The pleural space is bounded by two membranes, the visceral pleura covering the lung and the parietal pleura covering the chest wall and diaphragm. Into this space, normal liquid and protein enter from the systemic circulation and are removed by the parietal pleural lymphatics. Pleural pressure is subatmospheric and ensures inflation of the lung. Because the mesothelial boundaries are leaky, excess liquid can move across into this lower-pressure, high-capacitance space and accumulate as a pleural effusion. Thus, pleural effusions are common and of highly diverse etiologies; effusions can arise from the nearby pleural membranes or from more distant thoracic or abdominal organs. Depending on the protein and lactate dehydrogenase (LDH) concentrations of the liquid, these effusions can be categorized initially as exudates or transudates. Exudative pleural effusions meet at least one of the following criteria, whereas transudative effusions meet none (Light’s criteria): pleural fluid protein–to–serum protein ratio of more than 0.5, pleural fluid LDH–serum LDH ratio of more than 0.6, and pleural fluid LDH more than two thirds of the upper normal limit for serum. In this chapter, we discuss both the physiology and pathophysiology of liquid movement in the pleural space.
Other chapters cover the anatomy of the pleural membranes (see Chapter 1 ) and the pleural diseases related to pleural infections (see Chapter 80 ) and pleural tumors (see Chapter 82 ). In addition, pneumothorax, chylothorax, hemothorax, and fibrothorax are covered in Chapter 81 . This introductory chapter covers effusions in general with attention to transudates and those exudative effusions not caused by malignancy and infection.
Pleura: Form and Function
The two pleural membranes meet at the hilar root of the lung. In the sheep, an animal with a pleural anatomy similar to humans, the surface area of the visceral pleura of one lung, including that invaginating into the lung fissures, is similar to that of the parietal pleura of one hemithorax, approximately 1000 cm 2 . The normal pleural space is approximately 18 to 20 µm in width, although it widens at its most dependent areas. It has been shown that the pleural membranes do not touch each other and that the pleural space is a real, not a potential, space (see Fig. 1-30 ).
It is likely that the primary function of the pleural membranes is to allow extensive movement of the lung relative to the chest wall. If the lung adhered directly to the chest wall, its expansion and deflation would be more limited. Encased in its slippery coat, the lung, although still coupled mechanically to the chest wall, is able to expand across a breadth of several intercostal spaces. Nonetheless, in clinical and experimental studies, pleural symphysis has not been associated with large abnormalities in lung function. The most common findings have been a decrease in volume of the affected lung and, in one study, of the opposite lung as well. If pleural thickening accompanies pleural symphysis, abnormalities of lung function may result more from fibrothorax than from obliteration of the pleural space alone.
The visceral pleura may also provide mechanical support for the lung: contributing to the shape of the lung, providing a limit to expansion, and contributing to the work of deflation. Because the submesothelial connective tissue is continuous with the connective tissue of the lung parenchyma, the visceral pleura may help to distribute the forces produced by negative inflation pressures evenly over the lung. In this way, overdistention of alveoli at the pleural surface may be avoided, lessening the chance of rupture and pneumothorax.
One recognized function of the pleural space is to provide a route by which edema can escape the lung. As has been shown in several experimental studies of either hydrostatic or increased-permeability lung edema, the pleural space can function as an additional safety factor protecting against the development of alveolar edema. The formation of transudative effusions in patients with congestive heart failure apparently reflects the movement of edema from the lung to a space where its effects on lung function are relatively small.
Embryology and Anatomy
By 3 weeks’ gestational age, the pleural, pericardial, and peritoneal spaces begin to form from the mesoderm and, by 9 weeks, the pleural cavity has become separated from both the pericardial and peritoneal spaces. Various cysts, diverticulae, and defects can result from incomplete partitioning of the three mesodermal spaces. As the lung develops, the lung buds invaginate into the visceral pleura and henceforth retain a pleural covering.
The pleural membranes are smooth, glistening coverings for the constantly moving lung. Overlying each pleural membrane is a single cell layer of mesothelial cells. These cells can vary in shape from flat to cuboidal or columnar, perhaps on the basis of the degree of stretching of the underlying submesothelial tissue. These cells, the most numerous of the pleural space, may have a variety of functions important to pleural biology. Mesothelial cells can secrete the macromolecular components of the extracellular matrix and organize them into mature matrix, phagocytose particles, produce fibrinolytic and procoagulant factors, and secrete neutrophil and monocyte chemotactic factors that may be important for inflammatory cell recruitment into the pleural spaces. The mesothelial cells also produce cytokines such as transforming growth factor-β, epidermal growth factor, and platelet-derived growth factor, cytokines that are important in pleural inflammation and fibrosis.
On their surface are microvilli, which are irregularly distributed over the pleural surfaces. Although microvilli presumably exist to increase surface area for metabolic activity, the function of these prominent features is unknown. Mesothelial cells produce hyaluronan but not mucin, express keratin microfilaments, stain negatively with epithelial-specific antibodies (Ber-EP4, B72.3, Leu.M1, and CEA), and stain positively for calretinin and mesothelin—all features that are important for histochemical and immunohistochemical identification of the cells in pleural effusions.
The cells lie on a thin basement membrane overlying a varying region of connective tissue containing mostly collagen and elastin. Although the parietal pleural connective tissue layer is of a consistent thickness, the visceral connective tissue layer varies greatly. In a single individual, the visceral pleura varies from a thinner layer at the cranial region to a thicker layer at the caudal region. Interestingly, among mammalian species, whereas the parietal pleura is constant, the visceral pleura varies greatly (see Fig. 1-31 ). Analysis of the constituents of the visceral pleura has shown that there is more collagen relative to elastin than is found in the lung parenchyma, a finding consistent with a mechanical role for the pleura. This connective tissue layer contains blood vessels and lymphatics and joins with the connective tissue of the lung. The submesothelial tissue has been shown to have mechanical strength and to contain various growth factors that can support cell growth, suggesting that mesothelium could function as a repair and regenerative material.
The parietal pleura is supplied by intercostal arteries ( Fig. 79-1A ). In humans and other large mammals, the visceral pleura is exclusively supplied by the bronchial circulation, which drains into pulmonary veins ( Fig. 79-1B ). The drainage route via pulmonary veins may have contributed to earlier confusion about whether the visceral pleural blood supply was from a systemic (bronchial) or pulmonary circulation. Both pleurae in humans therefore have a systemic circulation, although the visceral pleural bronchial circulation may have a slightly lower perfusion pressure than the parietal pleural intercostal circulation because of its drainage into a lower-pressure venous system.
If one injects carbon particles into the pleural space as a visible marker of lymphatic drainage pathways, one later finds that the black carbon has been taken up into lymphatics on the parietal side, not the visceral side ( Fig. 79–2 ) (see Fig. 1-32C ). The visceral pleura has extensive lymphatics, but they do not connect to the pleural space (see Fig. 1-27 ). Demonstrated in rabbits, sheep, and now in man, the parietal pleural lymphatics connect to the pleural space via stomas, holes of 8 to 10 µm in diameter that are formed by discontinuities in the mesothelial layer where mesothelium joins to the underlying lymphatic endothelium (see Fig. 1-32A and B ). The stomas can accommodate particles as large as erythrocytes. In various experimental studies, these lymphatics have been shown to be the major route of exit of liquid from the pleural space. From the stomas, liquid drains to lacunae, spider-like submesothelial collecting lymphatics, which then drain to infracostal lymphatics, to parasternal and periaortic nodes, to the thoracic duct, and into the systemic venous system. Lymphoid cells have been described lying within aggregates underneath morphologically different mesothelial cells, forming raised structures called Kampmeier foci that may have an immune function, as shown for the peritoneal space.
The parietal pleura contains sensory nerve fibers, supplied by the intercostal and phrenic nerves, and has long been thought to be the major site of pain sensation in the pleura. The costal and peripheral diaphragmatic regions are innervated by the intercostal nerves, and pain from these regions is referred to the adjacent chest wall. The central diaphragmatic region is innervated by the phrenic nerve, and pain from this region is referred to the ipsilateral shoulder. The visceral pleura has more recently been shown to have sensory nerve fibers that may participate in pain or other sensations such as dyspnea. In addition, pleural adhesions may contain pain fibers and contribute to post-thoracotomy or postpleurodesis pain.
Physiology of the Pleural Space
Normal Pleural Liquid and Protein Turnover
In the past 10 to 20 years, a consensus has developed that normal pleural liquid arises from the systemic pleural vessels in both pleurae, flows across the leaky pleural membranes into the pleural space, and exits the pleural space via the parietal pleural lymphatics ( Fig. 79-3 ). In this way, the pleural space is analogous to other interstitial spaces of the body. There are several lines of evidence for this view.
