FEVER AND PLEURAL EFFUSIONS
Anatomy and Pathophysiology of the Pleural Space
Evaluation of Pleural Effusions
Tests of Occasional Value
Pleural Biopsy and/or Thoracoscopy
Types of Pleural Effusion
Iatrogenic Pleural Effusions
Pleural effusions associated with fever constitute a common medical problem. The clinician must have a thorough understanding of the pathophysiology of pleural effusions and a realistic approach to diagnosis and treatment if optimal treatment is to be provided.
Anatomy and Pathophysiology of the Pleural Space
The pleural space, truly a “potential space” formed at the interface of the parietal and visceral pleurae, acts as a lubricant and normally contains only 7 to 14 mL of fluid. The area becomes a true space in disease states, when it may fill with air or fluid. Blood supply to the parietal pleura comes primarily from branches of the intercostal and superior phrenic arteries, whereas the visceral pleura is supplied by both pulmonary and pericardiophrenic arteries. Venous drainage of the parietal pleura is via the intercostal veins; the visceral pleura is drained primarily by pulmonary veins.
Pleural lymphatics, located in the connective tissues that underlie the mesothelial cells of the pleural surfaces, freely interconnect with those below the diaphragm. Materials placed in subdiaphragmatic lymphatics drain into the intercostal and mediastinal nodes. Drainage is extremely important for removal of erythrocytes and proteins from the pleural space. Gases and liquids are rapidly cleared from the pleural space. Particulate matter, including erythrocytes and proteins, is removed primarily through the lymphatics. Normally, 250 to 500 mL of fluids and contained materials can be removed daily in this fashion. Pleural effusions develop when discrepancies develop between the rates of production and absorption of fluid. Causes include disorders of hydrostatic or colloid oncotic pressure, lymphatics, and capillary permeability.
Pleural fluid is normally sterile, but it easily supports growth of pathogens. This is in part related to the fluid basis of effusion, which allows extreme mobility of bacteria and thus impairs early phagocytosis before the arrival of opsonins. Recently, reviews have documented the pathophysiology of empyema. The presence of bacteria within pleural effusions initiates a variety of host responses that involve cytokines. If the response fails to inhibit bacterial growth, opsonins and complement become deficient, and the fluid becomes hypoxic and acidic. The inflammatory process typical of empyema releases components capable of bacterial inhibition. In such a state, bacterial reproduction slows and may be reduced to every 24 hours. This may explain in part why antibiotics need to be administered for prolonged periods in patients treated for undrained empyema.
Evaluation of Pleural Effusions
History and physical examination provide important clues for both presence and etiology. Questions regarding the presence of pneumonia, subdiaphragmatic illness, or malignancy, and medicine use and epidemiology (e.g., travel and tuberculosis exposure) are indicated. Specific risks for HIV infection should be assessed, as recent literature has summarized some characteristic experiences with pleural effusions in this disease. Physical examination should be comprehensive. Patients housed in adult critical care units are often identified to have pleural effusions. Most are small and need not be sampled unless infection or another specific diagnosis is strongly suspected.
The chest roentgenogram is often insensitive (even with lateral decubitus views) until at least 500 mL of fluid is present. Ultrasonography may reveal as little as 5 to 50 mL, whereas computed tomography has the advantage of additionally revealing details of the pulmonary parenchyma and mediastinum and is better at differentiating pleural thickening from fluid. Either study may be employed to guide thoracentesis in selected cases.
Thoracentesis is indicated to assess the cause of most effusions and provides a diagnosis in about 75% of cases. There are no absolute contraindications, although bleeding tendencies, poor patient cooperation, and mechanical ventilation are considered relative contraindications. Well-documented congestive heart failure or generalized anasarca may be managed without this study. Occasionally, because of small size or difficult location, thoracentesis employing either ultrasound or computed tomographic guidance may be indicated. Recent prospective studies demonstrate that this procedure provides useful information more than 90% of the time but may be associated with both technical problems and adverse reactions. Table 6-1 lists studies to be considered once fluid has been obtained.
