Symposium on Pulmonary Disease
Pleural Effusions Richard W. Light, M.D.*
The pleural space and the fluid within it are not under static conditions. During each respiratory cycle the pleural pressures and the geometry of the pleural space fluctuate widely. Fluid constantly moves into and out of the pleural space. In normal subjects, a dynamic equilibrium is reached where there is a small amount of pleural fluid, and fluid formation equals fluid absorption. Many different conditions may alter the dynamics of the pleural space so that pleural fluid accumulates (Table 1). This article will review those factors influencing the formation and absorption of pleural fluid, and will discuss the characteristics of the pleural fluid resulting from various conditions.
PHYSIOLOGY OF THE PLEURAL SPACE Protein-Free Liquid Exchange The passage of protein-free liquid through the pleural membranes is dependent upon the hydrostatic and colloid-osmotic pressures across them (Fig. 1). The net hydrostatic pressure which forces the fluid from the parietal pleura into the pleural space is the systemic capillary pressure (30 cm H 2 0) minus the negative pleural pressure (5 cm H 2 0) or 35 cm H 2 0. Opposing this is the colloid-osmotic pressure in the blood (34 cm H 2 0) minus that in the pleural fluid (8 cm H 2 0) or 26 cm H 2 0. The net pressure difference of 9 cm H 2 0 favors movement of fluid from the parietal pleura into the pleural space. The only difference between the visceral and the parietal pleura in the above scheme is that the capillaries of the visceral pleura have the pressures of the pulmonary circulation. Therefore the hydrostatic pressure in these capillaries is much lower. The net force across the visceral pleura is 10 cm H 2 0, favoring absorption of pleural fluid. 1 In summary, protein-free liquid normally flows from the systemic capillaries in the parietal pleura to the pleural space and thence to the pulmonary capillaries in the visceral pleura. The amount of fluid that traverses the pleural space in 24 hours is high. Agostoni has shown that "Associate Professor, Department of Medicine, Louisiana State University Medical Center, Shreveport, Louisiana Medical Clinics of North America- Vol. 61, No. 6, November 1977 1339
1340
RICHARD
Table 1.
W.
LIGHT
Differential Diagnosis of Pleural Effusion
Transudative Pleural Effusions
Congestive heart failure Cirrhosis Nephrotic syndrome Acute glomerulonephritis Myxedema Peritoneal dialysis Hypoproteinemia Meigs' syndrome Sarcoidosis Exudative Pleural Effusions
Infectious diseases Tuberculosis Bacterial infections Viral infections Fungus infections Parasitic infections N eoplastic diseases Mesotheliomas Metastatic disease Collagen vascular diseases Systemic lupus erythematosus Rheumatoid pleuritis
Exudative Pleural Effusions (Continued)
Pulmonary infarction - embolization Gastrointestinal diseases Pancreatitis Esophagealrupture Subphrenic abscess Hepatic abscess Whipple's disease Diaphragmatic hernia Trauma Hemothorax Chylothorax Drug hypersensitivity Nitrofurantoin Methysergide Miscellaneous diseases Asbestos exposure Pulmonary and lymph node myomatosis Uremia Postmyocardial infarction syndrome Trapped lung Congenital abnormalities of the lymphatics Post radiation therapy
the conductance of the visceral pleura for isotonic saline in open-chest dogs and isolated lung lobes is 0.4 x 10-3 and 0.9 x 10-3 ml per hr per cm H 2 0 per sq cm of pleural surface. Assuming that the conductance of human pleura is similar to that of dogs' pleura, that the pressures in Figure 1 are valid, and that the surface area of the visceral pleura is 5000 square cm, between 5,000 and 10,000 ml of protein-free fluid pass through the pleural space in 24 hours. Stewart32 has shown that the average clearance of para-aminohippurate from the pleural space of five patients with congestive heart failure was 300 ml per hour. This is probably a good approximation for pleural blood flow, which ultimately limits the rate of pleural fluid formation and absorption.
Protein Exchange Normally, there are small amounts of fluid with a protein content approximating 1.5 gm per 100 ml present in the pleural space. 2 This protein leaks from the pleural capillaries into the pleural space. If the protein accumulated, the colloid-osmotic pressure of the pleural fluid would increase and the pressure gradients (Fig. 1) would favor fluid movement from both the visceral and the parietal pleura to the pleural space. It would be expected that pleural fluid would continue to accumulate until the protein was diluted and/or the pleural pressure increased to restore conditions favorable for fluid absorption. Since the serum protein concentration is greater than the pleural fluid protein concentration, the
1341
PLEURAL EFFUSIONS
PARIETAL PLEURA
PLEURAL SPACE
SYSTEMIC CAPILLARY HYDROSTATIC PRESSURE
PLEURAL PRESSURE
30
VISCERAL PLEURA PULMONARY CAPILLARY HYDROSTATIC PRESSURE
-----t;.
~----11
COLLOID OSMOTIC PRESSURE COLLOID OSMOTIC PRESSURE
COLLOID OSMOTIC PRESSURE
34 ~----+
NET
4
~(---
34
----~) 13~(----
-----,)~23
-:l> Figure 1. Diagrammatic representation of the pressures involved in the formation and absorption of pleural fluid. RESULTANT
-9~
-10
pleural fluid protein would never be removed by simple diffusion and pleural effusions would never resolve spontaneously. Nevertheless, some pleural effusions with high protein contents resolve spontaneously. How is the protein removed from the pleural space? Courtice and Simmonds8 ligated the thoracic duct and the right lymphatic duct in cats with pleural effusions. When labeled protein was placed in the effusion, none of it reached the systemic circulation. Stewart32 subsequently demonstrated that in patients with transudative pleural effusions, the lymphatic flow from the pleural space averages 0.2 ml per kg per hr at night and nearly doubles during the day. Therefore, for a 60 kg individual, the lymphatic clearance of the pleural space is in the range of 250 to 500 ml per 24 hours; this is less than 10 per cent of the clearance of protein-free liquid. Red blood cells and white blood cells are also removed from the pleural space by the lymphatics. 33 Thus, protein-free fluid is constantly entering the pleural space from the parietal pleura and leaving through the visceral pleura. Protein enters the pleural space from both pleural surfaces and leaves via the lymphatics. Pleural fluid accumulates when the normal dynamic equilibrium is upset and continues to accumulate until another equilibrium is reached.
TRANSUDATIVE VERSUS EXUDATIVE PLEURAL EFFUSIONS Pleural effusions have classically been divided into transudates and exudates. A transudate occurs when the systemic factors influencing the formation or absorption of pleural fluid are altered. Decreased plasma colloid-osmotic pressure or elevated hydrostatic pressure in the systemic or pulmonary circulation are alterations that produce transudates. The pleural surfaces are not involved by the primary pathologic process when there is a transudative pleural effusion. In contrast, an exudate results from disease of the pleural surface. The two main mechanisms by which pleural disease leads to pleural fluid accumulation are: (1) increased permeability of the capillaries for
1342
RICHARD
W.