Intrapleural pressure is lower than the interstitial pressure of either of the pleural tissues. This pressure difference constitutes a gradient for liquid movement into but not out of the pleural space.
The pleural membranes are leaky to liquid and protein. Whether tested in vitro or in situ, the pleura offers little resistance to liquid or protein movement.
Mesothelial cells express various transporters and aquaporins, but these have not been shown to have a role in reabsorption of effusions. Although normal pleural liquid has been reported to be alkaline with a higher bicarbonate than plasma, there is no evidence for mesothelial participation in generating a bicarbonate gradient. If indeed the mesothelial layer is leaky, it is difficult to explain how mesothelial cells could maintain such a gradient. (The bicarbonate difference could also be explained by a passive distribution of ions across a semipermeable membrane, a phenomenon called the Donnan equilibrium. )
The entry of pleural liquid is slow and compatible with known interstitial flow rates. By noninvasive studies of the equilibration of radiolabeled albumin, the entry rate of pleural liquid is approximately 0.01 mL/kg per hour in a sheep, or about 0.5 mL hourly or 12 mL a day in a grown man. The half-time of pleural liquid turnover in sheep and rabbits is 6 to 8 hours.
The protein concentration of normal pleural liquid is low in sheep and probably in humans, which implies sieving of the protein across a high pressure gradient. The protein concentration of sheep pleural liquid (10 g/L, 1 g/dL) and pleural-to-plasma protein concentration ratio (0.15) are similar to those of filtrates from high-pressure systemic vessels. By comparison, a filtrate from low-pressure pulmonary vessels has a higher protein concentration (45 g/L, 4.5 g/dL) and ratio (lymph-to-plasma protein concentration ratio 0.69).
The majority of liquid exits the pleural space by bulk flow, not by diffusion or active transport. This is evident because the protein concentration of pleural effusions remains constant as the effusion is absorbed, as is expected with bulk flow. If liquid were absorbed by diffusion or active transport, proteins would exit at a slower rate and the protein concentration would progressively increase. In addition, erythrocytes instilled into the pleural space are absorbed intact and in almost the same proportion as the liquid and protein. This indicates that the major route of exit is via holes large enough to accommodate sheep erythrocytes (6 to 8 µm diameter). The only possible exit is via the parietal pleural stomas (10 to 12 µm diameter) into the pleural lymphatics. Of note, these lymphatics have a large capacity for absorption. When artificial effusions were instilled into the pleural space of awake sheep, the exit rate (0.28 mL/kg per hour) was nearly 30 times the baseline exit rate (0.01 mL/kg per hour).
The pleural pressure in humans is approximately −5 cm H 2 O at midchest at functional residual capacity and −30 cm H 2 O at total lung capacity. If the compliance of the lung decreased, pleural pressures at the same lung volumes would be more negative. In one study of patients undergoing thoracentesis, those with more negative pleural pressures had a smaller improvement in lung volume than those with less negative pressures, presumably reflecting the presence of underlying diseased, noncompliant lung.
Although the pleural space pressure is subatmospheric, gases do not accumulate there. The sum of all partial pressures of gases in capillary blood is approximately 700 mm Hg, or 60 mm Hg below atmospheric (P h 2 o = 47, P co 2 = 46, P n 2 = 570, and P o 2 = 40 mm Hg). The subatmospheric pressure of dissolved gases in capillary blood helps to maintain the pleural space free of gas and facilitates absorption of any gas that does enter the pleural space. Of note, to increase the gradient favoring absorption of gas, one can lower the partial pressure of nitrogen in the blood by having a patient breathe increased concentrations of inspired oxygen. The oxygen displaces alveolar nitrogen, thereby lowering the partial pressure of nitrogen in capillary blood; because of the limited absorption of oxygen due to the plateau of the oxygen-hemoglobin dissociation curve, however, the increase in inspired oxygen does not add greatly to the partial pressure of oxygen in capillary blood.
Pathophysiology of the Pleural Space
For pleural liquid to accumulate to form an effusion, it is likely that both the entry rate of liquid must increase and the exit rate must decrease. If only the entry rate increased, it would require a sustained rate more than 30 times normal to exceed the reserve lymphatic removal capacity; if the exit rate decreased, it would take more than a month at the normal entry rate of 12 mL per day to produce an effusion detectable by chest radiograph. Thus, in the clinical setting, it is most likely that excess pleural liquid accumulates due to changes in both entry and exit rates.
Increased entry of liquid may result from increased filtration across systemic or pulmonary capillaries or entry of another liquid (e.g., chyle, cerebrospinal fluid, urine, intravenous fluid from a displaced catheter). Decreased exit of liquid may result from interference with lymphatic function (e.g., obstruction of the parietal pleural stomas, inhibition of lymphatic contractility, infiltration of draining parasternal lymph nodes, or elevation of the systemic venous pressure into which the lymph drains). There are few studies on the rate of removal of liquid in humans; however, decreases in lymphatic clearance have been confirmed in patients with tuberculous and malignant effusions, and in those with the yellow nail syndrome, a disease of lymphatic function. In some cases, it is likely that the same disease process acts to increase entry and reduce exit of liquid whereas, in other cases, different disease processes may act cooperatively to produce an effusion.
To determine the origin of effusions, a classic and useful distinction is between transudates and exudates (see “ Separation of Exudates from Transudates ” later). Transudates form by leakage of liquid across an intact capillary barrier, owing to increases in hydrostatic pressures or decreases in osmotic pressures across that barrier. Transudates generally indicate that the pleural membranes are not themselves diseased. Exudates generally form from leakage of liquid and protein across an altered capillary barrier with increased permeability. The protein ratio, LDH ratio, and absolute pleural LDH concentration constitute Light’s criteria.
Transudates include various low protein liquids that arise from noninjured capillary beds. The majority of transudates and, in fact the majority of all effusions, are caused by congestive heart failure. These transudates have been shown to form mainly from leakage of edema across normal pulmonary capillaries into the pulmonary interstitium; this interstitial edema can then move toward the pleural space and across the leaky visceral pleura into the pleural space. Other transudates, those associated with the nephrotic syndrome or atelectasis, may form because of altered pressures (osmotic or hydrostatic) across the pleural capillaries. Some transudates, usually small, may develop primarily because of an isolated decrease in exit rate. Hepatic hydrothorax and effusions from peritoneal dialysis develop when liquid flows from the peritoneal space into the lower pressure pleural space across macroscopic holes in the diaphragm. Finally, other very low protein fluids such as urine or cerebrospinal fluid or intravenous liquids may find their way to the pleural space if their normal course is disrupted.
Exudates arise from injured capillary beds, either in the lung, pleura, or adjacent tissues. Most exudates, such as those associated with pneumonia or pulmonary embolism (PE), probably form following lung inflammation and injury when a high-protein lung edema leaks into the pleural space. Another large category of exudates arises from pleural injury due to inflammation, infection, or malignancy. Exudates can also form when exudative liquid in the mediastinum (esophageal rupture or chylothorax), retroperitoneum (pancreatic pseudocyst), or peritoneum (ascites with spontaneous bacterial peritonitis or Meig syndrome) finds its way into the lower-pressure pleural space.
As stated, for either transudates or exudates, lymphatic obstruction may contribute to the accumulation of the effusion. Nonetheless, because lymphatic clearance does not alter the pleural fluid protein concentration, the protein concentration gives information about the formation of the fluid, not its removal .
Effects of Pleural Effusions on Lung and Cardiac Function
In the presence of a space-occupying liquid in the pleural space, the lung recoils inward, the chest wall recoils outward, and the diaphragm is depressed inferiorly and is sometimes inverted. If the lung and chest wall have normal compliances, the decrease in lung volume accounts for approximately a third of the volume of the pleural effusion, and the increase in the size of the hemithorax accounts for the remaining two thirds. As a result, lung volumes are reduced by less than the pleural effusion volume. If the lung is otherwise normal, there is no evidence that an effusion causes significant hypoxemia, presumably because ventilation and perfusion decrease similarly. In fact, in one study, mild hypoxemia present before thoracentesis worsened significantly after thoracentesis, when perfusion presumably was restored while ventilation remained inadequate. In another study using multiple inert gas techniques to quantify V/Q distributions, pleural effusion was associated with a small intrapulmonary perfusion shunt (6.9%) that did not change significantly when measured again 30 minutes after thoracentesis of approximately 700 mL (6.1%). Draining pleural effusions in patients with refractory hypoxemia on mechanical ventilation may improve oxygenation, although there is no consensus on the indications for thoracentesis in this setting. It appears therefore that the effects of pleural effusion and thoracentesis on oxygenation are variable and may depend on the underlying lung function.
