C-Reactive Protein in Lymphocytic Pleural Effusions: A Diagnostic Aid in Tuberculous PleuritisGarcia-Pachon E. · Soler M.J. · Padilla-Navas I. · Romero V. · Shum C.
Section of Pneumology, Hospital General Universitario, Elche, Spain Corresponding Author
Background:C-reactive protein (CRP) pleural fluid levels have been found to be higher in tuberculosis and parapneumonic effusions than in other causes of pleural effusion. Objective:The aim of this study was to analyze whether CRP (a simple and inexpensive test) may be a diagnostic aid for tuberculosis in lymphocytic pleural effusions. Methods:One hundred and forty-four patients with a lymphocytic pleural effusion (more than 50% lymphocytes in the differential white blood cell count) were included. The patients were 93 men (65%) and 51 women (35%), aged 64 ± 18 years (mean ± SD). The diagnoses were as follows: tuberculosis, 20; pleural effusion associated with malignancy, 69; transudates, 38; other benign exudates, 17. Results: The CRP pleural fluid level was higher in tuberculous pleuritis (54 ± 24 mg/l) than in lymphocytic effusions of other origin (21 ± 16 mg/l; p < 0.001). High CRP levels (≥50 mg/l) have a high specificity for tuberculosis (95%), and low levels (<30 mg/l) have a high sensitivity (95%) for excluding disease. Conclusions: CRP pleural fluid level determination is useful in the diagnostic workup of lymphocytic pleural effusions. High CRP levels are very suggestive of tuberculous pleuritis, and low CRP levels make this diagnosis unlikely.
Copyright © 2005 S. Karger AG, Basel
In many areas of the world, tuberculosis remains the most common cause of pleural effusion in the absence of demonstrable pulmonary disease . To obtain tubercle bacilli is infrequent in pleural fluid or pleural tissue ; consequently, the analysis of the pleural fluid is useful in the diagnosis of tuberculous pleuritis. Several biochemical parameters are of interest in diagnosing tuberculous pleural effusion, especially adenosine deaminase (ADA) [3,4,5]. Since almost all patients with tuberculosis pleural fluid have more than 50% lymphocytes in the pleural white blood cell count [2,6,7,8], the use of ADA has been evaluated in the subgroup of patients with lymphocytic pleural effusions, showing higher diagnostic value [9,10,11,12].
Unfortunately, the ADA test is not available worldwide; most importantly, the test is not accessible in most developing countries where the incidence of tuberculosis is very high. For this reason, alternative tests with lower cost and greater availability are of great interest.
C-reactive protein (CRP) is an acute-phase protein widely used as a marker of inflammation and tissue injury . Its determination is simple, quick and inexpensive. In pleural fluid, CRP levels have been found higher in tuberculosis and parapneumonic effusions than in other causes of pleural effusions . The aim of this study was to analyze whether CRP may be a diagnostic aid for tuberculosis in lymphocytic pleural effusions.
Material and Methods
A diagnostic thoracocentesis was performed in 205 consecutive patients. Sixty-one patients were excluded, 41 with nonlymphocytic pleural effusions and 20 with lymphocytic pleural effusions in whom no diagnosis could be obtained. The remaining 144 patients were included in the study. A pleural fluid was considered to be lymphocytic when the differential nucleated cell count revealed more than 50% lymphocytes. The patients were 93 men (65%) and 51 women (35%), aged 64 ± 18 years (mean ± SD).
The final diagnoses of the pleural effusions were established on the basis of well-defined diagnostic criteria usually accepted in the literature and previously detailed by our group [14, 15]. The diagnosis of tuberculous pleural effusion was based on the presence of caseous granulomata in the pleural biopsy and/or positive culture for Mycobacterium tuberculosis in pleural fluid or biopsy material, or positive sputum culture for M. tuberculosis with an exudative pleural effusion. In patients aged 35 years or younger, diagnosis was also made on the basis of a positive skin test and an ADA value >47 U/l, as well as successful treatment with antituberculosis chemotherapy .
CRP in pleural fluid was determined by the turbidimetric immunoassay method with a commercial kit (Dimension, Du Pont, Wilmington, Del., USA). The cost of each determination of CRP was lower than USD 1. The normal value of CRP in serum is <10 mg/l.
Statistical analysis was made by a two-tailed unpaired ttest for the differences between CRP pleural fluid levels in tuberculous and nontuberculous pleuritis. A pvalue of <0.05 was considered statistically significant. Sensitivity, specificity, positive likelihood ratio and negative likelihood ratio with 95% confidence intervals of CRP at different cut-off points were calculated.
Of the 144 patients included in the study, 20 were diagnosed as having tuberculous pleuritis (14%). Fifteen of them were diagnosed by biopsy (11 patients), culture (1 patient) or both biopsy and culture (3 patients), and the remaining 5 patients on the basis of an ADA level >47 U/l and response to therapy. Diagnoses and CRP levels are detailed in table 1 and in figure 1. On mean, patients with tuberculous pleural effusion had a significantly higher CRP level in pleural fluid. The mean CRP level ± SD in the 20 patients with lymphocytic effusion in whom no diagnosis was obtained was 28 ± 10 mg/l; none of them had a level >50 mg/l.
Table 1. Diagnoses and values of C-reactive protein in lymphocytic pleural fluids
Fig. 1. Individual values of C-reactive protein in each diagnostic group.
