prompt further investigation as to an underlying disorder leading to the development of noncirrhotic ascites.

Physical exam findings are typically readily apparent. Abdominal distention, bulging flanks, especially in the setting of known liver disease should prompt further investigation and work-up. Typically flank dullness to percussion is noted, often appreciated higher up the abdominal wall than normally expected and accompanied by shifting when the patient is repositioned. Generally 1500 mL of ascitic fluid is required to manifest as shifting dullness [3]. The presence of intraluminal gastrointestinal gas, increased central adiposity, and intra-abdominal masses can mimic ascites and make clinical investigation difficult. If exam findings are equivocal and there remains a high degree of suspicion to suspect underlying ascites, abdominal ultrasonography can be useful and is able to detect as little as 100 mL of free fluid within the abdominal cavity.

Once the presence of ascites is confirmed based on exam or imaging, abdominal paracentesis is recommended for further evaluation with appropriate diagnostic fluid testing. Per American Association for the Study of Liver Disease (AASLD) guidelines this should be performed in every case of newly diagnosed ascites whether inpatient or outpatient, and also during each hospitalization for patients with cirrhosis and ascites who are admitted to the hospital to exclude the presence of spontaneous bacterial peritonitis (SBP) [4]. This, in addition to patients who display clinically features suggestive of SBP (altered mental status, confusion, abdominal pain, fevers, peripheral leukocytosis, etc.).

The abdominal paracentesis procedure itself is quite safe and relatively few true contraindications exist. Coagulopathy is frequently cited as a precluding factor for safe performance of the procedure, as elevated international normalized ratio (INR) values are common in patients with cirrhosis and evidence of hepatic synthetic dysfunction. In reality, due to balanced deficiencies of both procoagulant and anticoagulant factors, the elevations of INR seen in cirrhotic patients likely overestimate a truly elevated bleeding risk, and paracentesis has been shown to be quite safe in this patient population [5]. Valid exclusionary criteria would, however, include the presence of clear evidence of fibrinolysis or disseminated intravascular coagulation (DIC). In general the routine use of plasma transfusion prior to performing paracentesis is not recommended and has never been shown to prevent bleeding or improve outcomes. Serious complications including bowel perforation, hemoperitoneum, or death are exceedingly rare. The most common complication is abdominal wall hematoma occurring in 2 % of cases, and significant enough to require a blood transfusion in only 1 % of procedures [5]. Complications are typically less common when more experienced providers perform the procedure, and the use of ultrasonography guidance is strongly encouraged especially in obese patients or patients with abdominal surgical scars.

Routine laboratory analysis should include cell count, gram stain, total protein and albumin from ascitic fluid and serum (in order to calculate a serum-ascites albumin gradient), bacterial cultures, glucose, and lactate dehydrogenase (LDH). Depending on gross appearance of the ascitic fluid, or on other clinical factors which may be present, such as concern for malignancy or thoracic duct injury, additional values may be run as appropriate such as cytology, lipase and amylase levels, triglycerides, and/or bilirubin levels. The cutoff value for total PMN count in order to make a diagnosis of SBP is generally accepted to be 250 PMN’s per mm³. The total leukocyte count can be elevated in other conditions besides SBP, including in peritoneal carcinomatosis and in tuberculous peritonitis, though these entities typically display a lymphocyte predominance on differential [6]. If the specimen is grossly bloody, which typically occurs following a traumatic “tap,” the total PMN count may be elevated due to the increased number of neutrophils within blood itself. A correction factor is employed to calculate the total number of polymorphonuclear cells (PMN’s) by subtracting 1 PMN for every 250 red blood cells (RBC’s) identified on the cell count.