Large pleural effusions may impair cardiac function, most likely by decreasing the distending pressures on the cardiac chambers and cardiac filling. In a study of 27 patients with large effusions occupying more than half the hemithorax, clinical and echocardiographic findings of cardiac tamponade were identified in most patients. These findings, including elevated jugular venous pressure, pulsus paradoxus, right ventricular diastolic collapse, or flow velocity paradoxus, resolved in all patients when studied again 24 hours after thoracentesis of more than 1 L. Large pleural effusions, especially left-sided effusions, should be considered as potentially reversible causes of cardiac dysfunction. Thoracentesis of a unilateral effusion (mean value 1.5 L) has also been shown to improve exercise tolerance.
Common symptoms of patients with effusions are pleuritic chest pain, cough, and dyspnea. It appears that the three symptoms are due to different causes. Pleuritic chest pain derives from inflammation of the parietal pleura and possibly the visceral pleura. Occasionally, this symptom is accompanied by an audible or palpable pleural rub, reflecting the movement of abnormal pleural tissues ( ). Cough may be induced by distortion of the lung, in the same way as cough follows lung collapse from a pneumothorax. Dyspnea is most likely caused by the mechanical inefficiency of the respiratory muscles that are stretched by the outward displacement of the chest wall and the downward displacement of the diaphragm. After the removal of large amounts of pleural liquid, dyspnea is generally relieved promptly, although the reduction in pleural liquid volume is associated with only small increases in lung volume and little improvement, or an actual decrease, in P o 2 . In one study, nine patients underwent removal of more than 1800 mL of pleural liquid and, despite increases in vital capacity of only 300 mL, all patients experienced immediate relief of dyspnea. Although the vital capacity changed little, patients could generate a more negative pleural pressure at the same lung volume after thoracentesis than before, indicating an improved efficiency of the respiratory muscles following the return of the chest wall and diaphragm to a more normal position after thoracentesis. Another related explanation is that dyspnea is due to the inversion of the diaphragm caused by the weight of the pleural effusion, and that dyspnea is promptly relieved when thoracentesis allows the restoration of a dome-shaped diaphragm. It appears that mechanical effects of a pleural effusion account for dyspnea and for the rapid relief of dyspnea after removal of pleural liquid.
Approach to Patients with Pleural Effusion
Physical findings of a pleural effusion include dullness to percussion, decreased breath sounds, egophony at the upper level of the effusion, and decreased tactile fremitus. With large effusions, signs can include asymmetrical chest expansion or even bulging of the intercostal spaces. A systematic review concluded that the most useful physical findings were dullness to percussion and decreased tactile fremitus (see Chapter 16 ).
The possibility of a pleural effusion should be considered in any patient with an abnormal chest radiograph. Increased opacity on chest radiography is frequently attributed to a parenchymal process when it actually represents pleural fluid. Free pleural fluid gravitates to the most dependent part of the thoracic cavity, which is the posterior costophrenic sulcus when the patient is upright. Therefore, if the posterior costophrenic angle is blunted or if the posterior part of the diaphragm is not visible on the lateral chest radiograph, bilateral decubitus chest radiographs (see Figs. 18-7 and 18-9 ) or an ultrasonic examination (see eFig. 18-21 ) of the pleural space should be obtained to ascertain whether free pleural fluid is present. If the distance between the inside of the thoracic cavity and the outside of the lung is less than 10 mm, the pleural effusion is not likely to be clinically significant and, in any case, will be difficult to obtain by thoracentesis. If the distance is greater than 10 mm, an effort should be made to determine the cause of the pleural effusion.
Differential Diagnosis of Pleural Effusion
Many different diseases may have an accompanying pleural effusion ( Table 79-1 ). The vigor with which various diagnoses are pursued should depend on the likelihood that the person has that particular disease. Table 79-2 shows rough estimates of the annual incidence of the most common causes of pleural effusions. Congestive heart failure and cirrhosis are responsible for almost all transudative pleural effusions. Pneumonia, malignant pleural disease, PE, and gastrointestinal disease account for at least 90% of all exudative pleural effusions.
|TRANSUDATIVE PLEURAL EFFUSIONS|
|Congestive heart failure|
|Central venous occlusion|
|Bone marrow transplantation|
|EXUDATIVE PLEURAL EFFUSIONS|
|Neoplastic diseases (see Chapter 82 )|
|Primary effusion lymphoma|
|Infectious diseases (see Chapter 80 )|
|Pyogenic bacterial infections|
|Actinomycosis and nocardiosis|
|Collagen vascular diseases|
|Systemic lupus erythematosus|
|Eosinophilic granulomatosis with polyangiitis (Churg-Strauss)|
|Granulomatosis with polyangiitis (Wegener)|
|Post–cardiac injury syndrome|
|Post–coronary artery bypass surgery|
|Ovarian hyperstimulation syndrome|
|Yellow nail syndrome|
|Drug-induced pleural disease|
|Hemothorax (see Chapter 81 )|
|Chylothorax (see Chapter 81 )|
|Type of Effusion||Incidence|
|Congestive heart failure||500,000|
|Post–coronary artery bypass surgery||60,000|
|Cirrhosis with ascites||50,000|
|Collagen vascular disease||6000|
Separation of Exudates From Transudates
A diagnostic thoracentesis should be performed on nearly every patient with free pleural fluid that measures more than 10 mm on the decubitus radiograph, ultrasound, or chest computed tomography (CT) scan. If the patient has obvious congestive heart failure, consideration can be given to postponing the thoracentesis until the heart failure is treated. However, if the patient is febrile or has pleuritic chest pain or if the effusions are not of comparable size on both sides, thoracentesis should be performed without delay.
Thoracentesis is a safe procedure when performed by an experienced operator. Because of the small-bore needle required, it can be safely performed in patients with coagulopathies and thrombocytopenia and in patients on positive mechanical ventilation. Descriptions of technique emphasize proper positioning of the patient, identification of the area of decreased tactile fremitus that is a sensitive physical finding for the level of the effusion, and adequate local anesthesia of parietal pleura and the skin. The needle should run over the top of the rib to avoid the neurovascular bundle that travels in each intercostal space. Of note, this bundle travels in the middle of the intercostal space from the spine for approximately 5 to 6 inches (13 cm) before taking its safer position beneath the upper rib ( eFig. 79-1 ; also see eFig. 19-1E ). Thus, one should avoid thoracentesis medial to the midclavicular line.
Complications from thoracentesis include pneumothorax and hemothorax (see eFig. 19-1 ). Estimates of each complication from prospective studies are low (2% to 6% for pneumothorax; 1% for hemothorax), with only half of the pneumothoraces requiring chest thoracostomy. Pneumothorax is often associated with procedural events such as aspiration of air, multiple needle passes, and development of new symptoms. The risk for pneumothorax appears to be higher in patients with prior radiotherapy to the chest, multiple prior thoracenteses, or the use of vacuum bottles. In an uncomplicated thoracentesis that is well tolerated by the patient, there appears to be no value for routine chest radiography postprocedure. In addition, routine post-thoracentesis radiographs rarely show new findings. Thoracentesis may be safer when guided by ultrasound, although, given the safety of the procedure in uncomplicated cases and the cost or inaccessibility of ultrasound, ultrasound can be selected for higher-risk procedures. For example, it would be appropriate to use ultrasound guidance for thoracentesis on small or multiloculated effusions, in patients with poor lung function or bullous lung disease, and in patients on positive pressure ventilation (see Chapter 20 ).
The first question that should be answered with the diagnostic thoracentesis is whether the patient has a transudative or an exudative pleural effusion (see “ Pleural Effusions ” earlier). The identification of transudates or exudates is made by analysis of the levels of protein and lactate dehydrogenase (LDH) in the pleural fluid and the serum. Exudative pleural effusions meet at least one of the following criteria, whereas transudative pleural effusions meet none : (1) pleural fluid protein/serum protein greater than 0.50; (2) pleural fluid LDH/serum LDH greater than 0.60; and (3) pleural fluid LDH greater than two thirds of the upper normal limit for serum. If none of these criteria is met, the patient has a transudative pleural effusion and the pleural surfaces can be ignored while the congestive heart failure, cirrhosis, or nephrosis is treated. In the rare cases where malignancy has been associated with a transudate, extrapleural effects of the tumor or other causes such as concurrent congestive heart failure are the most likely cause as evidenced by the rarity of a positive cytology in those effusions.
