Evaluation and management of pleural effusions following cardiac surgery

INTRODUCTION — Postoperative pleural effusions are common in patients who undergo cardiac surgery [1-12]. Most are benign, nonspecific, do not require any investigation, and resolve spontaneously. However, a small proportion require attention, typically thoracentesis, and occasionally other investigations.

In this review, we discuss the prevalence, etiology, evaluation, and management of patients who present with a pleural effusion after cardiac surgery. Initial evaluation of a pleural effusion in non-surgical setting is discussed separately. (See "Pleural fluid analysis in adults with a pleural effusion".)

PREVALENCE — Pleural effusions most often occur following coronary artery bypass grafting (CABG) and cardiac transplantation, but can also be seen following mitral and aortic valve replacement surgery [2-11,13-16]. Pleural effusions occur in up to 90 percent of patients during the early postoperative period (≤30 days after surgery) and in up to two-thirds of patients in the late postoperative period (>30 days after surgery). Large, clinically significant pleural effusions occur in a minority of patients (5 to 10 percent at 30 days). However, accurate estimates have been hampered by differences in timing, methods, and frequency of postoperative imaging examinations used to detect pleural fluid in the individual studies.

●Early – The vast majority of pleural effusions that occur postoperatively, develop within the first 30 days and are typically small and nonspecific [3,17].

•One study of 47 patients noted that 89 percent had pleural effusions when examined by thoracic ultrasound on the sixth postoperative day; only 60 percent of these patients had evidence of pleural fluid when evaluated with standard chest radiography, which have a lower diagnostic sensitivity [3]. (See "Bedside pleural ultrasonography: Equipment, technique, and the identification of pleural effusion and pneumothorax".)

•In contrast, in another study, serial imaging with echocardiography and/or lung ultrasonography was performed in 91 patients following cardiac surgery [17]. One-third had a pleural effusion during their postoperative course that dwindled to 11 percent at discharge.

●Late – Pleural effusions are also common in the late (>30 days) postoperative course following cardiac surgery [3,18].

•In a prospective case series of 389 patients evaluated 28 days after cardiac surgery [18], two-thirds of patients had a pleural effusion after a CABG, 62 percent had an effusion after CABG plus valve surgery, and 45 percent had an effusion after valve surgery alone.

•In contrast to early pleural effusions, which are generally small, 10 percent of late pleural effusions are large (ie, an effusion that occupies more than 25 percent of the hemithorax) [18].

•Rates may be higher in those who have an internal mammary artery (IMA) graft CABG compared with a saphenous vein graft CABG (73 versus 29 percent) [3]. (See 'Risk factors' below.)

The prevalence does not appear to have changed as more advanced surgery techniques are employed. A study that accrued 76 patients from 2013 to 2015 undergoing CABG and aortic valve replacement reported that over one-third had a pleural effusion that required thoracentesis [19]. These effusions also appear clinically important in that among patients undergoing CABG, 16 percent require readmission in 30 days and 23 percent of these patients have pleural effusions as the readmission diagnosis [20].

ETIOLOGIES AND PATHOGENESIS — Most pleural effusions are nonspecific in nature (ie, hemorrhagic and related to the surgery itself) and are less commonly due to a complication of surgery such as heart failure, hemothorax, or infection (table 1).

Nonspecific pleural effusion — Most of the pleural effusions that follow cardiac surgery develop as a consequence of the surgical procedure itself (ie, "nonspecific pleural effusion").

It is thought that nonspecific pleural effusions, which occur early following cardiac surgery (≤30 days following surgery), are due to inflammation induced by hemorrhage and the trauma of surgery, while late effusions (>30 days following surgery) may be due to an immune or inflammatory phenomenon induced by surgery.

●Early nonspecific pleural effusion – Suggested mechanisms include the following:

•Topical cardiac cooling – Several studies have reported a higher incidence of early pleural effusions in patients who receive topical cooling during cardiac surgery compared with those who do not receive cooling [5,21]. In one retrospective study, early pleural effusions occurred in 50 percent of patients who underwent topical cardiac cooling compared with 18 percent of patients who did not receive topical cooling [5]. It is speculated that cooling may injure pericardial and pleural membranes or the left phrenic nerve resulting in an atelectasis-related pleural effusion [22].

•Pleural injury from pleurotomy – Pleurotomy-induced pleural injury may contribute to nonspecific post-cardiac surgery effusions [9,23,24]. Pleurotomy is frequently performed in patients undergoing internal mammary artery (IMA) coronary artery bypass graft (CABG), and pleural effusions are more common after IMA-CABG than after saphenous vein (SV) graft CABG [3,6,14,15].

•Pre-clinical presentation of post-cardiac injury syndrome – It has been proposed that early effusions may represent the initial phase in the development of pleural effusions associated with post-cardiac injury syndrome. However, there are no data to support this theory. (See "Post-cardiac injury syndromes".)

