The hallmarks of pleural effusion on physical examination are
diminution of breath sounds, dullness to percussion, and decreased
tactile fremitus over the area of the effusion, but such physical findings
generally occur only with effusions > 300 mL. Other findings such
as a friction rub may accompany the development of pleural effusion
secondary to pleuritis, and are dependent on the underlying pathology.
In cases of large effusions, egophony and crackles may be heard
above the effusion, reflecting compressed lung overlying the effusion.
Signs of tension (increased pressure affecting one hemithorax) such
as midline shift of the trachea and bulging of the intercostal spaces may
accompany massive effusions (see Table 22–2).
Causes of Massive Effusion with Mediastinal Shift. |Favorite Table|Download (.pdf)
Causes of Massive Effusion with Mediastinal Shift.
|Metastatic disease of the pleura (75% of massive
|Congestive heart failure|
Symptoms of pleural effusion may be absent, but often include
dyspnea, cough, or pleuritic chest pain. Interestingly, dyspnea
is often disproportionate to measurable hypoxemia or reduction in lung
volume, and appears to result more from mechanical inefficiency
of the respiratory muscles due to distortion of the chest wall and
diaphragm by the effusion. Dyspnea markedly out of proportion to
the size of the effusion should suggest the possibility of pulmonary
embolism as the underlying diagnosis.
Although as little as 5 mL of pleural fluid can be detected using
lateral decubitus position chest radiography, the more traditional
posteroanterior and lateral views are significantly less sensitive. At
least 50–75 mL of fluid must accumulate before blunting
of the posterior costophrenic angle may be identified on the lateral
view, and > 175–200 mL must be present to cause visible
blunting of the lateral costophrenic angles on the posteroanterior
view. With greater accumulation, the apex of the hemidiaphragm may
be laterally displaced, then obscured (>500 mL), with subsequent
opacification of the lung base. The appearance of a fluid meniscus
along the chest wall or mediastinum is frequently seen, as is thickening
of the major and minor fissures, indicative of superior tracking
Trapping or loculation of pleural fluid may lead to the appearance
of pleural or even parenchymal lung masses. These have a typically
lenticular or rounded appearance and smooth-contoured interface
with the lung. Such collections usually appear along the chest wall
or within the major or minor fissures of the lung, and lack air
bronchograms within the “mass.” Loculations are
best identified by either ultrasound or computed tomography (CT)
Ultrasound may be very useful in two situations: (1) diagnosis
and sampling of loculated fluid collections and (2) guided sampling
of small effusions or those difficult to tap (failing two or three attempts).
Ultrasound is not generally warranted for the routine diagnosis/sampling
of moderate or large effusions and has not been shown to decrease
the incidence of pneumothorax as a complication of the procedure.
CT is the most sensitive radiographic study for the detection
and delineation of pleural fluid collections. Free flowing fluid
appears as a sickle-shaped opacity in the most dependent part of
the thorax posteriorly, whereas loculations appear as lenticular
or rounded opacities in a fixed position (nonflowing). CT may be
extremely useful in the sampling of loculated fluid or the placement
of drainage catheters in complex collections. Although CT is also
helpful in distinguishing pleural fluid from parenchymal and extrapleural
disease due to its ability to distinguish these anatomic compartments,
CT density coefficients are not specific enough to definitively
discriminate among parenchymal lesions, solid pleural masses, and
pleural collections of serous fluid, blood, or pus.
Magnetic resonance imaging is of limited use at present in the
evaluation of pleural effusions, with CT being the preferred modality.
