Effect of Repeated Thoracenteses on Fluid Characteristics, Cytokines, and Fibrinolytic Activity in Malignant Pleural Effusion

Effect of Repeated Thoracenteses on Fluid Characteristics, Cytokines, and Fibrinolytic Activity in Malignant Pleural Effusion* Chi-Li Chung, MD; Yi-Chu Chen, BS; and Shi-Chuan Chang, MD, PhD, FCCP

Objective: To evaluate the effect of repeated thoracenteses on the fluid characteristics and the levels of various cytokines, including tumor necrosis factor (TNF)-␣, interleukin (IL)-1␤, IL-5, IL-6, and IL-8, and of plasminogen activator inhibitor type 1 (PAI-1) and tissue type plasminogen activator in malignant pleural effusion and its clinical significance. Design: A prospective study. Patients and methods: Twenty-six patients with symptomatic and a large amount of free-flow malignant pleural effusions were studied. Thoracentesis with drainage of 500 mL of pleural fluid per day was performed for 3 continuous days (days 1 to 3). The effusion samples were collected to evaluate the changes of fluid characteristics, cytokine levels, and fibrinolytic activity. Chest ultrasonography was done on day 6 to observe the presence of fibrin strands. The result of pleurodesis was evaluated in the patients classified into groups based on chest ultrasonographic findings. Results: The values of TNF-␣, PAI-1, IL-8, and neutrophil count in pleural fluid increased significantly during repeated thoracenteses in 26 patients studied. A positive correlation was found between the concentrations of TNF-␣ and PAI-1 and between the values of IL-8 and neutrophils. On day 6, fibrin strands were observed in the pleural effusion on chest ultrasonography in 11 patients (42%, fibrinous group) but were absent in the remaining 15 patients (nonfibrinous group). During repeated thoracenteses, a significant increase of effusion PAI-1 and TNF-␣ was observed in the fibrinous group but not in the nonfibrinous group. In addition, the levels of effusion PAI-1 and TNF-␣ obtained from day 2 and day 3 were significantly higher in the fibrinous group than in the nonfibrinous group. The success rate of pleurodesis was significantly higher in the fibrinous group (11 of 11 patients, 100%) than in the nonfibrinous group (8 of 12 patients, 67%). Conclusions: Repeated thoracenteses may cause pleural inflammation and induce local release of proinflammatory cytokine as TNF-␣, which may subsequently enhance the release of PAI-1 and lead to fibrin formation in malignant effusion. The presence of fibrin strands after repeated thoracenteses may be of considerable value in predicting the success of subsequent pleurodesis in patients with malignant pleural effusions. (CHEST 2003; 123:1188 –1195) Key words: fibrinolysis; malignancy; pleural effusion; proinflammatory cytokines; thoracentesis Abbreviations: IL ⫽ interleukin; LDH ⫽ lactate dehydrogenase; PAI ⫽ plasminogen activator inhibitor; PAI-1 ⫽ plasminogen activator inhibitor type 1; TNF ⫽ tumor necrosis factor; tPA ⫽ tissue type plasminogen activator

is a common cause of exudative pleuM alignancy ral effusion. Carcinoma arising from any organ can metastasize to the pleura and result in pleural effusion. The patients may present with dyspnea and *From the Department of Chest Medicine (Dr. Chung), Taipei Medical University Hospital, Graduate Institute of Medical Sciences, Taipei Medical University; Chest Department (Ms. Chen), Taipei Veterans General Hospital; and School of Medicine (Dr. Chang), National Yang-Ming University, Taipei, Taiwan, ROC. This work was supported by a grant from the Taipei Medical University (TMC89-Y05-A134). 1188

nonproductive cough when their pleural effusions increase to significant amount. To relieve respiratory distress caused by malignant pleural effusion, theraManuscript received January 22, 2002; revision accepted October 7, 2002. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: [email protected]). Correspondence to: Shi-Chuan Chang, MD, PhD, FCCP, Chest Department, Taipei Veterans General Hospital, # 201, Sec. 2, Shih-Pai Rd, Shih-Pai, Taipei, Taiwan 112, ROC; e-mail: scchang@ vghtpe.gov.tw Clinical Investigations