•Lymphatic disruption – Surgical interruption of mediastinal lymphatic channels, immobility, and decreased ventilation may interfere with drainage of pleural fluid resulting in the development of an early pleural effusion following cardiac surgery. A randomized study suggested that early mobilization may mitigate the risk of developing pleural effusions and atelectasis [25].

Late nonspecific pleural effusion – Suggested mechanisms include the following:

●Pericarditis – The predominance of isolated left-sided pleural effusions following cardiac surgery has led to the suggestion that underlying pericarditis is an important causative factor [3,11]. One study demonstrated that patients who had persistent pericardial effusions after cardiac surgery were also more likely to have late-onset pleural effusions [11]. In another report of 47 patients undergoing elective CABG surgery, there was no correlation between pericardial and pleural effusions seven days after surgery but there was a relationship after 14 and 30 days [3].

●Pleural inflammation – Even in the absence of pericarditis, the presence of elevated pleural fluid protein content, inflammatory cells, and vascular endothelial growth factor supports the role of pleural inflammation in late nonspecific effusions that follow cardiac surgery [26].

●Incomplete resolution of post-cardiac injury syndrome – Some believe that late pleural effusions may be a manifestation of resolving post-cardiac injury syndrome, although typical clinical and serologic features are absent. (See "Post-cardiac injury syndromes".)

Pleural effusions due to specific etiologies — Specific causes of pleural effusion following cardiac surgery include the following (table 1):

●Post-cardiac injury syndrome (ie, the postpericardiotomy syndrome variant) (see "Post-cardiac injury syndromes")

●Serious complicating conditions of cardiac surgery including:

•Heart failure including constrictive pericarditis (see "Approach to diagnosis and evaluation of acute decompensated heart failure in adults" and "Constrictive pericarditis: Diagnostic evaluation")

•Pulmonary embolism (PE) (see "Epidemiology and pathogenesis of acute pulmonary embolism in adults", section on 'Pathophysiologic response to PE' and "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism", section on 'Clinical presentation')

•Hemothorax (see "Initial evaluation and management of penetrating thoracic trauma in adults", section on 'Hemothorax and vascular injury')

•Parapneumonic effusion or empyema (includes trapped lung) (see "Epidemiology, clinical presentation, and diagnostic evaluation of parapneumonic effusion and empyema in adults" and "Diagnosis and management of pleural causes of nonexpandable lung")

•Chylothorax (table 2) (see "Etiology, clinical presentation, and diagnosis of chylothorax")

•Infectious mediastinitis (see "Postoperative mediastinitis after cardiac surgery")

•Intravascular catheter erosion into the pleural space (see "Central venous catheters: Overview of complications and prevention in adults")

Although all of these conditions typically occur within the early postoperative phase, late presentations are plausible for some of these conditions (eg, chylothorax, PE, parapneumonic effusion).

RISK FACTORS — Several factors that increase the risk of pleural effusion following cardiac surgery have been suggested. However, patients without these risk factors can still develop pleural effusion.

Common risk factors for the development of nonspecific pleural effusion include the following [3,5,6,9,14,15,21-23]:

●Cardiac cooling (particularly with ice)

●Pleurotomy

●Internal mammary artery coronary artery grafting (IMA-CABG)

Other less commonly cited risk factors include the following [27-32]:

●Female sex

●Heart failure

●Anticoagulant therapy

●Antiarrhythmic agents

●Low body weight

●Posterior pericardiotomy approach

●Previous asbestos exposure

CLINICAL MANIFESTATIONS

Clinical features — The majority of pleural effusions that occur after cardiac surgery are small, asymptomatic, and often present as an incidental pleural effusion on a routine postoperative chest radiograph or on lung ultrasound [3,15,33].

●History and examination – Following cardiac surgery, only 7 to 10 percent of nonspecific pleural effusions are large enough to cause symptoms [2,7,34]. Large symptomatic pleural effusions are more likely to occur during the late postoperative period (>30 days following surgery) than the early postoperative period (≤30 days). The most common symptom is dyspnea; chest pain or fever is rare. Examination may reveal dullness and reduced air entry into the base of the affected lung.

The timing and rapidity of symptoms may help differentiate some etiologies. Early effusions tend to be nonspecific but may also be due to pneumonia, pulmonary embolism (PE), and heart failure. Rapid onset of dyspnea should prompt a chest radiograph looking for supportive evidence of a large pleural effusion that might suggest acute hemothorax, chylothorax, or catheter displacement. Late effusions are also more likely to be nonspecific or due to post-cardiac injury syndrome, and less commonly, a slowly developing chylothorax or trapped lung.