It has been said that “the sun should never set on an
undiagnosed effusion.” Although this imperative continues
to be largely true when evaluating new pleural effusions, exceptions
have been suggested. First, in cases of negligible pleural fluid
volume (< 1 cm of fluid present on lateral decubitus film or
the absence of blunting of the posterior costophrenic angles on
lateral chest radiograph) sampling may be deferred. In the case
of suspected pneumonia, a more aggressive approach may be warranted,
as detailed in Chapter 23: Empyema. Second, patients with congestive heart
failure and bilateral pleural effusions of similar size, in the
absence of chest pain or fever, may be observed during a 72-h trial
of diuresis, after which those with persistent effusions should undergo
prompt thoracentesis. Otherwise, new effusions should be sampled
on presentation. If the patient is dyspneic at rest, therapeutic
drainage should be performed to alleviate symptoms. This may be
done as one or more thoracenteses of 1–1.5 L each (or more
in the case of massive effusion), as long as the patient does not
develop worsening dyspnea, chest pain, or severe cough during fluid
removal. Although relative contraindications to thoracentesis including
bleeding diathesis [as indicated by platelets <50,000,
prothrombin time (PT) or partial thromboplastin time (PTT) >2 mid-normal
range, or serum creatinine >6] and mechanical ventilation
should be weighed, no absolute contraindication exists. Small or
loculated effusions may require ultrasound- or CT-guided thoracentesis.
Important clues to the cause of pleural effusion may be gained
by simple inspection of the fluid. Frank pus indicates a pleural
space infection, or empyema, the management of which is detailed in
Chapter 23: Empyema. Bloody fluid raises the possibility of hemothorax, and
a spun fluid hematocrit should be performed. A fluid hematocrit
>50% of the measured peripheral blood hematocrit is diagnostic
of hemothorax (Chapter 21: Pneumothorax/Hemothorax), whereas a fluid hematocrit 1–50% of
the peripheral blood typically suggests cancer. An elevated hematocrit
may also occur with pulmonary embolism, trauma, or even pneumonia.
A fluid hematocrit <1% of peripheral blood is considered
insignificantly bloody. Milky or turbid pleural fluid should also
be spun: clearing indicates the presence of cellular debris, whereas
continued turbidity suggests chylous effusion (lipid-containing
lymph from an obstructed or disrupted thoracic duct) or pseudochylous
effusion (cholesterol-rich fluid forming in a chronic setting).
Fluid triglycerides and cholesterol should be sent for testing to
pursue these diagnoses (see below). Clear yellow pleural fluid,
particularly with an odor of urine, suggests urinary obstruction
with urinothorax, and fluid creatinine (Cr) confirms this diagnosis (pleural
fluid Cr > serum Cr).
Essential to the diagnosis of pleural effusions is the distinction
between exudative effusions (those arising from disease affecting
the pleura and leading to increased vascular permeability and/or decreased
lymphatic drainage of the pleural space) and transudative effusions
(those arising from derangement of hydrostatic and oncotic forces
within the thorax in the setting of normal pleura). This distinction
is best made using pleural fluid (PF) and serum lactate dehydrogenase
(LDH) and total protein measurements to apply “Light’s
criteria.” Pleural fluid is likely exudative if one or more
of the following criteria are met:
- 1. PF protein/serum
protein ratio > 0.5.
- 2. PF LDH/serum LDH
RATIO > 0.6.
- 3. PF LDH > two-thirds the upper
limits of the normal serum range.
Occasionally, these criteria may be met in the setting of strong
clinical evidence to suggest a transudative etiology (most commonly
in the setting of recent diuretic therapy). In this situation, determination
of the PF–serum albumin gradient is useful: a gradient
>1.2 g/dL is indicative of a transudate.
Examination of pleural fluid may be limited to LDH and total
protein (±albumin, as above) in the setting of
a presumed transudative etiology [eg, congestive heart
failure (CHF)]. Once established as a transudate, diagnosis
and treatment of the underlying condition, such as CHF (> 90% of
cases), cirrhosis, or nephrotic syndrome, should be pursued. Further
evaluation is warranted for exudative effusions, however (Figure
22–1). This should include fluid cell count and
differential, glucose, Gram stain and culture (best sent in bedside-inoculated
blood culture bottles), and cytology (several hundred milliliters,
if a malignancy is suspected). Findings in the most common causes
of pleural effusion are listed in Table 22–3.
