September - December 2003: 
Volume 16, Issue 3

Click on the image to download the Issue in PDF format.


New advances in the investigation of pleural diseases
Pulmonary Medicine, Intensive Care Unit, Pulmonary Medicine Department, Regional General Hospital of Athens “Evangelismos”
Full text

Sometimes the cause of a pleural effusion is apparent, as it is for example in the case of a parapneumonic effusion or a small effusion caused by left heart failure. However, there have also been cases in which a definite diagnosis could not be made even though a diagnostic thoracoscopy or open pleural biopsy had been performed. Although a variety of clinical conditions may be the cause of a pleural effusion, the possibility of a malignant effusion should always be examined in difficult to diagnose cases. Moreover, neoplastic disease is the second most common cause of exudative pleural effusion after parapneumonic effusion1 and is usually diagnosed without difficulty at the first encounter with the patient. Tuberculous pleural effusion is also a quite common cause of exudative pleural effusion in our country and should at all times be included in the differential diagnosis process. More often than not, a tuberculous pleural effusion is impossible to differentiate from a malignant pleural effusion on clinical grounds alone. Diagnostic methods traditionally used in the diagnosis of these two conditions are either not adequately precise or more invasive than they should. In consequence, the following question is constantly raised during the clinical investigation of pleural diseases: how can differentiation between tuberculous and malignant pleural effusion be reliably made and in the least invasive way possible? Some interesting articles relating to this question were published in the first half of the year 2003 and will be discussed below. In addition, we will refer to two publications dealing with eosinophilic pleural effusions. Eosinophilic pleural effusion, which is defined as the presence of more than 10 percent eosinophils in pleural fluid, is often a diagnostic challenge since it may occur due to a variety of diseases and clinical conditions. Among them, the most common cause is the presence of air or/and blood in the pleural space. Furthermore, infections, neoplastic diseases, inflammatory pleural effusion, drug toxicity and pulmonary embolism may also manifest an eosinophilic pleural effusion. The cause of the effusion remains undiagnosed in a significant percentage of patients. It is believed that the presence of eosinophils in pleural fluid obtained from a patient with a recent history of thoracentesis should be attributed to bleeding or small pneumothorax caused during the procedure. But, how often does thoracentesis cause an increase in pleural-fluid eosinophils? A study dealing with this question will be presented in the follow ing sections. Furthermore, it is essential to determine whether the presence of eosinophils in pleural fluid indicates the presence of a benign or a malignant condition, a subject treated in the last review we will present.

Diagnosis of tuberculous pleural effusion is generally based on microbiologic evidence of the presence of mycobacteria in pleural fluid or in sputum. Pleural fluid culture has a sensitivity of about 30%, whereas the absence of radiographic evidence of pulmonary tuberculosis is associated with a sputum culture sensitivity rate of less than 10%.2 The presence of a granuloma in closed needle biopsy of the pleura, ideally in combination with demonstrating the presence of mycobacteria in the tissue has a higher sensitivity rate. In the previous year, Frank reviewed the relevant literature and concluded that histological studies have an average sensitivity of 69% (range 28‑88%) in diagnosing tuberculous pleural effusion. The first paper we will discuss examines the sensitivity of induced sputum culture in patients with tuberculous pleural effusion.3 Conde et al studied prospectively 113 patients with probable tuberculous pleural effusion. The diagnosis of tuberculous pleural effusion was confirmed by conventional diagnostic methods in 84 of these patients. Histopathological examination of closed biopsy samples of the pleura had a sensitivity equal to 78%, whereas M. tuberculous mycobacteria were cultured in tissue in 62% and in pleural fluid in 12% of the patients. Induced sputum culture was positive in 52% of the patients. It is worthy of note that induced sputum culture was positive at about the same rate in both patients with radiographically evident parenchymal disease (9/20, 45%) and patients with pleural effusion as the only radiographic finding (35/64, 55%). Direct staining for M. tuberculous was positive in 17% of pleural fluid samples and in 12% in induced sputum samples. The higher yield of pleural fluid staining as regards recovery of mycobacteria is most probably due to the high pecentage of HIV patients (15.5%) in the study population. Apart from this, the findings of this study fit our clinical routine very well. The higher sensitivity of induced sputum studies warrants their inclusion in the array of tests performed in order to diagnose tuberculous pleuritis. In addition, this study pinpoints how often parenchymal disease is also present in patients with tuberculous pleuritis, even though no signs of it are found in chest radiograph.

