Hepatic hydrothorax

INTRODUCTION — 

Hepatic hydrothorax refers to the presence of a pleural effusion (usually >500 mL) in a patient with cirrhosis in the absence of other underlying causes (eg, cardiac, pulmonary, kidney, pleural or malignant disease) [1-3]. Hepatic hydrothorax occurs in approximately 5 to 15 percent of patients with cirrhosis and is thought to account for approximately 2 percent of all pleural effusions [4]. Patients who develop hepatic hydrothorax are more likely to have ascites, hepatic encephalopathy, acute kidney injury (AKI), and increased risk of mortality. However, hepatic hydrothorax occasionally occurs in the absence of ascites [5]. In a retrospective analysis of 495 patients with cirrhosis and pleural effusion, 16 percent had hepatic hydrothorax [6]. While patients with ascites can often tolerate up to 5 to 10 L of fluid with only mild symptoms, those with a pleural effusion can have severe symptoms (such as shortness of breath, cough, and hypoxemia) with as little as 500 mL of fluid [7].

This topic will review the clinical manifestations, diagnosis, and management of hepatic hydrothorax. Other complications of cirrhosis, including ascites and variceal hemorrhage, are discussed elsewhere. (See "Cirrhosis in adults: Overview of complications, general management, and prognosis" and "Ascites in adults with cirrhosis: Initial therapy" and "Ascites in adults with cirrhosis: Diuretic-resistant ascites" and "Primary prevention of bleeding from esophageal varices in patients with cirrhosis" and "Prevention of recurrent bleeding from esophageal varices in patients with cirrhosis".)

PATHOGENESIS — 

Although the exact mechanisms involved in the development of hepatic hydrothorax are incompletely understood, it probably results from the passage of ascites from the peritoneal cavity into the pleural cavity through small diaphragmatic defects. These defects are typically less than 1 cm (and may be microscopic) and are generally located in the tendinous portion of the diaphragm [8-13]. Hepatic hydrothorax becomes apparent when the absorptive capacity of the pleural space is exceeded. The pathologic mechanisms behind the formation of ascites are discussed in detail elsewhere. (See "Pathogenesis of ascites in patients with cirrhosis".)

The diaphragmatic defects are more often found in the right hemidiaphragm, likely due in part to the fact that the left hemidiaphragm is thicker and more muscular. Hepatic hydrothorax develops on the right side in approximately 73 to 85 percent of patients, on the left side in approximately 13 to 17 percent, and bilaterally in approximately 8 to 24 percent [3,6,14,15].

The negative intrathoracic pressure generated during inspiration promotes the passage of fluid from the abdominal cavity to the pleural space. The combination of increased intraabdominal pressure and a negative intrathoracic pressure facilitates movement of fluid from the peritoneal cavity into the pleural space through defects in the diaphragm [16]. This could explain why some patients with hepatic hydrothorax do not have apparent ascites [5,17,18]. This theory is supported by studies using 99mTc-human albumin or 99mTc-sulphur colloid, which demonstrate unidirectional passage of these markers from the abdominal cavity to the pleural cavity [5,19-23].

Similar to spontaneous bacterial peritonitis (SBP) in patients with ascites, patients with hepatic hydrothorax may develop spontaneous bacterial empyema (SBEM). It is thought to occur due to seeding of the pleural effusion with bacteria that spread directly from the abdominal cavity in a patient with SBP or from bacteremia. The etiologic agents in most cases are Escherichia coli, Streptococcus species, Enterococcus, Klebsiella, or Pseudomonas [24]. Approximately one-half of episodes are associated with SBP [25].

CLINICAL MANIFESTATIONS — 

Common clinical manifestations in patients with hepatic hydrothorax include dyspnea, a nonproductive cough, pleuritic chest pain, nausea, and fatigue due to hypoxemia [1,2,6,14,26,27]. Patients often have signs of ascites (eg, abdominal distension) as well as other findings associated with cirrhosis (eg, palmar erythema, spider angioma and/or jaundice) [3]. Rarely, a pleural effusion may be an incidental finding on chest radiograph in an asymptomatic patient. (See "Cirrhosis in adults: Etiologies, clinical manifestations, and diagnosis", section on 'Clinical manifestations'.)