LIGHT
protein such as occurs with pulmonary embolization, pneumonia, and tuberculosis and (2) lymphatic obstruction such as occurs with lymphoma. The first question that must be answered when a pleural effusion is discovered is whether that effusion is a transudate or an exudate. If the effusion is a transudate, no further diagnostic procedures are necessary and therapy is directed toward the underlying congestive heart failure, cirrhosis, or nephrosis. On the other hand, if the effusion proves to be an exudate, more extensive diagnostic procedures are in order to delineate the cause of the pleural disease. The Separation of Transudates from Exudates Classically, a pleural fluid has been classified as an exudate when the protein level exceeds 3.0 gm per 100 ml or its specific gravity is above 1.016.5. 15 ,25 Unfortunately, the use of either of these criteria results in wrongly classifying over 10 per cent of effusions. 5, 15, 18 It has subsequently been shown that the simultaneous use of protein and lactic acid dehydrogenase (LDH) levels in pleural fluid and serum effectively separates transudative from exudative effusions. In a prospective study of 150 pleural effusions,18 102 of 103 exudates had at least one of the following characteristics, while only one of 47 transudates had any of these characteristics: 1. Pleural fluid protein divided by serum protein greater than 0.5 2. Pleural fluid LDH divided by serum LDH greater than 0.6 3. Pleural fluid LDH greater than two thirds the upper limit of normal for the serum LDH. Therefore, when evaluating a patient with a pleural effusion, the lactic dehydrogenase and protein level of the pleural fluid and serum should be measured. If the fluid has none of the above three characteristics, it is a transudate and therapy need only be directed toward the heart failure, nephrosis, cirrhosis, and so forth. However, if the fluid has one or more of the above characteristics, it is an exudate and attention must be focused on methods to elucidate the etiology of the pleural disease. The specific gravity of the pleural fluid was used in the past to separate transudates from exudates because it was a simple and rapid method of estimating the protein content of the fluid. 24 A specific gravity of 1.016 corresponds to a protein content of 3.0 gm per 100 ml and each deviation of ±0.003 in specific gravity represents approximately 1.0 gm per 100 ml protein. For example, a specific gravity of 1.022 corresponds to a protein level of 5.0 gm per 100 ml. Recently, refractometers have been used to estimate the specific gravity of pleural fluid in many institutions. The refractometer has two distinct advantages over the hydrometer-it is essentially independent of temperature and it only takes 0.2 ml rather than 20 ml of pleural fluid. There is a linear relationship between the refractive index and the protein content of pleural fluid. Unfortunately, the scale on the commercially available refractometers is calibrated for the specific gravity of urine rather than pleural fluid. A reading of 1.020 on the urine specific
PLEURAL EFFUSIONS
1343
gravity scale corresponds to a protein of 3.0 gm per 100 ml, and each deviation of 0.004 ml represents 1 gram of protein. For example, a refractometer reading of 1.028 on the urine specific gravity scale corresponds to a pleural fluid protein of 5 gm per 100 ml. If the above facts are kept in mind, the refractometer is the simplest way to get an estimate of the pleural fluid protein. 4
TESTS FOR DIFFERENTIATING VARIOUS EXUDATIVE EFFUSIONS Appearance of the Fluid The gross appearance of the pleural fluid frequently yields useful diagnostic information. The color, turbidity, viscosity, and odor should be described. Most transudative and many exudative pleural effusions are clear, straw-colored, quite fluid, and odorless. Any deviations are noteworthy. A white, milky appearance indicates a chylothorax, a chyliform pleural effusion, or a pyothorax. The pyothorax can be distinguished from the chylothorax and chyliform effusion in that after centrifugation there is a clear supernatant only with pyothorax. A feculant odor is highly suggestive of an empyema secondary to anaerobic organisms. A clear or bloody fluid that is quite viscous is highly suggestive of malignant mesothelioma; the high viscosity is secondary to elevated pleural fluid hyaluronic acid levels."7 The fluid from frank empyemas is viscid and opaque. Red Cell Count- Bloody Pleural Effusion It requires only 5,000 to 10,000 red blood cells per cu mm to impart a red color to a pleural effusion. Assuming that a given pleural effusion has a total volume of 500 cc and the red cell count in the peripheral blood is 5 million per cu mm, a leak of only 1 cc of blood into the pleural space at the time of thoracentesis will result in a blood-tinged pleural effusion. For this reason, the mere fact that a pleural effusion is bloodtinged has limited diagnostic implications. Over 15 per cent of transudative and over 40 per cent of all types of exudative pleural fluids will be blood-tinged, Le., pleural fluid red cell counts between 5,000 and 100,000 per cu mm!! The question as to whether the blood was introduced via the thoracentesis or had been there previously can frequently be resolved with a Wright stain of the sediment. If there has been blood in the pleural space for a matter of hours, the macrophages in the pleural fluid will contain hemoglobin inclusion bodies that will stain pink. Grossly bloody pleural effusions have pleural fluid red cell counts above 100,000 per cu mm. This finding is very suggestive of one of three disease processes: trauma, malignancy, or pulmonary embolization. 21
1344
RICHARD
W.
LIGHT
White Cell Count The pleural fluid white blood cell count is of limited value. A pleural fluid white cell count of 1,000 per cu mm roughly separates transudative from exudative pleural effusions, but not nearly as effectively as do the protein and lactic dehydrogenase. Pleural fluid white cell counts above 10,000 per cu mm are most common with parapneumonic effusions but are also seen with pancreatitis, pulmonary infarction, collagen vascular diseases, malignancy, and tuberculosis. 2! Differential White Cell Count Examination of a Wright stain of pleural fluid is one of the most useful tests to be performed on pleural fluid. Eight different types of cells should be looked for in the pleural fluid: mesothelial cells, macrophages, plasma cells, lymphocytes, polymorphonuclear neutrophils, eosinophils, basophils, and malignant cells. 3 ! The stain is made most easily by centrifuging about 10 cc of fluid and then resuspending the button of cells in about 0.5 cc of supernatant. Slides of the pleural fluid are then made and stained just as are slides for peripheral blood smears. Occasionally, large amounts of fibrinogen adhere to the cells. In this case, resuspension in saline followed by centrifugation is indicated in order to evaluate cellular morphology. MESOTHELIAL CELLS AND MACROPHAGES. Mesothelial cells3 ! are the cells which line the pleural cavities. They are usually 12 to 30 microns in diameter, but multinucleated forms may have diameters up to 75 microns. The cytoplasm is light blue and is homogeneous. The nucleus stains purplish, is relatively large, is round and has a uniform appearance. There are usually 1 to 3 blue nucleoli. Collections of mesothelial cells (sometimes up to 12 or more) are occasionally seen adhering to one another. Macrophages, by definition, are cells which store vital dyes. The nuclei of macrophages are irregular, and the chromatin is lacy rather than stippled. The cytoplasm is gray, cloudy, and full of vacuoles. When there is active phagocytosis, these macrophages often contain red blood cells and neutrophils in various stages of digestion. Most pleural fluid macrophages probably evolve from mesothelial cells, although some probably evolve from circulating monocytes. Mesothelial cells are significant in two different respects. (1) They may be confused with malignant cells; frequently an experienced pathologist is required to make this differentiation. (2) Their presence or absence is often useful diagnostically. Mesothelial cells are uncommon in tuberculous effusions. Spriggs and Boddington3 ! analyzed 65 tuberculous effusions and found that only one had more than one mesothelial cell per 1,000 cells. They concluded that the presence of numerous mesothelial cells nearly excludes a diagnosis of tuberculosis. This has subsequently been confirmed by other workers.21. 35 However, this finding is not diagnostic of tuberculosis, as a paucity of mesothelial cells also occurs with empyema and some cases of malignancy. PLASMA CELLS. Plasma cells are cells with eccentric nuclei, deeply basophilic cytoplasm, and a clear area at the center of the cell. The
PLEURAL EFFUSIONS
1345
nuclear chromatin is usually clumped. The presence of numerous plasma cells suggests multiple myeloma, but in a study of 16 cases of pleural effusions with plasmacytosis greater than 5 per cent, their presence was of no diagnostic importance. Four were associated with malignancies, 3 with tuberculosis, 3 with pulmonary infarcts, 2 with pneumonia, 1 with rheumatoid arthritis, 1 with sepsis, and the basis for one was unknown. 31 LYMPHOCYTES. The discovery that more than 50 per cent of the white cells in pleural fluid are smalllymphocytes is important diagnostically. In two series,21. 35 96 of 211 exudative pleural effusions had more than 50 per cent smalllymphocytes. Of these 96 effusions, 90 were due to tuberculosis or malignancy. Since these are the two diseases that can be diagnosed with pleural biopsy, the finding of predominantly small lymphocytes in an exudative pleural effusion is a strong indication for pleural biopsy. When the same two series are combined, almost all the effusions secondary to tuberculosis (43/46) but only about half the effusions secondary to malignancy (47/90) had predominantly smalllymphocytes. POL YMORPHONUCLEAR LEUKOCYTES. Since neutrophils are the cellular component of the acute inflammatory response, they predominate in pleural fluid, resulting from acute inflammation of the pleura. Examples are effusions associated with pneumonia, pancreatitis, pulmonary emboli, peritonitis, and early tuberculosis. Some neutrophils are found in almost every pleural effusion; they appear in enormous numbers in empyemas. In this case, the neutrophils degenerate and their nuclei become blurred and lose their normal purple-staining characteristics. Their cytoplasm shows toxic granulation and fat vacuolation, and their normal granulation is lost. In all other effusions, the neutrophils appear much as they do in the peripheral blood. If neutrophils predominate in an effusion attributed to congestive heart failure, the possibility of pulmonary emboli should be investigated. EOSINOPHILS. The presence of pleural fluid eosinophilia is of very little use in the differential diagnosis of pleural effusionY Most effusions with significant eosinophilia (>10 per cent) are either bloody or associated with a pneumothorax. If the effusion is not bloody and there is no accompanying pneumothorax, the most likely etiology is a viral pleuritis or a resolving parapneumonic effusion. The presence of pleural fluid eosinophilia in a parapneumonic effusion is a good prognostic sign as these effusions virtually never become purulent. BASOPHILS. Basophilic pleural effusions are uncommon. There are usually a few basophils in eosinophilic pleural effusions; their presence has no known diagnostic significance. Protein and Lactic Dehydrogenase In general, measurements of the pleural fluid protein and lactic dehydrogenase are not useful in differentiating various types of exudative pleural effusions. IS All exudates tend to contain increased amounts of both. However, if the lactic dehydrogenase is elevated and the protein is not, the effusion is probably due to malignancy. Conversely, most pa-
1346
RICHARD
W.