The clinical features of an underlying cause may be present. For example, patients with pneumonia may have cough and sputum production, patients with post-cardiac injury syndrome may have fever and pericardial chest pain, and patients with pulmonary embolism may have pleuritic pain and hemoptysis. Patients with heart failure may have evidence of bibasilar crackles and fluid overload and patients with infectious mediastinitis may have anterior chest pain and sternal wound infection. (See "Sputum cultures for the evaluation of bacterial pneumonia" and "Overview of community-acquired pneumonia in adults", section on 'Clinical presentation' and "Epidemiology, pathogenesis, microbiology, and diagnosis of hospital-acquired and ventilator-associated pneumonia in adults" and "Post-cardiac injury syndromes", section on 'Clinical features' and "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism", section on 'Clinical presentation' and "Approach to diagnosis and evaluation of acute decompensated heart failure in adults", section on 'Clinical manifestations' and "Postoperative mediastinitis after cardiac surgery", section on 'Clinical features'.)

●Laboratory tests – Routine laboratory tests are also nonspecific but may reveal an elevated white blood cell count and erythrocyte sedimentation rate or C-reactive protein (from inflammation and/or infection). A rapid decrease in hematocrit may suggest hemothorax. D-Dimer may be elevated in patients with PE, but as it is almost always elevated after surgery, D-dimer offers limited utility in this setting. Brain natriuretic peptide may be elevated in patients with heart failure.

Chest radiograph — Pleural effusions that develop after cardiac surgery are typically unilateral and left-sided (since pleurotomy is mostly left-sided in cardiac surgery) [35]. In one study three-quarters of pleural effusions in this setting were left-sided or larger on the left, while 20 percent were bilateral and less than 10 percent were right-sided or larger on the right [35].

Most pleural effusions in this setting occur early (ie, within the first 30 days postoperatively) and are small (ie, occupy less than 25 percent of the hemithorax or less than two intercostal spaces). Approximately, 10 percent are large (ie, occupy more than 25 percent of the hemithorax or more than two intercostal spaces), and these tend to occur later in the course following surgery.

The chest radiograph may also show evidence of previous surgery or of complications of the surgery such as pulmonary edema to suggest heart failure, a newly enlarged heart that may suggest a pericardial effusion, pneumonia that may suggest a parapneumonic effusion, a wide mediastinum to suggest infectious mediastinitis, or vascular catheter displacement to suggest erosion into the pleural space.

EVALUATION

Estimate symptoms due to the effusion and their time course — For patients who develop a pleural effusion following cardiac surgery, we assess for the presence or absence of symptoms due to the pleural effusion as well as for the time course of symptoms (ie, early [≤30 days after surgery] versus late [>30 days after surgery]), since this information can dictate management. For example, patients who are asymptomatic who develop early, small pleural effusions do not typically need thoracentesis, while patients who are symptomatic or have large effusions generally need further investigation, typically thoracentesis.

Determining whether symptoms are due to the effusion can be challenging and is discussed separately. (See "Management of malignant pleural effusions", section on 'Treatment goals'.)

Assess for nonspecific or specific etiology — For patients who develop a pleural effusion following cardiac surgery, we re-take the history and re-examine the patient.

We inquire about the nature of the surgery (eg, patients who undergo cardiac bypass surgery are more likely to develop pleural effusion than patients who undergo valve surgery), perioperative risk factors for nonspecific pleural effusion (see 'Risk factors' above) as well as examine surgical notes for complexity and complications that may have occurred during surgery. We assess fluid balance and look for evidence of heart failure (eg, elevated jugular venous pressure and peripheral edema). We look for evidence of previous bleeding via the chest tube or any periods of perioperative hemodynamic instability to suggest hemothorax and for signs of infection including cough and sputum production to suggest pneumonia. We inquire about pericardial and/or pleuritic pain and examine the patient for a low-grade fever and a pericardial or pleural rub to suggest post-cardiac injury syndrome. Pleuritic chest pain, hemoptysis, and hypoxemia may suggest pulmonary embolism. Dyspnea that is disproportionate to the size of the effusion may also suggest pneumonia, PE or heart failure.

Estimate size by ultrasonography — For most patients who develop a pleural effusion following cardiac surgery, we estimate pleural effusion size using bedside ultrasonography [33,36]. The estimation is somewhat subjective. A pleural effusion that encompasses 25 percent or more of the hemithorax is considered large enough to undergo diagnostic and/or therapeutic thoracentesis. In contrast one that encompasses less than 25 percent of the hemithorax may not necessarily need to be investigated, unless it is associated with symptoms or a specific etiology is suspected (eg, chylothorax). (See "Bedside pleural ultrasonography: Equipment, technique, and the identification of pleural effusion and pneumothorax".)

If bedside ultrasonography is not available, size may be estimated using plain chest radiography and if the pleural effusion occupies more than two intercostal spaces, then this is also considered sufficient for thoracentesis and may require further investigation.

