Proinflammatory and Antiinflammatory Cytokine Levels in Complicated and Noncomplicated Parapneumonic Pleural Effusions

Transcription

1 CHEST Proinflammatory and Antiinflammatory Cytokine Levels in Complicated and Noncomplicated Parapneumonic Pleural Effusions Evaldo Marchi, MD, FCCP; Francisco S. Vargas, MD; Milena M. Acencio, BS; Rosa M. S. Sigrist, MD; Marjourie D. A. Biscaro, MD; Leila Antonangelo, MD; Lisete R. Teixeira, MD, FCCP; and Richard W. Light, MD, FCCP Original Research DISORDERS OF THE PLEURA Objectives: This study aimed to evaluate a panel of proinflammatory and antiinflammatory cytokines in noncomplicated and complicated parapneumonic pleural effusions and to correlate their levels with pleural fluid biochemical parameters. Methods: Serum and pleural effusion were collected from 60 patients with noncomplicated (n 5 26) or complicated (n 5 34) parapneumonic effusions and assayed for cytologic, biochemical, and proinflammatory and antiinflammatory cytokines. Student t test was used to compare serum and pleural fluid values, Spearman correlation to analyze the relationship between pleural fluid cytokines and biochemical parameters, and accuracy of pleural fluid cytokine levels to determine the optimal cutoff value for identification of complicated Corrections for multiple comparisons were applied and a P value,.05 was accepted as significant. Results: Serum and pleural fluid cytokine levels of IL-8, vascular endothelial growth factor (VEGF), IL-10, and tumor necrosis factor (TNF) soluble receptor (sr) II were similar between groups. In contrast, complicated effusions had higher levels of pleural fluid IL-1 b, IL-1 receptor antagonist (ra), and TNF sri. Negative correlations were found between pleural fluid glucose with IL-1 b and TNF sri and positive correlations between lactic dehydrogenate (LDH) with IL-1 b, IL-8, and VEGF. Pleural fluid levels of IL-1 b, IL-1ra, and TNF sri were more accurate than IL-8, VEGF, IL-10, and TNF srii in discriminating complicated Conclusions: Both proinflammatory and antiinflammatory cytokine levels in pleural fluid are elevated in complicated in comparison with noncomplicated parapneumonic pleural effusions, and they correlate with both pleural fluid glucose and LDH levels. IL-1 b, IL-1ra, and TNF sri had higher sensitivity and specificity than IL-8, VEGF, IL-10, and TNF srii in discriminating complicated CHEST 2012; 141(1): Abbreviations: IL-1ra 5 IL-1 receptor antagonist; LDH 5 lactic dehydrogenate; sr 5 soluble receptor; TNF 5 tumor necrosis factor; VEGF 5 vascular endothelial growth factor Pleural effusions may occur in 20% to 57% of patients following pneumonia. Prompt diagnosis of parapneumonic effusions is desirable, since treatment modalities vary from a conservative approach in noncomplicated effusions to an aggressive approach in complicated effusions and empyema. 1,2 Normally, pleural fluid biochemical parameters are used to differentiate complicated from noncomplicated effusions, namely ph, glucose, and lactic dehydrogenase (LDH) levels. However, none of these parameters has a high sensitivity, and differentiation between noncomplicated and complicated parapneumonic effusions may be a challenge in clinical practice.3 Other parameters have been investigated as mark ers of complicated parapneumonic effusions, especially pleural fluid inflammatory cytokines IL-8 and tumor necrosis factor (TNF). 4,5 However, the role of other inflammatory cytokines, such as IL-1 b, a potent TNF stimulant known to be involved in the acute phase of inflammation by recruiting neutrophils and activating lymphocyte proliferation, and vascular endothelial growth factor (VEGF), a cytokine involved in cell proliferation and angiogenesis, considered CHEST / 141 / 1 / JANUARY,