Other, specialized tests are indicated by the clinical setting,
as noted in Table 22–3 and discussed below.
Algorithm for the initial evaluation of pleural effusion.
(Modified from Light RW: Pleural effusion. N Engl J Med 2002;346:1971.)
Table 22–3. Causes
of Pleural Effusions and Their Typical Findings (Ranked by Relative
Incidence).1 |Favorite Table|Download (.pdf)
Table 22–3. Causes
of Pleural Effusions and Their Typical Findings (Ranked by Relative
|Etiology||Gross Appearance||Exudative (Ex)/Transudative (Trans)||White Blood Cell Count and Differential||Red Blood Cell Count||Glucose||Chest Radiograph||Comment|
|Congestive heart failure||Serous||Trans||<1k, lymphocytes and mesothelial cells||<1k||Serum||Most often (80–90%) bilateral and symmetric;
pulmonary vascular congestion, pulmonary edema||May appear “exudate” by Light’s
criteria, especially after ≥ 3 day diuresis|
|Uncomplicated parapneumonic||Serous to turbid||Ex||5k–25k, PMNs; mononuclear in viral disease||<5k||Serum||Loculated or large effusions (>50% hemithorax) may require
thoracostomy (see Chapter 23: Empyema)||Fluid pH <7.30
or glucose <60 suggest complicated parapneumonic, whereas frank pus,
positive Gram stain or culture is diagnostic of empyema (see Chapter 23: Empyema)|
|Empyema||Turbid to purulent (pus)||Ex||25k–100k, PMNs (may be < 1k due to lysis)||<5k||Low||May cause massive effusion with mediastinal shift|
|Cancer (other than mesothelioma)||Turbid to bloody; may be serous||Ex||1k–10k, lymphocytes or mesothelial cells predominantly; eosinophilia
rare||100–100k||Usually serum; < 60 mg/dL in 15%||Effusion ipsilateral to cancer in lung ± breast cancer; effusions
typically moderate/large; 10% have massive effusion||Lung > breast > lymphoma; low glucose and pH may correlate with
large tumor burden and poor prognosis; fluid cytology ± amylase should be obtained|
|Pulmonary embolism||Serous to grossly bloody||Ex > trans||1k–100k, PMNs and lymphocytes; often > 10% eosinophils
and/or mesothelial cells||100–100k||Variable||30–40% PE have effusion, usually small;
>50% have concomitant parenchymal infiltrate||55% of all patients with pleuritic chest pain and
effusion have PE; 25% of patients with PE and effusion have
no chest pain|
|Acute (< 30 days)||Bloody||Ex||Frequently eosinophilic||>100k||Serum||Usually small and left-sided||Up to 90% of postCABG patients; resolves spontaneously|
|Chronic||Serous > bloody||Ex||Lymphocytic||<5k||Serum||Larger than acute; 40% bilateral or right-sided||10% of patients; no correlation to history of acute effusion; ?autoimmune|
|Cirrhosis with ascites||Serous||Trans||<1k, lymphocytes and mesothelial cells||<1k||Serum||70% are right-sided, but may be left (15%) or
bilateral (15%); small to massive (10%); may have small
pneumothorax after paracentesis||5–10% of patients with ascites develop hepatic hydrothorax; may
be hemorrhagic in setting of coagulopathy; detectable ascites may
be absent; may be complicated by spontaneous bacterial pleuritis
|Pancreatitis||Serosanguineous to turbid||Ex||1k–50k, PMNs||1k–10k||Serum||Typically small and bilateral; left-sided predominance in larger
effusions||High amylase in fluid; persistent
effusions (> 2 weeks) suggest pseudocyst or abscess|
|Esophageal rupture||Turbid to purulent; red-brown||Ex||1k–50k, PMNs; squamous epithelial cells||1k–10k||Usually low||Usually left-sided; pneumothorax in 25%; occasionally mediastinal