Assess need for further imaging — In many patients, there is no need for further imaging since most have small asymptomatic pleural effusions which resolve spontaneously. However, we typically perform additional imaging when an etiology other than nonspecific pleural effusion related to cardiac surgery is suspected or the effusion appears loculated on ultrasonography. As examples, chest computed tomography (CT) may be useful when pneumonia or parapneumonic effusion is suspected, when the pleural effusion is large or loculated, or if the effusion is recurrent or refractory despite initial percutaneous drainage. CT pulmonary angiography may be performed when PE is suspected. (See "Epidemiology, clinical presentation, and diagnostic evaluation of parapneumonic effusion and empyema in adults", section on 'Chest computed tomography' and "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism", section on 'Computed tomography pulmonary angiography'.)

Assess need for diagnostic thoracentesis — In most cases, we do not perform diagnostic thoracentesis since most effusions are nonspecific, small, asymptomatic, and self-resolving. However, in select patients, we perform diagnostic ultrasound-guided thoracentesis the details of which are discussed in this section.

Large, progressive, symptomatic, or right-sided nonspecific pleural effusion — In patients with pleural effusion following cardiac surgery, we generally perform diagnostic and/or therapeutic ultrasound-guided thoracentesis when the effusion is large, progressive, symptomatic, and/or predominantly right-sided. This can help confirm the diagnosis, identify a specific etiology, and be symptomatically therapeutic. (See "Ultrasound-guided thoracentesis".)

Suspect specific etiology — We also generally perform diagnostic thoracentesis when we suspect specific diagnoses including hemothorax, chylothorax, parapneumonic effusion, empyema, or post-cardiac injury syndrome, provided the pleural effusion is large enough to safely undergo thoracentesis. We do not typically obtain pleural fluid when heart failure is suspected. (See 'Management nonspecific pleural effusions' below.)

Pleural fluid characteristics — Expected pleural fluid characteristics are shown in the table (table 1) [13,27,37]. Nonspecific pleural effusions are exudates. Early pleural effusions tend to be bloody and exudative with a high eosinophil count and neutrophil predominance, while late pleural effusions are typically serous and exudative with a lymphocytic predominance.

Additional testing — Further testing is determined when pleural fluid characteristics differ from expected and/or when symptoms that suggest a specific etiology are present. As examples:

●For patients with features suggestive of heart failure (eg, transudative pleural fluid) or pericarditis, further evaluation with serum brain natriuretic peptide (BNP) electrocardiography (ST changes to support pericarditis), and echocardiography (low ejection fraction, pericardial effusion, constrictive physiology) is appropriate. (See 'Post-cardiac injury/postpericardiotomy syndrome' below and 'Heart failure' below.)

●Turbid fluid may suggest a parapneumonic effusion or empyema which should prompt additional culture and drainage of pleural fluid, or a chylothorax which can be diagnosed by measuring the pleural fluid triglycerides. (See "Epidemiology, clinical presentation, and diagnostic evaluation of parapneumonic effusion and empyema in adults" and "Management and prognosis of parapneumonic pleural effusion and empyema in adults" and "Etiology, clinical presentation, and diagnosis of chylothorax".)

●Tests for tuberculosis are prudent in lymphocytic predominant pleural effusions such that measuring adenosine deaminase levels and performing acid fast staining and culture for mycobacterium is appropriate. (See "Tuberculous pleural effusion", section on 'Diagnosis'.)

●Milky fluid may prompt a triglyceride level to diagnose chylothorax or prompt an evaluation for catheter displacement (eg, lipids transfusing through a central venous catheter). Occasionally, in postoperative patients with limited oral intake, chylous fluid may not have this characteristic milky appearance. Rarely, lymphatic imaging may be needed if chylothorax is diagnosed on thoracentesis. (See "Etiology, clinical presentation, and diagnosis of chylothorax", section on 'Lymphatic imaging in select patients'.)

●Bloody fluid should prompt a formal hematocrit on the fluid, and if consistent with a hemothorax (eg, pleurocrit to hematocrit ratio >50 percent), chest tube drainage or surgical exploration may be indicated. (See 'Hemothorax' below.)

●In rare cases when the diagnosis is unclear, thoracoscopic biopsy may be warranted. (See "Diagnostic evaluation of the hemodynamically stable adult with a pleural effusion", section on 'Pleural biopsy'.)

MANAGEMENT NONSPECIFIC PLEURAL EFFUSIONS — In most patients who have a nonspecific pleural effusion following cardiac surgery, management is largely dictated by pleural effusion size, location, and symptoms.

Asymptomatic and/or small pleural effusion

Observation — For those with a small (<25 percent of the hemithorax), asymptomatic, nonspecific pleural effusion following cardiac surgery, no intervention is needed. This is particularly true for patients whose pleural effusion is left-sided and develops during the first one to two days after surgery and does not progress. We typically follow such patients clinically for the development of symptoms, and radiologically with chest radiography or ultrasonography (eg, every six weeks) for the enlargement or resolution of the effusion. While most pleural effusions resolve within two to six weeks, some pleural effusions may take two to 20 months to clear completely [7,9]. For patients whose pleural effusion increases in size or becomes symptomatic during the observation period, we perform thoracentesis. (See "Bedside pleural ultrasonography: Equipment, technique, and the identification of pleural effusion and pneumothorax".)