Table 6-1. Common tests useful in evaluating thoracentesis fluid
Pleural effusions may be exudative or transudative. The latter type tends to be benign and occurs when mechanical factors alter pleural fluid formation or resorption. Thus, the identification of transudative fluid generally truncates the evaluation. Patient position may modify fluid characteristics. An upright position may result in documentation of exudates, whereas borderline transudates may have been noted while the patient is supine. Exudates result from inflammation or malignancy that interferes with pleural surfaces or lymphatic drainage. Differentiation is important because of the broad types of illness that fall into the two categories. Some diagnostic categories may be either; significant diuresis may alter pleural fluid protein and lactate dehydrogenase (LDH) levels so that the fluid mimics exudate. Simultaneous measurements of serum and pleural fluid LDH and protein content allow for more accurate assessment. A pleural fluid-to-serum LDH ratio above 0.6 or protein ratio above 0.5, or a pleural fluid LDH level above 200 IU/L, generally documents an exudate. Virtually all exudates exhibit at least one of these characteristics, and transudates typically lack all three. Pleural fluid cholesterol above 45 mg/dL appears specific for exudative effusion but may not be as sensitive as the prior criteria. The combination of pleural fluid cholesterol above 45 mg/dL and LDH above 200 IU/L appears to be both highly sensitive and specific for exudate and does not require simultaneous blood sampling.
Pleural fluid glucose levels below 40 mg/100 mL are generally seen in effusions caused by bacterial infection, tuberculosis, malignancy, or rheumatoid arthritis. The value of low glucose levels lies more in documenting the need for further evaluation than in providing a specific diagnosis. Pleural fluid amylase levels are elevated in pancreatitis and esophageal rupture, and with amylase-producing tumors. Amylase may be further divided into that of salivary or pancreatic origin. Markedly elevated levels of pancreatic origin almost always result from pancreatitis and are usually associated with a pleural fluid-serum amylase ratio above 1. Esophageal rupture may be suspected by the presence of amylase of salivary origin.
Pleural fluid pH is useful in defining parapneumonic effusions that must be treated by tube thoracostomy. Low pleural fluid pH values occur primarily in malignancy, tuberculosis, and bacterial infections. Pleural fluid acidosis is defined by a pH below 7.3 or (if acidemia is present) by a value more than 0.15 below that of blood pH. Some authors now recommend immediate chest tube drainage for parapneumonic effusions associated with pH below 7.1, glucose below 40 mg/dL, LDH above 1,000 IU/L, or evidence of loculation. Parapneumonic effusions with pH values above 7.1, especially if accompanied by glucose levels above 40 mg/100 mL, may be successfully treated without tube placement. Repeated thoracentesis to document trends is recommended
Exudates often contain more than 1,000 cells per milliliter, but this parameter is not as useful as LDH and protein ratios for distinguishing between exudates and transudates. An elevated percentage of polymorphonuclear leukocytes, seen primarily in bacterial infections, is also noted in pancreatitis, connective tissue disease, and pulmonary infarction. Tuberculosis of short duration has been associated with this cell type as well. A predominance of lymphocytes is seen in more than 80% of tuberculous and malignant pleural effusions. Tuberculosis is also associated with a relative absence of mesothelial cells in the differential count. Newer tests to define lymphocyte type further may be important in defining the etiology of pleural effusions. As an example, recent studies have demonstrated elevated levels of helper T cells in a patient with pleural effusion resulting from sarcoidosis.
Eosinophilic pleural effusions contain more than 10% eosinophils and may account for 2% to 9% of all pleural effusions. A recent investigation documented only idiopathic cases and effusions following thoracic surgery as statistically correlated with eosinophilia, and survivorship was longer in patients with eosinophilic effusions.
Erythrocytes are common but of uncertain importance. Historically, the presence of blood-tinged fluid was considered indicative of tuberculosis, pulmonary infarction, or malignancy. However, less than 2 mL of blood per 1,000 mL of fluid in an effusion creates this appearance. An RBC level of more than 100,000/mL is generally associated with malignancy, trauma, and pulmonary infarction.
Malignant cells should be sought in exudative effusions not otherwise diagnosed. Cytologic examination establishes a diagnosis in about 50% of malignant pleural effusions. Fresh samples must be used, and several techniques may be necessary to demonstrate malignant cells.
All pleural effusions should be analyzed by Gram’s stain and aerobic and anaerobic culture. Although only 5% of bacterial pneumonias are complicated by infected pleural fluid, laboratory information is valuable. Not all effusions need to be evaluated for tuberculosis or fungal infection. Smears for acid-fast bacilli (AFB) and fungal smears and cultures should be obtained with lymphocytic exudative effusions or when the diagnosis of an exudate is elusive. Smears for AFB are positive in fewer than 25% of cases and increase both patient and laboratory costs.