2 markers of acute inflammation, have not been studied in detail in parapneumonic Likewise, there are no studies in the literature addressing the role of antiinflammatory cytokines in complicated and noncomplicated parapneumonic pleural Among those, IL-10 is known to downregulate the expression of several cytokines and, thus, counteract the excessive immune and inflammatory effects produced by these mediators. In addition, the antagonist of the cell membrane receptor of IL-1 (IL-1ra) is known to be a natural blocker of the proinflammatory cytokine IL-1, and the soluble receptors (srs) of TNF (TNF sri and srii) act as blockers of the binding of TNF in cell membranes. As the dynamics of the inflammatory process in the pleural cavity with parapneumonic effusions are still not completely clarified, in this study we aimed to determine the levels of proinflammatory and antiinflammatory cytokines in noncomplicated and complicated parapneumonic pleural effusions and to correlate their levels with pleural fluid biochemical parameters. We hypothesized that noncomplicated effusions would have higher levels of antiinflammatory cytokines that could play a role in the dynam ics of inflammation underlying parapneumonic Materials and Methods This is a prospective study approved by the ethics committee of the institutions involved (approval number 998/05 by the ethics committee for research projects of Hospital das Clinicas, University of São Paulo, Brazil). Sixty patients with documented pneumonia and pleural effusion were recruited from January 2008 to December 2009 and managed by the same medical team. Parapneumonic effusions were classified as noncomplicated (n 5 26) or complicated (n 5 34). Patients who received chest tubes for the treatment of their parapneumonic effusions were classified as having complicated parapneumonic The decision to place a chest tube was left to the discretion of the attending physician, although it is likely that this decision was influenced by the biochemical characteristic of the fluid. As only Manuscript received December 15, 2010; revision accepted June 1, Affiliations: From the Pulmonary Division (Drs Marchi, Vargas, Antonangelo, and Teixeira and Ms Acencio), Heart Institute (InCor), University of São Paulo Medical School; the Jundiaí Medical College (Drs Marchi, Sigrist, and Biscaro), São Paulo, Brazil; and Vanderbilt University (Dr Light), Nashville, TN. Funding/Support: Supported by the Foundation to Support Research from the State of São Paulo, Brazil (FAPESP) [Grant ]. Correspondence to: Evaldo Marchi, MD, FCCP, Pleura Laboratory, University of São Paulo Medical School, Pulmonary Division, Heart Institute (Incor), and Medical College of Jundiaí, Alameda das Castanheiras, 196-Cd. Terras de São Carlos, Jundiaí , São Paulo, Brazil; evmarchi@uol. com.br. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( site/misc/reprints.xhtml ). DOI: /chest a few patients had chest ultrasonography and/or CT scans, these methods were not used to differentiate complicated from noncomplicated Cytologic and Biochemical Analysis Serum and pleural effusion samples were collected in ethylenediaminetetraacetic acid for cytologic and cytokine analysis and in dry tubes for biochemical analysis. Pleural fluid samples were placed in a hemocytometer for determination of total nucleated cell count and onto slides for Leishman preparation and determination of neutrophil percent. Other samples were centrifuged at 1,000 rpm for 10 min at 4 C, withdrawn immediately for LDH analysis, and stored at 280 C for cytokine determination. Cytokine Measurement Proinflammatory cytokines IL-1 b, IL-8, and VEGF (R&D System Inc), and antiinflammatory cytokines IL-1ra, IL-10, and TNF sri and TNF srii (R&D System Inc) were measured by enzyme-linked immunoabsorbent assay according to the manufacturer s directions. Quantification of the cytokines was done by comparison of the optical density in the enzyme-linked immunoabsorbent assay reader (Power Wave; Bio-Tek). Lower level detection values for IL-1 b were 3.9 pg/ml; levels for IL-8, VEGF, IL-1ra, IL-10, and TNF sri and srii were 31.3 pg/ml. Statistical Analysis Statistical analyses were performed using statistical computer software (Sigma Stat; Jandel Scientific). Student t test was used to compare the blood and pleural fluid values, and if the logtransformed data did not satisfy normality tests, the Mann-Whitney rank sum test was used. Data are expressed as the mean SD if the data set was normally distributed and the median and the 25th and 75th percentiles if the data were not normally distributed. Spearman correlation was used