LIGHT
tients with pleural fluid proteins above 6.0 gm per 100 ml have tuberculous or parapneumonic effusions. The measurement of isoenzymes of lactic dehydrogenase in pleural fluid is also of limited value. The pleural fluid LDH isoenzyme pattern for all benign exudates is characterized by a higher percentage of LDH4 and LDH-5 in the pleural fluid than in the serum. Most malignant exudates have this same pattern but a few have predominantly LDH-2. The presence of the latter pattern is highly suggestive that the patient has a malignant pleural effusion. 20 The other situation in which the LDH isoenzymes are of value is with bloody pleural effusions. At times when there is hemolysis, the pleural fluid lactic dehydrogenase will be elevated from the hemolysis even though the effusion is a transudate. Since most of the lactic dehydrogenase secondary to hemolysis is LDH1, subtraction of the LDH-1 from the total lactic dehydrogenase will give a pleural fluid LDH which can be used in differentiating transudates from exudates. It should be emphasized that the LDH-1 is not elevated in most bloody pleural effusions. 20 Simultaneous protein electrophoresis of the pleural fluid and serum is without value in the differential diagnosis of pleural effusions because both the serum and the pleural fluid have the same pattern. Likewise, measurements of mucoprotein or immunoglobulin levels in pleural effusions are not useful diagnostically.3o.36
Pleural Fluid Cytology A pleural fluid specimen from every patient with pleural effusion should be sent for cytopathologic studies. The first pleural fluid cytology will be positive for malignant cells in about 60 per cent of the effusions secondary to pleural malignancy.21 If three separate specimens are submitted, up to 90 per cent of those associated with pleural malignancy will have positive cytopathology. The later samples add diagnostic information because they contain fresher cells, since the older degenerated cells were removed during the earlier thoracenteses. Two points should be emphasized about pleural fluid cytology.31 (1) Malignancies can cause pleural effusions by mechanisms other than direct pleural involvement, e.g., lymphatic or bronchial obstruction and hypoproteinemia. Therefore, not all pleural effusions associated with malignancies have positive pleural fluid cytopathology. (2) Because of the marked variation that mesothelial cells undergo in response to inflammation, pleural fluid cytology is difficult. All cases in which the results are questionable should be reviewed by an experienced cytopatholo gis t. Glucose The pleural fluid glucose level of transudates and most exudates parallels that of the serum. However, there are four categories of exudative pleural effusions in which the pleural fluid glucose may be low «60 mg per 100 ml).22 1. Parapneumonic effusions. The lower the glucose is in this situation, the higher the probability that the effusion is infected and that in-
PLEURAL EFFUSIONS
1347
sertion of chest tubes will be necessary for its resolution. However, a decreased pH is more sensitive than a decreased glucose in predicting which parapneumonic effusions will become complicated (see below). 2. Rheumatoid pleural effusion. Carr and Powers first made the observation that pleural effusions secondary to rheumatoid disease have a low glucose concentration. In reviewing 74 pleural effusions in which glucose measurements had been made, they found that 5 had a glucose level of less than 30 mg per 100 ml; 4 of these 5 were secondary to rheumatoid disease. The diagnosis of pleural effusion secondary to rheumatoid disease is doubtful if the pleural fluid glucose is above 30 mg per 100 ml. The explanation for the low pleural fluid glucose level in this condition appears to be a selective block to the entry of glucose into the pleural effusionY' Interestingly, the pleural fluid glucose level is not depressed with systemic lupus erythematosus. 3. Tuberculous pleural effusion. Although an occasional tuberculous pleural effusion will have a very low pleural fluid glucose level,3 the mean glucose level in these effusions is about 80 mg per 100 ml,22 Therefore, the pleural fluid glucose is usually not helpful in making the diagnosis of tuberculous pleuritis. 4. Malignant pleural effusion. In approximately 15 per cent of malignant effusions, the pleural fluid glucose is below 60 mg per 100 ml. 22 Malignant pleural effusions with low pleural fluid glucose levels have one of two characteristics: they have myriads of malignant cells in the fluid 7 or they are very large, occupying the entire hemithorax. 22 Amylase An amylase determination should be made on every pleural fluid examined. Pancreatitis is complicated by pleural effusions nearly 10 per cent of the time.!'] In effusions secondary to pancreatic disease, an elevated pleural fluid amylase may be the first clue to the diagnosis. 22 There are only two other conditions in which the pleural fluid amylase is elevated. In pleural effusions secondary to esophageal rupture, the amylase is markedly elevated. The origin of the amylase is salivary rather than pancreatic in this situation. 28 The saliva enters the pleural space through the tear in the esophagus. The pleural fluid amylase is elevated in approximately 10 per cent of patients with malignant pleural effusions. 22 The primary tumor in such instances is usually not in the pancreas.
Pleural Fluid pH The pleural fluid pH is a very useful prognostic index in patients with acute bacterial pneumonia and pleural effusion (parapneumonic effusion). A low pleural fluid pH «7.20) in this situation, strongly suggests that the effusion will not resolve withou t chest tubes. HI, 26 Since drainage of the pleural space can become very difficult if loculation develops, we recommend thoracentesis and measurement of pleural fluid pH in all patients with parapneumonic effusions as soon as the effusion is recognized. If the pleural fluid pH is below 7.20, and the arterial pH is above 7.35, tube thoracostomy should be instituted immediately. If the
1348
RICHARD
W.
LIGHT
arterial pH is below 7.35, chest tubes should be placed only if the pleural fluid pH is more than 0.15 units lower than the arterial pH. If the pleural fluid pH is between 7.20 and 7.30, daily thoracentesis with pleural fluid pH measurements should be performed until the pleural fluid pH approaches 7.30. If the pleural fluid pH falls to or below 7.20, tube thoracostomy should be instituted. In this situation the pleural fluid pH becomes lowered before the pleural fluid glucose falls or before organisms become visible on Gram stain. 19 The use of a low pleural fluid pH as an indication for chest tube placement is applicable only with parapneumonic effusions. With all transudative and most exudative pleural effusions, the pleural fluid pH parallels that of arterial blood. However, the pleural fluid pH is less than 7.20 in most rheumatoid pleural effusions and some effusions caused by malignancy or tuberculosis, none of which should necessarily be treated with tube thoracostomy.19 It should be emphasized that the determination of pleural fluid pH must be made with the same care as with arterial pH. The fluid must be withdrawn anaerobically into a heparinized syringe and maintained at 0° C until the pH is measured. Cultures and Bacteriologic Stains When a diagnostic thoracentesis is performed, the fluid obtained should be cultured for both anaerobic and aerobic organisms. Mycobacterial and fungal cultures should also be obtained. The chances for obtaining a positive mycobacterial culture are enhanced by a generous specimen. A Gram stain of the fluid should be examined to rule out bacterial infection. Since the probability of detecting tubercular organisms with an acid-fast stain for tube,rcle bacilli is minimal even in tuberculous effusions, the author employs this stain on the sediment only if mesothelial cells compose less than 1 per cent of the total cells on the Wright stain. Other Tests Lupus ERYTHEMATOSUS CELLS. Pleural fluid LE preparations should be obtained on all patients who have puzzling exudative effusions. When the effusion is secondary to lupus erythematosus, the pleural fluid will usually contain LE cells sometimes even when there are no LE cells in the peripheral blood. LE cells in the pleural fluid are thought to be pathognomonic of systemic lupus erythematosus. 34 COMPLEMENT. Low levels of complement in pleural fluid are found in effusions caused by either rheumatoid pleuritis or lupus erythematosus. Hunder et al. 12 measured the pleural fluid complement levels of 50 patients with pleural effusions of various etiologies. The pleural fluid complement was low «10 units per ml) in 11 of 12 patients with systemic lupus erythematosus or rheumatoid pleuritis but exceeded this level in 37 of 38 patients with other diseases. With the collagen vascular diseases, the pleural fluid complement levels are depressed substantially more than the serum complement levels. If collagen vascular disease is suspected, pleural fluid complement levels should be obtained.
1349
PLEURAL EFFUSIONS
Table 2. DESTINATION
Processing Pleural Fluid AMOUNT
TESTS ORDERED
Bacteriology
5 cc
Bacterial cultures Gram Stain
Tuberculosis and mycology
5 cc
Tuberculosis fungal cultures Acid-fast bacillus stain
Cytology
5 cc
Cytology
Hematology
5 cc
Red cell count White cell count Wright stain
Chemistry
5 cc
Glucose Amylase Lactic dehydrogenase Protein
Blood gas laboratory
5 cc
pH
RHEUMATOID FACTOR- Although the rheumatoid factor is high in effusions secondary to rheumatoid arthritis, it is also frequently elevated in exudates secondary to pneumonia, tuberculosis, and carcinoma. Therefore, it is not useful diagnostically,11 HYALURONIC ACID. A pleural fluid hyaluronic acid level is useful in the diagnosis of malignant mesothelioma. Only patients with mesotheliomas have a markedly elevated hyaluronic acid level. Unfortunately, not all patients with mesothelioma have elevated levels of pleural fluid hyaluronic acid.27 LIPID ANALYSIS. Pleural fluid is occasionally found to be milky or opalescent. If the fluid remains cloudy after centrifugation, the patient has either a chylous or a chyliform pleural effusion. A chylous pleural effusion arises when the thoracic duct is severed or obstructed. The cloudiness in this situation is caused by chylomicra and lipid analysis reveals high levels of triglycerides but low levels of cholesterol. A chyliform (cholesterol) pleural effusion develops when a pleural effusion has been present for a long time, and for unknown reasons cholesterol accumulates. The fluid in this situation is also milky, but pleural fluid lipid analysis reveals elevated cholesterol and normal triglycerides. 2 " CHROMOSOME ANALYSIS. There are abnormalities in the number and structure of the chromosomes in malignant pleural effusions."' 11 It appears that with the combination of cytology and chromosomal analysis, a definite diagnosis of malignancy can be established in more patients than with either method alone. However, since chromosomal analysis is not routinely available and has not yet been applied to large numbers of benign exudates, it should still be regarded as experimental. CARCINOEMBRYONIC ANTIGENS. A preliminary report'4 on 6 patients has shown that the pleural fluid from some patients with malignancy has a high level of carcinoembryonic antigen-like substances (CEA-LS), while the CEA-LS is not elevated in benign effusions. If this
1350
RICHARD
W.