Less common presentations include an acute tension hydrothorax with associated severe dyspnea and hypotension [28] or spontaneous bacterial empyema (SBEM). SBEM has been described in 13 to 16 percent of patients with hepatic hydrothorax [25,29]. Infection should be suspected in a patient with hepatic hydrothorax who develops fever, new or worsening pleuritic chest pain, encephalopathy, or an unexplained worsening of kidney function. (See 'Thoracentesis and fluid testing' below.)

In most patients, hepatic hydrothorax occurs on the right side, reflecting the underlying mechanisms of fluid accumulation. However, in 17 percent of patients, hydrothorax occurs on the left side, and in up to 10 percent of patients, it is bilateral (table 1) [4,6].

DIAGNOSIS — 

The diagnosis of hepatic hydrothorax includes documentation of a pleural effusion and exclusion of alternative causes for the effusion. A pleural effusion can usually be demonstrated on a frontal chest radiograph, although lateral views may be required if the effusion is small. Pleural effusions in patients with hepatic hydrothorax are typically found in the right hemithorax, though they may be left-sided or bilateral. A thoracentesis as well as additional imaging (eg, chest computed tomography and echocardiogram) should be performed to confirm the diagnosis and to exclude infection or an alternate cause of a pleural effusion. Up to 20 percent of patients will have evidence of infection or a cause of pleural effusion other than hepatic hydrothorax [30]. (See "Pleural fluid analysis in adults with a pleural effusion" and 'Pathogenesis' above.)

Thoracentesis and fluid testing — Patients with suspected hepatic hydrothorax should undergo thoracentesis with testing of the pleural fluid. (See "Ultrasound-guided thoracentesis".)

Diagnostic tests that should be performed on the pleural fluid include (table 1) [1,14,31-33]:

●Cell count and differential

●Gram stain

●Culture

●Protein, albumin, lactate dehydrogenase (LDH), and bilirubin concentrations

●Pleural fluid pH

In addition, serum albumin, LDH, and bilirubin concentrations should be determined.

When obtaining cultures, the fluid should be inoculated directly into a blood culture bottle at the bedside. In one study, the sensitivity for detecting spontaneous bacterial empyema (SBEM) using this method was 77 percent, compared with 33 percent for inoculations done once the fluid was received by the microbiology laboratory [25,32].

Other tests may be useful if certain diagnoses are suspected (see "Pleural fluid analysis in adults with a pleural effusion"):

●Triglyceride level (chylothorax)

●Polymerase chain reaction for Mycobacterium (tuberculosis)

●Amylase concentration (pancreatitis)

●Cytology (malignancy)

Pleural effusions deriving from portal hypertension are transudative in nature and therefore similar to the ascitic fluid, with a low protein concentration (<2.5 g/dL) [1] and with a serum-to-pleural fluid albumin gradient >1.1 g/dL [6].

However, there may be some differences in fluid composition between the pleural fluid and the ascitic fluid because the mechanisms of fluid absorption from the pleural space are different than those in the peritoneal cavity [1,26,31]. For instance, total protein and albumin may be slightly higher in hepatic hydrothorax compared with levels in the ascitic fluid [12].

As an example of pleural fluid findings in patients with hepatic hydrothorax, a retrospective study of 41 patients with hepatic hydrothorax found that 33 patients had solitary hydrothorax, and of these, 31 (94 percent) were transudative [32]. Sixteen (48 percent) had a total protein level <1.5 g/dL in the pleural fluid, none had a serum albumin value <1.5 g/dL, and microbiologic cultures were negative in 31 patients. The median pleural fluid pH was 7.49, total protein was 1.5 g/dL, and lactate dehydrogenase (LDH) was 65 international unit/L. The median pleural fluid:serum protein ratio and median pleural fluid LDH:upper limit of normal LDH ratio were 0.25 and 0.27, respectively.

In uncomplicated hepatic hydrothorax, the polymorphonuclear cell (PMN) count is low (<250 cells/mm³), whereas in the setting of SBEM the PMN count is elevated [1,25].