These criteria may misidentify a transudative effusion as an exudative effusion in up to 25% of cases. If a patient appears to have a transudative effusion clinically, additional tests can be assessed to verify its transudative etiology. If the difference between the protein concentration of serum and pleural exceeds 3.1 g/dL or if the difference between the albumin concentration of serum and pleural exceeds 1.2 g/dL, the patient in all probability has a transudative effusion. If pleural concentrations of N-terminal probrain natriuretic peptide (NT-BNP) are elevated (>1300 pg/mL), the patient likely has a transudate from a cardiac cause.
Most studies dichotomize pleural effusions into transudates or exudates on the basis of a single cutoff point. An alternative approach recommended by Heffner and coworkers is to use likelihood ratios for identifying whether a pleural fluid is a transudate or an exudate.The idea behind this approach is that the higher a value (e.g., the pleural fluid LDH), the more likely the effusion is to be an exudate, and the lower the value, the less likely the effusion is to be an exudate. When these likelihood ratios are used in conjunction with pretest probabilities using Bayes theorem, posttest probabilities can be derived. The difficulty in using this approach is that estimates of pretest probabilities vary significantly from physician to physician. Moreover, most physicians do not understand the mathematics involved. Nonetheless, this approach does emphasize that it is important to take into consideration the absolute value of the measurements. Very high or very low measurements are almost always indicative of exudates and transudates, respectively, while values near the cutoff levels can be associated with either transudates or exudates and can be considered indeterminate .
When dealing with a patient who has a high likelihood of having a transudative pleural effusion, the most cost-effective use of the laboratory is to order only protein and LDH levels on the pleural fluid and obtain other laboratory tests only if the fluid turns out to be an exudate. In one study, of 320 pleural fluid specimens submitted for analysis, 83 were found to be transudative. For these 83 effusions, 725 additional laboratory tests had been ordered, increasing both cost and the incidence of false-positive tests (7/9). If one suspects initially that the patient has an exudate or if the fluid turns out to be an exudate, specimens can be sent for cytology, amylase, glucose, cell count and differential, and cultures.
Differentiating Exudative Pleural Effusions
Once it has been determined that the patient has an exudative pleural effusion, one should attempt to determine which of the diseases listed in Table 79-1 is responsible for the effusion, remembering that pneumonia, malignancy, and PE account for the great majority of all exudative pleural effusions. In all patients with undiagnosed exudative pleural effusions, the appearance of the fluid should be noted and the pleural fluid protein and LDH levels (if not already obtained), glucose level, differential cell count, and microbiologic and cytologic studies should be obtained. In selected patients, other tests on the pleural fluid, such as pH, amylase level, antinuclear antibody level, rheumatoid factor level, adenosine deaminase (ADA), lipid analysis, and so forth, may be of value. It is certainly not cost effective, however, to obtain all these tests routinely on patients with undiagnosed exudative pleural effusions.
Appearance of Pleural Fluid
The gross appearance of the pleural fluid should always be described and its odor noted. If the pleural fluid smells putrid, the patient has a bacterial infection (probably anaerobic) of the pleural space. If the fluid smells like urine, the patient probably has a urinothorax. If the pleural fluid is bloody, a pleural fluid hematocrit should be obtained. If the pleural fluid hematocrit is greater than 50% that of the peripheral blood, the patient has a hemothorax and the physician should strongly consider inserting chest tubes to monitor the rate of bleeding. If the pleural fluid hematocrit is less than 1%, the blood in the pleural fluid has no clinical significance. If the pleural fluid hematocrit is between 1% and 50%, the patient most likely has malignant pleural disease, a PE, or a traumatically induced pleural effusion.
If the pleural fluid is turbid, milky, or bloody, the fluid should be centrifuged and the supernatant examined. If the pleural fluid is turbid when originally obtained and the turbidity clears with centrifugation, the turbidity is due to cells or debris in the pleural fluid; if the turbidity persists after centrifugation, the patient probably has a chylothorax (see Fig. 81-8 , eFig. 81-8 , ) or a pseudochylothorax ( eFig. 79-2A ). These two entities can be differentiated by the patient’s history, examination of the sediment for cholesterol crystals (see eFig. 79-2B ), and lipid analysis of the supernatant. With chylothorax, the disease process is acute, the pleural surfaces are not thickened, there are no cholesterol crystals present, and the pleural fluid triglyceride level is usually above 110 mg/dL (1.24 mmol/L). With pseudochylothorax, the disease process is usually chronic, the pleural surfaces are usually thickened, there may be cholesterol crystals, and the pleural fluid triglyceride level is usually not elevated. (For further discussion of chylothorax and pseudochylothorax, see Chapter 81 .)
Pleural Fluid Protein
The pleural fluid protein level tends to be elevated to a comparable degree with all exudative pleural effusions and is therefore not generally useful in the differential diagnosis of an exudative pleural effusion. However, if the protein level is above 5.0 g/dL, the likelihood of the diagnosis of tuberculous pleurisy is increased. If the pleural fluid protein level is very low (<0.5 g/dL), the patient probably has a urinothorax, an effusion secondary to peritoneal dialysis, a leak of cerebrospinal fluid into the pleural space, or an effusion secondary to the misplacement of a central intravascular line.
Pleural Fluid Lactate Dehydrogenase
Whereas pleural liquid protein and LDH arise from filtration from serum and thus serve as indicators of vascular permeability, LDH, as an intracellular enzyme, also may indicate the degree of cell turnover within the pleural space. Nonetheless, the pleural fluid LDH level is increased to a comparable degree in patients with all categories of exudative pleural effusions and therefore is of no utility in the differential diagnosis of exudative pleural effusion. Likewise, the pleural fluid LDH isoenzymes are of limited use in the differential diagnosis of exudative pleural effusions. However, pleural fluid LDH concentration should be measured every time a diagnostic thoracentesis is performed, because the level of LDH in the pleural fluid reflects the degree of inflammation in the pleural space. If the pleural fluid LDH concentration increases with serial thoracentesis, the degree of inflammation in the pleural space is worsening and the physician should be more aggressive in pursuing the diagnosis. Alternatively, if the pleural fluid LDH level decreases with serial thoracentesis, the pleural disease is resolving and observation of the patient is indicated. When an effusion meets exudative criteria on the basis of LDH but not protein, the effusion is usually malignant or parapneumonic.
Pleural Fluid Glucose
A low glucose concentration probably indicates the coexistence of two abnormalities: a thickened, infiltrated pleura leading to an impaired diffusion of glucose into the pleural space plus increased metabolic activity leading to increased glucose utilization within the pleural space. The glucose level should be measured in all undiagnosed exudative pleural effusions because the demonstration of a reduced pleural fluid glucose level (<60 mg/dL, 3.3 mmol/L) narrows the diagnostic possibilities to seven: parapneumonic effusion, malignant effusion, tuberculous effusion, rheumatoid effusion, hemothorax, paragonimiasis, or eosinophilic granulomatosis with polyangiitis (EGPA, Churg-Strauss). If a patient with a parapneumonic effusion has a pleural fluid glucose level below 40 mg/dL (2.2 mmol/L), tube thoracostomy should be considered. Many patients with rheumatoid pleural effusions have a pleural fluid glucose level below 30 mg/dL (1.7 mmol/L). In contrast, most patients with pleural effusion secondary to systemic lupus erythematosus (SLE) will have a pleural fluid glucose level above 80 mg/dL (4.4 mmol/L). Patients with malignant pleural disease and a low pleural fluid glucose level usually have a positive pleural fluid cytology. In addition, their prognosis is poor, with a mean survival of less than 2 months.
Pleural Fluid White Cell Count and Differential
Pleural liquid that is submitted for white cell count and differential should be sent in a tube with an anticoagulant to prevent the cells from clumping. In the normal pleural space, the cell count has been reported to be 1700 cells/µL. In effusions, the cell count has limited diagnostic value. A pleural fluid white blood cell count of 1000/µL roughly separates transudative from exudative pleural effusion, whereas a pleural fluid white blood cell count above 10,000/µL is most common with empyemas and parapneumonic effusions but is also seen with pancreatitis, PE, and collagen vascular diseases and, occasionally, with malignancy and tuberculosis.