peutic thoracentesis is usually indicated. Malignant pleural effusion, however, may reaccumulate rapidly, and repeated thoracenteses are needed.1 Clinically, the presence of fibrin strands in pleural fluid shown on chest ultrasonography is suggestive of an inflammatory exudate.2 The fibrin strands, nevertheless, could also be found in malignant pleural effusion.3 Furthermore, based on our observations and those of other clinical investigators,2 repeated thoracenteses may induce the generation of fibrin strands in malignant effusion and result in loculation of the pleural spaces. As a result, the drainage of malignant effusion may become more difficult. Although this phenomenon is not uncommon, the underlying mechanisms of fibrin formation in malignant effusion induced by repeated thoracenteses remain unknown. By and large, fibrin turnover in the pleural cavity is greatly affected by the activity of fibrinolysis. The formation of key enzyme in fibrinolysis, plasmin, is based mainly on the equilibrium between plasminogen activators and plasminogen activator inhibitors (PAIs).4 Agrenius et al5 indicated that intrapleural injection of quinacrine, an irritative agent, in patients with malignant pleural effusions could increase the concentrations of PAI type 1 (PAI-1) and reduce fibrinolytic activity in pleural fluid. The levels of proinflammatory cytokines such as tumor necrosis factor (TNF)-␣ and interleukin (IL)-1␤ in pleural fluid were reported to be elevated markedly after intrapleural injection of quinacrine in patients with malignant pleural effusions.6,7 These findings suggest a strong relationship between the fibrinolytic activity and proinflammatory cytokines in the pleural cavity. The levels of TNF-␣ were reported to be significantly higher in tuberculous than in malignant pleural effusions.8 –11 Increased levels of pleural fluid TNF-␣ appeared to be an important indicator in patients with pleural tuberculosis who might acquire residual pleural thickening.11,12 Philip-Joe¨t et al13 indicated that pleural PAI-1 levels were greatly enhanced in exudates due to the inflammatory or infection process; however, the levels of plasminogen activators were increased in malignant effusions. In agreement with Philip-Joe¨t et al,13 we observed significantly higher levels of tissue type plasminogen activator (tPA) and lower values of PAI-1 in malignant than in tuberculous effusions.11 Taken together, these findings suggest that the enhanced release of TNF-␣ and/or IL-1␤ caused by pleural inflammation may result in an imbalance between PAI-1 and tPA in the pleural cavity, which may subsequently lead to fibrin formation and loculation of the pleural effusion. Because our investigations and the clinical observations of others2 indicated that repeated thoracenteses might induce www.chestjournal.org

generation of fibrin strands in malignant effusion, we speculate that repeated thoracenteses may cause pleural inflammation, enhanced release of proinflammatory cytokines, and an imbalance between PAI-1 and tPA, which may lead to formation of fibrin strands in malignant pleural effusion. This study was conducted prospectively to evaluate the effect of repeated thoracenteses on the fluid characteristics, the levels of cytokines related to inflammation and fibrinolytic activity in malignant pleural effusion and its clinical significance. Materials and Methods The study was approved by the Institutional Review Board of Taipei Medical University Hospital, and informed consent was obtained from all patients to participate in this study. The patients with large amounts of free-flow pleural effusion, the effusion occupying at least half of one hemithorax, of unknown causes and clinical symptoms such as dyspnea at rest or on exertion and/or cough ascribed to pleural effusion were eligible for this study. Chest ultrasonography was done, and pleural fluid was collected using a standard thoracentesis technique immediately or within 24 h after hospitalization. The pleural fluid samples were mixed with 3.8% sodium citrate in a 9:1 ratio of pleural fluid to citrate. The sodium citrate-mixed pleural fluid specimens were immersed in ice immediately and then centrifuged at 2,500g for 10 min. The cell-free supernatants of pleural fluid were frozen at ⫺ 70° immediately after centrifuge until later examinations. Analyses of pleural fluid for total leukocytes, cell differentials of leukocytes, pH value, and levels of protein, glucose, and lactate dehydrogenase (LDH) were performed in addition to cytologic and microbiologic examination of pleural fluid. In addition, approximately 500 mL of pleural fluid was drained out to relieve respiratory distress. The patients were included subsequently if cytologic examination of pleural fluid obtained from the first tap (day 1) showed malignant cells. Patients were excluded from this study if they had any of the following: (1) invasive procedures directed into the pleural cavity, or chest trauma within 3 months prior to hospitalization; (2) loculated pleural effusion or fibrin strands detected in pleural effusion shown on ultrasonography before first tap; or (3) no malignant cells in the pleural fluid obtained from the first tap. The patients who met the selection criteria were subjected to repeated thoracenteses with a drainage of 500 mL of pleural fluid at 24 (day 2) and 48 h (day 3) after the first tap.14 During each tap, the pleural fluid samples were collected and processed as the first tap. Chest ultrasonography was performed before each tap and on the day 6 to observe the presence of fibrin strands in pleural effusion. The sonograms were printed out and interpreted by another physician. When the two readers could not reach consensus, the case was presented to a third expert reader, and the adjudicated reading became final. In fact, no disagreement on the presence of fibrin strands was noted in this study. Based on the chest ultrasonographic findings observed on day 6, the patients were classified into fibrinous and nonfibrinous groups. The patients were classified into the fibrinous group if fibrin strands floating in pleural effusion were shown on chest ultrasonography.3 Those who had no fibrin strands detected in the pleural fluid were classified into nonfibrinous group. In the nonfibrinous group, when a significant amount of pleural effusion was still noted on day 6, a small-bore catheter was inserted into pleural cavity under the guidance of chest ultrasonography to drain out pleural fluid as completely as possible and to prepare for CHEST / 123 / 4 / APRIL, 2003