to measure the strength of association between biochemical parameters and proinflammatory and antiinflammatory cytokines in pleural fluid of all patients together. A P value,.05 was accepted as significant. Corrections for multiple comparisons were applied using Holm-Bonferroni method where appropriate. The optimal sensitivity and specificity of pleural fluid cytokine levels for distinguishing noncomplicated from complicated effusions were evaluated by the receiver operating characteristics by analyzing the area under the curve (AUC) and determining the optimal cutoff value. Results Blood and Pleural Fluid Cytologic and Biochemical Parameters The peripheral WBC counts were significantly higher in complicated in comparison with noncomplicated effusions (18,940 5,990 cells/m L vs 12,754 4,115 cells/ m L), whereas blood neutrophil percentages were similar in both groups. In pleural fluid, median total nucleated cell counts and neutrophil percents were significantly higher in complicated effusions (respectively, 2,300 cells/ m L vs 624 cells/ ml and 78% vs 44%) ( Table 1 ). The median pleural fluid LDH level was significantly higher (2,369 IU/L vs 459 IU/L) and the median pleural fluid glucose 184 Original Research

3 Table 1 Serum and Pleural Fluid Values of WBC and Neutrophils and Pleural Fluid Values of LDH and Glucose in Noncomplicated and Complicated PPE Parameter Media Noncomplicated PPE Complicated PPE P Valuea Corrected WBC, cells/ m L Serum 12,754 4,115 18,940 5,990, Pleural 624 (160-1,936) 2,300 (832-8,580) N, % Serum NS NS Pleural 44 (23-79) 78 (68-87) LDH, IU/L Pleural 459 ( ) 2,369 (1,931-3,430), Glucose, mg/dl Pleural 103 (88-136) 16 (10-28), Data presented as mean SD or median (25th-75th percentile). LDH 5 lactic dehydrogenate; N 5 neutrophil; NS 5 not significant; PPE 5 parapneumonic pleural effusion. level significantly lower (16 mg/dl vs 103 mg/dl; P 5.006) in complicated in comparison with noncomplicated effusions ( Table 1 ). Serum and Pleural Fluid Cytokines IL-1b, IL-8, and IL-10 levels were undetectable in serum of patients with either complicated or noncomplicated No significant difference was found in serum proinflammatory VEGF and antiinflammatory IL-1ra, TNF sri, and TNF srii cytokine levels between noncomplicated and complicated In addition, pleural fluid levels of IL-8, VEGF, IL-10, and TNF srii did not differ between groups ( Table 2 ). In contrast, noncomplicated effusions had significantly lower median levels of pleural fluid proinflammatory IL-1 b (3.9 pg/ml vs 246 pg/ml) as well as of antiinflammatory cytokine levels of IL-1ra (852 pg/ml vs 3,168 pg/ml) and TNF sri (8,962 pg/ml vs 10,319 pg/ml) ( Table 2 ). The median values of pleural fluid/serum ratio of proinflammatory and antiinflammatory cytokines did not differ between groups ( Table 2 ). Relationship Between Cytokines and Biochemical Levels in Pleural Effusion When all patients were considered, significant negative correlations were found between glucose and proinflammatory cytokine IL-1 b ( r ) and antiinflammatory cytokine TNF sri ( r ). In addition, levels of LDH were significantly positively correlated with proinflammatory cytokines IL-1 b ( r ), IL-8 ( r ), and VEGF ( r ) ( Table 3 ). Table 2 Serum and Pleural Fluid Levels and Pleural Fluid/Serum Ratios of Proinflammatory and Antiinflammatory Cytokines in Noncomplicated and Complicated PPE Cytokine Media Noncomplicated PPE Complicated PPE P Valuea Corrected Proinflammatory IL-1 b Serum ND ND NS NS Pleural ND 246 (31-697), IL-8 Serum ND ND NS NS Pleural 371 (213-1,738) 1,804 (678-2,088).019 NS VEGF Serum 363 ( ) 410 ( ) NS NS Pleural 580 (242-1,283) 1,355 (1,008-4,373).015 NS Ratio P/S 1.3 ( ) 3.8 ( ).012 NS Antiinflammatory IL-1ra Serum 341 (74-2,130) 286 (134-2,330) NS NS Pleural 852 (92-2,867) 3,168 (1,821-4,132) Ratio P/S 1.5 ( ) 8.2 (1.1-26).018 NS IL-10 Serum ND ND NS NS Pleural 179 ( ) 212 ( ) NS NS TNF sri Serum 1,771 (1,485-2,977) 2,770 (1,661-2,273) NS NS Pleural 8,962 (8,260-9,926) 10,319 (9,730-13,161) Ratio P/S 4.3 ( ) 4.3 ( ) NS NS TNF srii Serum 7,000 (5,745-13,630) 10,147 (7,928-12,970) NS NS Pleural 16,951 (15,202-20,260) 20,159 (18,214-21,291) NS NS Ratio P/S 2.3 ( ) 2.3 ( ) NS NS Data presented as median (25th-75th percentiles). When serum and/or pleural fluid levels of cytokines were undetectable, pleural fluid/serum ratios were not calculated, and comparisons between groups were done using minimum detectable levels for IL-1 b (3.9 pg/ml) and for IL-10 and IL-8 (31.3 pg/ml). IL-1ra 5 IL-1 receptor antagonist; ND 5 nondetectable; P/S 5 pleural/serum; sr 5 soluble receptor; TNF 5 tumor necrosis factor; VEGF 5 vascular endothelial growth factor. See Table 1 legend for expansion of other abbreviation. CHEST / 141 / 1 / JANUARY,