The differential cell count on the pleural fluid is much more useful than the white cell count itself. The normal pleural space contains predominantly macrophages (75%) followed by lymphocytes (23%). A change in differential from this normal distribution provides a clue to the underlying disease process. For the pleural fluid differential cell count, the cells should be partitioned into the following categories: polymorphonuclear leukocytes, eosinophils, small lymphocytes, mesothelial cells, and other mononuclear cells. Pleural effusions due to an acute disease process such as pneumonia, PE, pancreatitis, intra-abdominal abscess, or early tuberculosis contain predominantly polymorphonuclear leukocytes. Pleural effusions due to a chronic disease process contain predominantly mononuclear cells.
Pleural fluid eosinophilia (10% or more eosinophils by differential count) is most commonly due to air or blood in the pleural space. Interleukin-5 (IL-5) appears to be an important factor because the number and percentage of eosinophils in the pleural space are closely correlated with the pleural liquid IL-5 levels. Occasionally, no pleural fluid eosinophils are found in the initial thoracentesis, but many eosinophils are seen in a subsequent thoracentesis most likely due to entry of air or blood caused by the initial thoracentesis. With traumatic hemothorax, pleural fluid eosinophilia does not appear until the second week. The eosinophilia appears to be due to production of IL-5 by CD4 + T cells within the pleural space and has been associated with a type 2 innate immune response following a pneumothorax. At times, the pleural fluid eosinophilia associated with a hemothorax can lead to eosinophilia in the peripheral blood. The bloody pleural effusion complicating PE frequently contains many eosinophils. With pneumothorax, pleural eosinophilia appears within 3 days of the pneumothorax and reaches a peak after 6 days.
After excluding air or blood as the underlying cause, the etiologies of 392 cases of eosinophilic pleural effusions have been reported as follows: idiopathic 40%, malignancy 17%, parapneumonic 13%, tuberculosis 6%, PE 4%, transudates 8%, and other 13%. If the etiology of the eosinophilia is not evident, several unusual diagnoses should be considered. Benign asbestos pleural effusions are frequently eosinophilic. In one series, 15 of 29 patients (52%) with benign asbestos pleural effusions had pleural fluid eosinophilia. Patients with pleural effusions secondary to drug reactions (nitrofurantoin or dantrolene) typically have pleural fluid eosinophilia. The pleural fluid of patients with pleural paragonimiasis is typically eosinophilic with low glucose, low pH, and high LDH level. EGPA is the only other disease that produces this constellation of pleural fluid findings.
Mesothelial cells line the pleural cavities. It is unusual to find mesothelial cells in effusions due to tuberculosis. However, the absence of mesothelial cells is also common with other conditions in which the pleura becomes coated with fibrin, such as a complicated parapneumonic effusion.
Lymphocytic pleural effusions by definition contain more than 50% small lymphocytes. Most lymphocytic pleural effusions are due to malignancy or tuberculosis. Ninety of 96 exudative pleural effusions (94%) in two series with more than 50% small lymphocytes were due to tuberculosis or malignant disease. Because these two diseases can be diagnosed with needle biopsy of the pleura, the presence of pleural fluid lymphocytosis should alert the physician to consider needle biopsy of the pleura for diagnosis. In general, separation of pleural lymphocytes into T and B lymphocytes has not been useful diagnostically because most lymphocytic effusions contain a predominance of T cells (CD4 + ) whether the diagnosis is malignancy or tuberculosis. Such partitioning can be useful diagnostically, however, when a diagnosis of chronic lymphocytic leukemia or lymphoma is suspected. With these diseases the pleural lymphocytes are usually of B-cell origin.
Pleural Fluid Cytology
A pleural fluid specimen from every patient with an undiagnosed exudative pleural effusion should be sent for cytopathologic studies. The first pleural fluid cytologic study is positive for malignant cells in up to 60% of the effusions caused by pleural malignancy. If three separate specimens are submitted, up to 90% with pleural malignancy have positive cytopathology. The percentage of cases in which cytologic study of the pleural fluid establishes the diagnosis of a malignant pleural effusion ranges from 40% to 87%. The frequency of positive pleural fluid cytologic tests is dependent on the tumor type. For example, less than 25% of patients with Hodgkin disease have positive cytology, whereas most patients with adenocarcinomas have positive cytology. The percentage of positive diagnoses is higher if both cell blocks and smears are prepared by standard protocols and examined by an experienced cytologist. Each additional sample may increase diagnostic yield in part by providing a higher percentage of fresher cells as older degenerated cells are largely removed by the earlier thoracenteses. During thoracoscopy, pleural lavage has been found to increase the diagnostic yield, perhaps by harvesting more fresh cells for analysis. The percentage of positive diagnoses is obviously dependent on the skill and experience of the cytologist. Immunohistochemical stains of malignant cells are used to confirm a diagnosis and to specify tumor type, with many new markers available and being used in panels for optimal diagnostic efficacy. Cytology can provide enough DNA to assess mutational analysis for epidermal growth factor receptor (EGFR) activating mutations when extremely sensitive techniques including next-generation sequencing are used.
Other Diagnostic Tests for Malignancy
Cytology may be nondiagnostic either because of a problem of specificity (e.g., the malignant cells cannot be differentiated from reactive mesothelial cells and “atypical” benign cells) or because of a problem of sensitivity (e.g., the malignant cells are rare). Several assays are being evaluated for their ability to increase the specificity of cytology for diagnosis of malignancy. Fluorescent in situ hybridization (FISH) with chromosome-specific probes can confirm abnormal numbers of specific chromosomes (aneuploidy), thereby confirming that abnormal cells are indeed malignant. Early findings of malignancy including DNA methylation can be detected by methylation-specific polymerase chain reaction (PCR) and gene expression patterns can help distinguish mesothelioma and adenocarcinoma. EGFR mutations can predict response to EGFR tyrosine kinase inhibitors. On the other hand, biomarkers have generally been disappointing due to nonspecificity. However, pleural fluid mesothelin measurements hold promise for the diagnosis of mesothelioma; pleural mesothelin is more accurate than serum for mesothelioma, and a low value can be useful in excluding the diagnosis. However, routine use of mesothelin is not useful (see Chapter 82 ).
Culture and Bacteriologic Stains
Pleural fluid from patients with undiagnosed exudative pleural effusions should be cultured for bacteria (both aerobically and anaerobically), mycobacteria, and fungi. Gram stain should also be obtained. In the case of a probable complicated parapneumonic effusion with an initial negative Gram stain, the sediment of the pleural fluid should be stained because the bacteria will be precipitated in the sediment along with the white blood cells and debris. The yield of bacterial culture can be increased by using blood culture bottle cultures in addition to standard cultures; in 62 patients with pleural infection, the added inoculation of a set of aerobic and anaerobic blood culture bottles increased the identification of pathogens from 38% to 59%.
One potentially useful adjunct is molecular detection of bacterial antigens or DNA. This may be especially useful in children for whom the yield from culture is often poor, presumably due to antibiotics given before the thoracentesis. Such molecular tests can be rapid and more accurate. Antigen-based assays have shown promise in diagnosing empyema caused by Streptococcus pneumoniae and Streptococcus pyogenes . Amplification and sequencing of bacterial 16S ribosomal RNA have identified bacteria in pleural empyema, showing in one study of adults that the bacteriology of pleural infections differed from that of pneumonia. It is likely that use of species-specific PCR will become a more standard molecular test for a panel of organisms; it improves detection when compared with 16S PCR and has the potential to make specific diagnoses, although false-positive results are a potential limitation.
Other Diagnostic Tests for Pleural Fluid
Pleural Fluid pH and PCO 2
The pleural fluid pH can be reduced to less than 7.20 with 10 different conditions: (1) complicated parapneumonic effusion, (2) esophageal rupture, (3) rheumatoid pleuritis, (4) tuberculous pleuritis, (5) malignant pleural disease, (6) hemothorax, (7) systemic acidosis, (8) paragonimiasis, (9) lupus pleuritis, or (10) urinothorax. The decreased pleural fluid pH appears to result from lactic acid and carbon dioxide accumulation in the pleural fluid. The pleural fluid pH is most useful in determining whether chest tubes should be inserted in patients with parapneumonic effusions. A fall in the pleural fluid pH appears to be a sensitive indicator that the patient has a highly inflammatory parapneumonic pleural effusion that will require drainage.