1189

subsequent pleurodesis. In the fibrinous group, intrapleural injection of streptokinase, 250,000 IU, in 30 mL of saline solution via small-bore catheter was done. After each instillation of streptokinase, the catheter was clamped for 2 h and then de-clamped.15 Intrapleural injection of fibrinolytic agent was done twice daily until the drain yield was ⬍ 150 mL/d and insignificant amount of pleural effusion was evidenced by chest radiograph or chest ultrasonography. Then the patients, if clinically permitted, were subjected to pleurodesis with minocycline, 600 mg, dissolved in 100 mL of saline solution. The catheter was clamped for 2 h and then declamped. Negative pressure suction (⫺ 15 to ⫺ 20 cm H2O) was applied for up to additional 24 h, and the catheter was removed if drained yield was ⬍ 150 mL/d.1 The result of pleurodesis was evaluated 30 days after removal of the catheter. The success of pleurodesis was defined as no recurrence of pleural effusion, presence of insignificant amount of pleural effusion, or loculated pleural effusion without clinical symptoms. Recurrence of symptomatic pleural effusion requiring thoracentesis during the whole follow-up period was defined as failure.16,17 The levels of cytokines and fibrinolytic enzymes in the supernatants of pleural fluid were measured by the commercially available enzyme-linked immunosorbent assay kits: tPA and PAI-1 (American Diagnostica; Greenwich, CT), TNF-␣, IL-1␤, IL-5, IL-6, and IL-8 (R&D Systems; Minneapolis, MN). Data were expressed as mean ⫾ SEM. Nonparametric tests were used to analyze pleural fluid variables, since these variables were not normally distributed. The Friedman nonparametric two-way analysis of variances and multiple comparisons on ranks of several related samples were performed. The correlations between variables were determined by Spearman rank correlation coefficients. Specific comparisons of data between two groups were made using the Mann-Whitney U test or ␹2 method, when appropriate. Significance was defined as p ⬍ 0.05.

Results A total of 26 patients who were admitted to the Department of Chest Medicine and fulfilled the patient selection criteria were included for this study between October 1999 and April 2001. There were

10 men and 16 women (age range, 37 to 85 years; mean age, 65 years). The underlying malignant tumors were lung cancer in 12 patients, gastric carcinoma in 4 patients, hepatocellular carcinoma in 3 patients, breast cancer in 2 patients, and oropharyngeal carcinoma, esophageal carcinoma, pancreas cancer, colon cancer, and ovarian cancer in one patient each. The clinical symptoms improved after repeated thoracenteses in all patients. Complications such as traumatic tapping, pneumothorax, and reexpansion pulmonary edema were not found during or after thoracentesis in all patients. Pleural fluid characteristics including total and differential counts of leukocytes, and the values of pH, LDH, glucose, and protein obtained from day 1 to day 3 are shown in Table 1. Repeated thoracenteses had no significant effect on the values of pH, glucose, and LDH in pleural fluid. However, significant decrease of protein and increase of neutrophils were observed during repeated thoracenteses (Table 1). Repeated thoracenteses showed no effect on other subpopulations of leukocytes. The changes of effusion levels of cytokines, tPA, and PAI-1 from day 1 to day 3 are summarized in Table 2. The effusion levels of TNF-␣ and IL-8 increased significantly during repeated thoracenteses. The values of TNF-␣ and IL-8 on day 2 and day 3 were significantly higher than those on day 1. Repeated thoracenteses, however, showed no effect on the effusion levels of other cytokines including IL-1␤, IL-5, and IL-6, although a modest increase of IL-1␤ was found. The values of PAI-1 in pleural fluid increased gradually during repeated thoracenteses and were significantly higher on day 2 and day 3 than