Large, symptomatic, progressive, or right-sided pleural effusion

Therapeutic thoracentesis — For patients with a symptomatic, large (>25 percent of the hemithorax), or progressive pleural effusion following cardiac surgery, or a predominantly right-sided pleural effusion occurring in the absence of heart failure, we perform therapeutic thoracentesis with pleural fluid analysis under ultrasound guidance; chest tube or catheter drainage is an alternative and may be preferred for those in whom empyema or hemothorax is suspected or found on pleural fluid analysis. (See "Ultrasound-guided thoracentesis" and "Thoracostomy tubes and catheters: Placement techniques and complications".)

In some cases, asymptomatic patients who have a small right-sided effusion, observation is appropriate. (See 'Observation' above.)

Follow up — Following therapeutic thoracentesis, we immediately inquire about improvement in symptoms and typically perform imaging with chest radiography or ultrasonography to evaluate reduction in size of the pleural effusion. Most patients with nonspecific pleural effusions who undergo therapeutic thoracentesis improve and do not need another intervention.

Thereafter, we follow patients periodically with chest radiography or ultrasonography over the ensuing 2 to 12 months for continued improvement in their symptoms and reduction in the size of the pleural effusion.

However, among those who undergo a single thoracentesis, up to one-third can recur [35].

Recurrent or refractory pleural effusions — For patients with a recurrent large symptomatic nonspecific pleural effusion, we typically perform a second or third thoracentesis. The rate of accumulation may vary but in our experience ranges from days (eg, trapped lung) to weeks.

For patients with nonspecific pleural effusions that are refractory to serial thoracentesis, we typically prescribe an oral nonsteroidal anti-inflammatory agent (NSAID) at a dose and duration similar to patients with post-cardiac injury syndrome, although this approach is unproven. (See "Post-cardiac injury syndromes", section on 'Prevention and treatment'.)

As an alternative, patients may undergo a trial of an oral glucocorticoid (eg, prednisone 0.5 to 1mg/kg daily for two to four weeks; a taper may be necessary). Glucocorticoids have not been shown to prevent long-term sequelae of nonspecific effusions, such as a trapped lung, and are not appropriate for patients in whom a parapneumonic effusion or empyema is suspected.

For patients in whom conservative therapies have failed, we typically refer for video assisted thoracoscopic surgery (VATS; eg, pleurodesis, decortication) [38-40]. In one review of 11 observational studies of VTAS for refractory pleural effusions following cardiac surgery, no recurrences were reported. In rare cases, trapped lung due to dense fibrosis of the visceral pleura was found at VATS, which required conversion to a thoracotomy for decortication. These findings suggest that persistence of late onset pleural effusions may be due to trapped lung. (See "Diagnosis and management of pleural causes of nonexpandable lung", section on 'Trapped lung' and "Overview of minimally invasive thoracic surgery".)

MANAGEMENT PLEURAL EFFUSION DUE TO SPECIFIC ETIOLOGIES — Several specific conditions can cause pleural effusion after cardiac injury. These include post-cardiac injury syndrome, heart failure, pulmonary embolus (PE), hemothorax, pneumonia or empyema, infectious mediastinitis, chylothorax, and central venous catheter erosion into the pleural space (table 1).

Post-cardiac injury/postpericardiotomy syndrome — Postpericardiotomy syndrome (PPCS) is one of the three conditions classified under the more general term of post-cardiac injury syndrome. The other two conditions are postmyocardial infarction syndrome (Dressler syndrome) and post-traumatic pericarditis. (See "Post-cardiac injury syndromes".)

PPCS is a common cause of late pleural effusions following cardiac surgery, although it may begin to develop as early as two to three weeks following injury [41]. PPCS may be more common in patients undergoing aortic and mitral valve replacement, compared with patients undergoing coronary artery bypass grafting (CABG) [42]. PPCS is also a rare complication of PE, traumatic hemopericardium, radiofrequency ablation for tachyarrhythmias, percutaneous coronary intervention, implantation of pacemakers, and multiple other interventions that cause pleuropericardial injury [43-51].

The pleural effusions in PPCS are usually serous or hemorrhagic in appearance, exudative, primarily left-sided (85 percent of patients), and large (eg, >25 percent of the hemithorax in 75 percent of patients) [52-54].

Clinical features, diagnosis, and treatment of post-cardiac injury syndromes are discussed separately. (See "Post-cardiac injury syndromes".)