Historically, characterization of ascites fluid was regarded similarly to pleural effusions, as either exudative or transudative using mainly total protein levels, as well as LDH, however, this was shown to be notoriously inaccurate [2]. A more effective classification scheme is based on calculation of the serum-ascites albumin gradient, or SAAG, which provides a representative estimation of hydrostatic-oncotic balance. The preferred classification terms now for ascites are high-albumin gradient and low-albumin gradient in place of transudative and exudative respectively. Calculation is accomplished by subtracting the level of albumin from the ascitic fluid from the serum albumin level at the time the paracentesis was performed (keeping in mind exogenous administration of intravenous albumin products could lead to inaccuracies). Serum and ascitic fluid samples should be obtained at least on the same day; preferably within the same hour. In patients with portal hypertension, an elevated hydrostatic pressure gradient is present between the portal system and the ascitic fluid. This requires balance in terms of colloid oncotic pressure, of which albumin serves as a representative marker for, therefore leading to a direct correlation of the SAAG value and portal pressures. A generally accepted cutoff SAAG value of ≥1.1 g/dL has been shown to be consistent with portal hypertension with an accuracy of 97 % [2]. Conditions which lead to falsely low SAAG value include systemic hypotension by reduction of the portal pressure gradient, as well as with hyperglobulinemia where oncotic pressures are elevated thus narrowing the gradient. Falsely elevated SAAG values can be seen in conditions such as chylous ascites where lipids interfere with the laboratory analysis of albumin.

Ascitic fluid culture is used most commonly to detect the presence of a specific organism in cirrhotic patients with SBP in order to tailor antibiotic therapy and track resistance patterns for various organisms. TSBP infections are typically monomicrobial and characterized by overall low bacterial concentration as compared to infections of the urine or stool. As a result, typical culture methods tend to yield lower diagnostic results. It has been shown that, as with blood cultures for the detection of bacteremia, the practice of bedside inoculation of ascitic fluid samples into culture tubes dramatically improves diagnostic capabilities and this practice is recommended as standard of care when performing diagnostic paracentesis [7].

Total protein levels from ascitic fluid are useful in several circumstances. In patients with ascites of unclear etiology, and an elevated SAAG ≥1.1 g/dL, a low total protein level is mores suggestive of underlying cirrhosis whereas an elevated total protein level >2.5 g/dL is more suggestive of ascites from underlying heart failure [8]. This cutoff is less absolute and there are a significant number of patients with ascites due to cirrhosis who also have an ascitic fluid total protein of >2.5 g/dL. SAAG values tend to narrow with diuresis in patients with ascites from right-sided heart failure, whereas SAAG is generally unaffected by diuresis in cirrhotic patients. In patients with ascites and a SAAG value <1.1 g/dL, a low total protein level <2.5 g/dL is most consistent with nephrotic syndrome. Total protein levels are also inversely related to risk of development of SBP, through deficiency in opsonins, among other factors. Conversely, infections resulting from secondary bacterial peritonitis, which are often polymicrobial, tend to correlate with higher levels of ascitic fluid total protein. Highly sensitive diagnostic criteria exist to aid in identification of patients with intra-abdominal infection resulting from secondary bacterial peritonitis from perforated viscous [9]. The criteria include total protein >1 g/dL, glucose <50 mg/dL, and LDH > upper limit of normal for serum, and are associated with sensitivity of nearly 100 % and specificity of 45 % in identifying secondary bacterial peritonitis.

Glucose levels from ascitic fluid can be helpful for confirming a suspected diagnosis of SBP. Generally sterile ascitic fluid should have similar concentrations of glucose to serum; however, in the case of bacterial infection, levels of glucose typically fall to very low or undetectable levels. This can happen with either SBP or with secondary bacterial peritonitis from intestinal perforation, and is not generally helpful at distinguishing the two. LDH levels are typically reduced in sterile ascites, as the molecule is not easily released from the bloodstream. Levels of LDH in SBP are generally increased due to release of LDH from neutrophils. These levels are even more significantly elevated in cases of secondary bacterial peritonitis.

In cases where pancreatic ascites is suspected, amylase levels are generally found to be significantly elevated above serum levels as compared to uncomplicated ascites, where amylase values are generally around half of serum values.