The diagnostic criteria for SBEM are:

●Positive pleural fluid culture and a PMN cell count >250 cells/mm³

●Negative pleural fluid culture and a PMN cell count >500 cells/mm³

●No evidence of pneumonia on a chest imaging study

For patients with SBEM, fluid analysis may indicate a transudative or exudative effusion.

Data suggest that prompt testing of pleural fluid was associated with better outcomes. In an observational study including 54 patients with SBEM, early thoracentesis (ie, <24 hours from presentation) was associated with lower rates of mortality and intensive care unit (ICU) admission compared with delayed thoracentesis (7 versus 41 percent and 26 versus 56 percent, respectively) [34].

Imaging studies — Hepatic hydrothorax can typically be detected on a frontal chest radiograph. However, we generally recommend that patients also undergo a computed tomographic scan of the chest to exclude mediastinal, pulmonary, or pleural lesions. An abdominal ultrasound with Doppler study should be performed to examine the liver, rule out liver masses, ascertain the patency of the portal and hepatic veins, and detect ascites. Finally, echocardiography should be performed to rule out a cardiac cause for the pleural effusion. (See "Imaging of pleural effusions in adults" and "Diagnostic evaluation of the hemodynamically stable adult with a pleural effusion", section on 'Consider additional imaging'.)

In cases where the diagnosis is uncertain, an intraperitoneal injection of 99mTc-sulphur colloid or 99mTc-human serum albumin can be helpful. The radioisotopes migrate from the peritoneal cavity into the pleural space, therefore establishing a communication between both spaces and confirming the diagnosis of hepatic hydrothorax [5,19,35] (image 1). The optimal time to do an isotope study is shortly after therapeutic thoracentesis, when fluid is reaccumulating in the pleural cavity. When the chest is maximally filled with fluid, there may be minimal transfer of the isotope across the diaphragm, resulting in a false-negative study. Failure of the marker to show up in the pleural space (when the study has been performed shortly after thoracentesis) indicates an alternate cause of the pleural effusion.

DIFFERENTIAL DIAGNOSIS — 

The differential diagnosis of pleural effusions is broad and includes cardiopulmonary disease, chylothorax, tuberculosis, malignant effusions, and hemothorax (table 2 and table 3). Differentiating among the causes involves evaluating the patient's history, fluid analysis from a thoracentesis, and imaging test results. (See "Pleural fluid analysis in adults with a pleural effusion".)

Factors that suggest hepatic hydrothorax as the diagnosis include a history of cirrhosis (required), a transudative effusion, a right-sided effusion, and the simultaneous presence of ascites. However, even in the appropriate clinical setting, alternative diagnoses need to be excluded prior to making a diagnosis of hepatic hydrothorax. (See 'Diagnosis' above.)

MANAGEMENT — 

The management of patients with hepatic hydrothorax is typically multidisciplinary and includes input from hepatology, pulmonology, interventional radiology, and surgery. Treatment of hepatic hydrothorax is similar to the treatment of ascites [33]. Patients who are actively consuming alcohol should be encouraged to abstain completely. Even if alcohol is not the only cause of their liver disease (eg, in patients with cirrhosis due to chronic hepatitis C), the alcohol-related component of their liver disease may improve dramatically with abstinence. Management then includes dietary sodium restriction and diuretics (algorithm 1). In addition, patients with confirmed hepatic hydrothorax should be referred for liver transplantation if they are otherwise suitable candidates. (See "Ascites in adults with cirrhosis: Initial therapy", section on 'Alcohol abstinence' and "Liver transplantation in adults: Patient selection and pretransplantation evaluation".)

Patients who are severely symptomatic should undergo a therapeutic thoracentesis followed by management with a sodium-restricted diet and diuretics (furosemide 40 mg and spironolactone 100 mg daily to start). Patients who are mildly to moderately symptomatic can be treated initially with sodium restriction and diuretics. Management options for patients who are refractory to sodium restriction and diuretics include serial thoracenteses, transjugular intrahepatic portosystemic shunt (TIPS) placement, pleurodesis, thoracoscopic surgery to repair diaphragmatic defects, and liver transplantation.

