The Etiology of Pleural Effusions in an Area With High Incidence of Tuberculosis*

Transcription

1 The Etiology of Pleural Effusions in an Area With High Incidence of Tuberculosis* Luis Valdes, MD; David Alvarez, MD; Jose Manuel Valle, MD; Antonio Pose, MD; and Esther San Jose, MD To investigate the etiology of pleural effusions in our region, we undertook a prospective study of patients with this condition in our centers. During a 5-year period, we studied 642 pleural effusion patients aged 57.1::'::21.1 years,ofwhom401 were menaged56.5::'::21 years and 241 were women aged 57.8::'::21.4 years; the male/female ratio was 1.6:1. The most frequent cause of pleural effusion was tuberculosis (25% ), followed by neoplasia (22.9%) and congestive heart failure (17.9% ). The etiology of 48 cases (7.5%) remained uncertain. In the neoplastic effusion group, the most frequent locations of the primary tumor were lung (32.6% ), breast (11.5%), lymphoma (10.8%), and ovary (7.5%); in 21 cases (14.3% of the neoplastic group), it was not possible to identify the primary tumor. The Ill patients aged younger than 40 years with tuberculous effusions made up 69.4% of tuberculous effusion cases and the same percentage of patients younger than 40 years; the proportion. of effusions that were tuberculous peaked in the 11- to 30-year-old age group and declined steadily thereafter. Of the patients with neoplastic effusions, 83% were older than 50 years; the proportion of effusions that were neoplastic rose steadily from zero in the 0- to 30-year-old age group to a peak among 60- to 70-year-olds. The age-wise distribution of effusions secondary to congestive heart failure was similar to that of neoplastic effusions. Of the effusions secondary to congestive heart failure, 86% (99/115) affected the right pleura or both, and 83% of effusions secondary to pulmonary thromboembolism (15/18) affected the right side. Neoplastic, tuberculous, parapneumonic, empyematous, and other exudative effusions showed no preference for either side. Of the 97 bilateral effusions, 77 (79.4%) were secondary to heart failure (59, 60.8%) or neoplasia (18, 18.6%). We conclude that in our region, the most frequent cause of pleural effusion is tuberculosis, followed by neoplasia and congestive heart failure. We suggest that all those interested in pleural disease should detennine the etiologic pattern of pleural effusion in their region with a view to the adoption of regionally optimized diagnostic and therapeutic attitudes. (CHEST 1996; 109:158-62) I CHF=congestive heart failure Key words: congestive heart failure; pleural effusions; tuberculosis The difficulty in determining the cause of a pleural effusion, which is shown by the "unknown etiology" rates of up to 20% in some published case series, 1 2 is largely' due to the great variety of diseases that can bring about this condition. 3 The etiologic distribution of pleural effusions depends on geographic region, patient age, and advances in the diagnosis and treatment of the underlying causes. The distribution that is observed and reported in any one study may also depend on the type of subject covered by that study (all primary care patients in a given region; patients referred to specialistpneumologic services; etc). In this work, we investigated the etiology of pleural effusions among patients referred to the Pneumology and Internal Medicine Services of two hospitals in Santiago de Compostela (NW Spain) with a view to establishing *From the Seccion de Neumologfa (Drs. Valdes, Alvarez, and Valle), SeiVicio de Medicina Interna (Dr. Pose), and SeiVicio de Bioqufmica (Dr. San Tose), Hospital de Corum, Compleio Hospitalano U niversitario d'e Santiago, Santiago de Composte1a, Spam. Manuscript received February ~ 41995;. revision accepted August 3. Reprint Requests: Dr. Valdes, C/. Emilia Pardo Bazan, 26, Urbanizaci6n Parque Montouto, 15894, Teo, La Coruiia, Spain. locally optimized diagnostic and therapeutic attitudes. Those studied included practically all patients with pleural effusion referred to specialists in a region with a population of approximately 530,000 and a high incidence of tuberculosis. MATERIALS AND METHODS Following a predefined protocol, between January l, 1989 and December 31, 1993, we