LIGHT
can be confirmed, measurement of pleural fluid CEA-LS will be useful. However, at this time it must still be regarded as experimental.
PLEURAL BIOPSY Samples of the parietal pleura can easily be obtained with a Cope or Abrams pleural biopsy needle. Some authors have advocated pleural biopsy each time a thoracentesis is performed. However, pleural biopsy is associated with greater morbidity than is a simple diagnostic thoracentesis, and is really only useful when the patient has granulomatous disease or malignancy of the pleura. Therefore, I only recommend pleural biopsy if granulomatous disease or malignancy is suspected. There are two contraindications to pleural biopsy: a bleeding diathesis and an empyema. In order to do a pleural biopsy, the prothrombin time should be at least 50 per cent of the control value, the platelet count should be greater than 100,000, and the bleeding and clotting times should be normal. The percentage of specific diagnoses obtained with pleural biopsy depends on patient selection. If several bits of parietal pleura are obtained on three separate occasions, tuberculosis can be diagnosed in over 80 per cent and malignancy in over 50 per cent.23 If culture of a bit of the biopsy material is combined with pleural biopsy, the diagnosis of tuberculosis can be made in nearly 95 per cent of the cases. 16 Since pleural fluid cultures are positive in less than 25 per cent of patients with tuberculous pleuritis, culture of a pleural biopsy specimen should definitely be done.
SUMMARY Many different conditions result in the accumulation of pleural fluid. A diagnostic thoracentesis should be performed on all patients with pleural effusion from whom pleural fluid can be easily obtained. Empirically we have found that when the pleural effusion is more than 10 mm thick on the lateral decubitus roentgenogram, pleural fluid is easily obtained. At least 30 cc fluid should be obtained and distributed to the various laboratories as outlined in Table 2. The results of these tests will show whether the fluid is a transudate or an exudate. If the fluid is a transudate, no further diagnostic procedures need be directed towards the pleura. If the fluid is an exudate, the diagnosis will frequently be made by these original tests and therapy for the pleural disease can be instituted. If the diagnosis has not been made, the results of these tests should lead to a rational diagnostic attack.
1351
PLEURAL EFFUSIONS
REFERENCES 1. Agostoni, E., Taglietti, A., and Setnikar, I.: Absorption force of the capillaries of the visceral pleura in determination of the intrapleural pressure. Amer. J. PhysioI., 191 :277282,1957. 2. Agostoni, E.: Mechanics of the pleural space. Physiol. Rev., 52:57-128,1972. 3. Barber, L. M., Mazzadi, L., Deakins, D. D., et aI.: Glucose level in pleural fluid as a diagnostic aid. Dis. Chest, 31 :680-687, 1957. 4. Briggs, M. S., George, R. B., and Light, R. W.: Use of the refractive index to estimate protein concentration in pleural fluid. Clin. Res., 25:36A, 1977. 5. Carr, D. T., and Power, M. H.: Clinical value of measurements of concentration of protein in pleural fluid. New Eng. J. Med., 259:926-927, 1958. 6. Carr, D. T., and Power, M. H.: Pleural fluid glucose with special reference to its concentration in rheumatoid pleurisy with effusion. Dis. Chest, 37 :321-324, 1960. 7. Clarkson, B.: Relationship between cell type, glucose concentration and response to treatment in neoplastic effusions. Cancer, 17:914-928, 1964. 8. Courtice, F., and Simmonds, W.: Absorption of fluids from the pleural cavities of rabbits and cats. J. PhysioI., 109:117-130, 1949. 9. Dewald, G., Dines, D. E., Weiland, L. H., et al.: Usefulness of chromosome examination in the diagnosis of malignant pleural effusions. New Eng. J. Med., 295:1494-1500, 1976. 10. Dodson, W., and Hollingsworth, J.: Pleural effusion in rheumatoid arthritis: Impaired transport of glucose. New Eng. J. Med., 275:1337-1342, 1966. 11. Hansson, A., and Korsgaard, R.: Cytogenetical diagnosis of malignant pleural effusions. Scand. J. Resp. Dis., 55:301-308,1974. 12. Hunder, G. G., McDuffie, F. C., and Hepper, N. G.: Pleural fluid complement in systemic lupus erythematosus and rheumatoid arthritis. Ann. Intern, Med., 76:357-363, 1972, 13. Kaye, M.: Pleuropulmonary complications of pancreatitis. Thorax, 23:297-306,1968. 14. Kim, Y. D., and Hirata, A. A.: Carcinoemhryonic antigen-like substances in human cavity fluids. ImmunoI. Commun., 5:619-629,1976. 15. Leuallen, E., and Carr, D.: Pleural Effusion: A statistical study of 436 patients. New Eng. J. Med., 252 :79-83, 1955. 16. Levine, H., Metzger, W., Lacera, D., et a!.: Diagnosis of tuberculous pleurisy by culture of pleural biopsy specimen. Arch. Intern. Med., 126:269-271, 1970. 17. Levine, H., Szanto, M., Grieble, H., et al.: Rheumatoid factor in nonrheumatoid pleural effusions. Ann. Intern. Med., 69:487-492,1968. 18. Light, R. W., MacGregor, M. I., Luchsinger, P. C., et aI.: Pleural effusions: The diagnostic separation of transudates and exudates. Ann. Intern. Med., 77:507-514, 1972. 19. Light, R. W., MacGregor, M. I., Ball, W. C., Jr., et al.: Diagnostic significance of pleural fluid pH and Pco 2 . Chest, 64:591-596, 1973. 20. Light, R. W., and Ball, W. C., Jr.: Lactate dehydrogenase isoenzymes in pleural effusion. Amer. Rev. Resp. Dis., 108:660-665,1973. 21. Light, R. W., Erozan, Y. C., Ball, W. C. Jr.: Cells in pleural fluid: Their value in differential diagnosis. Arch. Intern. Med., 132 :854-860, 1973. 22. Light, R. W., and Ball, W. C. Jr.: Glucose and amylase in pleural effusions. J.A.M.A., 225:257-260, 1973. 23. Mestitz, P., Purves, M., and Pollard, A.: Pleural biopsy in the diagnosis of pleural effusion: A report of 200 cases. Lancet, 2: 1349-1353, 1958. 24. Paddock, F.: The relationship between the specific gravity and the protein content in human serous effusions. Amer. J. Med. ScL, 201 :569-574, 1941. 25. Paddock, F. K.: The diagnostic significance of serous fluids in disease. New Eng. J. Med., 223:1010-1015, 1940. 26. Potts, D. E., Levin, D. C., and Sahn, S. A.: Pleural fluid pH in parapneumonic effusions. Chest, 70 :328-331, 1976. 27. Rasmussen, L., and Faber, V.: Hyaluronic acid in 247 pleural fluids. Scand. J. Resp. Dis., 48:366-371,1967. 28. Sherr, H. P., Light, R. W., Merson, M. H., et al.: Origin of pleural fluid amylase in esophageal rupture. Ann. Intern. Med., 76:985-986, 1972. 29. Seriff, N. S., Cohen, M. L., Samuel, P., et al.: Chylothorax: diagnosis by lipoprotein electrophoresis of serum and pleural fluid. Thorax, 32:98-100, 1977. 30. Shallenberger, D. W., and Daniel, T. M.: Quantitative determinations of several pleural fluid proteins. Amer. Rev. Resp. Dis., 106: 121-122, 1972. 31. Spriggs, A., and Boddington, M.: The Cytology of Effusions. New York, Grune & Stratton, Inc., 1968. 32. Stewart, P. B.: The rate of formation and lymphatic removal of fluid in pleural effusions. J. Clin. Invest., 42:258-262, 1963.
1352
RICHARD
W.
LIGHT
33. Stewart P., and Burgen, A.: The turnover offiuid in the dog's pleural cavity. J. Lab. Clin. Med., 52:212-230,1958. 34. Winslow, W. A., Ploss, L. N., and Loitman, B.: Pleuritis in systemic lupus erythematosus: its importance as an early manifestation in diagnosis. Ann. Intern. Med., 49:70-88, 1958. 35. Yam, L.: Diagnostic significance of lymphocytes in pleural effusions. Ann. Intern. Med., 66:972-982, 1967. 36. Zinneman, H. H., Johnson, J J., and Lyon, R. H.: Proteins and mucoproteins in pleural effusions. Amer. Rev. Tuberc., 76:247-255,1957. Department of Medicine Louisiana State University Medical Center P.O. Box 33932 Shreveport, Louisiana 71130
Pleural Effusions Richard W. Light, M.D.*
The pleural space and the fluid within it are not under static conditions. During each respiratory cycle the pleural pressures and the geometry of the pleural space fluctuate widely. Fluid constantly moves into and out of the pleural space. In normal subjects, a dynamic equilibrium is reached where there is a small amount of pleural fluid, and fluid formation equals fluid absorption. Many different conditions may alter the dynamics of the pleural space so that pleural fluid accumulates (Table 1). This article will review those factors influencing the formation and absorption of pleural fluid, and will discuss the characteristics of the pleural fluid resulting from various conditions.