Chest tubes should not be placed for the treatment of hepatic hydrothorax (though they may be needed for patients with spontaneous bacterial empyema (SBEM) and frank pus, or in patients undergoing pleurodesis). Placement of chest tubes in patients with hepatic hydrothorax can result in massive protein and electrolyte depletion, infection, kidney failure, and bleeding [36-38]. A review of cases from a tertiary care center over a 10-year period indicated that 16 out of 17 patients developed serious complications following chest tube placement [39]. These complications included acute kidney injury (AKI), pneumothorax, and empyema. Furthermore, once inserted, it may be impossible to remove the tube because of the continuous reaccumulation of fluid [38]. Other complications include subcutaneous emphysema; lung, spleen, liver, and stomach lacerations; hemothorax resulting from intercostal artery laceration; unilateral pulmonary edema from rapid removal of fluid; and placement of the tube into the abdominal cavity.

Chest tube insertion for management of hepatic hydrothorax has also been associated with an increased risk of mortality [40]. In a study of 1981 patients with hepatic hydrothorax, patients who had chest tube insertion had higher risk of mortality compared with patients who underwent thoracentesis (odds ratio [OR] 2.1, 95% CI 1.4-3.1) [40].

Sodium restriction — Education about the importance of dietary sodium restriction is a central component of the management of hepatic hydrothorax. Dietary sodium should be moderately restricted (ie, 80 to 120 mEq [80 to 120 mmol] per day, corresponding to 1840 mg to 2760 mg of sodium) (figure 1 and table 4) [41]. This level of sodium restriction can be achieved in an outpatient setting without the purchase of special foods [1,26,42]. Patients should also be instructed to avoid nonsteroidal anti-inflammatory drugs, which can cause sodium retention and kidney failure. (See "Ascites in adults with cirrhosis: Initial therapy", section on 'Dietary sodium restriction'.)

Diuretics — Sodium restriction alone will be effective only in the small subset of patients. Most patients will also require therapy with diuretics. Patients should be started on furosemide 40 mg daily and spironolactone 100 mg daily. If there is no response, the doses of the diuretics may be increased in a stepwise fashion every three to five days by doubling the doses (maintaining a ratio of 40 mg:100 mg). The maximum doses are furosemide 160 mg daily and spironolactone 400 mg daily. (See "Ascites in adults with cirrhosis: Initial therapy", section on 'Diuretic therapy'.)

Refractory hydrothorax — Patients who have persistent hydrothorax despite a sodium-restricted diet and diuretic therapy or who develop diuretic-related complications are considered to have refractory hydrothorax. Small series suggest that 20 to 40 percent of patients have refractory hydrothorax [43-45].

It is important to exclude noncompliance with dietary sodium restriction and/or diuretics prior to diagnosing refractory hydrothorax. In addition to the history, measurement of 24-hour urinary sodium excretion can be helpful for determining adherence to management recommendations. (See "Ascites in adults with cirrhosis: Initial therapy", section on 'Diuretic resistance'.)

Management options for patients with refractory hydrothorax include repeated thoracenteses, TIPS placement, pleurodesis, surgical repair of defects in the diaphragm, and liver transplantation. Our approach is to start with repeated thoracenteses. However, for patients who require thoracentesis every two to three weeks, alternative treatments should be considered because these patients are at increased risk of adverse events from repeated thoracentesis. For such patients, the decision to perform TIPS placement is individualized and informed by multidisciplinary assessment and by patient comorbidities that may be a contraindication to elective TIPS (eg, congestive heart failure) [46]. (See "Overview of transjugular intrahepatic portosystemic shunts (TIPS)".)

For patients who are not candidates for TIPS placement, we will consider pleurodesis or thoracoscopic repair of the diaphragmatic defect.

Data have suggested that refractory hepatic hydrothorax was associated with increased mortality risk. In a study including 47 patients with refractory hepatic hydrothorax who were compared with 47 patients with refractory ascites who were matched for age, sex, and model for end-stage liver disease (MELD)-Na score, refractory hepatic hydrothorax was associated with higher mortality rates at one year (51 versus 19 percent) [47]. In a cohort study of 784 adults with cirrhosis who were evaluated for liver transplantation, patients with refractory hepatic hydrothorax had a median transplant-free survival of 0.26 years [45]. Compared with non-refractory hepatic hydrothorax, patients with refractory hydrothorax had higher risk of mortality without liver transplantation, after adjusting for MELD-Na score (HR 1.73; 95% CI 1.06-2.81). Among those who received liver transplantation, patients with refractory hydrothorax had a higher one-year survival benefit compared with non-refractory hepatic hydrothorax or other complications of cirrhosis (0.48 years versus 0.28 and 0.19 years, respectively) [45].