4 Table 3 Correlation Between Pleural Fluid Levels of Proinflammatory and Antiinflammatory Cytokines and Biochemical Parameters in All Patients Cytokine Glucose LDH r ValueP Valuer ValueP Value IL-1 b 20.64, ,.001 IL ,.001 VEGF ,.001 IL-1ra IL TNF sri 20.60, TNF srii When correction for multiple comparisons was applied, a corrected for significant values were: glucose vs IL-1 b , glucose vs TNF sri , LDH vs IL-1 b , LDH vs IL , and LDH vs VEGF For IL-1ra and LDH, after correction for multiple comparisons, P value was not significant. See Table 1 and 2 legends for expansion of abbreviations. Optimal Sensitivity and Specificity of Pleural Effusion Cytokines to Identify Complicated Parapneumonic Effusions Glucose, 58 mg/dl had the best sensitivity and specificity as reflected in a receiver operating characteristics curve to predict complicated effusions (sensitivity and specificity 100%). In addition, LDH. 1,158 IU/L had 90% sensitivity and 96% specificity in identifying complicated parapneumonic effusions, with AUC of 0.93 ( Table 4 ). Among cytokines, IL-1 b cutoff levels. 3.9 pg/ml had the highest sensitivity and specificity to discriminate between noncomplicated and complicated effusions (AUC, 0.88; sensitivity, 100%; specificity, 71%), followed by IL-1ra. 1,252 pg/ml (AUC, 0.78; sensitivity, 100%; specificity, 61%) and TNF sri. 9,272 pg/ml (AUC, 0.78; sensitivity, 86%; spec ificity, 62%). Proinflammatory IL-8 and VEGF pleural fluid levels had lower sensitivity and specificity to discriminate between noncomplicated and com- plicated parapneumonic effusions (IL pg/ml; AUC, 0.72; sensitivity, 77%; specificity, 67%; and VEGF. 1,283 pg/ml; AUC, 0.71; sensitivity, 67%; specificity, 76%) as well as antiinflammatory TNF srii. 17,785 pg/ml (sensitivity, 77%; specificity, 67%; AUC, 0.62) ( Table 4 ). Discussion Parapneumonic effusions are a common clinical problem, and the differentiation of complicated from noncomplicated effusions is a cornerstone in their therapeutic management. Pleural fluid ph, 7.2 has been suggested by guidelines as the most accurate marker to differentiate complicated from noncomplicated 3 However, pleural ph measurement may not be reliable because it is not usually collected in heparin-coated syringes and is frequently measured with ph indicator strips or ph meters rather than blood gas machines. 6,7 Therefore, in this article we used the classic definition of complicated parapneumonic effusion: a parapneumonic effusion that requires tube thoracostomy or thoracoscopy for its resolution. 1 Pleural fluid glucose, 60 mg/dl and LDH. 1,000 IU/L (or levels of LDH three times higher than the upper normal level of the method) have been suggested as other biochemical markers to make this differentiation. 8 We found that pleural fluid glucose levels, 58 mg/dl and LDH levels. 1,158 IU/L had a high sensitivity and specificity in predicting complicated effusions, implying that both parameters are good markers to differentiate complicated from noncomplicated parapneumonic Although these biochemical parameters are considered standards to define complicated effusions and are currently used in the clinical practice, the pathophysiology of inflammation in parapneumonic effusions is still not completely clarified. To better understand the dynamics of the inflammatory process Table 4 Optimal Sensitivity and Specificity of Pleural Effusion Biochemical Parameters and Proinflammatory and Antiinflammatory Cytokines for the Diagnosis of Complicated Parapneumonic Effusions Parameter Cutoff Sensitivity, % Specificity, % AUC 95% CI Biochemical Glucose, 58 mg/dl LDH. 1,158 IU/L Proinflammatory cytokines IL-1 b. 3.9 pg/ml IL pg/ml VEGF. 1,283 pg/ml Antiinflammatory cytokines IL-1ra. 1,252 pg/ml IL pg/ml TNF sri. 9,272 pg/ml TNF srii. 17,785 pg/ml AUC 5 area under the curve. See Table 1 and 2 legends for expansion of other abbreviations. 186 Original Research