Heart failure — The pleural effusions in heart failure, including those associated with constrictive pericarditis, are typically bilateral or right-sided and transudative, (as opposed to non-specific postoperative pleural effusions typically left sided and exudative) and patients may have associated clinical features of pulmonary edema or fluid overload. In most patients, we typically obtain an echocardiogram to look for left ventricular systolic function as well as a pericardial effusion or constrictive pericarditis. Evaluation and management of decompensated heart failure and constrictive pericarditis are discussed separately. (See "Approach to diagnosis and evaluation of acute decompensated heart failure in adults" and "Treatment of acute decompensated heart failure: General considerations" and "Treatment of acute decompensated heart failure: Specific therapies" and "Constrictive pericarditis: Diagnostic evaluation".)

Pulmonary embolism — Pleural effusion is a rare manifestation of PE and is mostly due to peripheral lung infarction, which can occur in approximately 10 percent of patients with PE.

The effusion is typically small, exudative and bloody, predilects to the affected side and is often associated with pleuritic pain. Imaging, typically computed tomographic pulmonary angiography, is required to make the diagnosis. Evaluation and management of PE are discussed separately. (See "Epidemiology and pathogenesis of acute pulmonary embolism in adults", section on 'Pathophysiologic response to PE' and "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism" and "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults".)

Hemothorax — Hemothorax is potential complication of cardiac surgery in which the pleural space is violated. Hemothorax can develop in any patient following cardiac surgery particularly those who are receiving anticoagulant and/or antiplatelet therapy.

The pleural effusion is typically left-sided (ie, the most common side for surgical intervention) and most often occurs within the first few hours or days following surgery. Some patients may experience an acute drop in their hematocrit and/or blood pressure while others develop a slow decline in hematocrit without hemodynamic instability.

When hemothorax is suspected, we immediately obtain a chest radiograph or ultrasound and perform diagnostic thoracentesis. A pleurocrit:hematocrit ratio >50 percent strongly supports the diagnosis.

Patients typically require chest tube drainage to avoid the future development of trapped lung [12,55,56]. We consider surgical exploration if acute blood loss is hemodynamically significant, large volume blood transfusion is required, or if blood loss is significant (eg, >200 mL/hour, >1500 mL in 24 hours [57]). We also consider intervention when hemothorax is incompletely drained with a chest tube (ie, thoracoscopic evacuation). Principles garnered from post-traumatic retained hemothorax are broadly applicable in post-cardiac surgery patients. (See "Thoracostomy tubes and catheters: Indications and tube selection in adults and children", section on 'Hemothorax' and "Initial evaluation and management of penetrating thoracic trauma in adults", section on 'Hemothorax and vascular injury'.)

Parapneumonic effusion or empyema — Patients who undergo cardiac surgery are at increased risk of developing pneumonia postoperatively that can be complicated by a parapneumonic effusion or empyema.

Pneumonia is an early complication of cardiac surgery. However, the diagnosis may be missed in some cases, and patients may present later in the course with a complicated pleural effusion or trapped lung. The pleural fluid is typically exudative with neutrophilic predominance (as opposed to lymphocytic predominance in nonspecific pleuritis), an elevated white blood cell count and lactate dehydrogenase; the glucose and pH may be low and the gram stain and/or culture may be positive for micro-organisms. The diagnosis and treatment of a parapneumonic effusion are (table 3) are discussed separately. (See "Overview of the management of postoperative pulmonary complications", section on 'Pneumonia' and "Epidemiology, clinical presentation, and diagnostic evaluation of parapneumonic effusion and empyema in adults" and "Management and prognosis of parapneumonic pleural effusion and empyema in adults".)

Infectious mediastinitis — Infectious mediastinitis is a serious complication of cardiac surgery that requires prompt surgical intervention such as debridement and irrigation [58-60].

Infectious mediastinitis generally occurs a few weeks after surgery and is typically associated with sternal wound infection. Chest computed tomographic (CT) imaging classically reveals mediastinal fluid collections that are difficult to differentiate from aseptic, postoperative mediastinal hematomas. Infectious mediastinitis, however, more commonly produces bilateral pleural effusions and/or CT evidence of mediastinal soft tissue swelling compared to patients with mediastinal fluid collections that do not require drainage [58]. (See "Postoperative mediastinitis after cardiac surgery".)

Chylothorax — Chylothoraces develop in less than 1 percent of patients undergoing cardiac surgery and result from surgical disruption of the thoracic duct or its intrathoracic lymphatic tributaries.

Chylothorax is more common in children than adults undergoing cardiac surgery and is most often seen after surgical correction of congenital heart disorders that involve the regions of the aorta or pulmonary arteries near the thoracic duct. Chylothorax is rarely seen in adults after intrapericardial operations, such as CABG surgery, although it develops most commonly when the internal mammary arteries (IMA) are used as a coronary graft, especially the left IMA [61,62].