Table 1—Effect of Repeated Thoracenteses on Pleural Fluid Characteristics in 26 Patients With Malignant Pleural Effusions* Effusion Variables pH Glucose, mg/dL Protein, g/dL LDH, IU/dL Total leukocytes, cells/␮L Neutrophils, cells/␮L % Lymphocytes, cells/␮L % Macrophages, cells/␮L % Eosinophils, cells/␮L %

Day 1

Day 2

Day 3

7.36 ⫾ 0.02 153 ⫾ 18 3.8 ⫾ 0.2† 584 ⫾ 122 1,434 ⫾ 393 628 ⫾ 268‡ 37 ⫾ 9 492 ⫾ 329 39 ⫾ 8 224 ⫾ 140 10 ⫾ 2 68 ⫾ 27 3⫾1

7.39 ⫾ 0.02 150 ⫾ 12 3.7 ⫾ 0.2 688 ⫾ 125 1,913 ⫾ 193 938 ⫾ 308 40 ⫾ 9 465 ⫾ 104 36 ⫾ 7 153 ⫾ 40 10 ⫾ 2 64 ⫾ 21 4⫾1

7.39 ⫾ 0.01 132 ⫾ 121 3.6 ⫾ 0.2† 646 ⫾ 120 1,795 ⫾ 494 1,027 ⫾ 416‡ 43 ⫾ 8 457 ⫾ 127 38 ⫾ 8 114 ⫾ 26 10 ⫾ 2 132 ⫾ 82 7⫾4

*Values are presented as mean ⫾ SEM. Day 1 ⫽ the time of first thoracentesis; day 2 ⫽ 24 h after first thoracentesis; day 3 ⫽ 48 h after first thoracentesis. †Day 3 vs day 1, p ⬍ 0.001. ‡Day 3 vs day 1, p ⬍ 0.05. 1190

Clinical Investigations

Table 2—Effect of Repeated Thoracenteses on Cytokines and Fibrolytic Enzymes in 26 Patients With Malignant Pleura Effusions* Variables

Day 1

Day 2

Day 3

PAI-1, ng/mL tPA, ng/mL TNF-␣, pg/mL IL-1␤, pg/mL IL-5, pg/mL IL-6, pg/mL IL-8, pg/mL

87.4 ⫾ 9.5†‡ 21.7 ⫾ 4.3 15.5 ⫾ 0.7†‡ 2.2 ⫾ 0.2 25.3 ⫾ 11.2 5,285.3 ⫾ 430.6 280.1 ⫾ 67.7‡§

99.1 ⫾ 9.9† 22.6 ⫾ 4.2 20.5 ⫾ 3.0† 3.2 ⫾ 0.6 35.7 ⫾ 14.9 5,507.4 ⫾ 428.1 710.6 ⫾ 366.9§㛳

99.9 ⫾ 9.8‡ 20.9 ⫾ 4.1 21.0 ⫾ 4.8‡ 5.6 ⫾ 3.3 27.7 ⫾ 8.5 5,465.1 ⫾ 381.7 770.4 ⫾ 375.7‡㛳

*Values are presented as mean ⫾ SEM. Day 1 ⫽ the time of first thoracentesis; day 2 ⫽ 24 h after first thoracentesis; day 3 ⫽ 48 h after first thoracentesis. †Day 2 vs day 1, p ⬍ 0.01. ‡Day 3 vs day 1, p ⬍ 0.05. §Day 2 vs day 1, p ⬍ 0.05. 㛳Day 3 vs day 2, p ⬍ 0.05.