studied all patients in the Pneumology and Internal Medicine SeiVices of the Hospital de Conxo and Hospital Xeral de Galicia (Santiago de Compostela, Spain) who during this period were admitted because of pleural effusion or had pleural effusion diagnosed following admission for some other cause. Only patients with a sure diagnosis of pleural effusion on the basis of clinical, analytical, radiologic, and anatomicopathologic findings were included in the study. A detailed history was obtained from each patient, and thoracentesis4 was performed on each following physical examination. If at this point the history and clinical and biochemical findings were suggestive of a transudate, 5 the appropriate therapy was begun and no further diagnostic tests were performed. If any of the findings were suggestive of an exudate, pleural biopsy was performed with an Abrams or Cope needle. 6 7 No patient underwent thoracoscopy. The causes of the pleural effusions were diagnosed in accordance with the following criteria: congestive heart failure (CHF)---com- 158 Clinical Investigations

2 patible clinical and radiologic findings, remission upon appropriate treatment, and absence of pulmonary inhltrates, chest pain, thrombophlebitis, and purulent sputum; parapneumonic origin-association with pneumonia, lung abscess, or bronchiectasis; empyemapresence of purulent fluid or positive culture of parapneumonic effusion; tuberculosis-presence of acid-fast bacilli in pleural fluid or biopsy tissue or presence of necrotic caseous granulomas in biopsy tissue; neoplasia-neoplastic tissue in the pleural cavity (biopsy ancl!or cytology); bilateral effusions were deemed neoplastic only if neoplastic tissue was found on both sides; pulmonary thromboembolism-angiographic evidence or perfusion gammagraphy showing deficient perfusion affecting at least the whole lobe 8 ; miscellaneous transudates and exudates-cirrhosis of the liver, nephrotic syndrome, hypoalbuminemia, massive atelectasis, superior vena cava syndrome, Meigs's syndrome, pancreatitis, systemic lupus erythematosus, chylothorax, hemothorax, rheumatoid arthritis, subphrenic pus collection, pleuropericarditis, yellow nail syndrome, hypothyroidism, and benign asbestosis were also identified as causes of pleural effusions following preestablished criteria 9 ; and effusions appearing within the 72 h following abdominal surgery when chest radiographs had been normal at admission were classified as of postsurgical origin. Pleural effusions whose causes were not established by any of the diagnostic tests (including three pleural biopsies) were classified as idiopathic. RESULTS Of the 642 patients aged 57.1±21.1 years who satisfied the criteria for inclusion in the study, 401 (62.5%) were men aged 56.5±21 years and 241 (37.5%) were women aged 57.8±21.4 years (male/female ratio: 1.6). The percentage of patients in each etiologic group is listed in Table 1 together with the size and age of the group. The most frequent cause of pleural effusion was tuberculosis (25% of all cases), followed by cancer (22.9%) and CHF (17.9%). In 48 cases (7.5%), all involving exudates, we were unable to identify the cause of the effusion. Table 2 lists the primary tumors responsible for neoplastic effusions. The most common, responsible for 32.6% of all ntoplastic effusions, was bronchogenic carcinoma, followed by breast cancer ( 11.5%) and lymphoma (10.8%). In 21 cases (14.3%), we were unable to identify the primary tumor. Table 3 lists the etiologies of the miscellaneous exudates and transudates. Of the 39 miscellaneous exudates, 15 (38.4%) were of postsurgical origin. Nine of the miscellaneous transudates (36%) were due to cirrhosis of the liver. Tuberculous effusions were most frequent in the first four decades (111/160, 69.4%) and were the most frequent type of pleural effusion (111, 69.4%) among the 160 patients in this age range. Neoplastic effusions were most common among the elderly, 83% (122/147) of the patients with neoplastic effusions being older than 50 years. The age-wise distributions of tuberculous and neoplastic effusions are shown in Figure 1: the number of tuberculous effusions peaked in the 11- to 30-year-old period, fell sharply in the fourth decade, and declined steadily thereafter until a small rise in the Table!