The routine measurement of pleural fluid pH is recommended only in patients with parapneumonic effusions. In general, pleural fluids with a low pH also have a low glucose ; thus, pleural fluid glucose can be used as an alternative to the pH measurement. When the pleural fluid pH is used as a diagnostic test, it must be measured with the same care as arterial pH. The fluid should be collected anaerobically in a heparinized syringe and placed on ice. If the sample is left open to air, a spuriously high pH value can be obtained because of the rapid loss of carbon dioxide. The pH must be measured with a blood gas machine; a pH meter or indicator paper is not sufficiently accurate. Pleural fluid pH, but not glucose, may be significantly altered by residual air or lidocaine in the syringe. In a 2012 study, 40% of pulmonologists did not appreciate that blood gas analysis was the only accurate method to measure pleural fluid pH and nearly 40% of pulmonologists incorrectly believed their laboratory was using blood gas analysis. The pleural fluid glucose may be a preferable test when the accuracy of pH measurements cannot be assured.
Pleural Fluid Amylase
The pleural fluid amylase is elevated in patients with pleural effusions secondary to esophageal perforation, pancreatic disease, or malignant disease. However, because such a small percentage of effusions are due to esophageal perforation or pancreatic disease, the routine measurement of pleural fluid amylase is not indicated. In the case of esophageal rupture, the origin of the amylase is the salivary glands. In animal models of esophageal rupture, the pleural fluid amylase concentration is elevated within 2 hours of esophageal rupture. In effusions due to pancreaticopleural fistulas, the amylase concentration is extremely high (>4000 IU/mL), reflective of the concentrations in pancreatic secretions. In approximately 10% of malignant effusions, the pleural fluid amylase level is mildly elevated. The site of the primary tumor in such patients is usually not the pancreas. Malignancy can be differentiated from pancreatic disease with amylase isoenzymes because the amylase with malignant effusions is primarily of the salivary type. Because lipase originates only from the pancreas, finding lipase in the pleural effusion should help identify the source as the pancreas.
Tests for Collagen Vascular Diseases
About 5% of patients with rheumatoid arthritis and 50% of patients with SLE have a pleural effusion sometime during the course of their disease. At times, the effusions may be the first manifestation of the disease; therefore, these diagnostic possibilities should be considered in patients with undiagnosed exudative pleural effusion.
Measurement of the antinuclear antibody (ANA) titer is the best screening test for lupus pleuritis, although it is now evident that a positive pleural fluid ANA is not specific for the diagnosis. Although all patients with lupus pleuritis have a positive pleural liquid ANA (>1 : 40), the finding of a positive ANA has been found in between 11% and 27% of all other effusions. Neither the titer of ANA, ratio of pleural-to-plasma ANA, or the pattern of staining has been found to increase the specificity for SLE. In fact, a positive pleural fluid ANA in patients without SLE may be associated with malignancy. In patients with SLE, the lack of ANA in pleural liquid may have a high negative predictive value; in patients with SLE and a pleural effusion of uncertain etiology, a lack of ANA (dsDNA, extractable nuclear antigens [ENA]) in the pleural fluid argues against the diagnosis of lupus pleuritis.
When a rheumatoid pleural effusion is suspected, the clinical picture usually establishes the diagnosis. If any question exists, the level of rheumatoid factor in the pleural fluid should be measured. Only patients with rheumatoid pleuritis have a pleural fluid rheumatoid factor titer equal to or greater than 1 : 320 and equal to or greater than the serum titer.
ADA, a product of activated lymphocytes, catalyzes the conversion of adenosine to inosine and is important for normal immune function. The pleural fluid ADA levels are elevated in almost all patients with tuberculous pleuritis but not with other conditions even when associated with lymphocytic effusions. Despite earlier concerns about false-negative values in HIV-positive patients, ADA remains a sensitive marker for tuberculous pleurisy in patients with HIV. ADA levels can be elevated in other conditions and also in neutrophilic effusions; in one study, ADA above the cutoff (35 U/L) was seen in up to 40% of parapneumonic effusions and in half of effusions due to lymphoma. Because it is a highly sensitive test, the ADA can be a useful test to exclude the diagnosis of tuberculosis when the ADA level is low (<40 U/L).
Interferon gamma, a T-cell lymphokine, may play a critical role in the effective clinical response to Mycobacterium tuberculosis. Pleural liquid interferon gamma is elevated almost exclusively in tuberculous effusions. Interferon gamma appears to be as useful as ADA; because ADA is less expensive, ADA is generally preferred. Compared with ADA and interferon gamma, the interferon gamma release assays of pleural fluid, however, have not been shown to be of use in the diagnosis or exclusion of active pleural tuberculosis; these tests are not currently recommended for pleural fluid.
Molecular Techniques for Diagnosis of Mycobacteria tuberculosis
Four molecular techniques are now available: PCR to detect specific mycobacterial DNA sequences in clinical specimens (pleural liquid or biopsy), nucleic acid probes to identify the organism in culture, restriction fragment length polymorphism to compare strains in epidemiologic studies, and gene-based susceptibility studies to screen for known genes associated with drug resistance.
In theory, PCR tests have great potential for providing a rapid, highly sensitive and specific diagnosis of mycobacterial infection. In practice, PCR amplification of clinical samples has been limited by a low sensitivity, which may be due to degradation of the target DNA by sample processing or by inhibitors of amplification in clinical fluids. PCR amplification assays show a high specificity and thus, when positive, can help in making the diagnosis; however, because of their low sensitivity, they are not useful in excluding the disease. Before wide application of PCR assays, considerations will also include its cost and current inability to identify antibiotic resistance of the organism. In the future, molecular techniques for diagnosis of tuberculosis and for other slow-growing microorganisms will likely prove increasingly important.
Useful Radiographic Tests in Patients with Suspected Pleural Disease
The possibility of a pleural effusion should be considered whenever a patient with an abnormal chest radiograph is evaluated. Two main factors influence the distribution of free fluid in the pleural space. First, the fluid collects in the most dependent part of the thoracic cavity because the lung is less dense than the pleural fluid. Second, because of their elastic recoil, the lobes of the lung generally maintain their traditional shape at all stages of collapse.
When the patient is upright, the fluid first accumulates between the inferior surface of the lower lobe and the diaphragm. If there is less than 75 mL of fluid, it may occupy only this position without overflowing into the costophrenic sinuses. When more fluid accumulates, it spills over into the posterior costophrenic angle and obliterates the posterior part of the diaphragm on the lateral projection. The possibility of a pleural effusion should be suspected whenever the posterior part of one or both diaphragms is obscured (see Fig. 18-9 ). The presence of a clinically significant amount of free pleural fluid can be excluded if both posterior costophrenic angles are clear. Pleural effusions on radiography can be missed in the setting of lower lobe consolidation and, in the setting of pneumonia, consideration should be given to seeking the presence of effusions with additional imaging.
When there are larger amounts of pleural fluid, the lateral costophrenic angle on the posteroanterior radiograph becomes blunted. Collins and associates demonstrated that at least 175 mL of pleural fluid had to be injected into the pleural space of cadavers before the lateral costophrenic angle was blunted. In some of their cases, more than 500 mL of pleural fluid could be present without blunting the lateral costophrenic angle. As more fluid accumulates, the entire outline of the diaphragm on the affected side is lost, and the fluid extends upward around the anterior, lateral, and posterior thoracic walls, producing opacification of the lung base and the typical meniscus shape of the fluid.
The changes just discussed are suggestive rather than diagnostic of the presence of pleural fluid. Lateral decubitus radiographs (see Figs. 18-7 and 18-9 ) or ultrasonic examination should be obtained in most instances when free pleural fluid is suspected. If the entire hemithorax is opacified, decubitus radiographs are of no use, because there is no air-containing lung in the hemithorax. The basis for the use of the lateral decubitus view is that free fluid gravitates to the most dependent part of the pleural space. When a patient is placed in the lateral recumbent position, the free pleural fluid on the dependent side accumulates between the chest wall and the lung ( Fig. 79-4 ). As little as 5 mL of pleural fluid can be demonstrated with properly exposed decubitus radiographs. The amount of pleural fluid can be semiquantitated by measuring the distance between the inner border of the chest wall and the outer border of the lung (see Fig. 79-4 ). As already stated, when this distance is less than 10 mm, the amount of pleural fluid is small and a diagnostic thoracentesis is usually not attempted.
Pleural fluid may become encapsulated by adhesions anywhere between the parietal and the visceral pleurae or in the interlobar fissures. Pleural fluid loculates most frequently in association with conditions that cause intense pleural inflammation, such as with a complicated parapneumonic effusion or tuberculous pleuritis. When the loculation is situated between the lung and the chest wall, there is a characteristic radiographic picture. The loculation is “D” shaped with the base of the D against the chest wall and the smooth convexity protruding inward toward the lung (see Fig. 18-39 ). The absence of air bronchograms helps differentiate a loculated pleural effusion from a parenchymal process. A definite diagnosis of loculated pleural effusion is best established by ultrasonography or CT.