on day 1. In contrast, repeated thoracenteses had no effect on the effusion levels of tPA. During repeated thoracenteses, a positive correlation was found between the effusion values of IL-8 and neutrophils (day 1, r ⫽ 0.43, p ⬍ 0.05; day 2, r ⫽ 0.51, p ⬍ 0.01; day 3, r ⫽ 0.68, p ⬍ 0.001). The values of TNF-␣ and IL-1␤ in pleural fluid were highly correlated (day 1, r ⫽ 0.46, p ⬍ 0.02; day 2, r ⫽ 0.63, p ⬍ 0.001; day 3, r ⫽ 0.52, p ⬍ 0.01). The effusion levels of PAI-1 on day 2 and day 3 correlated positively with those of TNF-␣ (day 2, r ⫽ 0.49, p ⬍ 0.05; day 3, r ⫽ 0.41, p ⬍ 0.05). Chest ultrasonography was performed to observe the presence of fibrin strands before each tap and on day 6 in all patients. On day 6, 11 of 26 patients (42%) showed fibrin strands in pleural effusion and were classified into the fibrinous group. The remaining 15 patients who had no fibrin strands were classified into the nonfibrinous group. In both groups, repeated thoracenteses had no significant effect on the pleural fluid characteristics, except that a significant decrease of protein was found. Furthermore, no significant difference in the pleural fluid characteristics was observed between the fibrinous and nonfibrinous groups (data not shown). The changes of effusion levels of cytokines and fibrinolytic enzymes during repeated thoracenteses in fibrinous and nonfibrinous groups are summarized in Table 3. In fibrinous group, the effusion PAI-1 and TNF-␣ increased during repeated thoracenteses and the values on day 2 and day 3 were significantly higher than those on day 1 (Table 3, Fig 1). In contrast, repeated thoracenteses had no effect on effusion levels of cytokines and fibrinolytic enzymes in the nonfibrinous group. Furthermore, the effusion levels of PAI-1 and TNF-␣ obtained from day 2 and day 3 were significantly higher in the fibrinous than in nonfibrinous groups (Table 3, Fig 1). www.chestjournal.org

The clinical features and outcomes of pleurodesis in fibrinous and nonfibrinous groups are summarized in Table 4. The clinical features were comparable between two groups, except that more women were in the fibrinous group than in the nonfibrinous group (p ⬍ 0.01). All patients underwent indwelling of a small-bore catheter to drain pleural fluid as completely as possible. The pleural fluid in the nonfibrinous group (n ⫽ 15) was completely drained out within 2 to 3 days, except for three patients with trapped lungs. Diffuse nodular pleura thickening and encasement of underlying lung were evidenced by thoracic CT in all three patients. Except for the patients with trapped lung, 12 patients were subjected to pleurodesis. In the fibrinous group, following intrapleural administration with streptokinase, all patients (n ⫽ 11) were subjected to pleurodesis with same protocol. The successful rate of pleurodesis was significantly higher in the fibrinous group (11 of 11, 100%) than in the nonfibrinous group (8 of 12 patients, 67%) [Table 4].

Discussion The present study demonstrated that the effusion values of TNF-␣, PAI-1, IL-8 and neutrophils increased significantly during repeated thoracenteses in 26 patients with malignant pleural effusions. A positive correlation was found between the concentrations of TNF-␣ and PAI-1, and between the levels of IL-8 and neutrophils. On day 6, fibrin strands developed in pleural fluid shown on chest ultrasonography in 11 patients (42%, fibrinous group) but were absent in the remaining 15 patients (nonfibrinous group). During repeated thoracenteses, the effusion TNF-␣ and PAI-1 increased in the fibrinous group but not in the nonfibrinous group. In addition, CHEST / 123 / 4 / APRIL, 2003

1191

Table 3—Effect of Repeated Thoracenteses on Effusion Levels of Cytokines and Fibrinolytic Enzymes in Fibrinous and Nonfibrinous Groups* Variables Fibrinous group PAI-1, ng/mL tPA, ng/mL TNF-␣, pg/mL IL-1␤, pg/mL IL-5, pg/mL IL-6, pg/mL IL-8, pg/mL Nonfibrinous group PAI-1, ng/mL tPA, ng/mL TNF-␣, pg/mL IL-1␤, pg/mL IL-5, pg/mL IL-6, pg/mL IL-8, pg/mL

Day 1

Day 2

Day 3

94.5 ⫾ 13.7†‡ 17.5 ⫾ 3.0 16.9 ⫾ 1.4§㛳 2.0 ⫾ 0.1 37.9 ⫾ 23.8 5,751.2 ⫾ 600.1 151.3 ⫾ 31.5

129.3 ⫾ 10.8†¶ 19.9 ⫾ 3.5 28.7 ⫾ 6.5§** 2.8 ⫾ 2.5 31.2 ⫾ 18.0 5,856.1 ⫾ 674.3 256.9 ⫾ 77.0