-Etiologic Distribution of Pleural Effusions Kind of Effusion No. Age, yr, or Cause (%) mean::'::sd Tuberculosis 160 (25) 34::'::18.4 Neoplasia 147 (22.9) 65::'::14 CHF 115 (17.9) 71.9::'::9.4 Parapneumonic 75 (11.7) 69.5::'::17.2 PTE* 18 (2.8) 64.4::'::17.5 Empyema 15 (2.3) 53::'::18 Other exudates 39(6) 53::'::16.8 Other transudates 25(3.9) 62.7 ::':: 13.6 Idiopathic 48 (7.5) 66.8::'::18.1 Total 642 (100) 57.1::'::21.1 *PTE=pulmonary thromboembolism. eighth decade; neoplastic effusions first appeared in the fourth decade and rose thereafter to a peak in the eighth. Effusions secondary to CHF showed an agewise distribution similar to that of neoplastic effusions, first appearing in the fifth decade and peaking in the eighth, in which 53% (61/115) of this kind of effusion occurred. Of the patients aged 71 to 80 years, 60% had effusions due to either CHF or neoplasia. In 300 patients (46.7%), pleural effusion occurred only on the right side, in 245 (38.2%) only on the left, and in 78 (15.1%) both sides were affected (Table 4). Of the 115 effusions due to CHF, 99 (86%) were either on the right side or bilateral, and 83% (15/18) of the effusions due to pulmonary thromboembolism were on the right side. Neoplastic, parapneumonic, empyematous, and other exudates showed no preference for either side. Of the 97 bilateral effusions, 77 (79.4%) were due to CHF or neoplasia. DISCUSSION In this study of patients referred to the Pneumology Table 2-Location or Type of Primary Tumors Associated With Neoplastic Pleural Effusion Origin Lung Breast Lymphoma Ovary Stomach Colon Mesothelioma Chronic lymphatic leukemia Esophagus Liver Myeloma Kaposi's sarcoma Neurofibroma Pancreas Kidney Uterus Thymoma Undetermined Total No.(%) 48 (32.6) 17 (11.5) 16 (10.8) 11 (7.5) 7(4.7) 5 (3.4) 5 (3.4) 3 (2) 21 (14.3) 147 (100) CHEST I 109 I 1 I JANUARY,

3 Table 3-Etiologic Distributions of Other Exudates and Other Transudates* Other Exudates Other Transudates (n=39) No. (n=25) 0. Postsurgery 15 Liver cirrhosis 9 Pancreatitis 6 Nephrotic syndrome 6 SLE 4 Hypoalbuminemia 4 Chylothorax 3 Massive a telectasis 3 Hemothorax 2 Sup vena cava S 2 Rheumatoid arthritis 2 Meigs's syndrome 1 Subphrenic pus collection 2 Pleuropericarditis 2 Yellow nail syndrome 1 Hypothyroidism 1 Benign asbestosis 1 *SLE=systemic lupus erythematosus; Sup vena cava S=superior vena cava syndrome. and Internal Medicine Services of our hospitals, the most frequent cause of pleural effusions was tuberculosis, which accounted for 25% of all cases, followed by cancer (22.9%) and CHF (17.9% ). The greater frequency of tuberculous effusions was even more marked among patients younger than 40 years, 69% of these cases being due to tuberculosis. The etiologic distribution of pleural effusions that is Table 4-Left!Right Distribution of Pleural Effusions in Each Etiologic Group* Bilateral, 0. (%) Total CHF 40 (34.8) 16 (13.9) 59 (51.3) 115 NEO 65 (44.2) 64 (43.5) 18 (12.2) 147 TB 81 (50.6) 76 (47.5) 3 (1.8) 160 PAR 37 (49.3) 35 (46.6) 3 (4) 75 EMP 7 (46.6) 8 (53.3) 15 PTE 15 (83.3) 3 (16.6) 18 OEX 17 (43.5) 16 (41) 6 (15.3) 39 OTRA 15 (60) 3 (12) 7 (28) 25 IDIO 23 (47.9) 24 (50) 1 (2) 48 Total 300 (46.7) 245 (38.2) 97 (15.1) 642 *NEO =neoplastic; TB=tuberculous; PAR=parapneumonics; EMP= empyema; PTE=pulmonary thromboembolism; OEX=other exudates; OTRA=other transudates; IDIO=idiopathic observed in any study evidently depends on the type of subject covered by the study, on the etiologic classification used, and on the diagnostic methods employed. Marel et al 10 involved primary care practitioners in a study aimed at determining the incidence and etiologic distribution of pleural effusion in the general population of a well-defined geographic region. One of their main conclusions was that the percentage of cases with no identifiable cause (0% in their case) must be much n > 80 (years) n TUBERCULOUS 0 NEOPLASTIC FIGU RE l. Age-wise distributions of patie nts with tube rculous and neoplastic pleural effusions. 160 Clinical Investigations