Tests of Occasional Value
Pleural effusions in systemic lupus erythematosus may be associated with antinuclear antibody titers above 1:160 and pleural fluid-to-serum antinuclear antibody ratios of more than 1. Positive lupus erythematosus preparations are noted in at least 85% of cases. Determination of levels of both complement and rheumatoid factor in pleural fluid may help diagnose rheumatoid arthritis. Measurement of adenine deaminase has become an effective means of diagnosing tuberculous pleural effusion, and in the presence of a lymphocytic effusion, an elevation is virtually pathognomonic. Investigators have considered an elevation in pleural effusion to levels above 45 to 55 U/L to be highly suggestive of tuberculosis.
Tests for the detection of bacterial antigens may document a bacterial etiology in the absence of viable organisms. Thus, partially treated infections may be diagnosed despite negative results on cultures. Organisms detectable include Haemophilus influenzae type b, Streptococcus pneumoniae, and several types of Neisseria meningitidis. Results with these techniques have been comparable to those of routine cultures. Other studies of occasional value include neutrophil elastase and a1-proteinase inhibitor (malignancy) and flow cytometry with immunochemistry (malignancy).
Pleural Biopsy and/or Thoracoscopy
Thoracoscopy, often with pleural biopsy, should be performed in difficult cases of exudative pleural effusion. Thoracoscopy has become increasingly popular and now can be performed under video guidance, which allows better visibility and fewer complications. Most authors will employ it if routine pleural biopsy findings are nondiagnostic. A recent review demonstrated that more than 90% of cases of elusive exudative effusions can be diagnosed by the use of this procedure. It is especially useful for finding nodular pleural lesions, which can then be sampled. Biopsy specimens should be submitted for bacteriology (aerobic/anaerobic, mycobacteriology, and mycology) and histopathology.
Despite best efforts, for some patients (especially those with exudative pleural effusions), an etiologic diagnosis cannot be made. Most of these cases are benign and remain undiagnosed. The mean time to resolution is 5 to 6 months; however, some cases relapse. Etiologies eventually discovered in a minority of these patients include asbestosis, rheumatoid arthritis, congestive heart failure, cirrhosis, and adenocarcinoma.
Types of Pleural Effusion
Table 6-2 lists some of the common causes of fever and pleural effusion. Selected ones are discussed below.
Table 6-2. Common pleural effusions associated with fever
Parapneumonic effusion is that associated with pneumonia, lung abscess, or bronchiectasis. Between 30% and 70% are associated with positive results on pleural fluid cultures. Regarding bacterial pneumonia, the likelihood of encountering parapneumonic effusions is as follows: Staphylococcus aureus, 75%; S. pneumoniae, 57%; viruses, 15% to 25%; H. influenzae, 50% to 75%; and Streptococcus pyogenes, 90%. The yield may be related to the duration of effusion, as organisms such as S. pneumoniae may undergo autolysis. Mycoplasma pneumoniae may be associated with parapneumonic effusions in up to 20% of cases. The relative frequency with Legionella species or gram-negative enteric bacilli remains unknown.
The management of parapneumonic effusions includes thoracentesis and antimicrobials. b-Lactams penetrate the pleural space well and achieve therapeutic levels early in therapy. Levels of parenteral aminoglycosides are decreased in the face of empyema in comparison with other effusions. As mentioned earlier, characteristics of the fluid help determine the need for tube or other forms of drainage.
Pleural empyema is defined as pus in the pleural space and can be demonstrated only by direct sampling. Often, pH values below 7 and glucose levels under 40 mg/100 mL are observed. The presence of pleural empyema requires parenteral antimicrobials and definitive drainage. Although tube thoracostomy has been traditionally employed, some persons may now benefit from thoracoscopy with repeated irrigations. Anaerobes may be noted in up to 40% of cases. Gram’s stain, culture, and other standard tests usually provide information sufficient to initiate therapy. Antimicrobial doses higher than those commonly used for uncomplicated pneumonia are necessary to ensure adequate drug levels. Length of therapy is variable, but generally therapy should be continued until the patient is afebrile, the peripheral WBC count approaches normal, and tube thoracostomy drainage is meager. If complicated pneumonia or lung abscess is simultaneously present, prolonged therapy may be needed. Patients who fail to defervesce with appropriate antimicrobials and closed tube thoracostomy should be evaluated for loculated pus. Either ultrasound or computed tomography can be used, and open surgical drainage or thoracoscopy may be indicated.