127.3 ⫾ 9.1‡# 17.0 ⫾ 2.8 29.6 ⫾ 11.2㛳†† 2.3 ⫾ 0.7 21.1 ⫾ 7.6 5,767.8 ⫾ 612.5 314.8 ⫾ 61.5

82.3 ⫾ 13.4 24.8 ⫾ 7.2 14.5 ⫾ 0.7 2.4 ⫾ 0.6 16.1 ⫾ 8.8 4,943.7 ⫾ 604.7 347.6 ⫾ 110.3

77.0 ⫾ 12.9¶ 24.5 ⫾ 7.0 14.4 ⫾ 0.7** 3.5 ⫾ 1.0 39.4 ⫾ 22.7 5,251.7 ⫾ 564.0 1,036.7 ⫾ 629.1

80.5 ⫾ 14.0# 23.8 ⫾ 6.9 14.7 ⫾ 0.6†† 8.1 ⫾ 5.7 32.5 ⫾ 13.9 5,243.5 ⫾ 496.2 1,104.5 ⫾ 645.1

*Values are presented as mean ⫾ SEM. Day 1 ⫽ the time of first thoracentesis; day 2 ⫽ 24 h after first thoracentesis; day 3 ⫽ 48 h after first thoracentesis. †Day 2 vs day 1, p ⬍ 0.01. ‡Day 3 vs day 1, p ⬍ 0.01. §Day 2 vs day 1, p ⬍ 0.05. 㛳Day 3 vs day 1, p ⬍ 0.05 (comparisons within fibrinous group). ¶Day 2, p ⬍ 0.05. #Day 3, p ⬍ 0.05. **Day 2, p ⬍ 0.01. ††Day 3, p ⬍ 0.05 (comparisons between two groups).

the effusion levels of TNF-␣ and PAI-1 obtained from day 2 and day 3 were significantly higher in the fibrinous group than in the nonfibrinous group. The successful rate of pleurodesis was significantly higher in the fibrinous group (11 of 11 patients, 100%) than in nonfibrinous group (8 of 12 patients, 67%). The fibrinolytic activity is enhanced in malignant effusion as evidenced by higher levels of tPA in malignant effusion than in inflammatory exudates.11,13,18,19 In contrast, fibrinolytic activity is depressed as shown by an increase of PAI-1 in the inflammatory effusions, which may subsequently lead to fibrin formation.11,13,18,19 In patients with malignant pleural effusions, the intrapleural administration of quinacrine could induce pleural inflammation and reduce pleural fibrinolytic activity, as evidenced by increased effusion levels of PAI-1.5–7 The changes of fibrinolytic activity in malignant pleural fluid after intrapleural administration of irritative agents might be regulated by proinflammatory cytokines, because TNF-␣ and IL-1␤ increased markedly in pleural fluid as did PAI-1.6, 7 In in vitro studies, TNF-␣ and IL-1␤ were shown to have an effect on the release of PAI-1 by human mesothelial cells, and a synergetic effect exerted by these two cytokines was observed.20,21 Furthermore, the levels of TNF-␣ were reported to be significantly higher in tuberculous than in malignant and transudative ef1192

fusions,8 –11 and correlated positively with those of PAI-1.11, 18 Taken together, these findings strongly suggest that pleural inflammation may reduce fibrinolytic activity in pleural fluid via proinflammatory cytokines such as TNF-␣ and IL-1␤. In supporting this concept, our results indicated that during repeated thoracenteses, TNF-␣, PAI-1, IL-8, and neutrophils increased significantly in malignant effusions (Tables 1, 2). Fibrin strands developed in the pleural fluid in 11 patients (42%, fibrinous group) but were absent in the remaining 15 patients (nonfibrinous group). The effusion TNF-␣ and PAI-1 increased during repeated thoracenteses in the fibrinous group but not in the nonfibrinous group. In addition, the effusion levels of TNF-␣ and PAI-1 obtained from day 2 and day 3 were significantly higher in the fibrinous group than in nonfibrinous group (Table 3, Fig 1). These findings highly suggested that TNF-␣ play a pivotal role in the regulation of fibrinolytic activity in malignant effusion during repeated thoracenteses. Recently, chest ultrasonography has been widely used to evaluate pleural disorders and has proved useful in determining the nature of pleural effusion.2,3,22,23 The presence of fibrin strands in the pleural effusion shown on real-time chest ultrasonography is of considerable significance. It is a specific ultrasonographic sign for pleural effusion, Clinical Investigations