Chylothorax can be right- or left-sided and is typically milky in appearance (although this typical appearance may be absent is the postoperative period), is exudative, and has a triglyceride level >110mg/dL. The evaluation and management of chylothorax are discussed separately. (See "Etiology, clinical presentation, and diagnosis of chylothorax" and "Management of chylothorax".)

Intravascular catheter erosion — In patients with a central venous catheter, erosion of the catheter through venous structures should be considered when pleural fluid accumulates in association with radiographic evidence of mediastinal widening [63,64].

Catheters can erode at any time after placement. In a case series, the mean time to erosion after catheter insertion was 3.6 days [63]. The diagnosis can be confirmed by pleural fluid-characteristics that are similar to the infusate (eg, milky and high triglycerides if lipids are being infused, low glucose if saline is being infused, or high glucose if total parenteral nutrition is being infused) and imaging that confirms inappropriate placement (eg, chest CT with contrast or chest radiograph). In some cases when uncertainty exists, the diagnosis may need to be confirmed by demonstrating extravascular extravasation of contrast infused through the catheter. The diagnosis and treatment of catheter displacement is discussed separately. (See "Pleural effusion of extra-vascular origin (PEEVO)", section on 'Extravascular migration of central venous catheter'.)

PROGNOSIS — The prognosis of most pleural effusions following cardiac surgery is good and the majority resolve over time. However, the development of a late effusion is one of the most common reasons for readmission after coronary artery bypass grafting and may be associated with a more limited prognosis, especially if the effusion is recurrent or refractory [20].

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: Pleural effusion".)

SUMMARY AND RECOMMENDATIONS

●Prevalence – Pleural effusions are frequent following coronary artery bypass grafting (CABG), cardiac transplantation, and, less often, mitral and aortic valve replacement surgery. They can occur in up to 90 percent of patients during the early postoperative period (≤30 days) and in up to two-thirds in the late postoperative period (>30 days). (See 'Prevalence' above.)

●Etiologies and pathogenesis – Most pleural effusions following cardiac surgery are nonspecific in nature (ie, related to the surgery itself) but can also be due to a complication of surgery such as post-cardiac injury syndrome (also known as postpericardiotomy syndrome [PPCS]), heart failure, pulmonary embolism (PE), hemothorax, infection, mediastinitis, chylothorax, or intravascular catheter erosion into the pleural space (table 1). The pathogenesis of nonspecific pleural effusions is unclear, although several mechanisms have been proposed including trauma of the surgery itself (early effusions) and immune mediated inflammation (late effusions) (table 4). (See 'Etiologies and pathogenesis' above.)

●Risk factors – Common factors that increase the risk of nonspecific pleural effusion following cardiac surgery include cardiac cooling (particularly with ice), pleurotomy, and internal mammary artery grafting. Others include female sex, heart failure, anticoagulant therapy, antiarrhythmic agents, low body weight, and a posterior pericardiotomy approach. (See 'Risk factors' above.)

●Clinical features

•Most pleural effusions that occur following cardiac surgery are small, nonspecific, and asymptomatic. Only 7 to 10 percent of nonspecific pleural effusions are large enough to cause symptoms and these are more likely to occur during the late postoperative period (ie, >30 days following surgery). The most common symptom is dyspnea; chest pain or fever is rare. (See 'Clinical manifestations' above.)

•The clinical features of a specific underlying cause may be present. For example, bibasilar crackles and fluid overload may suggest heart failure, cough and sputum production may suggest pneumonia, fever and pericardial chest pain may suggest PPCS, pleuritic pain and hemoptysis may suggest PE, and anterior chest pain and sternal wound infection may suggest infectious mediastinitis. Laboratory findings are typically nonspecific but may have supportive evidence of an underlying cause (eg, a sudden decrease in hematocrit may suggest hemothorax, elevated D-Dimer may suggest PE, and elevated brain natriuretic peptide may suggest heart failure). (See 'Clinical features' above.)

•Pleural effusions that develop after cardiac surgery are typically unilateral and left-sided. Approximately 20 percent are bilateral and less than 10 percent are right-sided. The chest radiograph may also show evidence of previous surgery or of complications of the surgery such as pulmonary edema to suggest heart failure, an newly enlarged heart that may suggest a pericardial effusion, pneumonia to suggest parapneumonic effusion, wide mediastinum to suggest infectious mediastinitis, or vascular catheter displacement to suggest erosion into the pleural space. (See 'Chest radiograph' above.)

●Evaluation – For patients who develop a pleural effusion following cardiac surgery, we assess for the presence or absence of symptoms due to the pleural effusion, the time course of symptoms (ie, early versus late), and features that may suggest an underlying specific etiology. We also estimate the size of the pleural effusion, typically using bedside ultrasonography or chest radiography. Additional imaging is generally only needed when an etiology other than a nonspecific pleural effusion is suspected, the effusion appears loculated on ultrasonography, or the effusion is refractory to initial percutaneous drainage. (See 'Evaluation' above.)