Refractory hydrothorax has also been associated with higher risk of liver-related mortality compared with refractory ascites in patients with comparable MELD-Na values [48]. In addition, MELD 3.0 may better predict risk of liver-related mortality in patients with refractory hydrothorax.

Thoracentesis — Therapeutic thoracentesis is the most effective way to reduce large effusions (greater than 1.5 L) [1,14,31]. After thoracentesis, diuretics should be continued to prevent reaccumulation of fluid. As a general rule, no more than 2 L of fluid should be removed because of the risk of pulmonary edema and hypotension [49]. However, this "2 L rule" was developed in patients with causes other than cirrhosis for their effusions; patients with cirrhosis may be able to tolerate the removal of larger volumes of fluid. The amount of fluid present prior to thoracentesis can be estimated by chest radiograph or computed tomographic scan. (See "Large volume (therapeutic) thoracentesis: Procedure and complications" and "Imaging of pleural effusions in adults".)

Complications of thoracentesis include pain at the puncture site, pneumothorax, hemothorax, vasovagal episodes, hemoptysis, air embolism, laceration of the liver or spleen, empyema, hemoptysis, and subcutaneous emphysema [50]. The probability of complications increases in patients who require frequent therapeutic thoracenteses (eg, every two to three weeks). The alternative approaches outlined below should be considered for these patients. While performing frequent thoracentesis may relive symptoms, it is not generally used as a long-term strategy because the benefit of such an approach is uncertain [51].

Coagulopathy is not considered a contraindication to therapeutic thoracentesis. One study demonstrated no increased risk of bleeding in patients with prothrombin times up to twice the midpoint of the normal range, or platelet counts of more than 50,000/microL; the authors concluded that prophylactic transfusions are not necessary [52]. Patients with a serum creatinine more than 6 mg/dL had greater blood loss compared with those with lower values.

Transjugular intrahepatic portosystemic shunt — Several groups have reported a beneficial effect of TIPS placement in patients with hepatic hydrothorax. Unfortunately, few of the studies describe how the refractoriness of the hydrothorax to medical therapy was determined. Despite this shortcoming, most series have demonstrated response rates in the range of 70 to 80 percent [53-59]. However, TIPS placement is associated with serious complications including technical complications (eg, cardiac arrhythmias, traversal of the liver capsule), hepatic encephalopathy, and TIPS stenosis. Because of the potential risks of TIPS placement, the decision to perform TIPS for refractory hepatic hydrothorax is individualized and informed by multidisciplinary assessment and patient comorbidities [46]. The pre-TIPS evaluation is discussed separately. (See "Overview of transjugular intrahepatic portosystemic shunts (TIPS)".)

A study that reviewed the course after TIPS placement in 28 patients with hepatic hydrothorax revealed that the 30-, 90-, and 150-day mortality rates were 14, 25, and 53 percent, respectively [60]. One-year survival without liver transplantation was 41 percent. In 68 percent of patients there was complete resolution of symptoms, and in 57 percent of patients there was disappearance of effusions on radiograph. TIPS placement was successful in 15 of 20 patients with a Child-Pugh score ≤10 and in only one of eight patients with a Child-Pugh score >10. In a meta-analysis of six studies that included a total of 198 patients with a mean duration of follow-up of 10 months, complete response to TIPS was seen in 56 percent (95% CI 45-67 percent) and a partial response in 18 percent (95% CI 11-24 percent) [61]. The incidence of hepatic encephalopathy was 12 percent (95% CI 6.3-17 percent), and the 45-day mortality was 18 percent (95% CI 11-24 percent). The mortality rate in refractory hepatic hydrothorax is comparable with those seen in patients with refractory ascites. The most important predictors of poor outcomes after TIPS in this population are older age, severe underlying liver disease, and/or associated kidney dysfunction. In an observational study including 51 patients with hepatic hydrothorax who underwent TIPS, the mean portal pressure gradient was reduced from 23 mmHg to 7 mmHg, while at 6 months, 10 patients (20 percent) had complete resolution of hepatic hydrothorax [62]. Serious complications reported within 30 days after TIPS placement included portosystemic encephalopathy (eight patients [16 percent]), ischemic hepatitis (four patients [8 percent]), and mortality (9 patients [18 percent]).