PHYSIOLOGY OF THE PLEURAL SPACE Protein-Free Liquid Exchange The passage of protein-free liquid through the pleural membranes is dependent upon the hydrostatic and colloid-osmotic pressures across them (Fig. 1). The net hydrostatic pressure which forces the fluid from the parietal pleura into the pleural space is the systemic capillary pressure (30 cm H 2 0) minus the negative pleural pressure (5 cm H 2 0) or 35 cm H 2 0. Opposing this is the colloid-osmotic pressure in the blood (34 cm H 2 0) minus that in the pleural fluid (8 cm H 2 0) or 26 cm H 2 0. The net pressure difference of 9 cm H 2 0 favors movement of fluid from the parietal pleura into the pleural space. The only difference between the visceral and the parietal pleura in the above scheme is that the capillaries of the visceral pleura have the pressures of the pulmonary circulation. Therefore the hydrostatic pressure in these capillaries is much lower. The net force across the visceral pleura is 10 cm H 2 0, favoring absorption of pleural fluid. 1 In summary, protein-free liquid normally flows from the systemic capillaries in the parietal pleura to the pleural space and thence to the pulmonary capillaries in the visceral pleura. The amount of fluid that traverses the pleural space in 24 hours is high. Agostoni has shown that "Associate Professor, Department of Medicine, Louisiana State University Medical Center, Shreveport, Louisiana Medical Clinics of North America- Vol. 61, No. 6, November 1977 1339
1340
RICHARD
Table 1.
W.
LIGHT
Differential Diagnosis of Pleural Effusion
Transudative Pleural Effusions
Congestive heart failure Cirrhosis Nephrotic syndrome Acute glomerulonephritis Myxedema Peritoneal dialysis Hypoproteinemia Meigs' syndrome Sarcoidosis Exudative Pleural Effusions
Infectious diseases Tuberculosis Bacterial infections Viral infections Fungus infections Parasitic infections N eoplastic diseases Mesotheliomas Metastatic disease Collagen vascular diseases Systemic lupus erythematosus Rheumatoid pleuritis
Exudative Pleural Effusions (Continued)
Pulmonary infarction - embolization Gastrointestinal diseases Pancreatitis Esophagealrupture Subphrenic abscess Hepatic abscess Whipple's disease Diaphragmatic hernia Trauma Hemothorax Chylothorax Drug hypersensitivity Nitrofurantoin Methysergide Miscellaneous diseases Asbestos exposure Pulmonary and lymph node myomatosis Uremia Postmyocardial infarction syndrome Trapped lung Congenital abnormalities of the lymphatics Post radiation therapy
the conductance of the visceral pleura for isotonic saline in open-chest dogs and isolated lung lobes is 0.4 x 10-3 and 0.9 x 10-3 ml per hr per cm H 2 0 per sq cm of pleural surface. Assuming that the conductance of human pleura is similar to that of dogs' pleura, that the pressures in Figure 1 are valid, and that the surface area of the visceral pleura is 5000 square cm, between 5,000 and 10,000 ml of protein-free fluid pass through the pleural space in 24 hours. Stewart32 has shown that the average clearance of para-aminohippurate from the pleural space of five patients with congestive heart failure was 300 ml per hour. This is probably a good approximation for pleural blood flow, which ultimately limits the rate of pleural fluid formation and absorption.
Protein Exchange Normally, there are small amounts of fluid with a protein content approximating 1.5 gm per 100 ml present in the pleural space. 2 This protein leaks from the pleural capillaries into the pleural space. If the protein accumulated, the colloid-osmotic pressure of the pleural fluid would increase and the pressure gradients (Fig. 1) would favor fluid movement from both the visceral and the parietal pleura to the pleural space. It would be expected that pleural fluid would continue to accumulate until the protein was diluted and/or the pleural pressure increased to restore conditions favorable for fluid absorption. Since the serum protein concentration is greater than the pleural fluid protein concentration, the
1341
PLEURAL EFFUSIONS
PARIETAL PLEURA
PLEURAL SPACE
SYSTEMIC CAPILLARY HYDROSTATIC PRESSURE
PLEURAL PRESSURE
30
VISCERAL PLEURA PULMONARY CAPILLARY HYDROSTATIC PRESSURE
-----t;.
~----11
COLLOID OSMOTIC PRESSURE COLLOID OSMOTIC PRESSURE
COLLOID OSMOTIC PRESSURE
34 ~----+
NET
4
~(---
34
----~) 13~(----
-----,)~23
-:l> Figure 1. Diagrammatic representation of the pressures involved in the formation and absorption of pleural fluid. RESULTANT
-9~
-10
pleural fluid protein would never be removed by simple diffusion and pleural effusions would never resolve spontaneously. Nevertheless, some pleural effusions with high protein contents resolve spontaneously. How is the protein removed from the pleural space? Courtice and Simmonds8 ligated the thoracic duct and the right lymphatic duct in cats with pleural effusions. When labeled protein was placed in the effusion, none of it reached the systemic circulation. Stewart32 subsequently demonstrated that in patients with transudative pleural effusions, the lymphatic flow from the pleural space averages 0.2 ml per kg per hr at night and nearly doubles during the day. Therefore, for a 60 kg individual, the lymphatic clearance of the pleural space is in the range of 250 to 500 ml per 24 hours; this is less than 10 per cent of the clearance of protein-free liquid. Red blood cells and white blood cells are also removed from the pleural space by the lymphatics. 33 Thus, protein-free fluid is constantly entering the pleural space from the parietal pleura and leaving through the visceral pleura. Protein enters the pleural space from both pleural surfaces and leaves via the lymphatics. Pleural fluid accumulates when the normal dynamic equilibrium is upset and continues to accumulate until another equilibrium is reached.
TRANSUDATIVE VERSUS EXUDATIVE PLEURAL EFFUSIONS Pleural effusions have classically been divided into transudates and exudates. A transudate occurs when the systemic factors influencing the formation or absorption of pleural fluid are altered. Decreased plasma colloid-osmotic pressure or elevated hydrostatic pressure in the systemic or pulmonary circulation are alterations that produce transudates. The pleural surfaces are not involved by the primary pathologic process when there is a transudative pleural effusion. In contrast, an exudate results from disease of the pleural surface. The two main mechanisms by which pleural disease leads to pleural fluid accumulation are: (1) increased permeability of the capillaries for
1342
RICHARD
W.
