Pulmonary embolism is a common disorder associated. Biochemical and Cytologic Characteristics of Pleural Effusions Secondary to Pulmonary Embolism*

Transcription

1 Biochemical and Cytologic Characteristics of Pleural Effusions Secondary to Pulmonary Embolism* Santiago Romero Candeira, MD; Luis Hernández Blasco, MD; Maria J. Soler, MD; Alejandro Muñoz, MD; and Ignacio Aranda, MD Study objectives: To characterize the biochemical and cytologic constituents of pleural effusions secondary to pulmonary embolism. Design: A descriptive clinical study. Setting: A community teaching hospital with 750 beds, which acts as a tertiary referral center for several subspecialties. Patients and interventions: Patients with pleural effusions secondary to pulmonary embolism who underwent diagnostic thoracentesis during the last 7 years were retrospectively studied. Pleural fluid mesothelial hyperplasia was revised and compared with that found in patients with pleural effusions of different origin. Results: Pleural effusions from all 60 patients with pulmonary embolism were exudates, and in 40 patients (67%) contained erythrocyte counts > 10,000/ L. A bloody appearance was not related to the use of anticoagulant therapy before thoracentesis. Polymorphonuclear leukocytes were predominant in 36 patients (60%); in 11 patients (18%), a proportion of eosinophils > 10% was found. Mesothelial hyperplasia was significantly higher in patients with pulmonary embolism than in patients in the control group (p < 0.01). Conclusions: In the absence of trauma, a bloody or eosinophilic effusion with a marked mesothelial hyperplasia should prompt a workup to rule out embolism. The finding of transudative pleural fluid chemistries in these patients should not be assumed to be secondary to embolism before ruling out other causes of transudative effusion. (CHEST 2002; 121: ) Key words: pleural effusion; pulmonary embolism; mesothelial hyperplasia Abbreviations: CA carbohydrate antigen; CEA carcinoembryonic antigen; LDH lactate dehydrogenase *From the Servicios de Neumología y Anatomía Patológica (Dr. Aranda), Hospital General Universitario de Alicante (Drs. Romero Candiera, Hernández Blasco, Soler, and Muñoz), Alicante, Spain. Manuscript received March 15, 2001; revision accepted August 16, Correspondence to: Santiago Romero Candeira, C/Italia, No. 30, Esc 2A, 1 DCHA, Alicante, Spain; romero_san@gva.es Pulmonary embolism is a common disorder associated with considerable morbidity and mortality. The overall mortality rate for patients with untreated pulmonary embolism is approximately 30%; however, if the diagnosis is made promptly and the appropriate therapy is instituted, the mortality rate can be reduced to 10%. 1,2 Therefore, to identify its clinical manifestations seems important to improve the prognosis. Although it has been estimated that pleural effusions occur in 30 to 50% of patients with pulmonary embolism, 3 in most large series, pulmonary embolism accounts for 5% of the pleural effusions. This underestimated incidence probably occurs because the diagnosis of pulmonary embolism is frequently not considered in patients with undiagnosed pleural effusions and, because the size of the effusion is usually small, patients with pulmonary embolism do not always undergo thoracentesis. Moreover, in these patients, the analysis of the pleural fluid has not been considered helpful for establishing the diagnosis, because it can have the features of either a transudate or an exudate, 4 and it is grossly bloody in only a minority of cases. 5 However, in patients suspected of having pulmonary embolism, a thoracentesis has been considered indicated to exclude other causes of pleural effusion, such as tuberculosis, malignant disease, or pneumonia with a parapneumonic effusion. 5 Moreover, a therapeutic thoracentesis has been recommended prior to obtaining a lung scan whenever feasible. 5 During the last 4 years, 226 patients received a CHEST / 121 / 2/ FEBRUARY,