widening and/or pneumomediastinum||Uncommon, but a surgical emergency; high amylase (salivary),
low pH, sometimes food particles;
esophagogram to diagnose|
|Connective tissue disease|
|Rheumatoid disease||Turbid—greenish-yellow||Ex||1k–20k, mononuclear cells
or PMNs, multi-nucleated macrophages||<1k||< 40 mg/dL||Small/moderate effusion; 25% bilateral; may
wax/wane and/or alternate sides over time||5% RA pts: 80% male, 80% have subcut
nodules; usually older with longstanding RA; may have chest pain
± fever; fluid pH < 7.20, high RF (>1:320) and LDH, often cholesterol
|Lupus (systemic or drug induced)||Serosanguineous/yellow||Ex||1k–20k, mononuclear cells or PMNs||<1k||Usually serum||Small effusion; 50% bilateral; may wax/wane
and/or alternate sides over time||40–50% develop effusion; most have chest
pain ± fever; may precede other lupus findings; pH >
7.35 (80%); fluid ANA or LE cell preparation not
helpful in diagnosis|
|Tuberculosis||Serous to serosanguineous||Ex||1k–10k, lymphocytic (PMNs if early); eosinophils
>10% or mesothelial cells >5% make TB an unlikely
diagnosis||<10k||May be serum or low||Small/moderate effusion, almost always unilateral;
one-third with concomitant parenchymal infiltrate (ipsilateral)||Two-thirds of patients have acute illness with cough and
chest pain, one-third chronic with low grade fever, weakness and weight
loss; fluid protein >5.0 g/dL suggests diagnosis;
see text for other tests|
|Asbestos||Serous to serosanguineous||Ex||1k–20k, mononuclear cells or PMNs; often (25–50%)
of patients eosinophilic||< 10k||Serum||Small/moderate unilateral (90%) effusion;
often with pleural plaques and/or parenchymal asbestosis
(50%)||Diagnosis of exclusion; one-half to two-thirds are asymptomatic;
chest pain most common complaint (one-third); may develop within years
to decades following exposure to asbestos; 80% benign/recurrent
but 20% develop massive pleural fibrosis (and ~5% mesothelioma)|
|Mesothelioma||Serous to bloody; may be viscid||Ex||1k–5k; mesothelial cells (normal and malignant) and
varying PMNs and lymphocytes||100–100k||May be serum or low (33%)||Effusion present in 75–90% of mesothelioma
patients; often large (and obscuring tumor); one-third have evident
pleural plaques in opposite hemithorax; may produce ipsilateral
mediastinal shift and loculation as encases lung; CT scan useful in
evaluation||Most patients experience insidious onset chest pain or dyspnea; 10–70%
with history of asbestos exposure; although fluid cytology may indicate malignancy, VATS
is often required to establish diagnosis of mesothelioma|
Although fluid cell count is relatively nonspecific, the cell
type predominance and pattern may be helpful in establishing a diagnosis.
Neutrophilic effusions [>50% polymorphonuclear
cells (PMNs)] indicate an acute process affecting the pleural
surfaces. In association with infiltrates, a neutrophilic effusion
typically reflects a parapneumonic response (including early viral,
early tuberculous, or more commonly bacterial infections) or pulmonary
embolism. In the absence of infiltrates, pulmonary embolism, acute
viral infection, or pancreatitis should be considered. Cancer and
acute tuberculous pleuritis are rare causes of neutrophilic effusions.