LIGHT
protein such as occurs with pulmonary embolization, pneumonia, and tuberculosis and (2) lymphatic obstruction such as occurs with lymphoma. The first question that must be answered when a pleural effusion is discovered is whether that effusion is a transudate or an exudate. If the effusion is a transudate, no further diagnostic procedures are necessary and therapy is directed toward the underlying congestive heart failure, cirrhosis, or nephrosis. On the other hand, if the effusion proves to be an exudate, more extensive diagnostic procedures are in order to delineate the cause of the pleural disease. The Separation of Transudates from Exudates Classically, a pleural fluid has been classified as an exudate when the protein level exceeds 3.0 gm per 100 ml or its specific gravity is above 1.016.5. 15 ,25 Unfortunately, the use of either of these criteria results in wrongly classifying over 10 per cent of effusions. 5, 15, 18 It has subsequently been shown that the simultaneous use of protein and lactic acid dehydrogenase (LDH) levels in pleural fluid and serum effectively separates transudative from exudative effusions. In a prospective study of 150 pleural effusions,18 102 of 103 exudates had at least one of the following characteristics, while only one of 47 transudates had any of these characteristics: 1. Pleural fluid protein divided by serum protein greater than 0.5 2. Pleural fluid LDH divided by serum LDH greater than 0.6 3. Pleural fluid LDH greater than two thirds the upper limit of normal for the serum LDH. Therefore, when evaluating a patient with a pleural effusion, the lactic dehydrogenase and protein level of the pleural fluid and serum should be measured. If the fluid has none of the above three characteristics, it is a transudate and therapy need only be directed toward the heart failure, nephrosis, cirrhosis, and so forth. However, if the fluid has one or more of the above characteristics, it is an exudate and attention must be focused on methods to elucidate the etiology of the pleural disease. The specific gravity of the pleural fluid was used in the past to separate transudates from exudates because it was a simple and rapid method of estimating the protein content of the fluid. 24 A specific gravity of 1.016 corresponds to a protein content of 3.0 gm per 100 ml and each deviation of ±0.003 in specific gravity represents approximately 1.0 gm per 100 ml protein. For example, a specific gravity of 1.022 corresponds to a protein level of 5.0 gm per 100 ml. Recently, refractometers have been used to estimate the specific gravity of pleural fluid in many institutions. The refractometer has two distinct advantages over the hydrometer-it is essentially independent of temperature and it only takes 0.2 ml rather than 20 ml of pleural fluid. There is a linear relationship between the refractive index and the protein content of pleural fluid. Unfortunately, the scale on the commercially available refractometers is calibrated for the specific gravity of urine rather than pleural fluid. A reading of 1.020 on the urine specific
PLEURAL EFFUSIONS
1343
gravity scale corresponds to a protein of 3.0 gm per 100 ml, and each deviation of 0.004 ml represents 1 gram of protein. For example, a refractometer reading of 1.028 on the urine specific gravity scale corresponds to a pleural fluid protein of 5 gm per 100 ml. If the above facts are kept in mind, the refractometer is the simplest way to get an estimate of the pleural fluid protein. 4
TESTS FOR DIFFERENTIATING VARIOUS EXUDATIVE EFFUSIONS Appearance of the Fluid The gross appearance of the pleural fluid frequently yields useful diagnostic information. The color, turbidity, viscosity, and odor should be described. Most transudative and many exudative pleural effusions are clear, straw-colored, quite fluid, and odorless. Any deviations are noteworthy. A white, milky appearance indicates a chylothorax, a chyliform pleural effusion, or a pyothorax. The pyothorax can be distinguished from the chylothorax and chyliform effusion in that after centrifugation there is a clear supernatant only with pyothorax. A feculant odor is highly suggestive of an empyema secondary to anaerobic organisms. A clear or bloody fluid that is quite viscous is highly suggestive of malignant mesothelioma; the high viscosity is secondary to elevated pleural fluid hyaluronic acid levels."7 The fluid from frank empyemas is viscid and opaque. Red Cell Count- Bloody Pleural Effusion It requires only 5,000 to 10,000 red blood cells per cu mm to impart a red color to a pleural effusion. Assuming that a given pleural effusion has a total volume of 500 cc and the red cell count in the peripheral blood is 5 million per cu mm, a leak of only 1 cc of blood into the pleural space at the time of thoracentesis will result in a blood-tinged pleural effusion. For this reason, the mere fact that a pleural effusion is bloodtinged has limited diagnostic implications. Over 15 per cent of transudative and over 40 per cent of all types of exudative pleural fluids will be blood-tinged, Le., pleural fluid red cell counts between 5,000 and 100,000 per cu mm!! The question as to whether the blood was introduced via the thoracentesis or had been there previously can frequently be resolved with a Wright stain of the sediment. If there has been blood in the pleural space for a matter of hours, the macrophages in the pleural fluid will contain hemoglobin inclusion bodies that will stain pink. Grossly bloody pleural effusions have pleural fluid red cell counts above 100,000 per cu mm. This finding is very suggestive of one of three disease processes: trauma, malignancy, or pulmonary embolization. 21
1344
RICHARD
W.
LIGHT
White Cell Count The pleural fluid white blood cell count is of limited value. A pleural fluid white cell count of 1,000 per cu mm roughly separates transudative from exudative pleural effusions, but not nearly as effectively as do the protein and lactic dehydrogenase. Pleural fluid white cell counts above 10,000 per cu mm are most common with parapneumonic effusions but are also seen with pancreatitis, pulmonary infarction, collagen vascular diseases, malignancy, and tuberculosis. 2! Differential White Cell Count Examination of a Wright stain of pleural fluid is one of the most useful tests to be performed on pleural fluid. Eight different types of cells should be looked for in the pleural fluid: mesothelial cells, macrophages, plasma cells, lymphocytes, polymorphonuclear neutrophils, eosinophils, basophils, and malignant cells. 3 ! The stain is made most easily by centrifuging about 10 cc of fluid and then resuspending the button of cells in about 0.5 cc of supernatant. Slides of the pleural fluid are then made and stained just as are slides for peripheral blood smears. Occasionally, large amounts of fibrinogen adhere to the cells. In this case, resuspension in saline followed by centrifugation is indicated in order to evaluate cellular morphology. MESOTHELIAL CELLS AND MACROPHAGES. Mesothelial cells3 ! are the cells which line the pleural cavities. They are usually 12 to 30 microns in diameter, but multinucleated forms may have diameters up to 75 microns. The cytoplasm is light blue and is homogeneous. The nucleus stains purplish, is relatively large, is round and has a uniform appearance. There are usually 1 to 3 blue nucleoli. Collections of mesothelial cells (sometimes up to 12 or more) are occasionally seen adhering to one another. Macrophages, by definition, are cells which store vital dyes. The nuclei of macrophages are irregular, and the chromatin is lacy rather than stippled. The cytoplasm is gray, cloudy, and full of vacuoles. When there is active phagocytosis, these macrophages often contain red blood cells and neutrophils in various stages of digestion. Most pleural fluid macrophages probably evolve from mesothelial cells, although some probably evolve from circulating monocytes. Mesothelial cells are significant in two different respects. (1) They may be confused with malignant cells; frequently an experienced pathologist is required to make this differentiation. (2) Their presence or absence is often useful diagnostically. Mesothelial cells are uncommon in tuberculous effusions. Spriggs and Boddington3 ! analyzed 65 tuberculous effusions and found that only one had more than one mesothelial cell per 1,000 cells. They concluded that the presence of numerous mesothelial cells nearly excludes a diagnosis of tuberculosis. This has subsequently been confirmed by other workers.21. 35 However, this finding is not diagnostic of tuberculosis, as a paucity of mesothelial cells also occurs with empyema and some cases of malignancy. PLASMA CELLS. Plasma cells are cells with eccentric nuclei, deeply basophilic cytoplasm, and a clear area at the center of the cell. The
PLEURAL EFFUSIONS
1345
nuclear chromatin is usually clumped. The presence of numerous plasma cells suggests multiple myeloma, but in a study of 16 cases of pleural effusions with plasmacytosis greater than 5 per cent, their presence was of no diagnostic importance. Four were associated with malignancies, 3 with tuberculosis, 3 with pulmonary infarcts, 2 with pneumonia, 1 with rheumatoid arthritis, 1 with sepsis, and the basis for one was unknown. 31 LYMPHOCYTES. The discovery that more than 50 per cent of the white cells in pleural fluid are smalllymphocytes is important diagnostically. In two series,21. 35 96 of 211 exudative pleural effusions had more than 50 per cent smalllymphocytes. Of these 96 effusions, 90 were due to tuberculosis or malignancy. Since these are the two diseases that can be diagnosed with pleural biopsy, the finding of predominantly small lymphocytes in an exudative pleural effusion is a strong indication for pleural biopsy. When the same two series are combined, almost all the effusions secondary to tuberculosis (43/46) but only about half the effusions secondary to malignancy (47/90) had predominantly smalllymphocytes. POL YMORPHONUCLEAR LEUKOCYTES. Since neutrophils are the cellular component of the acute inflammatory response, they predominate in pleural fluid, resulting from acute inflammation of the pleura. Examples are effusions associated with pneumonia, pancreatitis, pulmonary emboli, peritonitis, and early tuberculosis. Some neutrophils are found in almost every pleural effusion; they appear in enormous numbers in empyemas. In this case, the neutrophils degenerate and their nuclei become blurred and lose their normal purple-staining characteristics. Their cytoplasm shows toxic granulation and fat vacuolation, and their normal granulation is lost. In all other effusions, the neutrophils appear much as they do in the peripheral blood. If neutrophils predominate in an effusion attributed to congestive heart failure, the possibility of pulmonary emboli should be investigated. EOSINOPHILS. The presence of pleural fluid eosinophilia is of very little use in the differential diagnosis of pleural effusionY Most effusions with significant eosinophilia (>10 per cent) are either bloody or associated with a pneumothorax. If the effusion is not bloody and there is no accompanying pneumothorax, the most likely etiology is a viral pleuritis or a resolving parapneumonic effusion. The presence of pleural fluid eosinophilia in a parapneumonic effusion is a good prognostic sign as these effusions virtually never become purulent. BASOPHILS. Basophilic pleural effusions are uncommon. There are usually a few basophils in eosinophilic pleural effusions; their presence has no known diagnostic significance. Protein and Lactic Dehydrogenase In general, measurements of the pleural fluid protein and lactic dehydrogenase are not useful in differentiating various types of exudative pleural effusions. IS All exudates tend to contain increased amounts of both. However, if the lactic dehydrogenase is elevated and the protein is not, the effusion is probably due to malignancy. Conversely, most pa-
1346
RICHARD
W.