Thoracentesis:

•In most cases, we do not perform diagnostic thoracentesis since most effusions are nonspecific, small, asymptomatic, and self-resolving. (See 'Assess need for diagnostic thoracentesis' above.)

•For patients in whom the effusion is large (ie, encompasses 25 percent or more of the hemithorax or two intercostal spaces), progressive, symptomatic, or predominantly right-sided, or patients in whom a specific diagnosis is suspected, we suggest diagnostic and/or therapeutic ultrasound-guided thoracentesis (Grade 2C). Expected pleural fluid characteristics are shown in the table (table 1). (See 'Large, progressive, symptomatic, or right-sided nonspecific pleural effusion' above and 'Suspect specific etiology' above and 'Pleural fluid characteristics' above.)

•Further testing is determined when pleural fluid characteristics differ from expected and/or when symptoms suggest a specific etiology. (See 'Additional testing' above.)

●Management of nonspecific pleural effusions following cardiac surgery

•For patients with a small (<25 percent of the hemithorax), asymptomatic, pleural effusion following cardiac surgery, we suggest periodic observation rather than thoracentesis (Grade 2C). We follow patients clinically and radiologically until resolution (eg, two to six weeks or longer). (See 'Asymptomatic and/or small pleural effusion' above.)

•For patients with a large (>25 percent of the hemithorax), symptomatic, or progressive pleural effusion, or a predominantly right-sided pleural effusion occurring in the absence of heart failure, we suggest ultrasound-guided therapeutic thoracentesis with pleural fluid analysis (Grade 2C). Alternatively, chest tube drainage is an option and may be preferred for those in whom empyema or hemothorax is suspected or found on pleural fluid analysis. (See 'Large, symptomatic, progressive, or right-sided pleural effusion' above.)

•Most patients with nonspecific pleural effusions who undergo therapeutic thoracentesis improve and do not need another intervention, although up to one-third can recur. For those with recurrence, we typically perform a second or third therapeutic thoracentesis. (See 'Recurrent or refractory pleural effusions' above.)

-For patients with nonspecific pleural effusions that are refractory to serial thoracentesis, we suggest treatment with an oral nonsteroidal anti-inflammatory agent (NSAID) (Grade 2C). The dose and duration similar to patients with post-cardiac injury syndrome. As an alternative, patients may undergo a trial of an oral glucocorticoid (eg, prednisone 0.5 to 1mg/kg daily for two to four weeks; a taper may be necessary). Use of these treatments in this setting is based on indirect data from patients with post-cardiac injury syndrome. (See "Post-cardiac injury syndromes", section on 'Prevention and treatment'.)

-For patients in whom conservative therapies have failed, we suggest video assisted thoracoscopic surgical (VATS) pleurodesis. Decortication may be required if trapped lung is discovered during VATS.

●Management of pleural effusion caused by a specific etiology following cardiac surgery – Several specific conditions can cause pleural effusion after cardiac injury. These include post-cardiac injury syndrome (specifically PPCS), heart failure, PE, hemothorax, pneumonia or empyema, infectious mediastinitis, chylothorax, and central venous catheter erosion into the pleural space. Pleural fluid characteristics are shown in the table (table 1) and their management is discussed separately.

•Heart failure (see "Approach to diagnosis and evaluation of acute decompensated heart failure in adults" and "Treatment of acute decompensated heart failure: General considerations" and "Treatment of acute decompensated heart failure: Specific therapies")

•Pulmonary embolism (see "Epidemiology and pathogenesis of acute pulmonary embolism in adults", section on 'Pathophysiologic response to PE' and "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism" and "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults")

•Hemothorax (see "Thoracostomy tubes and catheters: Indications and tube selection in adults and children", section on 'Hemothorax' and "Initial evaluation and management of penetrating thoracic trauma in adults", section on 'Hemothorax and vascular injury')

•Parapneumonic effusion or empyema (see "Overview of the management of postoperative pulmonary complications", section on 'Pneumonia' and "Epidemiology, clinical presentation, and diagnostic evaluation of parapneumonic effusion and empyema in adults" and "Management and prognosis of parapneumonic pleural effusion and empyema in adults")

•Infectious mediastinitis (see "Postoperative mediastinitis after cardiac surgery")

•Chylothorax (see "Etiology, clinical presentation, and diagnosis of chylothorax" and "Management of chylothorax")

•Intravascular catheter erosion into the pleural space (see "Pleural effusion of extra-vascular origin (PEEVO)", section on 'Extravascular migration of central venous catheter')

●Prognosis – The prognosis of most pleural effusions following cardiac surgery is good and the majority resolve over time. However, the development of a late effusion may be associated with a more limited prognosis, especially if the effusion is recurrent or refractory. (See 'Prognosis' above.)