5 in the pleural cavity with parapneumonic effusions, cytokines have been studied as markers of inflammation and pleural fibrosis, 9-11 and as potential therapeutic targets. Many studies have addressed the role of cytokines in the differentiation of parapneumonic from effusions of other etiologies However, there are few reports addressing the role of proinflammatory cytokines in the pathophysiology of complicated and noncomplicated effusions, 4,5 and, to our knowledge, the role played by antiinflammatory cytokines in parapneumonic effusions has not been addressed previously in the literature. Our study compared the levels of proinflammatory and antiinflammatory cytokines in noncomplicated and complicated parapneumonic Our findings indicated that the levels of both proinflammatory and antiinflammatory cytokines were elevated in the pleural fluid of complicated parapneumonic Noncomplicated parapneumonic effusions had significantly lower levels of pleural fluid proinflammatory IL-1 b in comparison with complicated However, different from what we first hypothesized, antiinflammatory cytokine levels of IL-1ra and TNF sri, but not IL-10 and TNF srii, were lower in noncomplicated effusions in comparison with complicated Noncomplicated effusions had undetectable levels of pleural fluid IL-1 b in comparison with elevated levels of this cytokine in complicated Proinflammatory cytokines IL-1 b, TNF- a, and IL-8 have been described as markers of complicated 4,5,9-11 Pleural fluid TNF- a was reported to be significantly higher in complicated than in noncomplicated effusions (133.0 pg/ml vs 39.1 pg/ml), with 78% sensitivity and 89% specificity to identify complicated effusions at a cutoff value of 80 pg/ml. 5 In a study with 100 patients, 49 with noncomplicated and 51 with complicated parapneumonic effusions, pleural fluid IL-8 was reported to be higher in complicated in comparison with noncomplicated effusions (5,960 pg/ml vs 218 pg/ml, P,.001) and at a cutoff value of 1,000 pg/ml was able to predict complication in parapneumonic effusions with a sensitivity of 84% and a specificity of 82%. 4 In our study we found that pleural fluid IL-8 levels were higher, although not significantly, in complicated effusions (1,804 pg/ml vs 371 pg/ml), with 77% sensitivity and 67% specificity to identify complicated effusions at a cutoff value of 574 pg/ml. San José et al, 10 studying the role of cytokines IL-8 and IL-1 b in parapneumonic, empyema, tuberculous, and malignant effusions, found that IL-8 had a sensitivity of 73.5% and specificity of 65.1% at a cutoff of. 123 pg/ml, and IL-1 b had a sensitivity of 55.6% and specificity of 91.3% at a cutoff of pg/ml, and that the combination of IL-1 b and neutrophil count improved the yield to diagnose infectious effusions, with a sensitivity of 75.7% and a specificity of 83.1%. In our study we found pleural fluid IL-1 b levels higher in complicated effusions (246 pg/ml vs 3.9 pg/ml; P,.0028), with 100% sensitivity and 71% specificity to identify complicated effusions at a cutoff value. 3.9 pg/ml. Antiinflammatory cytokines are inhibitors of cytokines, and are known to be elevated in several inflammatory diseases, including rheumatoid arthritis, chronic inflammatory bowel disease, and sepsis In pleural effusions of septic patients the antiinflammatory cytokines IL-1ra and TNF sri and srii were reported to be elevated, suggesting that an antiinflammatory process probably occurs in order to oppose the effects of proinflammatory cytokines. 