A mononuclear predominance usually accompanies chronic processes
of the pleura. Such processes include cancer, resolving viral pleuritis,
tuberculosis, or pulmonary embolism (20%). A predominance
of small lymphocytes on differential (>50%) is nearly always
secondary to tuberculosis, cancer, or the occasional chronic postcoronary
artery bypass grafting (CABG) effusion, whereas the presence of
fluid eosinophils (>10%) or more than a few mesothelial
cells makes tuberculosis a less likely diagnosis and may suggest
a diagnosis of pulmonary embolism (PE).
Fluid eosinophilia (>10% eosinophils) is most often
secondary to air or blood in the pleural space following trauma
or other intrathoracic insult, and can confound interpretation of
the fluid differential following repeated thoracentesis. It may
also arise from drug-induced pleuritis (dantrolene, bromocriptine,
or nitrofurantoin), asbestos-related effusion, paragonimiasis, or
Churg–Strauss syndrome. Concomitant low fluid pH and glucose
suggest the latter two diagnoses. Frequently, no diagnosis is established
in the setting of eosinophilic effusion.
Pleural fluid glucose may be significantly lower than serum due
to impairment of glucose diffusion through the pleura and/or
increased consumption of glucose by cells within the pleural space. Causes
of low fluid glucose (<60 mg/dL) include complicated
parapneumonic effusion/empyema, cancer, tuberculous pleuritis,
and rheumatoid disease. Less common causes include esophageal rupture,
hemothorax, Churg–Strauss syndrome, paragonimiasis, and
occasionally lupus pleuritis. Although fluid pH is currently preferred
over fluid glucose in the evaluation of parapneumonic effusions,
in its absence a fluid glucose of <60 mg/dL is useful
in suggesting a complicated parapneumonic effusion that possibly
requires chest tube drainage (see Chapter 23: Empyema).
Causes of elevated pleural fluid glucose are few and result almost
exclusively from the leakage of high glucose solutions into the
pleural space either via a diaphragmatic defect (peritoneal dialysis with
high glucose dialysate) or iatrogenic misadventure (such as a misplaced
central venous line).
Although no longer recommended for the routine evaluation of
exudative pleural effusions, fluid amylase level is extremely useful
in the rapid detection of esophageal perforation and the diagnosis
of pancreatitis-related effusions. Elevated fluid amylase levels
(greater than the upper limit of normal serum) occur in esophageal
perforation (salivary), acute pancreatitis (pancreatic), chronic pancreatitis
with fistula (pancreatic; >4000 IU/mL), and about 10% of
malignant effusions (salivary), most commonly from adenocarcinoma.
Elevated fluid amylase may rarely be seen in tuberculous pleuritis,
parapneumonic effusion, and cirrhosis, as well. Effusions secondary
to acute pancreatitis occur in approximately 50% of cases
and tend to correlate with attack severity and subsequent development
of pancreatic pseudocysts.
Fluid pH (by blood gas machine) is indicated in the assessment
of parapneumonic effusions (see Chapter 23: Empyema), and may be useful in
malignant effusion. Causes of low fluid pH (<7.20) include complicated
parapneumonic effusion/empyema, esophageal perforation,
rheumatoid pleuritis, tuberculosis, cancer, hemothorax, paragonimiasis,
urinothorax, systemic acidosis, and occasionally lupus pleuritis.
The finding of low pH in malignant effusion indicates extensive
pleural disease and correlates with a high likelihood of fluid cytological
diagnosis, but poor response to chemical pleurodesis and poor overall
prognosis (life expectancy less than 3 months from thoracentesis).
Acidic malignant effusion alone should not, however, preclude therapeutic
pleurodesis in otherwise appropriate patients.
Stains, and Other Microbiological Tests
Gram’s stain and culture (aerobic and anaerobic) should
be sent on all exudative effusions for two important reasons: First,
a Gram’s stain or culture positive for organisms is diagnostic
of empyema, which most often requires immediate interventions (see
Chapter 23: Empyema). Second, stain and culture are necessary to exclude
pleural space infection in the setting of existing (and often confounding)
pleural disease, such as for rheumatoid pleuritis in which the fluid
glucose and pH are typically low even in the absence of infection.