2 diagnosis of pulmonary embolism at our hospital. Ninety-nine patients (44%) had radiographic evidence of pleural effusions, and 41 patients underwent diagnostic thoracentesis. The uniform exudative features found in the pleural fluid of those patients, together with the absence of large, recent, and complete reports about pleural fluid findings in patients with pulmonary embolism, have prompted us to report the results of the present series that, as far as we know, is the largest of those published. The aims of this study were to characterize the chemical and cytologic constituents of these effusions in a substantial number of well-demonstrated cases. moderate [26 to 50%], and severe [ 50%]). Pleural fluid from 60 consecutive patients with pleural effusions of different etiologies (15 neoplastic, 15 traumatic, 15 parapneumonic, and 15 tuberculosis) were analyzed following similar norms, and the results were used as controls. Statistical Analysis The data are reported as mean SD (95% confidence intervals of mean) or median (range) depending on their distribution. Differences in quantitative variables between groups were assessed by means of unpaired t test or the Mann-Whitney U test, when appropriate. The 2 test was used to assess differences in categorical variables between groups. Values of p 0.05 were considered as significant. Materials and Methods We retrospectively reviewed the medical records and chest radiographs of all patients with a definitive diagnosis of pulmonary embolism who underwent diagnostic thoracentesis during the last 7 years. The chest radiographs were reviewed to assess the size and location of the pleural effusion, and the associated findings. A definitive diagnosis of pulmonary embolism was considered when an abnormal angiographic finding (pulmonary angiography or contrast-enhanced spiral CT scanning) showed distinct filling defects or sharp arterial cutoffs (35 patients), or when there was a high clinical suspicion together with a highprobability ventilation-perfusion scan (19 patients), or the demonstration of deep vein thrombosis by ultrasonography or venography (6 patients). Thoracentesis was performed following norms previously communicated. 6 In patients with bilateral pleural effusions, the thoracentesis was performed in the hemithorax with the larger pleural effusion. In anticoagulated patients, heparin was discontinued for at least 4 h before the thoracentesis. The following studies were performed on pleural fluid samples: ph, glucose, protein, lactate dehydrogenase (LDH), cholesterol, amylase, carcinoembryonic antigen (CEA), carbohydrate antigen (CA) 15 3, differential cell count, bacterial and fungal culture, acid-fast bacillus smear and culture, and cytology. Only the results of the first thoracentesis were considered. Simultaneously to the thoracentesis, a sample of serum was obtained to measure glucose, protein, LDH, and cholesterol levels. Biochemical parameters were determined using methods previously described. 7 Pleural fluid WBC and RBC counts were performed manually in a Neubauer camera on pleural fluid anticoagulated with ethylenediaminetetra-acetic acid (Neubauer Improved; Marienfeld; Lauda-Königshofen, Germany), using prestained (methyleneblue-cresyl violet acetate), ready-to-use slides (Boehringer Mannheim UK, Ltd; Lewes, East Sussex, UK). Criteria used for separating exudates from transudates were those of Light et al. 8 Tumor markers (CEA and CA 15 3) were measured following norms previously published. Malignancy cutoffs of 5 ng/ml for CEA and 25 ng/ml for CA 15 3 were used. 9 Pleural fluid cytologic samples were routinely prepared with a cytocentrifuge. The slides were immediately fixed with alcohol ether solution and stained with the Papanicolaou and hematoxylin-eosin method. To evaluate mesothelial hyperplasia, all available cytologic samples were revised by two observers, blind of any clinical information, using a two-headed microscope. Cytologic smears were allocated a semiquantitative mesothelial score based on the profusion and arrangement of mesothelial cells, according to the percentage of total cellularity (absent or mild [ 25%], Results Sixty patients (37 men and 23 women; mean age, years; range, 31 to 88 years) complied with the conditions of the study. Twenty-five patients (42%) were smokers (mean, 42 pack/years). One or more associated diseases were evident in 47 patients (78%; Table 1). No patients presented with signs of congestive heart failure when thoracentesis was performed; however, seven patients were receiving diuretic therapy for hypertension or previous congestive heart failure. Pleural effusions were bilateral in eight patients (13%); 29 effusions (48%) were right sided and 23 effusions (38%) were left sided. The pleural effusion occupied less than one third of the hemithorax in 48 patients (80%), and from one third to one half of the hemithorax in the remaining 12 patients. A radiographic pulmonary infiltrate suggestive of pulmonary infarction 10 was evident in eight patients (13%). Table 1 Demographic and Clinical Characteristics of 60 Patients With Pulmonary Embolism* Characteristics Data Age, yr Men/women 37/23 Smokers, No. (%) 25 (42) Mean pack/years 42 Underlying illenss Surgery/traumatic 16 Diabetes 12 Chronic venous insufficiency 9 Heart disease Ischemic or hypertensive 11 Valvular 4 COPD 5 Collagen-vascular disease 4 Liver cirrhosis 1 No previous disease 13 Receiving diuretic therapy 7 *Data are presented as mean SD or No. unless otherwise indicated. 