LIGHT
tients with pleural fluid proteins above 6.0 gm per 100 ml have tuberculous or parapneumonic effusions. The measurement of isoenzymes of lactic dehydrogenase in pleural fluid is also of limited value. The pleural fluid LDH isoenzyme pattern for all benign exudates is characterized by a higher percentage of LDH4 and LDH-5 in the pleural fluid than in the serum. Most malignant exudates have this same pattern but a few have predominantly LDH-2. The presence of the latter pattern is highly suggestive that the patient has a malignant pleural effusion. 20 The other situation in which the LDH isoenzymes are of value is with bloody pleural effusions. At times when there is hemolysis, the pleural fluid lactic dehydrogenase will be elevated from the hemolysis even though the effusion is a transudate. Since most of the lactic dehydrogenase secondary to hemolysis is LDH1, subtraction of the LDH-1 from the total lactic dehydrogenase will give a pleural fluid LDH which can be used in differentiating transudates from exudates. It should be emphasized that the LDH-1 is not elevated in most bloody pleural effusions. 20 Simultaneous protein electrophoresis of the pleural fluid and serum is without value in the differential diagnosis of pleural effusions because both the serum and the pleural fluid have the same pattern. Likewise, measurements of mucoprotein or immunoglobulin levels in pleural effusions are not useful diagnostically.3o.36
Pleural Fluid Cytology A pleural fluid specimen from every patient with pleural effusion should be sent for cytopathologic studies. The first pleural fluid cytology will be positive for malignant cells in about 60 per cent of the effusions secondary to pleural malignancy.21 If three separate specimens are submitted, up to 90 per cent of those associated with pleural malignancy will have positive cytopathology. The later samples add diagnostic information because they contain fresher cells, since the older degenerated cells were removed during the earlier thoracenteses. Two points should be emphasized about pleural fluid cytology.31 (1) Malignancies can cause pleural effusions by mechanisms other than direct pleural involvement, e.g., lymphatic or bronchial obstruction and hypoproteinemia. Therefore, not all pleural effusions associated with malignancies have positive pleural fluid cytopathology. (2) Because of the marked variation that mesothelial cells undergo in response to inflammation, pleural fluid cytology is difficult. All cases in which the results are questionable should be reviewed by an experienced cytopatholo gis t. Glucose The pleural fluid glucose level of transudates and most exudates parallels that of the serum. However, there are four categories of exudative pleural effusions in which the pleural fluid glucose may be low «60 mg per 100 ml).22 1. Parapneumonic effusions. The lower the glucose is in this situation, the higher the probability that the effusion is infected and that in-
PLEURAL EFFUSIONS
1347
sertion of chest tubes will be necessary for its resolution. However, a decreased pH is more sensitive than a decreased glucose in predicting which parapneumonic effusions will become complicated (see below). 2. Rheumatoid pleural effusion. Carr and Powers first made the observation that pleural effusions secondary to rheumatoid disease have a low glucose concentration. In reviewing 74 pleural effusions in which glucose measurements had been made, they found that 5 had a glucose level of less than 30 mg per 100 ml; 4 of these 5 were secondary to rheumatoid disease. The diagnosis of pleural effusion secondary to rheumatoid disease is doubtful if the pleural fluid glucose is above 30 mg per 100 ml. The explanation for the low pleural fluid glucose level in this condition appears to be a selective block to the entry of glucose into the pleural effusionY' Interestingly, the pleural fluid glucose level is not depressed with systemic lupus erythematosus. 3. Tuberculous pleural effusion. Although an occasional tuberculous pleural effusion will have a very low pleural fluid glucose level,3 the mean glucose level in these effusions is about 80 mg per 100 ml,22 Therefore, the pleural fluid glucose is usually not helpful in making the diagnosis of tuberculous pleuritis. 4. Malignant pleural effusion. In approximately 15 per cent of malignant effusions, the pleural fluid glucose is below 60 mg per 100 ml. 22 Malignant pleural effusions with low pleural fluid glucose levels have one of two characteristics: they have myriads of malignant cells in the fluid 7 or they are very large, occupying the entire hemithorax. 22 Amylase An amylase determination should be made on every pleural fluid examined. Pancreatitis is complicated by pleural effusions nearly 10 per cent of the time.!'] In effusions secondary to pancreatic disease, an elevated pleural fluid amylase may be the first clue to the diagnosis. 22 There are only two other conditions in which the pleural fluid amylase is elevated. In pleural effusions secondary to esophageal rupture, the amylase is markedly elevated. The origin of the amylase is salivary rather than pancreatic in this situation. 28 The saliva enters the pleural space through the tear in the esophagus. The pleural fluid amylase is elevated in approximately 10 per cent of patients with malignant pleural effusions. 22 The primary tumor in such instances is usually not in the pancreas.
Pleural Fluid pH The pleural fluid pH is a very useful prognostic index in patients with acute bacterial pneumonia and pleural effusion (parapneumonic effusion). A low pleural fluid pH «7.20) in this situation, strongly suggests that the effusion will not resolve withou t chest tubes. HI, 26 Since drainage of the pleural space can become very difficult if loculation develops, we recommend thoracentesis and measurement of pleural fluid pH in all patients with parapneumonic effusions as soon as the effusion is recognized. If the pleural fluid pH is below 7.20, and the arterial pH is above 7.35, tube thoracostomy should be instituted immediately. If the
1348
RICHARD
W.
LIGHT
arterial pH is below 7.35, chest tubes should be placed only if the pleural fluid pH is more than 0.15 units lower than the arterial pH. If the pleural fluid pH is between 7.20 and 7.30, daily thoracentesis with pleural fluid pH measurements should be performed until the pleural fluid pH approaches 7.30. If the pleural fluid pH falls to or below 7.20, tube thoracostomy should be instituted. In this situation the pleural fluid pH becomes lowered before the pleural fluid glucose falls or before organisms become visible on Gram stain. 19 The use of a low pleural fluid pH as an indication for chest tube placement is applicable only with parapneumonic effusions. With all transudative and most exudative pleural effusions, the pleural fluid pH parallels that of arterial blood. However, the pleural fluid pH is less than 7.20 in most rheumatoid pleural effusions and some effusions caused by malignancy or tuberculosis, none of which should necessarily be treated with tube thoracostomy.19 It should be emphasized that the determination of pleural fluid pH must be made with the same care as with arterial pH. The fluid must be withdrawn anaerobically into a heparinized syringe and maintained at 0° C until the pH is measured. Cultures and Bacteriologic Stains When a diagnostic thoracentesis is performed, the fluid obtained should be cultured for both anaerobic and aerobic organisms. Mycobacterial and fungal cultures should also be obtained. The chances for obtaining a positive mycobacterial culture are enhanced by a generous specimen. A Gram stain of the fluid should be examined to rule out bacterial infection. Since the probability of detecting tubercular organisms with an acid-fast stain for tube,rcle bacilli is minimal even in tuberculous effusions, the author employs this stain on the sediment only if mesothelial cells compose less than 1 per cent of the total cells on the Wright stain. Other Tests Lupus ERYTHEMATOSUS CELLS. Pleural fluid LE preparations should be obtained on all patients who have puzzling exudative effusions. When the effusion is secondary to lupus erythematosus, the pleural fluid will usually contain LE cells sometimes even when there are no LE cells in the peripheral blood. LE cells in the pleural fluid are thought to be pathognomonic of systemic lupus erythematosus. 34 COMPLEMENT. Low levels of complement in pleural fluid are found in effusions caused by either rheumatoid pleuritis or lupus erythematosus. Hunder et al. 12 measured the pleural fluid complement levels of 50 patients with pleural effusions of various etiologies. The pleural fluid complement was low «10 units per ml) in 11 of 12 patients with systemic lupus erythematosus or rheumatoid pleuritis but exceeded this level in 37 of 38 patients with other diseases. With the collagen vascular diseases, the pleural fluid complement levels are depressed substantially more than the serum complement levels. If collagen vascular disease is suspected, pleural fluid complement levels should be obtained.
1349
PLEURAL EFFUSIONS
Table 2. DESTINATION
Processing Pleural Fluid AMOUNT
TESTS ORDERED
Bacteriology
5 cc
Bacterial cultures Gram Stain
Tuberculosis and mycology
5 cc
Tuberculosis fungal cultures Acid-fast bacillus stain
Cytology
5 cc
Cytology
Hematology
5 cc
Red cell count White cell count Wright stain
Chemistry
5 cc
Glucose Amylase Lactic dehydrogenase Protein
Blood gas laboratory
5 cc
pH
RHEUMATOID FACTOR- Although the rheumatoid factor is high in effusions secondary to rheumatoid arthritis, it is also frequently elevated in exudates secondary to pneumonia, tuberculosis, and carcinoma. Therefore, it is not useful diagnostically,11 HYALURONIC ACID. A pleural fluid hyaluronic acid level is useful in the diagnosis of malignant mesothelioma. Only patients with mesotheliomas have a markedly elevated hyaluronic acid level. Unfortunately, not all patients with mesothelioma have elevated levels of pleural fluid hyaluronic acid.27 LIPID ANALYSIS. Pleural fluid is occasionally found to be milky or opalescent. If the fluid remains cloudy after centrifugation, the patient has either a chylous or a chyliform pleural effusion. A chylous pleural effusion arises when the thoracic duct is severed or obstructed. The cloudiness in this situation is caused by chylomicra and lipid analysis reveals high levels of triglycerides but low levels of cholesterol. A chyliform (cholesterol) pleural effusion develops when a pleural effusion has been present for a long time, and for unknown reasons cholesterol accumulates. The fluid in this situation is also milky, but pleural fluid lipid analysis reveals elevated cholesterol and normal triglycerides. 2 " CHROMOSOME ANALYSIS. There are abnormalities in the number and structure of the chromosomes in malignant pleural effusions."' 11 It appears that with the combination of cytology and chromosomal analysis, a definite diagnosis of malignancy can be established in more patients than with either method alone. However, since chromosomal analysis is not routinely available and has not yet been applied to large numbers of benign exudates, it should still be regarded as experimental. CARCINOEMBRYONIC ANTIGENS. A preliminary report'4 on 6 patients has shown that the pleural fluid from some patients with malignancy has a high level of carcinoembryonic antigen-like substances (CEA-LS), while the CEA-LS is not elevated in benign effusions. If this
1350
RICHARD
W.