The aforementioned information relates to the microbiologic approach to the diagnosis of tuberculous pleural effusion. However, a biochemical approach has emerged in the last years: high levels of adenosine deaminase (ADA), an enzyme produced by lymphocytes, mononuclear cells and macrophages, have been found in the pleural fluid from patients with tuberculous pleural effusion. Is ADA measurement a reliable marker of tuberculosis? A classic international book on pleural diseases indicates that ADA levels higher than 40 U/L are suggestive of tuberculosis, whereas levels above 70 U/L establish the diagnosis.4 Goto et al have recently reviewed the relevant literature (1996-99). Forty studies were included in the meta-analysis of the reliability of ADA measurements and sensitivity and specificity rates were found to range from 47 to 100% and from 50 to 100%, respectively. Further processing of the meta-analysis data indicated that the best yield of ADA measurements as assessed by the ROC curve corresponded to sensitivity and specificity rates equal to 92.2%. This yield is considered very satisfactory, especially in the case of a condition that is often treated without having established a diagnosis.

It should be noted, however, that most studies on this topic have included both polymorphonuclear cell pleural effusion and lymphocytic pleural effusion cases. Tuberculous pleural effusion is known to very rarely present as a polymorphonuclear effusion. Even if these cases are taken into account, the specificity of ADA measurements in lymphocytic effusions is minimally influenced by the presence of small lymphocytes. A Spanish study published in the year we are reviewing showed that a positive ADA measurement (cutoff point 40 U/L) was obtained in 1.7% of 410 non-tuberculous lymphocytic pleural effusion cases.6 If the patients of an American and another Spanish series are added to the patients of the aforementioned study,7,8 we sum up a total of 809 patients with non-tuberculous lymphocytic (>50% lymphocytes) pleural effusion. Seventeen of them (2.2%) had ADA levels above 40 U/L. Pleural effusion in these patients was due to lymphoma in 6 cases, acute lymphoblastic leukemia in 1, metastatic carcinoma in 4, parapneumonic effusion in 4, mesothelioma in 1; the cause was unknown in 1 patient (idiopathic pleural effusion). Only two of these patients had ADA levels above 70 U/L, one with metastatic adenocarcinoma and the patient with acute lymphoblastic leukemia. I believe that, according to the available data, when investigating a lymphocytic pleural effusion, one should take seriously into account ADA levels: if the levels of this enzyme are above 40 U/L and no other condition is considered possible on clinical grounds, it is justified to initiate treatment of tuberculosis. I would reserve the escalation of the diagnostic investigation so as to include thoracoscopy and pleural biopsy for patients with lower ADA levels or patients with ADA levels > 40U/L who additionally present signs that strongly indicate malignancy. The next most essential issue in the management of an undiagnosed pleural effusion is to reliably exclude a neoplastic pleural process. The sensitivity of cytology testing varies according to the histological type of the neoplasia namely from 10% in mesothelioma and 20% in epidermoid carcinoma to 70% in adenocarcinoma.1 The sensitivity rates of closed needle biopsy reported in various studies range from 40 to 75%; in addition, biopsy is rarely positive in cases with negative cytology results.4 The standard test to establish the neoplastic origin of a pleural effusion is thoracoscopy which has a sensitivity of 95%.9 In other words, should a patient with a persistent lymphocytic or mononuclear pleural effusion whose cytology testing is negative and pleural fluid ADA levels are low, undergo thoracoscopy and biopsy of the pleura? Traditionally, the answer to this question is usually "yes". Challenging the validity of the affirmative answer, a study group from Oxford studied prospectively 50 patients with suspected neoplastic pleural effusion and negative results on cytologic examination of pleural fluid.10 Patients were randomized in two groups: subjects in one group underwent biopsy with an Abrams needle and subjects in the other group CT guided biopsy with cutting needle. Guided biopsy samples were obtained from the thicker areas of parietal pleura on CT images after infusion of a contrast agent. In the Abrams needle biopsy group, sensitivity, specificity, negative and positive predictive values were 47%, 100%, 44% and 100%, respectively. In the CT guided biopsy group, sensitivity, specificity, negative and positive predictive values were 87%, 100%, 80% and 100%, respectively. The authors concluded that CT guided biopsy is significantly more sensitive than closed biopsy of the pleura in the diagnosis of a malignant pleural effusion; moreover, the sensitivity of the former method is comparable to that of thoracoscopic biopsy. However, this study has a drawback that hinders its being widely accepted: Twenty of a total of 33 patients with malignant effusion had malignant mesothelioma, an unusually high percentage of presentation even for the United Kingdom where this condition is relatively common. This neoplasm affects the pleura in such a way that wide areas of abnormal thickness are formed, facilitating the acquisition of positive biopsy samples regardless of whether these samples are obtained blindly or under guidance. Although in the study of Maskell et al CT guided biopsy was found equally sensitive in mesothelioma and metastatic malignant pleural effusion cases, it may be suggested that an abnormally thick area of parietal pleura that would be appropriate for biopsy, may not be radiographically evident in patients with metastatic pleural cancer. Nevertheless, I see no reason why I should not request a CT imaging study with infusion of a contrast agent having informed the radiologist that biopsy samples should be obtained in case a parietal pleura thickness is found, before proceeding to thoracoscopic biopsy for the purposes of investigating a possible malignancy.