Risk factors for mortality after TIPS placement for hepatic hydrothorax included a Child-Pugh score >10, MELD score >15 to 17, and an elevated pre-TIPS creatinine [59,60]. However, there is insufficient published evidence to support a specific prognostic score cutoff as a contraindication to TIPS [46].

A lack of response in the hydrothorax after TIPS placement is associated with an increased mortality rate [59]. These findings have suggested that TIPS should be considered early given the high morbidity and mortality in this population.

Pleurodesis — Chemical pleurodesis is commonly used to treat patients with recurrent, symptomatic pleural effusions due to malignancy. However, little information is available to guide the clinician in the management of recurrent, symptomatic, nonmalignant pleural effusions, even though both chemical pleurodesis and talc poudrage have been used in this setting. (See "Chemical pleurodesis for the prevention of recurrent pleural effusion".)

Clinical experience suggests that hepatic hydrothorax is the most difficult form of nonmalignant pleural effusion to treat with chemical pleurodesis. Because of the rapid migration of fluid from the abdomen into the pleural space, it is often difficult to keep the two pleural surfaces apposed long enough for the inflammatory process to result in pleural symphysis [24]. Thus, it is rarely performed and is typically reserved for patients for whom no other options exist.

In addition to the challenge of achieving pleural symphysis, multiple adverse events (eg, fever, kidney failure, pneumothorax, hepatic encephalopathy) have been reported in case series of patients who have undergone pleurodesis for hepatic hydrothorax [63-66]. In a meta-analysis of 13 studies including 180 patients who underwent pleurodesis by chemical and/or mechanical means (ie, conventional thoracoscopy or video-assisted thoracic surgery), the pooled rate of complete response was 72 percent (95% CI 65-79 percent) [66]. Complications related to chemical pleurodesis were reported in six studies including 63 patients, and the pooled incidence of adverse events was 82 percent (95% CI 66-94 percent). (See "Medical thoracoscopy (pleuroscopy): Equipment, procedure, and complications".)

Thoracoscopic repair — Thorascopic repair of the diaphragmatic defects seen in patients with hepatic hydrothorax may be possible. A diaphragmatic repair involving a pleural flap and surgical mesh reinforcement has been described, but experience is limited [67]. A retrospective analysis of 63 patients with refractory hepatic hydrothorax who underwent thoracoscopic mesh onlay reinforcement to repair diaphragmatic defects found that after a median 20.5 months, only four patients experienced recurrence [68]. Surgical repair of diaphragmatic defects prevents unidirectional shifting of fluid into the thoracic cavity from the abdomen. However, surgical repair has been associated with high mortality in patients with decompensated cirrhosis; thus, this approach may be an option for patients who are not candidates for TIPS but have low MELD scores. The procedure can be performed with video-assisted thoracic surgery with or without mesh, while the type of diaphragmatic defect may influence the repair strategy [68].

Liver transplantation — Patients with hepatic hydrothorax typically have end-stage liver disease, and liver transplantation is the definitive treatment for hepatic hydrothorax, particularly for refractory hydrothorax [45]. Patients with hepatic hydrothorax who do not have contraindications to liver transplantation should be referred for a transplantation evaluation. The outcomes in patients with hepatic hydrothorax who undergo liver transplantation are similar [69,70] or better than those in patients who undergo liver transplantation for other reasons [45]. (See "Liver transplantation in adults: Patient selection and pretransplantation evaluation".)