466 Clinical Investigations

3 The average lag time between initial presentation to diagnosis of pulmonary embolism was days, and to thoracentesis was days. The average lag time between the diagnosis of pulmonary embolism and time to thoracentesis was 2 3 days. Pleural fluid was considered blood tinged or bloody in 34 patients (57%) for its macroscopic aspect. The results of pleural fluid determinations are showed in Table 2. The pleural fluid RBC count was 10,000/ L in 20 patients (33%) and was 100,000/ L in 11 patients (18%). Pleural fluid was considered blood tinged in one patient with an erythrocyte count 10,000/ L; conversely, six effusions with erythrocyte counts 10,000/ L were considered straw-yellow. At the moment of the thoracentesis, 29 patients (48%) were receiving anticoagulant therapy (heparin [n 27] and a derivative of acenocumarol [n 2]). The mean pleural fluid RBC count in this group was lower (55, ,734/ L) than in patients not receiving anticoagulants (107, ,813/ L), but the difference was not significant (p 0.3). Four of 29 patients (15%) receiving anticoagulant therapy and 7 of 31 patients (22%) without this type of therapy had 100,000 RBCs in pleural fluid (p 0.72). Five of the eight patients with radiographic evidence of pulmonary infarction had erythrocyte counts 10,000/ L. The differential WBC count revealed predominantly polymorphonuclear leukocytes in 36 patients (60%), while lymphocytes were predominant in 24 patients. Eleven patients (18%) presented with pleural fluid eosinophilia ( 10% eosinophils). Five of 10 eosinophilic effusions had erythrocyte counts 10,000/ L. Pleural effusions from all 60 patients were exudates. False-positive pleural fluid levels were found in 1 patient for CEA and in 9 of 50 patients (18%) for CA Microbiologic studies in pleural fluid, with findings always negative, were performed in 55 patients. Five patients underwent pleural biopsy with nonspecific findings. In 51 patients, cytologic smears were technically adequate to be subjected to evaluation. The mesothelial score was higher in patients with pulmonary embolism than in patients with tuberculosis (p 0.005), malignancy (p 0.04), and the total control group (p 0.01; Table 3). Three patients presented a self-limited vasovagal reflex during thoracentesis. No other complications occurred. Discussion This study demonstrates the uniform exudative character of pleural effusions due to pulmonary embolism. Until 1976, most authors classified effusions due to pulmonary embolism as exudates However, Bynum and Wilson 4 published that year a study in which more than one third of the 26 patients included could have transudates. Since then, this study has been repeatedly referred to and, as far as we know, never contested in spite of significant methodologic limitations: first, in an undetermined number of patients, pleural fluid was obtained for examination by tube thoracostomy or open thoracotomy, procedures that may change the composition of pleural fluid; second, some criteria used to consider a pleural effusion as an exudate (pleural fluid protein value 3 g/dl, or a specific gravity 1.016) are quite dated, and in no patient a simultaneous serum specimen was obtained for examination; third, because sufficient pleural fluid was not always available, only in 12 of the 26 patients studied was the LDH measured, and only in 9 patients was the specific gravity determined. An increase of the systemic venous pressure at the Table 2 Pleural Fluid Constituents in 60 Patients* Pleural Fluid Constituents Data Erythrocytes 26,600 (20 1,150,000) Leukocytes 3,200 (190 45,000) Proteins ( ) LDH (228 2,783) Glucose 106 (11 364) Cholesterol (82 101) Triglycerides (31 39) Amylase (n 54) 42 (10 276) CEA (n 51) 0.9 ( ) CA 15.3 (n 50) (4 132) ph (n 47) ( ) *Unless otherwise indicated, data are presented as median (range) or mean SD (95% confidence interval). Table 3 Pleural Fluid Mesothelial Scores in Pulmonary Embolism and Control Groups* Variables Hyperplasia Pulmonary embolism 11 (22) 11 (22) 29 (56) (n 51) Control group (n 60) 24 (40) 20 (33) 16 (27) Neoplasm (n 15) 5 (33) 7 (47) 3 (20) Traumatic (n 15) 5 (33) 3 (22) 7 (44) Parapneumonic (n 15) 6 (36) 4 (27) 5 (36) Tuberculosis (n 15) 8 (53) 6 (40) 1 (7) *Data are presented as No. (%). p p p with respect to pulmonary embolism. CHEST / 121 / 2/ FEBRUARY,