We may now take a look at the discussion about eosinophilic pleuritis. Martinez-Garcia et al studied 273 pleural fluid samples from 120 patients and found no significant difference in eosinophil count between the two samples obtained from each patient, even in the cases who underwent closed pleural biopsy.11 This study does not eliminate the possibility of a marked rise in eosinophils when thoracocentecis is followed by manifest pneumothorax and/or major bleeding. It weakens, however, the view supported by many clinicians that if thoracocentesis has been performed, a case of pleural effusion with more than 10% eosinophils should not be investigated as a case of eosinophilic pleural effusion. A further question that concerns all eosinophilic pleural effusion cases is whether malignancy should be considered in differential diagnosis or not. "It should not", responded Adelman in 1984 after reviewing all cases of eosinophilic pleural effusion reported until then, namely a total of 343 cases.12 In consequence, the view that eosinophilic pleural effusion indicates a benign condition prevailed. In the years that followed, however, three series, two prospective and one retrospective, including a total of 1058 patients (111 with eosinophilic pleural effusion) were published and indicated that benign and malignant causes of eosinophilic pleural effusion are almost equally possible.13-15 More recently, the author of the present paper and R. Light reviewed the characteristics of 1195 pleural fluid samples from patients with pleural effusion admitted to a tertiary care hospital in Tennessee (USA) specializing in cardiosurgery. One hundred seventy two patients had eosinophilic pleural effusion, the second most common cause of which was neoplasia. The most common cause was postoperative pleural effusion, a quite expected finding given the great number of cardiac operations in the study population.16 We proceeded to add all cases of pleural effusion published till the present day to those recovered by the data base of the aforementioned hospital. Cases with eosinophilic pleural effusion associated with trauma of any kind, physical or operative, as well as cases of spontaneous pneumothorax were excluded from our calculations, since the cause of pleural effusion is apparent in these cases. Among 392 non post-traumatic cases with eosinophilic pleural effusion, the most common cause of pleural effusion was malignancy (17%). Parapneumonic effusion accounted for 12.5%, transudates of any kind for 7.9%, tuberculous pleural effusion for 5.6% and pulmonary embolism for 4.3% of the cases. Various causes accounted for 12.8% of the cases, whereas idiopathic pleural effusion was the final diagnosis in 40%. It is not clear why later series contradicted the findings of Adelman et al. In the years that followed the report of Adelman et al, the occurrence of malignancy as a cause of pleural effusion has increased. As a result, according to the Bayes theorem, its occurrence as a cause of eosinophilic pleural effusion has increased as well. This, however, is nothing but an assumption. Whatever the cause of this controversy, the review of the available data has convinced us that when investigating an eosinophilic effusion, the possibility of malignancy should be seriously considered, particularly in patients at high risk for lung or breast cancer.