Figure 1. The changes of effusion PAI-1 (top, A) and TNF-␣ (bottom, B) during repeated thoracenteses in patients with malignant pleural effusions, who were classified into the fibrinous group (solid bars) and the nonfibrinous group (open bars).

highly suggestive of an exudate effusion, and valuable in distinguishing pleural effusion from a solid pleural mass.2,3,22–25 In addition, the presence of mobile fibrin strands in the pleural fluid could be www.chestjournal.org

easily identified on real-time chest ultrasonography. To our knowledge, chest radiography plays a limited role in this issue. Other imaging modalities such as CT and MRI may have some role in evaluating the CHEST / 123 / 4 / APRIL, 2003

1193

Table 4 —Clinical Features and Outcome of Pleurodesis in Fibrinous and Nonfibrinous Groups Characteristics Patients, No. (%) Mean age, yr Male/female gender, No. (%)* Primary tumor, No. (%) Lung Breast Liver Stomach Others Trapped lung, No. (%) Pleurodesis success, No. (%)†

Fibrinous Group

Nonfibrinous Group

11 (42) 67 1/10 (9/91)

15 (58) 68 9/6 (60/40)

5 (46) 1 (9) 0 (0) 3 (27) 2 (18) 0 (0) 11/11 (100)

7 (46) 1 (7) 3 (20) 1 (7) 3 (20) 3 (20) 8/12 (67)

*p ⬍ 0.01. †p ⬍ 0.05.

empyema, pleural adhesion, and/or complex exudate, but not in detecting fibrin strands in the pleural fluid.2,23,26 As a consequence, chest ultrasonography was used to detect the presence of fibrin strands in malignant effusion during and after repeated thoracenteses in this study. To our knowledge, the clinical significance of the development of fibrin strands in malignant effusions after repeated thoracenteses has not been well investigated. Our results showed that after repeated thoracenteses, fibrin strands might develop in malignant effusion, found in 11 of 26 patients (42%). Furthermore, an increase of effusion TNF-␣ and PAI-1 during repeated thoracenteses was found in the fibrinous group only, and the values of effusion TNF-␣ and PAI-1 obtained from day 2 and day 3 were significantly higher in the fibrinous than in nonfibrinous groups (Table 3, Fig 1). These findings strongly supported that the decreased fibrinolytic activity, as evidenced by an increase of PAI-1, in malignant effusion following repeated thoracenteses could be regulated by TNF-␣. The presence of fibrin strands in malignant effusion after repeated thoracenteses is of clinical significance because the successful rate of subsequent pleurodesis was significantly higher in the fibrinous group (11 of 11 patients, 100%) than in the nonfibrinous group (8 of 12 patients, 67%) [Table 4]. Accordingly, the presence of fibrin strands in malignant effusion after repeated thoracenteses may serve as a valuable indicator for predicting the outcome of subsequent pleurodesis. There was no significant difference in the clinical features and the data of pleural fluid analyses obtained from first tap between the fibrinous group and the nonfibrinous group, except for more women in the former group (Tables 3, 4). The reasons why 1194

the data of first tap failed to predict fibrin formation in malignant effusion after repeated thoracenteses remain unknown. These may be explained in part by the limited case number in this study. Another explanation might be due to patient selection criteria used in the present study. The patients who were excluded from this study due to the presence of fibrin strands in the pleural effusions shown on chest ultrasonography before first tap might have higher values of effusion TNF-␣ and PAI-1. Further studies with large population of patients are needed to verify this issue. In conclusion, repeated thoracenteses may cause pleural inflammation and increase local release of TNF-␣, which may subsequently enhance the activity of PAI-1 and result in fibrin formation in malignant effusion. The presence of fibrin strands in malignant effusion after repeated thoracenteses may be of considerable value in predicting the success of subsequent pleurodesis.