4 smaller in general population studies than in studies of patients referred to specialists. In our study, 7.5% of the cases seen were idiopathic, which is in keeping with the prediction of Marel et al 10 but is low in comparison with some other published case series The statistics of Marel et al 10 included effusions detected only on autopsy. This increased the observed incidence to 0.32% (from 0.21% calculated on the basis of effusions detected in vivo), but analysis of their findings does not suggest that the noninclusion of postmortem results introduces significant bias in the etiologic distributions observed in other studies. According to some sources, 3 10 CHF is the most frequent cause of pleural effusion, but in several studies, CHF has accounted for only 6 to 16% of cases; in ours, it accounted for 17.9%. These discrepancies are attributable to differences regarding the class of patient considered, to the adoption of a more or less conservative diagnostic attitude in clear cases ofchf, and to interregional differences regarding incidence of the causes of pleural effusion. In particular, the incidence of tuberculosis in our region is very high, 95/100,000 inhabitants. 13 In this study, pleural effusion due to CHF affected the left side in 13.9% of cases, the right in 34.8%, and was bilateral in 51.3%. These results are similar to those of Martfnez-Berganza and Cfa 14 but differ from those of Race et al, 15 who reported that 88% of these effusions were bilateral, only 8% being limited to the right side and 4% to the left. The ratio of men to women among effusions due to CHF was 2.1, whereas more women than men were affected in the study by Martfnez-Berganza and Cfa.l 4 Neoplastic effusions were the second most common type in our study (22.9% ), and the first most common (28.8%) among patients older than 50 years. The prevalence of neoplastic effusions in other published case series ranges from 14.7% 11 to 48%.1 The five types of cancer that most frequently metastasized in the pleura (Table 2) were the same as reported in the review by Sahn 16 of nine published case series with a total of 1,783 cases. As expected, these effusions showed no preference for either side and were the type of exudate with the highest proportion of bilateral cases (18/31, 58%). The ratio of male to female patients with neoplastic effusion was slightly smaller than in the whole series, 1.4. The most common cause of pleural effusion in our study was tuberculosis (25%). The mean age of these patients was 34± 18.4 years, the male/female ratio was 1.5, and unlike some other authors who have observed most tuberculosis effusions on the right side, 17 we found no such preference (though bilateral cases were rare). The finding that tuberculosis was the most common cause of effusion in our stud;;; is in keeping with the results of Fontan Bueso et al 8 concerning a population with the same characteristics as ours. In general, the proportion of pleural effusions due to tuberculosis varies widely, depending on the local incidence of tuberculosis. Marel et al 10 saw no tuberculous case in their study of a population in Central Bohemia (Czech Republic) with an incidence of tuberculosis of 18/100,000, and Storey et al 1 found only 1 among the 108 effusions whose cause they identified in a series of 133 at the Mayo Clinic, but 14 of 103 exudates were tuberculous in the study by Light et al 19 of a Baltimore suburb with a high incidence of tuberculosis, and Chretien et al 12 reported 30% of pleural effusions to be tuberculous in a study carried out at the Laennec Hospital, Paris. In keeping with the above-mentioned average age of the tuberculous effusion patients in this study, the frequency of these effusions was much greater, 69%, among patients younger than 40 years than among older patients. This high incidence among young patients, together with the high diagnostic efficiency of interferon gamma and adenosine deaminase for this kind of effusion, 20 has led us to the conclusion that diagnosis of the origin of exudative pleural effusion in patients younger than 36 years requires only pleural biopsy if pleural adenosine deaminase is low, cytologic studies have proved negative, and there is no positive indication of any nonneoplastic etiology. 21 The fourth largest etiologic group in our study comprised parapneumonic effusions, which together with empyemas made up 14% of all cases. This was not unexpected, since it is estimated that 40% of bacterial pneumonia cases involve pleural effusion. 