LIGHT
can be confirmed, measurement of pleural fluid CEA-LS will be useful. However, at this time it must still be regarded as experimental.
PLEURAL BIOPSY Samples of the parietal pleura can easily be obtained with a Cope or Abrams pleural biopsy needle. Some authors have advocated pleural biopsy each time a thoracentesis is performed. However, pleural biopsy is associated with greater morbidity than is a simple diagnostic thoracentesis, and is really only useful when the patient has granulomatous disease or malignancy of the pleura. Therefore, I only recommend pleural biopsy if granulomatous disease or malignancy is suspected. There are two contraindications to pleural biopsy: a bleeding diathesis and an empyema. In order to do a pleural biopsy, the prothrombin time should be at least 50 per cent of the control value, the platelet count should be greater than 100,000, and the bleeding and clotting times should be normal. The percentage of specific diagnoses obtained with pleural biopsy depends on patient selection. If several bits of parietal pleura are obtained on three separate occasions, tuberculosis can be diagnosed in over 80 per cent and malignancy in over 50 per cent.23 If culture of a bit of the biopsy material is combined with pleural biopsy, the diagnosis of tuberculosis can be made in nearly 95 per cent of the cases. 16 Since pleural fluid cultures are positive in less than 25 per cent of patients with tuberculous pleuritis, culture of a pleural biopsy specimen should definitely be done.
SUMMARY Many different conditions result in the accumulation of pleural fluid. A diagnostic thoracentesis should be performed on all patients with pleural effusion from whom pleural fluid can be easily obtained. Empirically we have found that when the pleural effusion is more than 10 mm thick on the lateral decubitus roentgenogram, pleural fluid is easily obtained. At least 30 cc fluid should be obtained and distributed to the various laboratories as outlined in Table 2. The results of these tests will show whether the fluid is a transudate or an exudate. If the fluid is a transudate, no further diagnostic procedures need be directed towards the pleura. If the fluid is an exudate, the diagnosis will frequently be made by these original tests and therapy for the pleural disease can be instituted. If the diagnosis has not been made, the results of these tests should lead to a rational diagnostic attack.
1351
PLEURAL EFFUSIONS
REFERENCES 1. Agostoni, E., Taglietti, A., and Setnikar, I.: Absorption force of the capillaries of the visceral pleura in determination of the intrapleural pressure. Amer. J. PhysioI., 191 :277282,1957. 2. Agostoni, E.: Mechanics of the pleural space. Physiol. Rev., 52:57-128,1972. 3. Barber, L. M., Mazzadi, L., Deakins, D. D., et aI.: Glucose level in pleural fluid as a diagnostic aid. Dis. Chest, 31 :680-687, 1957. 4. Briggs, M. S., George, R. B., and Light, R. W.: Use of the refractive index to estimate protein concentration in pleural fluid. Clin. Res., 25:36A, 1977. 5. Carr, D. T., and Power, M. H.: Clinical value of measurements of concentration of protein in pleural fluid. New Eng. J. Med., 259:926-927, 1958. 6. Carr, D. T., and Power, M. H.: Pleural fluid glucose with special reference to its concentration in rheumatoid pleurisy with effusion. Dis. Chest, 37 :321-324, 1960. 7. Clarkson, B.: Relationship between cell type, glucose concentration and response to treatment in neoplastic effusions. Cancer, 17:914-928, 1964. 8. Courtice, F., and Simmonds, W.: Absorption of fluids from the pleural cavities of rabbits and cats. J. PhysioI., 109:117-130, 1949. 9. Dewald, G., Dines, D. E., Weiland, L. H., et al.: Usefulness of chromosome examination in the diagnosis of malignant pleural effusions. New Eng. J. Med., 295:1494-1500, 1976. 10. Dodson, W., and Hollingsworth, J.: Pleural effusion in rheumatoid arthritis: Impaired transport of glucose. New Eng. J. Med., 275:1337-1342, 1966. 11. Hansson, A., and Korsgaard, R.: Cytogenetical diagnosis of malignant pleural effusions. Scand. J. Resp. Dis., 55:301-308,1974. 12. Hunder, G. G., McDuffie, F. C., and Hepper, N. G.: Pleural fluid complement in systemic lupus erythematosus and rheumatoid arthritis. Ann. Intern, Med., 76:357-363, 1972, 13. Kaye, M.: Pleuropulmonary complications of pancreatitis. Thorax, 23:297-306,1968. 14. Kim, Y. D., and Hirata, A. A.: Carcinoemhryonic antigen-like substances in human cavity fluids. ImmunoI. Commun., 5:619-629,1976. 15. Leuallen, E., and Carr, D.: Pleural Effusion: A statistical study of 436 patients. New Eng. J. Med., 252 :79-83, 1955. 16. Levine, H., Metzger, W., Lacera, D., et a!.: Diagnosis of tuberculous pleurisy by culture of pleural biopsy specimen. Arch. Intern. Med., 126:269-271, 1970. 17. Levine, H., Szanto, M., Grieble, H., et al.: Rheumatoid factor in nonrheumatoid pleural effusions. Ann. Intern. Med., 69:487-492,1968. 18. Light, R. W., MacGregor, M. I., Luchsinger, P. C., et aI.: Pleural effusions: The diagnostic separation of transudates and exudates. Ann. Intern. Med., 77:507-514, 1972. 19. Light, R. W., MacGregor, M. I., Ball, W. C., Jr., et al.: Diagnostic significance of pleural fluid pH and Pco 2 . Chest, 64:591-596, 1973. 20. Light, R. W., and Ball, W. C., Jr.: Lactate dehydrogenase isoenzymes in pleural effusion. Amer. Rev. Resp. Dis., 108:660-665,1973. 21. Light, R. W., Erozan, Y. C., Ball, W. C. Jr.: Cells in pleural fluid: Their value in differential diagnosis. Arch. Intern. Med., 132 :854-860, 1973. 22. Light, R. W., and Ball, W. C. Jr.: Glucose and amylase in pleural effusions. J.A.M.A., 225:257-260, 1973. 23. Mestitz, P., Purves, M., and Pollard, A.: Pleural biopsy in the diagnosis of pleural effusion: A report of 200 cases. Lancet, 2: 1349-1353, 1958. 24. Paddock, F.: The relationship between the specific gravity and the protein content in human serous effusions. Amer. J. Med. ScL, 201 :569-574, 1941. 25. Paddock, F. K.: The diagnostic significance of serous fluids in disease. New Eng. J. Med., 223:1010-1015, 1940. 26. Potts, D. E., Levin, D. C., and Sahn, S. A.: Pleural fluid pH in parapneumonic effusions. Chest, 70 :328-331, 1976. 27. Rasmussen, L., and Faber, V.: Hyaluronic acid in 247 pleural fluids. Scand. J. Resp. Dis., 48:366-371,1967. 28. Sherr, H. P., Light, R. W., Merson, M. H., et al.: Origin of pleural fluid amylase in esophageal rupture. Ann. Intern. Med., 76:985-986, 1972. 29. Seriff, N. S., Cohen, M. L., Samuel, P., et al.: Chylothorax: diagnosis by lipoprotein electrophoresis of serum and pleural fluid. Thorax, 32:98-100, 1977. 30. Shallenberger, D. W., and Daniel, T. M.: Quantitative determinations of several pleural fluid proteins. Amer. Rev. Resp. Dis., 106: 121-122, 1972. 31. Spriggs, A., and Boddington, M.: The Cytology of Effusions. New York, Grune & Stratton, Inc., 1968. 32. Stewart, P. B.: The rate of formation and lymphatic removal of fluid in pleural effusions. J. Clin. Invest., 42:258-262, 1963.
1352
RICHARD
W.
LIGHT
33. Stewart P., and Burgen, A.: The turnover offiuid in the dog's pleural cavity. J. Lab. Clin. Med., 52:212-230,1958. 34. Winslow, W. A., Ploss, L. N., and Loitman, B.: Pleuritis in systemic lupus erythematosus: its importance as an early manifestation in diagnosis. Ann. Intern. Med., 49:70-88, 1958. 35. Yam, L.: Diagnostic significance of lymphocytes in pleural effusions. Ann. Intern. Med., 66:972-982, 1967. 36. Zinneman, H. H., Johnson, J J., and Lyon, R. H.: Proteins and mucoproteins in pleural effusions. Amer. Rev. Tuberc., 76:247-255,1957. Department of Medicine Louisiana State University Medical Center P.O. Box 33932 Shreveport, Louisiana 71130