4 parietal pleural surface secondary to pulmonary hypertension and right ventricular failure has been considered the pathogenic mechanism responsible for transudates in pulmonary embolism. 5 The fact that patients with congestive heart failure demonstrate a close correlation between the presence of pleural effusion(s) and pulmonary venous pressure 15 suggests that the interstitial space of the lungs acts as the origin of this type of effusion in most patients rather than the parietal pleural surface. By interfering with lymphatic drainage from the lung, an elevated right atrial pressure may help sustain or worsen pleural effusions induced by elevated left atrial pressure, but not to cause pleural effusions in absence of left atrial hypertension. 16 However, pulmonary hypertension in patients with pleural effusion secondary to acute pulmonary embolism is usually not severe enough to cause right ventricular failure. 17 In fact, it has been suspected that most, if not all, transudates in patients with pulmonary embolism are caused by clinical unrecognized congestive heart failure. 18 The results of the present study confirm the changing character of the WBC (total number and differential pattern) in pleural effusions of pulmonary embolism and their frequent, although not invariable, bloody character. Pleural effusions with erythrocyte counts 10,000/ L were herein found in 67% of the patients, a proportion similar to that found by Bynum and Wilson. 4 However, in the present study, the macroscopic appearance was considered bloody in a lower proportion (56% of patients). The bloody appearance is one of the most characteristic features of this type of effusion, and in the absence of trauma or malignancy must prompt a workup to confirm or rule out the diagnosis of pulmonary embolism. However, the low sensitivity of this finding does not allow us to rule out the suspicion of embolism when a clear pleural fluid is found. The use of heparin therapy before performing the thoracentesis in some patients could change the initial characteristics of the effusion, because pleural hemorrhage secondary to anticoagulation has been reported in patients whose effusions progress during therapy. 19 However, a traumatic hemorrhage caused by the thoracentesis seems more likely in patients receiving anticoagulant therapy. In the present study, in which one half of the patients included were receiving heparin therapy when thoracentesis was performed, an influence of this therapy in the bloody character of the effusion could not be demonstrated. In fact, although not significantly, the number of erythrocytes was lower in patients receiving anticoagulant therapy. All of this suggests the beneficial therapeutic effect of heparin and demonstrates its exceptional bleeding effect over pulmonary infarction. Pulmonary embolism has been considered responsible for 4% of eosinophilic pleural effusions. 20 However, this percentage may underestimate the true incidence. In the same meta-analysis, 39% of the pleural fluid eosinophilias (78 of 200) were cryptogenetic, a group in which pulmonary embolism is presumed to be responsible for a substantial fraction of the effusions. 