22 This kind of effusion showed no preference for any particular age or location, bilateral cases were rare, and the male/female ratio was 2.9. Pulmonary thromboembolism was responsible for 2.8% of our pleural effusion cases, a figure identical to that reported by S t o r e ~ reported by Light et al 9 and the 2.1% found by Marel et al 1 and similar to the 3.3% et ajl 0 among effusions diagnosed in vivo. Since pulmonary thromboembolism can cause sudden death, this figure can be almost trebled if postmortem diagnoses are included, and this cause of pleural effusion is accordingly certainly underestimated in most published case series. In our study, these effusions occurred mainly on the right (83.3%) and among men (male/female ratio 3.5 to 1). In conclusion, the most common cause of the pleural effusions seen in the Pneumology and Internal Medicine Services of our hospitals is tuberculosis, followed by cancer and CHF. This etiologic distribution may be typical of regions which, like ours, have a high incidence of tuberculosis. However, the etiology of pleural effusion is so diverse and subject to so many factors that, in our opinion, all groups with an interest in pleural disease should themselves determine the etiologic distribution of pleural effusions in their region CHEST I I JANUARY,

5 with a view to adopting locally optimized diagnostic and therapeutic attitudes. REFERENCES 1 Storey DD, Dines DE, Coles DT. Pleural effusion: a diagnostic dilemma. JAMA 1976; 236: Hirsch A, Ruffie P, Nebut M, et al. Pleural effusion: laboratory test in 300 cases. Thorax 1979; 34: Light RW. Approach to the patient. In: Light RW, ed. Pleural diseases. Philadelphia: Lea & Febiger, 1990; Light RW. Thoracocentesis (diagnostic and therapeutic) and pleural biopsy. In: Light RW, ed. Pleural diseases. Philadelphia: Lea & Febiger, 1990; Valdes L, Pose A, Suarez J, et al. Cholesterol: a useful parameter for distinguishing between pleural exudates and transudates. Chest 1991; 99: Abrams LD. New inventions: a pleural biopsy punch. Lancet 1958; 1: Cope C. New pleural biopsy needle. JAMA 1958; 167: McNeil BJ. Ventilation-perfusion studies and the diagnostic of pulmonary embolism: concise communication. J Nucl Med 1980; 21: Sahn SA. State of the art: the pleura. Am Rev Respir Dis 1988; 138: Mare! M, Zrustova M, Stasny B, et al. The incidence of pleural effusion in a well-defined region: epidemiologic study in Central Bohemia. Chest 1993; 104: Peterman T A, Speicher CE. Evaluating pleural effusions: a twostage laboratory approach. JAMA 1984; 252: Chretien J, Hirsch A, Daniel C, et al. Diagnostic aspects. In: Chretien J, Hirsch A, eds. Diseases of the pleura. Chicago: Yearbook Medical Publishers, 1983; Salgueiro M, Zamarron C, Alvarez-Calder6n P, et al. Estudio epidemiol6gico de Ia tuberculosis en el area sanitaria de Santiago de Compostela durante los afios 89, 90 y 91. An Med Intern (Madrid) 1993; 9: Martinez-Berganza A, Cfa P. Epidemiologfa de las enfermedades de Ia pleura: a prop6sito de 562 casos. Med Clin (Bare) 1988; 90: Race GA, Scheifley CH, Edwards JE. Hydrothorax in congestive heart failure. Am J Med 1957; 22: Sahn SA. Malignant pleural effusions. In: Fishman A, ed. Pulmonary diseases and disorders. 2nd ed. New York: McGraw Hill, 1988; Sibley JC. A study of 200 cases of tuberculous pleurisy with effusion. Am Rev Tuberc 1950; 62: Fontan Bueso J, Verea Hernando H, Perez Garcia-Buela J, et al. Diagnostic value of simultaneous determination of pleural adenosine deaminase and pleural lysozyme/serum lysozyme ratio in pleural effusions. Chest 1988; 93: Light RW, Macgregor I, Luchsinger PC, et al. Pleural effusions: the diagnostic separation of transudates and exudates. Ann Intern Med 1972; 77: Valdes L, San Jose E, Alvarez D, et al. Diagnosis of tuberculous pleurisy using the biologic parameters adenosine deaminase, lysozyme and interferon gamma. Chest 1993; 103: Valdes L, Alvarez D, San Jose E, et al. The value of adenosine deaminase in the diagnosis of tuberculous pleural effusions in young patients in a region of rugh prevalence of tuberculosis. Thorax 1995; 50: Light RW, Girard WM, Jenkinson SG, et al. Parapneumonic effusions. Am J M ed 1980; 69: Clinical Investigations