One way to document and locate loculated pleural fluid is with ultrasound. In the presence of pleural fluid, the proximal echoes from the skin, intercostal muscles, and parietal pleura are separated from the distal echoes arising from the visceral pleura and the lung by a central echo-free space. The advantages of ultrasound over CT are the ease and speed with which the examination can be performed, the lack of ionizing radiation, the relatively low cost, and the ability to provide continuous guidance for thoracentesis or pleural biopsy.
The appropriate site for a thoracentesis can be identified using ultrasound. If a patient has a moderate or large effusion, the thoracentesis may be performed without ultrasound but large studies have shown a lower incidence of pneumothorax when ultrasound is used. Ultrasound should definitely be used if no fluid is obtained on an initial attempt or if the effusion is small. When ultrasound is used to identify the site for thoracentesis, it is important to perform the thoracentesis at the time of the ultrasonic examination. If the skin is marked and the patient returned to his or her room, thoracentesis may be unsuccessful because the relationship between the skin and the pleural fluid may have changed. In addition, when the thoracentesis is performed at the time of the ultrasonic examination, there is immediate feedback that is valuable in improving the skill of the ultrasonographer (see Chapter 20 ).
Chest CT is currently the best way to visualize the pleural space. Chest CT has its greatest utility in distinguishing parenchymal and pleural abnormalities. With current protocols involving the rapid injection of intravenous contrast medium, the unaerated perfused lung parenchyma will enhance, whereas the pleural fluid will not. The use of CT to discriminate transudates from exudates by their attenuation (Hounsfield units) has not been found to be clinically useful due to a great deal of overlap.
Chest CT is quite useful in distinguishing a parenchymal lung abscess located near the chest wall from an empyema with an air-fluid level. The most distinctive features are the margins of the abnormality. With empyema, the cavity walls are of uniform thickness both internally and externally (see eFigs. 80-2 , 80-5 ) and the adjacent lung is usually compressed. The angle of contact with the chest wall may be obtuse (see eFig. 33-7B ). In addition, most empyemas have a lenticular shape and demonstrate the “split pleura” sign ( Fig. 79-5A see also eFigs. 80-2 , 80-5 ). With lung abscess, the walls of the cavity are not of uniform thickness and the adjacent lung is not compressed (see eFigs. 33-4 , eFigs. 33-13 , eFigs. 33-21 ). The angle of contact with the chest wall may be acute ( Fig. 79-5B ).
In diffuse pleural disease, chest CT is useful in distinguishing malignant from benign causes. Features associated with malignancy include circumferential pleural thickening, pleural nodules (see eFig. 53-5 ), parietal thickening greater than 1 cm, and mediastinal pleural involvement (see eFig. 53-4 and ). The distinction of metastatic disease from mesothelioma can be difficult, although hilar adenopathy is more common with metastatic disease.
CT pulmonary angiography (CTPA) has become a first-line imaging test for the evaluation of PE (see Chapters 18 and Chapter 57 ). The evaluation of a patient with a pleural effusion for PE can begin with a Doppler ultrasound of the lower extremities. If the ultrasound identifies thrombus, the patient can then be treated for thromboembolic disease. If it is negative, the patient may still have a PE. In the past the standard approach was to proceed to lung ventilation-perfusion scanning. However, this has largely been replaced by CTPA. CTPA is highly sensitive and specific for pulmonary emboli in the proximal, segmental pulmonary arteries. In contrast to ventilation-perfusion lung scanning, CTPA can establish alternate diagnoses as well. Where CTPA is not available, lung scanning can be used, perhaps after efforts to improve accuracy by withdrawing as much pleural liquid as possible. A diagnostic scan (normal or high probability) can then be used either to exclude the diagnosis or initiate anticoagulation. If nondiagnostic (e.g., of low or intermediate probability), the scan should be followed by pulmonary angiography.
Magnetic Resonance Imaging
At present, magnetic resonance imaging (MRI) of the chest is less satisfactory than ultrasound or CT in identifying the presence of pleural fluid. It is possible that with improved MR technology, the characteristics of pleural fluid can be determined noninvasively. Respiratory and cardiac motion is the major limitation in evaluating specific intensity patterns of fluid collections of various compositions. MRI may have current value in delineating malignancy in the pleural space and may be better than chest CT for determining chest wall or diaphragmatic invasion.
Positron Emission Tomography and PET/CT
Positron emission tomography (PET) visualizes tissues that are metabolically active by their concentration of the radioisotope 18 F-fluorodeoxyglucose. Because most malignant cells have a higher metabolic rate than nonmalignant cells, PET can help differentiate malignant from benign lesions, stage patients with malignancy, and identify recurrence. Introduced in 1998, PET/CT integrates the metabolic information of PET with the detailed anatomic information of CT. Compared with CT or PET alone, PET/CT can allow detection of additional lesions, localize them more accurately, characterize them as highly likely to be malignant, and discriminate malignant from surrounding normal tissue (see Chapter 82 ).
Invasive Tests in Patients with Undiagnosed Exudative Pleural Effusions
In the patient with an undiagnosed exudative pleural effusion, several invasive tests might be considered, including blind or image-guided needle biopsy of the pleura, bronchoscopy, thoracoscopy or video-assisted thoracic surgery, and open biopsy of the pleura. It is important to remember that no diagnosis is ever established for approximately 20% of all exudative pleural effusions and that many resolve spontaneously, leaving no residua. In patients with undiagnosed exudative pleural effusions, three factors should influence the vigor with which one pursues the diagnosis with invasive tests. First, the symptoms and clinical course of the patient. If symptoms are minimal or improving with time, a less aggressive approach is indicated. Second, the trend of the pleural fluid LDH level with time. If the pleural fluid LDH increases with serial thoracenteses, a more aggressive approach is indicated. Third, the attitude of the patient. If the patient is anxious about the cause of the pleural effusion, an aggressive approach should be taken. Furthermore, when an undiagnosed exudate resolves without treatment, two diagnoses should still be considered and excluded: PE and tuberculosis.
Needle Biopsy of the Pleura
Small specimens of the parietal pleura can be obtained with needle biopsy, often called blind or closed needle biopsy. The needles most commonly used for this procedure are the Cope needle and the Abrams needle. Because needle biopsy of the pleura is useful mainly to establish the diagnosis of malignant or tuberculous pleural effusions, this procedure should be considered when one of these diagnoses is suspected.
In malignant pleural disease, the needle biopsy of the pleura will be positive in 40% to 60% of patients. Overall, the yield from pleural fluid cytology tends to be higher, probably because it samples cells dislodged from the entire pleura, while the needle biopsy can only sample from a localized area. In one series of 281 patients with malignant pleural effusions, the pleural biopsy was positive in 43%, whereas the pleural fluid cytology was positive in 58%. In 7% the pleural biopsy was positive and the pleural fluid cytology negative. In a more recent study of 66 patients, although cytology was more likely to be positive than biopsy (69% vs. 48%), closed pleural biopsy added diagnoses in some cytology-negative patients. A prudent approach to the patient with a suspected malignant pleural effusion is to obtain a pleural biopsy only if the cytology obtained at the time of the initial diagnostic thoracentesis is nondiagnostic. If the CT shows pleural thickening or nodularity, an image-guided pleural biopsy is an excellent option.
Needle biopsy of the pleura has greater utility for the diagnosis of tuberculous pleuritis than of malignancy. The initial biopsy is positive for granulomas in 50% to 80% of patients. The demonstration of granulomas on the pleural biopsy is virtually diagnostic of tuberculous pleuritis; caseous necrosis or acid-fast bacilli need not be demonstrated, although on rare occasions, fungal diseases, sarcoidosis, or rheumatoid pleuritis can produce granulomatous pleuritis. When tuberculous pleuritis is suspected, a portion of the pleural biopsy specimen should be cultured for mycobacteria. In one series of 21 patients with tuberculous pleuritis, either the microscopic examination or the biopsy culture was positive in 20 of the 21 patients (95%) ; in a more recent series of 113 patients with tuberculous pleural effusion, the sensitivity of closed pleural biopsy (by showing granulomas on pathology or culture) was 92%. If the initial biopsy is nondiagnostic and the patient has tuberculous pleuritis, a second biopsy will be diagnostic 10% to 40% of the time.