14 The antiinflammatory cytokine IL-1ra, a natural blocker of the proinflammatory cytokine IL-1 b, 15 was also found elevated in patients with tuberculous and parapneumonic effusions in comparison with transudates and malignant 16 It is known that the inflammatory response is complex, and the imbalance between proinflammatory and antiinflammatory cytokines is only one of the factors involved in inflammation. In our study we found that the antiinflammatory cytokine IL-1ra is elevated in the pleural fluid of complicated in comparison with noncomplicated We also studied the correlations between pleural fluid cytokines and biochemical parameters in all patients together. Significant negative correlations were found between glucose and proinflammatory cytokine IL-1 b and antiinflammatory cytokine TNF sri. In contrast, levels of LDH positively correlated with proinflammatory cytokines IL-1 b, IL-8, and VEGF, but with none of the antiinflammatory cytokines. As proinflammatory cytokines are known to be elevated in parapneumonic effusions, 17,18 negative correlations with low glucose levels and positive correlations with high LDH levels were expected. However, a correlation between the antiinflammatory cytokine TNF RI and glucose was also found, indicating that antiinflammatory cytokines may play a role in the dynamic of inflammation in parapneumonic Our study also indicated that among biochemical parameters, glucose, 58 mg/dl and LDH. 1,158 IU/L had a high sensitivity and specificity to predict complicated Among the pleural fluid cytokines, IL-1 b. 3.9 pg/ml had the highest sensitivity and specificity to predict complication in parapneumonic effusions, followed by IL-1ra. 1,252 pg/ml and TNF sri. 9,272 pg/ml. In addition, pleural fluid levels of IL pg/ml, VEGF. 1,283 pg/ml, and TNF srii. 17,785 pg/ml may be potentially CHEST / 141 / 1 / JANUARY,

6 useful to indicate complication of parapneumonic Antiinflammatory cytokines have been considered in strategies targeting inflammatory mediators in several inflammatory diseases In this study we sought to study the imbalance of proinflammatory and antiinflammatory cytokines in noncomplicated and complicated parapneumonic effusions, aiming to better understand the role of these inflammatory mediators in the pathogenesis of this disease. We do not think that determining the levels of the proinflammatory and antiinflammatory in order to differentiate complicated from noncomplicated parapneumonic effusions is useful in clinical practice, since biochemical parameters are readily available and practical tools to make this differentiation. However, based on the knowledge on how inflammation occurs in parapneumonic effusions, we speculate that future strategies can be used to target both the antiinflammatory and proinflammatory cytokines in order to better control and prevent complication in parapneumonic Among the limitations of this study, we point out that measuring cytokines is always a challenge; for some situations values may not be detectable, and in others there is a wide range of values. For this reason, the number of patients included should be higher to ensure better accuracy. In addition, as pleural fluids were sampled only once, several factors could influence the inflammatory response in pleural fluid of parapneumonic effusions, including the type of bacteria, comorbidities, and different times of disease course. We also point out that, ideally, the pleural fluid ph should be measured, since this biochemical parameter is considered the gold standard in the differentiation of complicated from noncomplicated In our study, however, we could not ensure that ph was collected and measured accurately using a blood gas machine, and for that reason we elected not to report the values for this parameter. Last, the definition of uncomplicated and complicated parapneumonic effusion is not clear-cut. In the present study we defined complicated parapneumonic effusions as those that received chest tubes. However, the decision to insert chest tubes was made with the knowledge of the pleural fluid glucose and LDH, and this could have influenced which patients were given chest tubes. In conclusion, our results show that pleural fluid proinflammatory cytokines IL-1 b, and the antiinflammatory cytokines IL-1ra and TNF sri, but not IL-10 and TNF srii, are elevated in complicated in comparison with noncomplicated parapneumonic pleural The strong correlations of pleural fluid cytokines IL-1 b and TNF sri with pleural fluid glu- cose, and of IL-1 b, IL-8, and VEGF with pleural fluid LDH indicate that cytokines may play