13 However, the frequency of pleural fluid eosinophilia in pleural effusions secondary to pulmonary embolism is largely unknown. Bynum and Wilson 4 found a predominance of polymorphonuclear leukocytes in pleural fluid in 61% of their 26 patients; however, they did not indicate the percentage of eosinophils. The number of effusions due to pulmonary embolism in series that compare the etiology of eosinophilic and not eosinophilic pleural effusions is small (four to nine patients), and the percentage of them with pleural fluid eosinophilia vary between 0% and 33% In the present study, a percentage of 18% was found and, given the larger number of patients studied, seems more realistic. A mesothelial reaction has been considered characteristic, or at least a frequent feature, in pleural effusions secondary to pulmonary embolism. The results of this study confirm that a mesothelial hyperplasia is a frequent feature of these kind of effusions and that its presence, mainly when the changes are severe, may help to distinguish pulmonary embolism from other disorders usually considered in the differential diagnosis. In conclusion, pleural effusions secondary to pulmonary embolism are always exudates, frequently hemorrhagic, and with a marked mesothelial hyperplasia. The presence of these characteristics in the absence of trauma or cytologic evidence of malignancy should prompt a workup to rule out pulmonary embolism. The finding of transudative pleural fluid chemistries in a patient with pulmonary embolism should not be assumed to be secondary to embolism before ruling out other causes of transudative effusion. Anticoagulant therapy is not associated to the bloody character of the effusion. References 1 Matsumoto AH, Tegtmeyer CJ. Contemporary diagnostic approaches to acute pulmonary embolism. Radiol Clin North Am 1995; 33: Dale JE, Alper JS. Natural history of pulmonary embolism. Prog Cardiovasc Dis 1975; 17: Worsley DF, Alavi A, Aronchick JM, et al. Chest radiographic findings in patients with acute pulmonary embolism: observation from the PIOPED study. Radiology 1993; 189: Clinical Investigations

5 4 Bynum LJ, Wilson JE. Characteristics of pleural effusions associated with pulmonary embolism. Arch Intern Med 1976; 136: Light RW. Pleural effusion due to pulmonary embolization. In: Pleural diseases, 3rd ed. Baltimore, MD: Williams & Wilkins, 1995; Romero-Candeira S, Fernández C, Martín C, et al. Influence of diuretics on the concentration of proteins and other components of pleural transudates in patients with heart failure. Am J Med 2001; 110: Romero S, Candela A, Martín C, et al. Evaluation of different criteria for the separation of pleural transudates from exudates. Chest 1993; 104: Light RW, Gregor MI, Luchsinger PC, et al. Pleural effusions: the diagnostic separation of transudates and exudates. Ann Intern Med 1972; 77: Romero S, Fernández C, Arriero JM, et al. CEA, CA 15 3 and CYFRA 21 1 in serum and pleural fluid of patients with pleural effusions. Eur Respir J 1996; 9: Shah AA, Davis SD, Gamsu G, et al. Parenchymal and pleural findings in patients with and patients without acute pulmonary embolism detected at spiral CT. Radiology 1999; 211: Hodgson CH. Pulmonary embolism and infarction. Dis Chest 1965; 47: Carr DT, Soule EH, Ellis FH. Management of pleural effusions. Med Clin North Am 1964; 48: Busey JF, Fenger EPK, Hepper NG, et al. Therapy of pleural effusion: a statement by the committee on therapy (American Thoracic Society). Am Rev Respir Dis 1968; 97: Light RW, Erozan, Ball WC. Cells in pleural fluid: their value in differential diagnosis. Arch Intern Med 1973; 132: Wiener-Kronish JP, Matthay MA, Callen PW, et al. Relationship of pleural effusions to pulmonary hemodynamics in patients with congestive heart failure. Am Rev Respir Dis 1985; 132: Szabo G, Magyar Z. Effect of systemic venous pressure on lymph pressure and flow. Am J Physiol 1967; 212: Stein PD, Athanasoulis C, Greenspan RH, et al. Relation of plain chest radiographic findings to pulmonary arterial pressure and arterial blood oxygen levels in patients with acute pulmonary embolism. Am J Cardiol 1992; 69: Fraser RS, Muller NL, Colman N, et al. Pleural effusion. In: Diagnosis of diseases of the chest. 4th ed. Philadelphia, PA: W.B. Saunders, 1999; Rostand RA, Feldman RL, Block ER. Massive hemothorax complicating heparin anticoagulation for pulmonary embolus. South Med J 1977; 70: Adelman M, Albelda SM, Gottlieb J, et al. Diagnostic utility of pleural fluid eosinophilia. Am J Med 1984; Díaz Nuevo G, Jimenez Castro D, Pérez-Rodriguez E, et al. Eosinofilia pleural: significado diagnóstico y pronóstico. Rev Clin Esp 1999; 199: Rubins JB, Bloomfield Rubins HB. Etiology and prognostic significance of eosinophilic pleural effusions: a prospective study. Chest 1996; 110: Martínez-García MA, Cases-Viedma E, Cordero-Rodriguez PJ, et al. Diagnostic utility of eosinophil in the pleural fluid. Eur Respir J 2000; 15: Kuhn M, Fitting JW, Lewenberger P. Probability of malignancy in pleural fluid eosinophilia. Chest 1989; 96: CHEST / 121 / 2/ FEBRUARY,