References 1 Light RW. Pleural diseases. 4th ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2001; 108 –134 2 McLoud TC, Flower CDR. Imaging the pleura: sonography, CT, MR imaging. AJR Am J Roentgenol 1991; 156:1145–1153 3 Yang PC, Luh KT, Chang DB, et al. Value of sonography in determining the nature of pleural effusion: analysis of 320 cases. AJR Am J Roentgenol 1992; 159:29 –33 4 Bithell TC. Blood coagulation. In: Lee GR, Bithell TC, Foerster J, et al, eds. Wintrobe’s clinical hematology. 9th ed. Philadelphia, PA: Lea & Febiger, 1993; 566 – 615 5 Agrenius V, Chmielewska J, Widstro¨ m O, et al. Pleural fibrinolytic activity is decreased in inflammation as demonstrated in quinacrine pleurodesis treatment of malignant effusion. Am Rev Respir Dis 1989; 140:1381–1385 6 Agrenius V, Gustafsson LE, Widstro¨ m O. Tumor necrosis factor-␣ and nitric oxide, determined as nitrate, in malignant pleural effusion. Respir Med 1994; 88:743–748 7 Agrenius V, Ukale V, Widstro¨ m O, et al. Quinacrine-induced pleural inflammation in malignant pleurisy: relation between drainage time of pleural fluid and local interleukin-1␤. Respiration 1993; 60:366 –372 8 Gu¨ rsel G, Go¨ kcora N, Elbeg S, et al. Tumor necrosis factor-␣ (TNF-␣) in pleural fluids. Tuberc Lung Dis 1995; 76:370 – 371 9 So¨ derblom T, Nyberg P, Teppo AM, et al. Pleural fluid interferon-␥ and tumor necrosis factor-␣ in tuberculous and rheumatoid pleurisy. Eur Respir J 1996; 9:1652–1655 10 Orphanidou D, Gaga M, Rasidakis A, et al. Tumor necrosis factor, interleukin-1 and adenosine deaminase in tuberculous pleural effusion. Respir Med 1996; 90:95–98 11 Hua CC, Chang LC, Chen YC, et al. Proinflammatory cytokines and fibrinolytic enzymes in tuberculous and malignant pleural effusions. Chest 1999; 116:1292–1296 12 de Pablo A, Villena V, Echave-Sustaeta J, et al. Are pleural fluid parameters related to the development of residual pleural thickening? Chest 1997; 112:1293–1297 13 Philip-Joe¨ t F, Alessi MC, Philip-Joe¨ t C, et al. Fibrinolytic and inflammatory processes in pleural effusions. Eur Respir J 1995; 8:1352–1356 Clinical Investigations

14 Colice GL, Rubins JB. Practical management of pleural effusions: when and how should fluid accumulation be drained? Postgrad Med 1999; 105:67–77 15 Davies CWH, Traill ZC, Gleeson FV, et al. Intrapleural streptokinase in the management of malignant multiloculated pleural effusions. Chest 1999; 115:729 –733 16 Rodriguez-Panadero F, Lopez Mejias J. Low glucose and pH levels in malignant pleural effusions: diagnostic significance and prognostic value in respect to pleurodesis. Am Rev Respir Dis 1989; 139:663– 667 17 Lan RS, Lo SK, Chuang ML, et al. Elastance of the pleural space: a predictor for the outcome of pleurodesis in patients with malignant pleural effusion. Ann Intern Med 1997; 126:768 –774 18 Good JT, Taryle DA, Hyers TM, et al. Clotting and fibrinolytic activity of pleural fluid in a model of pleural adhesions. Am Rev Respir Dis 1978; 118:903–908 19 Idell S, Girard W, Koenig KB, et al. Abnormality of pathways of fibrin turnover in the human pleural space. Am Rev Respir Dis 1991; 144:187–194

www.chestjournal.org

20 Idell S, Zweib C, Kumar A, et al. Pathways of fibrin turnover of human pleural mesothelial cells in vitro. Am J Respir Cell Mol Biol 1992; 7:414 – 426 21 Whawell SA, Thompson JN. Cytokine-induced release of plasminogen activator inhibitor-1 by human mesothelial cells. Eur J Surg 1995; 161:315–317 22 Wernecke K. Sonographic features of pleural disease. AJR Am J Roentgenol 1997; 168:1061–1066 23 Mu¨ ller NL. Imaging of the pleura. Radiology 1993; 186:297– 309 24 Marks WM, Filly RA, Callen PW. Real-time evaluation of pleural lesions: new observations regarding the probability of obtaining free fluid. Radiology 1982; 142:163–164 25 Hirsch JH, Roger JV, Mack LA. Real-time sonography of pleural opacities. AJR Am J Roentgenol 1981; 136: 297–301 26 Mason AC, Miller BH, Krasna MJ, et al. Accuracy of CT for the detection of pleural adhesions: correlation with videoassisted thoracoscopic surgery. Chest 1999; 115:423– 427

CHEST / 123 / 4 / APRIL, 2003

1195