a role in the modulation of inflammation in pleural fluid of parapneumonic Based on these results, we speculate that future therapies based on targeting proinflammatory and antiinflammatory mediators should be considered in the treatment of parapneumonic Acknowledgments Author contributions : Dr Marchi: contributed to patient care, data collection, writing, revision, and approval of the manuscript. Dr Vargas: contributed to writing, revision, and approval of the manuscript. Ms Acencio: contributed to laboratory data analysis, writing, revision, and approval of the manuscript. Dr Sigrist: contributed to patient care, data collection, writing, revision, and approval of the manuscript. Dr Biscaro: contributed to patient care, data collection, writing, revision, and approval of the manuscript. Dr Antonangelo: contributed to laboratory data analysis, writing, revision, and approval of the manuscript. Dr Teixeira: contributed to writing, revision, and approval of the manuscript. Dr Light: contributed to writing, revision, and approval of the manuscript. Financial/nonfinancial disclosures: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data, or in the preparation of the manuscript. Other contributions: The authors thank Gabriela G. Carnevale, PhD; Sirlei Siani Moraes, PhD; and Carlos S. R. Silva, PhD, for their valuable collaboration in this study. References 1. Light RW. Parapneumonic effusions and empyema. Proc Am Thorac Soc ;3(1): Rahman NM, Davies RJO. Effusion from infections: parapneumonic effusions and empyema. In: Light RW, Lee G. Textbook of Pleural Diseases. 2nd ed. London, England: Hodder Arnold : Davies HE, Davies RJ, Davies CW ; BTS Pleural Disease Guideline Group. Management of pleural infection in adults: British Thoracic Society Pleural Disease Guideline Thorax ;65(suppl 2 ):ii41-ii Porcel JM, Galindo C, Esquerda A, et al. Pleural fluid interleukin-8 and C-reactive protein for discriminating complicated non-purulent from uncomplicated parapneumonic effusions. Respirology ;13(1): Porcel JM, Vives M, Esquerda A. Tumor necrosis factoralpha in pleural fluid: a marker of complicated parapneumonic effusions. Chest ;125(1): Rahman NM, Mishra EK, Davies HE, Davies RJ, Lee YC. Clinically important factors influencing the diagnostic measurement of pleural fluid ph and glucose. Am J Respir Crit Care Med ;178(5): Bowling MR, Chatterjee A, Conforti J, Adair N, Haponik E, Chin R Jr. Perceptions vs. reality: measuring of pleural fluid ph in North Carolina. N C Med J ;70(1): Light RW, Girard WM, Jenkinson SG, George RB. The incidence and significance of parapneumonic effusions. Am J Med. 1980;69 (4): Original Research

7 9. Xirouchaki N, Tzanakis N, Bouros D, et al. Diagnostic value of interleukin-1alpha, interleukin-6, and tumor necrosis factor in pleural effusions. Chest ;121 (3 ): San José ME, Valdes L, Gonzalez-Barcala FJ, et al. Diagnostic value of proinflammatory interleukins in parapneumonic effusions. Am J Clin Pathol ;133 (6 ): Chung CL, Chen CH, Sheu JR, Chen YC, Chang SC. Proinflammatory cytokines, transforming growth factor-beta1, and fibrinolytic enzymes in loculated and free-flowing pleural exudates. Chest ;128 (2 ): Geyer M, Müller-Ladner U. Actual status of antiinterleukin-1 therapies in rheumatic diseases. Curr Opin Rheumatol ; 22 (3 ): Bosani M, Ardizzone S, Porro GB. Biologic targeting in the treatment of inflammatory bowel diseases. Biologics ;3 : Marie C, Losser MR, Fitting C, Kermarrec N, Payen D, Cavaillon JM. Cytokines and soluble cytokine receptors in pleural effusions from septic and nonseptic patients. Am J Respir Crit Care Med ;156 (5 ): Hannum CH, Wilcox CJ, Arend WP, et al. Interleukin-1 receptor antagonist activity of a human interleukin-1 inhibitor. Nature ;343 (6256 ): Yanagawa H, Yano S, Haku T, Ohmoto Y, Sone S. Interleukin-1 receptor antagonist in pleural effusion due to inflammatory and malignant lung disease. Eur Respir J ;9 (6 ): Porcel JM. Pleural fluid tests to identify complicated parapneumonic effusions. Curr Opin Pulm Med ; 16 ( 4 ): Porcel JM, Vives M, Cao G, et al. Biomarkers of infection for the differential diagnosis of pleural effusions. Eur Respir J ;34 (6 ): CHEST / 141 / 1 / JANUARY,