Involvement of the pleural space occurs in about 4% of patients diagnosed with tuberculous and most commonly is noted as an early complication of primary disease. Presentation may be acute and mimic bacterial pneumonia, or more chronic, in which case it is characterized by weight loss and anorexia. Approximately 75% to 80% of patients with tuberculous pleurisy are febrile. Thirty percent have simultaneous pulmonary parenchymal involvement. For obscure reasons, pleural effusions most commonly involve the right hemithorax and almost invariably are unilateral. Lymphocytes generally predominate; however, early in the course of disease polymorphonuclear leukocytes may be noted. Glucose levels may be normal or low, and AFB smears are generally negative. Cultures of pleural effusion are positive in about 50% of cases. Diagnosis should be suspected in lymphocyte-predominant exudative effusions, and pleural biopsy is generally the procedure of choice if the AFB smear of fluid is initially negative. Recommendations are for submission of three biopsy specimens, as yield from a pleural fluid increases from about 70% to above 90% with additional specimens. Adenosine deaminase has been recommended as a test for tuberculous pleurisy. In some hands, levels above 50 U/L were more than 90% sensitive and specific for tuberculosis, whereas levels below 45 U/L were 100% specific and sensitive for alternative diagnoses. This test should be performed when tuberculous pleurisy is suspected.
Fever occurs in up to 68% of angiographically documented cases of pulmonary thromboembolic disease, may reach levels of 39°C, and can last for many days. Observations suggesting pulmonary thromboembolic disease include (a) a history of embolic events, (b) fever, and (c) phlebitis. Pleural fluid evaluation is often nondiagnostic. In 33% of cases, fluid is transudative and contains fewer than 10,000 RBCs per milliliter. RBC counts above 100,000/mL suggest this diagnosis if trauma and malignancy can be excluded. WBC counts can reach 70,000/mL. Early on, polymorphonuclear leukocytes predominate, and lymphocytes are noted after several days.
Iatrogenic Pleural Effusions
Causes of iatrogenic pleural effusions include drugs (e.g., heparin, hydralazine, sulfa drugs, nitrofurantoin, albumin, ionic contrast dye) and procedures (e.g., sclerotherapy, surgery, misadventures with central intravascular lines, and peritoneal dialysis). All can be associated with fever. The diagnostic approach is similar to that with other effusions, and discontinuation of an offending medication often results in clinical improvement.
Pleural effusions associated with fever are a common problem and may have many causes. The physician should have a working knowledge of the mechanisms involved in the formation of fluid and be comfortable using tests available for diagnosis. In selected situations, small effusions that have been incidentally identified may be observed. When pleural fluid has been sampled, hematologic, chemical, and microbiologic studies generally provide a diagnosis. Occasionally, cases prove more frustrating, and biopsy or other analyses may be indicated. The cause of the effusion may remain elusive on occasion, and repeated assessments may be necessary. Fortunately, many cases of chronic exudative undiagnosed effusion appear to be benign. (R.B.B.)
Bartter T, et al. The evaluation of pleural effusion. Chest 1994; 106:1209–1214.
The authors present an excellent review of the roles of imaging, thoracentesis, and other studies in the evaluation of pleural effusions. They also provide a good framework for differentiating exudates from transudates and for the management of exudative pleural effusions.
Black LF. The pleural space and pleural fluid. Mayo Clin Proc 1972;47:493–506.
This article, now more than two decades old, is a superb review of the anatomy and physiology of the pleural space. Although somewhat technical, it provides an excellent basis for the understanding of pleural disease.
Bryant RE, Salmon CJ. Pleural empyema. Clin Infect Dis 1996;22:747–764.
This article is an excellent in-depth review of the history, pathophysiology, anatomy, diagnosis, and management of pleural empyema. It contains contemporary information about the role of intrapleural thrombolysis and video-assisted thoracoscopy. Recommendations regarding antibiotic therapy are basic and do not really address the role of newer agents, which may have a role for prolonged oral therapy in selected cases.
Collins TR, Sahn SA. Thoracentesis: clinical value, complications, technical problems, and patient experience. Chest 1987;91:817–822.
Eighty-nine patients undergoing 129 consecutive thoracenteses were evaluated. Ninety-two percent of procedures provided useful information. Twenty percent of procedures were associated with complications that included pneumothorax and cough. Subjective patient discomfort was seen in more than 20% of cases, and technical problems were encountered in more than 20% of cases.
Ferrer SJ. Pleural tuberculosis: incidence, pathogenesis, diagnosis and treatment. Opin Pulmon Med 1996;2:327–334.
The author presents an excellent overview of issues related to pleural tuberculosis and deals with the issue in patients with HIV/AIDS as well. Therapy is primarily with antituberculous agents, with very limited roles for either corticosteroids or repeated thoracenteses.
Ferrer JS, et al. Evolution of idiopathic pleural effusion. Chest 1996;109:1508–1513.