The greatest value of needle biopsy for a patient with tuberculosis is in obtaining material for culture of M. tuberculosis for the determination of drug susceptibility. Often, the presentation of a patient with a recent purified protein derivative conversion and an exudative pleural effusion with lymphocytosis is classic and unlikely to be due to any diagnosis other than tuberculous pleurisy. The diagnosis can be further supported by measurements, where available, of pleural fluid ADA or interferon gamma. In those cases, treatment for tuberculosis can be offered with confidence without needle biopsy confirmation. However, when the patient may have been exposed to drug-resistant organisms, needle biopsy will increase the likelihood of obtaining organisms for culture and is recommended. Sputum induction can also be useful for culturing the organism, even when only pleural involvement is suspected and the chest radiograph shows no pulmonary involvement; in this setting, sputum induction has been shown to have a yield similar to culture of the pleural biopsy (52% vs. 62%). Medical thoracoscopy has a higher yield than blind needle biopsy for obtaining pleural material for culture. When different approaches were compared in the same 51 patients, thoracoscopy obtained a positive culture of M. tuberculosis in 76%, whereas the Abrams needle obtained a culture in 48%. If the need for obtaining positive cultures is high, the more invasive diagnostic tests including medical thoracoscopy may be preferred.
The two major complications of needle biopsy of the pleura are pneumothorax and bleeding. Pneumothoraces require a chest tube in only about 1% of pleural biopsies. It is likely that many pneumothoraces develop because of leakage of air through the biopsy needle and do not necessarily indicate puncture of the lung. A hemothorax can result from inadvertent biopsy of an intercostal artery or vein. In one older series, a fatal hemothorax was a complication in 2 of 227 biopsy procedures. The pleural biopsy needle can also be mistakenly inserted into the liver, spleen, or kidney, which can lead to hemorrhage in these organs. However, in general, bleeding complications are rare.
If no diagnosis is obtained after routine laboratory tests including cytology and one needle biopsy of the pleura, what can be said concerning the patient? Poe and coworkers followed 143 such patients for 12 to 72 months, during which time 29 patients were diagnosed with malignant pleural disease and one patient with tuberculous pleuritis. In all 29 cases in which malignancy was eventually diagnosed, the diagnosis of malignant neoplasm was suggested by clinical criteria such as weight loss, constitutional symptoms, or a history of previous cancer. These authors concluded that most patients with undiagnosed exudative pleural effusions in whom the clinical picture does not suggest malignancy are best managed by observation. In those with symptoms suggestive of malignancy, image-guided pleural biopsy or thoracoscopy are probably the procedures of choice.
Image-Guided Pleural Biopsy
In patients with pleural abnormalities consistent with malignancy, CT-guided cutting-needle biopsies may supplant closed pleural biopsy for diagnosis, where image-guided technology is available. In a randomized study of 50 patients with suspected malignant pleural effusions with negative cytology, those randomized to image-guided biopsy were more likely to be diagnosed (87% sensitivity) than those randomized to blind pleural biopsy (47%). In situations where the effusion is small and there is loculation or pleural thickening without an effusion, CT image-guided pleural biopsy is an excellent choice. PET/CT may increase the yield further by improving interpretation and selection of biopsy targets (see earlier).
Another procedure that should be considered in the patient with an undiagnosed pleural effusion is bronchoscopy (see Chapter 22 ). If the patient has an associated parenchymal lesion or hemoptysis, fiberoptic bronchoscopy will provide a diagnosis in nearly 75%. On the other hand, if the patient has neither a parenchymal abnormality nor hemoptysis, a diagnosis for the pleural effusion is established less than 10% of the time. At present, chest CT should be performed in all patients with undiagnosed exudative pleural effusions. Bronchoscopy should then be performed only if the CT scan demonstrates parenchymal or obstructing airway abnormalities or if the patient has hemoptysis. At the time of bronchoscopy, special attention is paid to those portions of the lung in which the parenchymal abnormalities were demonstrated.
Thoracoscopy or Video-Assisted Thoracic Surgery
Thoracoscopy is discussed fully in Chapter 24 . Thoracoscopy may be useful diagnostically in patients in whom the origin of a pleural effusion remains unclear after routine fluid analysis and needle biopsy of the pleura. In many cases, especially for the evaluation of malignancy, thoracoscopy may supplant needle biopsy because of a greater diagnostic yield and the added ability to provide a pleurodesis. Thoracoscopy can be performed by pulmonologists using local anesthesia and conscious sedation for direct visualization of the pleural surfaces, tissue sampling, and pleurodesis; thoracoscopy performed by thoracic surgeons, generally referred to as video-assisted thoracic surgery (VATS), utilizes general anesthesia and single-lung ventilation by double-lumen intubation and allows greater access to the pleura and lung for surgical procedures.
Which patients with undiagnosed pleural effusions should undergo thoracoscopy? A diagnosis can be established in more than 90% of patients with malignancy including those with mesothelioma. Moreover, talc can be insufflated at the time of the procedure and this will control the effusion in the majority of patients. Nonetheless, there are minor risks of the procedure, a need for postprocedure chest tube(s), and a procedure cost that should be considered. Thoracoscopy is therefore recommended for the patient with an undiagnosed pleural effusion after diagnostic thoracentesis and needle biopsy of the pleura in whom the diagnosis of malignancy is strongly suspected and in whom one wishes to establish this diagnosis.
Open Biopsy of the Pleura
Thoracotomy with direct biopsy of the pleura provides the best visualization of the pleura and the best biopsy specimens. Nowadays, the less invasive thoracoscopy can replace thoracotomy in most instances. The main indication for open pleural biopsy is progressive undiagnosed pleural disease that cannot be approached by or has failed to be diagnosed by thoracoscopy. In the past, for example, the diagnosis of malignant mesothelioma was usually made with open biopsy of the pleura, but now the diagnosis can be established in the majority of the cases with thoracoscopy.
It should be emphasized that open pleural biopsy does not always provide a diagnosis in patients with pleural effusions. Over the 11-year period from 1962 to 1972, 51 patients with pleural effusion at the Mayo Clinic had no diagnosis after an open pleural biopsy. In 31 of these patients (61%), there was no recurrence of the pleural effusion, and no cause ever became apparent. However, 13 of the patients were eventually proved to have malignancy; 6 had lymphoma and 4 had malignant mesothelioma. Thus, observation of patients with undiagnosed pleural effusions is often warranted unless there is compelling reason to pursue the diagnosis of malignancy. Since the time of this study, there have been many improvements in diagnostic tests that may increase the diagnostic yield of invasive procedures and the treatment options available.
Transudative Pleural Effusions
Transudative pleural effusions frequently accompany many common clinical disorders. It is noteworthy that the primary abnormality in most cases of transudative pleural effusions originates in organs other than the pleura or lungs, especially the heart, liver, and kidneys. This association emphasizes the fact that although patients may visit their physicians for respiratory complaints, these symptoms may be caused by extrapulmonary disorders.
Congestive Heart Failure
Congestive heart failure (CHF) is probably the most common cause of pleural effusion. The incidence of pleural effusion in patients with CHF is high. In one series of 60 patients with an exacerbation of stable CHF, chest CT scans demonstrated that 50 patients (83%) had a right-sided pleural effusion and 46 patients (77%) had a left-sided effusion. Approximately one third of the effusions had a volume that exceeded 700 mL.
The pleural fluid that accumulates with CHF is related to the clearance of pulmonary interstitial fluid across a leaky mesothelium into the pleural space. In studies in which sheep lungs were isolated in situ, volume loading led to an increased transudation across the lung into the pleural space. The pleural fluid had the same protein concentration as that of the lung lymph and the interstitial edema liquid in the lung. The volume of pleural fluid constituted about 25% of all edema formed in the lung. In the clinical situation, patients with CHF are much more likely to have a pleural effusion if there is radiologically apparent pulmonary edema.
Patients with pleural effusion resulting from CHF usually have symptoms and signs of heart failure, such as dyspnea on exertion, orthopnea, nocturia, peripheral edema, distended neck veins, crackles, and a cardiac gallop. The chest radiograph almost always reveals cardiomegaly in addition to the pleural effusion.
The pleural effusions seen with CHF tend to be bilateral, with larger effusions on the right ( Fig. 79-6A ). On CT imaging, interstitial and alveolar edema can often be detected by the presence of thickened septae and patchy opacities ( Fig. 79-6B and ). Thickened septa represent interstitial edema, as seen in edematous lungs frozen to demonstrate the location of edema; the interstitial edema is continuous with the interlobular septa and the subpleural space, from which the edema has been shown to move across the visceral pleura to the pleural space ( Fig. 79-6C ).