The diagnostic accuracy (sensitivity, specificity, positive and negative likelihood ratios) of CRP in pleural fluid for diagnosing tuberculosis was calculated at different cut-off points (table 2). A CRP level <30 mg/l virtually ruled out the possibility of tuberculosis as the cause of a lymphocytic pleural effusion (sensitivity 95%). A specificity >90% for diagnosing tuberculosis was obtained at a CRP value ≥45 mg/l. When the pleural fluid CRP level was ≥50 mg/l, the specificity was 95%.
Table 2. Diagnostic accuracy of C-reactive protein for tuberculous pleuritis at different cut-off values
Since tuberculous pleurisy is always an exudative effusion, data of diagnostic value (specificity) excluding transudates have been analyzed. Results are not significantly different to those obtained in all patients as a whole; for CRP ≥45 mg/l, specificity was 88% (95% CI: 79–94%), and for CRP ≥50 mg/l, it was 93% (95% CI: 85–97%).
This study shows that the pleural fluid CRP level in lymphocytic pleural effusions has diagnostic value in tuberculous pleuritis. Depending on the CRP level, this diagnosis can be strongly suggested (>50 mg/l) or, on the contrary, can be considered highly unlikely (a CRP pleural fluid level <30 mg/l).
Ideally, the diagnosis of tuberculous pleuritis is made by the demonstration of M. tuberculosis in pleural fluid or in pleural tissue. However, smears of the pleural fluid for mycobacteria are frequently negative , and cultures, when positive, take several weeks until results can be obtained [8, 16]. For this reason, in addition to the clinical and radiological characteristics, the cytological and biochemical findings can aid in diagnosing tuberculous pleuritis. At present, several tests are of great interest for diagnosing tuberculous pleural effusion, such as ADA [3,4,5], interferon-=γ [4, 17, 18], lysozyme [4, 19], the polymerase chain reaction [20, 21] or specific antibodies . Lymphocytic predominance in pleural fluid is the rule in tuberculous effusions . This characteristic has been used in order to improve the efficiency of ADA [9,10,11], the most widely used biochemical test for detecting tuberculosis.
However, the tests mentioned above need specific and/or expensive equipment that is not available in most laboratories. From our study, we have determined that measurement of CRP level in pleural fluid can be useful, with the advantage that CRP can be determined in conventional autoanalyzers and is an inexpensive test. In patients with lymphocytic pleural effusion, values of CRP in the pleural fluid >50 mg/l should strongly suggest the possibility of tuberculous pleuritis. On the contrary, low CRP levels (<30 mg/l) virtually ruled out this possibility. Since a biochemical diagnostic test of pleural fluid cannot be considered in a dichotomous view, we have calculated the diagnostic values at different cut-off levels that are more appropriate for clinical purposes.
Previous studies on CRP in pleural fluid concluded that this parameter may differentiate transudates from exudates, but with worse results than those of classical tests [23,24,25]. In one study , the authors found that CRP levels were significantly higher in parapneumonic and tuberculous effusions, and our group  reported previously that in pleural exudates a CRP <20 mg/l suggests a malignant origin, whereas a level >45 mg/l virtually ruled out this possibility. Very recently, Chierakul et al.  studied CRP in 55 patients with tuberculous pleuritis and in 60 patients with malignant pleural effusions. They found that with a cut-off level of CRP ≥30 mg/l, sensitivity was 72 and specificity 93%. However, in the study by Chierakul et al. , pleural effusions due to other origins were excluded.
It is well known that other characteristics of the pleural fluid might be useful for diagnosing tuberculous pleurisy. Pleural fluid from patients with tuberculosis rarely contains more than 5% mesothelial cells [6, 7], but the absence of mesothelial cells is not diagnostic of tuberculosis, because in any condition in which the pleural surfaces are extensively involved by an inflammatory process, mesothelial cells are not found in the pleural fluid . A pleural fluid protein level >5 g/dl suggests the diagnosis , but is not specific. In the past, it was believed that in tuberculous pleural fluid, the glucose level was reduced and pH was lower than in malignant effusions [27, 28], but now it is known that these data were inaccurate and neither glucose nor pH pleural fluid levels add significant information in cases of tuberculous pleuritis [1, 29, 30]. In conclusion, CRP pleural fluid level determination is a simple and inexpensive test useful in the diagnostic workup of lymphocytic pleural effusions, especially in areas where more sophisticated tests are not available. High CRP levels (>50 mg/l) are very suggestive of tuberculous pleuritis, and low CRP levels make this diagnosis unlikely. Obviously, the exclusive use of this parameter is not enough for diagnosing tuberculous pleurisy, but the CRP pleural fluid level adds significant diagnostic information that should be considered together with all the other epidemiological, clinical, cytological and biochemical data.
Eduardo Garcia-Pachon, MD
Section of Pneumology, Hospital General Universitario
ES–03203 Elche, Alicante (Spain)
Tel. +34 96 667 90 00, Fax +34 96 667 91 08, E-Mail firstname.lastname@example.org
Received: July 9, 2004
Accepted after revision: October 25, 2004
Number of Print Pages : 4
Number of Figures : 1, Number of Tables : 2, Number of References : 30
Respiration (International Journal of Thoracic Medicine)
Vol. 72, No. 5, Year 2005 (Cover Date: September-October 2005)
Journal Editor: Bolliger, C.T. (Cape Town)
ISSN: 0025–7931 (print), 1423–0356 (Online)
For additional information: http://www.karger.com/res