This report of 40 patients followed for as much as 10 years demonstrates that many patients with exudative pleural effusions and no specific diagnosis did well. Mean time to resolution was less than 6 months, and the course of most patients was benign. One of the entry criteria for this study was an adenosine deaminase level below 43 IU/L, which in the opinion of the authors was valuable for ruling out pleural tuberculosis. Despite long-term follow-up, in 80% a diagnosis was never obtained; most of the remainder had nonmalignant conditions.
Harris RJ, et al. The diagnostic and therapeutic utility of thoracoscopy. Chest 1995;108:828–841.
An excellent review of the historical and current uses of thoracoscopy as a modality for diagnosing and treating pleural disease. This technique is actually not new, but it has undergone a renaissance in part because of the addition of video assistance, which allows easier imaging of the pleural space. The use of this technique needs to be studied better in controlled trials so that overuse will be avoided. However, it does appear to be extremely valuable as a tool to recognize specific intrapleural lesions.
Henschke CI, et al. Pleural effusions: pathogenesis, radiographic evaluation, and therapy. J Thorac Imaging 1989;4:49–60.
This excellent overview of the radiographic evaluation of pleural effusions describes clinical conditions that mimic effusions, difficulties with loculated collections, and clues to the presence of empyema. The roles of ultrasound, computed tomography, and magnetic resonance imaging are also discussed. The latter modalities are useful in distinguishing pleural from parenchymal disease, and magnetic resonance imaging may prove beneficial in distinguishing etiologies of effusions.
Leslie WK, Kinasewitz GT. Clinical characteristics of the patient with nonspecific pleuritis. Chest 1988;94:603–608.
This retrospective analysis of 119 patients who underwent pleural biopsy identifies variables associated with malignant or granulomatous disease. Patients with a diagnosis of nonspecific pleuritis can be managed conservatively if weight loss, a positive tuberculin test result, lymphocytosis above 95%, and a fluid level above half of the hemithorax are not demonstrated.
Light RW, et al. Parapneumonic effusions. Am J Med 1980;69:507–512.
The authors prospectively assessed 90 patients with parapneumonic effusion. A glucose level below 40 mg/100 mL or a pH below 7 predicted complicated effusions and indicated the need for tube thoracostomy. Patients with a pH above 7.2 and a pleural fluid LDH level below 1,000 mg/100 mL only rarely have a complicated course.
Mattison LE, et al. Pleural effusions in the medical ICU: prevalence, causes, and clinical implications. Chest 1997;111: 1018–1023.
The investigators assessed 100 patients admitted to a medical ICU. Of these, 62% were documented to have pleural effusions; about two thirds of these were present on admission. Most were small and of no clinical significance. Most were present on chest roentgenograms. If not clinically suspected to be infected, the authors feel that most of these can be observed prospectively without thoracentesis.
Poe RH, et al. Utility of pleural fluid analysis in predicting tube thoracostomy/decortication in parapneumonic effusions. Chest 1991;100:963–967.
The authors retrospectively evaluated 133 patients at three hospitals who underwent thoracentesis. Assessment included laboratory data from effusions and ultimate need for surgical drainage or decortication. They concluded that Light’s standard criteria for drainage (purulence, glucose 1,000 IU/L, pH <7) are specific but not sensitive in predicting the need for eventual chest tube drainage and decortication. Patients who do not meet the criteria must still be carefully followed.
Rubins JB, Rubins HB. Etiology and prognostic significance of eosinophilic pleural effusions. Chest 1996;110:1271–1274.
This is an interesting brief report of more than 470 patients with pleural effusions, of which almost 10% were eosinophilic. The authors conclude that the only statistical significance of eosinophilia in pleural effusions was either after thoracic surgery or in idiopathic cases. No correlations with malignancy were noted, and eosinophilic effusions generally resulted in more prolonged survival than others.
Sahn SA. The differential diagnosis of pleural effusions. West J Med 1982;13:99–108.
This excellent basic article by one of the giants in the field of pleural effusion reviews physiology, thoracentesis, and analysis of pleural effusion. Charts are provided to distinguish among causes of pleural effusion based on laboratory characteristics.
Trejo O, et al. Pleural effusion in patients infected with the human immunodeficiency virus. Eur J Clin Microbiol Infect Dis 1997;16:807–815.
A cohort of HIV-positive patients with pleural effusions was compared with a similar number of patients without HIV infection but with documented pleural effusion—either parapneumonic effusion or tuberculosis. Most HIV-positive persons were intravenous drug users, and this population had a high incidence of infection as a cause of the effusion. Tests demonstrated similar results between the two groups for a given diagnosis.