Other microbial stains and cultures, such as mycobacterial or fungal,
typically have low yields and should be sent only when a specific
diagnosis is suspected.
The diagnosis of tuberculous pleuritis should be pursued in all
patients with unexplained pleural fluid lymphocytosis, yet establishing
this diagnosis remains difficult. Stains for acid-fast bacilli are
positive in < 10% of cases, whereas mycobacterial culture
of fluid takes weeks and lacks sensitivity (<40%). Even
a positive skin test (purified protein derivative, PPD) may initially
be absent in up to 30% of patients with tuberculous pleuritis.
Several other tests have been evaluated for the diagnosis, including
fluid adenosine-deaminase (ADA), interferon-γ levels,
and polymerase chain reaction to detect mycobacterial DNA. These
tests remain controversial and unavailable at many centers. In the
absence of positive fluid acid-fast bacillus (AFB) or culture, and
lacking access to competent assays of these newer markers, the use
of pleural needle biopsy should be considered. Pleural biopsy is
diagnostic in 50–80% of patients, with subsequent
mycobacterial culture increasing diagnostic yield to >90%.
and Other Approaches to the Diagnosis of Cancer
Although variably successful in diagnosing cancer, pleural fluid
cytology is minimally invasive and may yield a ready diagnosis.
Cytology yield ranges from 40 to 87% and is influenced
by tumor type (adenocarcinomas highest) and extent of disease. Large
tumor burden, as indicated by low fluid pH and glucose, has cytological
yields approaching 95%. The collection of one or two additional
fluid samples for cytology several days after an initial large volume
(>300 mL) thoracentesis increases the diagnostic yield, as well.
Unfortunately, even the presence of malignant cells on cytology
may be insufficient to establish a specific diagnosis of cancer.
The use of a variety of immunohistochemical stains may be helpful
in this situation, whereas flow cytometry may detect clonal cell
populations and establish the diagnosis of lymphoma.
If no diagnosis can be established through repeated fluid cytology,
fiberoptic bronchoscopy may be considered if the patient has findings
of an airway lesion such as hemoptysis or atelectasis on imaging.
Either bronchoscopic or CT-guided biopsy may be appropriate if an
accessible lesion is present on chest radiography or CT. In the
absence of such a lesion, pleural needle biopsy (although of lower
yield than fluid cytology) may have a role, particularly if tuberculous
effusion is a consideration. However, thoracoscopic biopsy offers
the most definitive approach and should be considered both to establish
the diagnosis and possibly to perform palliative pleurodesis (see below).
Measurement of pleural fluid lipids and cholesterol is generally
reserved for cases of suspected chylothorax or pseudochylothorax,
usually on the basis of milky effusions that remain turbid even after
centrifugation. Often, the distinction between chylous and pseudochylous
effusions may be made on clinical history: pseudochylous fluid is
usually found in the setting of chronic pleural effusion with pleural
fibrosis, whereas chylous effusions are most often acute in nature.
Although the most definitive approach to diagnosing chylothorax
is the demonstration of chylomicrons in the pleural fluid by lipoprotein
analysis, a simpler and less costly initial approach employs measurement
of triglycerides (TG) and cholesterol. Chylothorax is diagnosed
when fluid TG is >110 mg/dL, fluid TG/serum TG
is >1, and fluid cholesterol/serum cholesterol is <1.
Fluid TG <50 mg/dL effectively rules out the diagnosis
of chylothorax, and in the presence of a fluid cholesterol >250
mg/dL identifies pseudochylothorax. Fluid TG of 50–110
mg/dL (or >110 in the setting of a fluid/serum
cholesterol ratio >1) warrants lipoprotein analysis for a definitive
assessment. Fluid glucose and potassium should be sent when central
venous total parenteral nutrition (TPN) is in use and leakage into
the thorax is suspected.