Therapies with uncertain or no benefit — Chest tube placement should been avoided due to the risk of complications and poor outcomes [36-38]. This procedure is associated with risk of protein loss, infection, hemothorax, and electrolyte abnormalities. Moreover, given the high rates of ongoing fluid production and accumulation within the pleural space, removal of the chest drain is often challenging. Indwelling pleural catheters (IPCs) have been studied as a possible treatment option for patients with refractory hepatic hydrothorax. However, the reported outcomes for patients undergoing chest IPC have been poor [71,72]. In a retrospective study comparing IPC placement with thoracentesis in 1278 patients with hepatic hydrothorax, the 30-day mortality rate was higher with IPC placement (23.5 versus 18.6 percent) [72]. Additional studies are needed to examine the outcomes and risks associated with IPC for patients with refractory hepatic hydrothorax.

Treatment of infection — SBEM has been described in 13 to 16 percent of patients with hepatic hydrothorax. The etiologic agents in most cases are E. coli, Streptococcus species, Enterococcus, Klebsiella, or Pseudomonas [24]. If infection is suspected, therapy with an intravenous antibiotic should be started after the pleural fluid has been sampled for cell count and differential, Gram stain, and culture [25]. (See 'Thoracentesis and fluid testing' above.)

The diagnostic criteria for SBEM are:

●Positive pleural fluid culture and a polymorphonuclear cell (PMN) cell count >250 cells/mm³

●Negative pleural fluid culture and a PMN cell count >500 cells/mm³

●No evidence of pneumonia on a chest imaging study

Criteria for selecting antibiotics include the severity of infection, local resistance patterns, and risk factors for multidrug resistant organisms (MDROs) such as previous antibiotic exposure, healthcare-associated infections, and bacterial colonization. For patients who are critically ill (eg, sepsis, acute on chronic liver failure), we initiate broad-spectrum antibiotic therapy promptly. Lack of response to initial treatment has been associated with higher risk of mortality.

The antibiotics used to treat SBEM are the same as those used for the treatment of spontaneous bacterial peritonitis (SBP). We use third-generation cephalosporins and other beta-lactam antibiotics for community-acquired SBEM (ie, typically ceftriaxone 2 g every 24 hours for 7 to 10 days) [73]. In areas with low prevalence of fluroquinolone resistance, a fluoroquinolone such as levofloxacin or ciprofloxacin is an option for patients who are allergic to penicillin. However, for patients with healthcare-associated SBEM, we typically initiate broader spectrum antimicrobial therapy because extended-spectrum beta-lactamase-producing bacteria reduce the efficacy of third-generation cephalosporins [74]. Antibiotic therapy can be tailored once results from antibiotic sensitivity testing are available. In patients in whom there is slow clinical recovery, a repeat thoracentesis is useful to document that the patient is responding. (See "Spontaneous bacterial peritonitis in adults: Treatment and prophylaxis", section on 'Selecting empiric therapy' and "Epidemiology, clinical presentation, and diagnostic evaluation of parapneumonic effusion and empyema in adults".)

Mortality rates in patients with SBEM are as high as 20 percent, even with antimicrobial therapy [25,27,29]. Independent risk factors associated with poor outcomes in SBEM include a high MELD-Na score, initial intensive care unit admission, and initial antibiotic treatment failure [29]. (See "Model for End-stage Liver Disease (MELD)".)

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Cirrhosis".)

SUMMARY AND RECOMMENDATIONS

●Background – Hepatic hydrothorax refers to the presence of a pleural effusion (usually >500 mL) in a patient with cirrhosis in the absence of other underlying causes (eg, cardiac, pulmonary, kidney, pleural or malignant disease). Hepatic hydrothorax occurs in approximately 5 to 15 percent of patients with cirrhosis. (See 'Introduction' above.)

●Pathogenesis – Although the exact mechanisms involved in the development of hepatic hydrothorax are incompletely understood, it probably results from the passage of ascites from the peritoneal cavity into the pleural cavity through small diaphragmatic defects. (See 'Pathogenesis' above.)

●Clinical features – Common clinical manifestations in patients with hepatic hydrothorax include dyspnea, a nonproductive cough, pleuritic chest pain, and fatigue due to hypoxemia. Patients often have signs of ascites (eg, abdominal distension) as well as other findings associated with cirrhosis (eg, palmar erythema). Rarely, a pleural effusion may be an incidental finding in an asymptomatic patient. (See 'Clinical manifestations' above.)