To sum up, clinical studies dealing with issues relating to the diagnosis of pleural diseases that have been published in the first 6 months of the year 2003 encourage those who trusted ADA measurements in the diagnosis of tuberculous pleuritis, but also support the recovery of M. tuberculous bacilli in induced sputum samples from these patients. Furthermore, these studies suggest a new, relatively non-invasive method to verify a neoplastic pleural effusion based on histology. Lastly, the view that a high eosinophil percentage in pleural fluid should not be rashly attributed to a recent thoracocentesis is enhanced; moreover, it is pinpointed that eosinophilic pleural effusion does not at all mean that we are not dealing with a malignant condition.


1. Light RW. Pleural effusion. N Engl J Med 2002; 346: 1971-77.

2. Tuberculous pleural effusions. Eur Respir Mon 2002; 22: 219-233.

3. Conde MB, Loivos AC, Rezende VM, et al. Yield of sputum induction in the diagnosis of pleural tuberculosis. Am J Respir Crit Care Med 2003; 167: 723-25.

4. Light RW. Pleural Diseases. 4th ed. Philadelphia: Lippincott Williams and Wilkins, 2001: 42-86.

5. Goto M, Noguchi Y, Koyama H, et al. Diagnostic value of adenosine deaminase in tuberculous pleural effusion: a meta-analysis. Ann Clin Biochem 2003; 40: 374-81.

6. Jimenez Castro D, Diaz Nuevo G, Perez-Rodriguez E, Light RW. Diagnostic value of adenosine deaminase in nontuberculous lymphocytic pleural effusions. Eur Respir J 2003; 21: 220-24.

7. Lee YC Rogers JT Rodriguez RM, et al. Adenosine deaminase levels in nontuberculous lymphocytic pleural effusions. Chest 2001; 120: 356-61.

8. Porcel JM, Vines M. Adenosine deaminase levels in nontuberculous lymphocytic pleural effusions [letter]. Chest 2002; 121: 1379-80.

9. Harris RJ, Kavuru MS, Rice TW, Kirby TJ. The diagnostic and therapeutic utility of thoracoscopy: a review. Chest 1995; 108: 828-41.

10. Maskell NA, Glesson FV, Davies RJO. Standard pleural biopsy versus CT-guided cutting-needle biopsy for diagnosis of malignant disease in pleural effusions: a randomized control study. Lancet 2003; 361: 1326-31.

11. Martinez-Garcia MA, Cases Viedma C, Perpina Tordera M, Sanchi Aldas JL. Repeated thoracentesis: an important risk factor for eosinophilic pleural effusion? Respiration 2003; 70: 82-6.

12. Adelman M, Albelda SM, Gottlieb J, Haponik EF. Diagnostic utility of pleural fluid eosinophilia. Am J Med 1984; 77: 915-20.

13. Kuhn M, Fitting JW, Leuenberger P. Probability of malignancy in pleural fluid eosinophilia. Chest 1989; 9: 992-4.

14. Martinez-Garcia MA, Cases-Viedma E, Cordero-Rodriguez PJ, et al. Diagnostic utility of eosinophils in the pleural fluid. Eur Respir J 2000; 15:166-9.

15. Rubins JB, Rubins HB. Etiology and prognostic significance of eosinophilic pleural effusions. A prospective study. Chest 1996; 110: 1271-4.

16. Kalomenidis I, Light RW. Eosinophilic pleural effusions. Current Opinion in Pulmonary Medicine 2003; 9: 254-260.