●Diagnosis – The diagnosis of hepatic hydrothorax includes documentation of a pleural effusion and exclusion of alternative causes for the effusion. A pleural effusion can usually be demonstrated on a frontal chest radiograph, although lateral views may be required if the effusion is small. Pleural effusions in patients with hepatic hydrothorax are typically found in the right hemithorax, though they may be left-sided or bilateral. A thoracentesis as well as additional imaging (eg, a chest computed tomographic scan and echocardiogram) should be performed to confirm the diagnosis and to exclude infection or an alternate cause of the effusion (table 2 and table 3). Up to 20 percent of patients will have evidence of infection or a cause of pleural effusion other than hepatic hydrothorax. (See 'Diagnosis' above.)

•Diagnostic tests that should be performed on the pleural fluid include (table 1) (see 'Thoracentesis and fluid testing' above):

-Cell count and differential

-Gram stain

-Culture (using blood culture bottles that are inoculated at the bedside)

-Protein, albumin, lactate dehydrogenase (LDH), and bilirubin concentrations

-Pleural fluid pH

In addition, serum albumin, LDH, and bilirubin concentrations should be determined.

•Other tests may be useful if certain diagnoses are suspected (see "Pleural fluid analysis in adults with a pleural effusion"):

-Triglyceride level (chylothorax)

-Polymerase chain reaction for mycobacterium (tuberculosis)

-Amylase concentration (pancreatitis)

-Cytology (malignancy)

●Management – Treatment of hepatic hydrothorax is similar to the treatment of ascites (algorithm 1) (see "Ascites in adults with cirrhosis: Initial therapy" and 'Management' above):

•Patients who are actively consuming alcohol should be encouraged to abstain completely. Even if alcohol is not the only cause of their liver disease (eg, in patients with cirrhosis due to chronic hepatitis C), the alcohol-related component of their liver disease may improve dramatically with abstinence. (See "Ascites in adults with cirrhosis: Initial therapy", section on 'Alcohol abstinence'.)

•We suggest that patients with hepatic hydrothorax are initially treated with a sodium-restricted diet and diuretics rather than a sodium-restricted diet alone (Grade 2C). Patients should be started on furosemide 40 mg daily and spironolactone 100 mg daily. If there is no response, the doses of the diuretics may be increased in a stepwise fashion every three to five days by doubling the doses (maintaining a ratio of 40 mg:100 mg). The maximum doses are furosemide 160 mg daily and spironolactone 400 mg daily. (See "Ascites in adults with cirrhosis: Initial therapy", section on 'Dietary sodium restriction' and "Ascites in adults with cirrhosis: Initial therapy", section on 'Diuretic therapy'.)

For patients who are severely symptomatic, a therapeutic thoracentesis can be performed along with the initiation of a sodium-restricted diet and diuretics to provide more rapid fluid removal. (See "Large volume (therapeutic) thoracentesis: Procedure and complications".)

•Options for patients who are refractory to sodium restriction and diuretics include serial thoracenteses, transjugular intrahepatic portosystemic shunt (TIPS) placement, pleurodesis, thoracoscopic surgery for diaphragmatic repair, and liver transplantation. Our approach is to start with repeated thoracenteses. However, for patients who require thoracentesis every two to three weeks, alternative treatments should be considered because these patients are at increased risk of having adverse events from repeated thoracentesis. For such patients, the decision to perform TIPS placement is individualized and informed by multidisciplinary assessment and by patient comorbidities that may be a contraindication to elective TIPS (eg, congestive heart failure). (See "Overview of transjugular intrahepatic portosystemic shunts (TIPS)".)

Alternative treatments for patients who are not good candidates for TIPS placement include pleurodesis or thoracoscopic repair of the diaphragmatic defect. (See 'Refractory hydrothorax' above.)

•Chest tubes should not be placed for the treatment of hepatic hydrothorax. Placement of chest tubes in patients with hepatic hydrothorax can result in massive protein and electrolyte depletion, infection, kidney failure, and bleeding. In addition, once inserted, it may be impossible to remove a chest tube because of the continuous reaccumulation of fluid.

•Patients with hepatic hydrothorax should be referred for liver transplantation if they are otherwise suitable candidates. (See "Liver transplantation in adults: Patient selection and pretransplantation evaluation".)