- Top of page
Background and objective: The aim of this study was to determine whether high levels of pleural adenosine deaminase (pADA) are predictive for tuberculosis when pleural effusions do not satisfy the criteria for lymphocytic effusions or show neutrophil predominance.
Methods: This was a retrospective observational study of 147 consecutive patients with exudative pleural effusions that were diagnosed by analysis of fluid samples during a 3-year period from 1 April 2007 to 31 March 2010. Multiple linear correlation tests were used to assess clinical variables as possible predictors of high pADA levels.
Results: High pleural LDH (pLDH) and pleural potassium (pK) levels were associated with high pADA levels (P < 0.0001). Although there was a linear correlation between pLDH and pADA levels in patients with parapneumonic effusions (PPE) (n = 75), tubercular effusions (n = 21), malignant effusions (n = 41) and miscellaneous effusions (n = 10), a significant linear correlation between pK and pADA levels was observed only in patients with PPE (ρ = 0.525, P < 0.0001). When the cut-off value for pK was set at 5.0 mEq/L, pADA levels were >50 IU/L and pK levels were >5.0 mEq/L in only one patient (5%) in the tuberculosis group (n = 21) and 15 patients (12%, all with PPE) in the non-tuberculosis group (n = 126).
Conclusions: When pK levels exceed 5.0 mEq/L, high pADA levels do not necessarily indicate the presence of tuberculous pleuritis.
- Top of page
Tuberculous pleuritis occurs in approximately 5% of patients with Mycobacterium tuberculosis infection and is the second most common type of extrapulmonary tuberculosis after lymphatic involvement.1,2 Tubercular pleural effusions are caused by a delayed hypersensitivity reaction to mycobacterial antigens in the pleural cavity and have a higher proportion of T-lymphocytes than blood. Measurement of adenosine deaminase (ADA) in lymphocytic exudative effusions is helpful in diagnosing tuberculous pleuritis because Ziehl–Neelsen staining and cultures are positive in less than 10% and 30%, respectively, of patients with tuberculous pleuritis,3 and because pleural biopsy and thoracoscopy are invasive procedures. Pleural ADA (pADA) levels are typically greater than 40–60 U/L,4–7 whereas pADA levels are rarely <40 U/L in tuberculosis patients.8
On the other hand, it is well known that pADA levels increase in patients with non-tuberculosis or non-malignant exudative effusions. Some studies have shown that pADA levels were high in patients with parapneumonic effusions (PPE).9–11 The elevation of pADA in tuberculosis is due to the ADA-2 isoenzyme, which is only present in monocytes and macrophages, and is released when those cells are stimulated by live microorganisms, whereas other types of pleuritis, including PPE or empyema, are associated with high levels of the ADA-1 isoenzyme, which is found in many kinds of tissues.12–14 Although these isoenzymes are useful for diagnosing tuberculous pleuritis or PPE, these isoenzymes cannot easily be measured in fluids using commercial kits.
PPE result from the accumulation of fluids in the pleural cavity as a consequence of bacterial pneumonia, and in many cases show a predominance of neutrophils. However, it may be difficult to distinguish tuberculous pleuritis from PPE using differences in the predominant cells because in some tuberculosis patients, the exudates show a predominance of neutrophils,15 whereas in some patients with PPE, the exudates show a predominance of lymphocytes.16 Furthermore, neutrophils predominate in the first few days of tuberculous effusions, whereas lymphocytes predominate thereafter.17 Measurement of ADA levels in lymphocytic exudative effusions is helpful for diagnosing tuberculous pleuritis; however, it is not clear whether high ADA levels are related to tuberculosis when exudates do not satisfy the criteria for lymphocytic effusions or show a predominance of neutrophils.
A new predictive factor is required to determine whether or not high pADA levels are related to tuberculosis in patients with pleural effusions. The purpose of this retrospective study was to assess potential predictive factors by evaluating differences in clinical data for exudative effusions between patients with or without high pADA levels.
- Top of page
This was a retrospective observational study of 147 consecutive patients with exudative pleural effusions that were diagnosed by analysis of fluid samples obtained by thoracentesis or chest drainage during the 3-year period from 1 April 2007 to 31 March 2010. Pleural effusions were defined as exudates when the analysis indicated that the fluid satisfied the criteria of Light et al.18 The study was approved by the institutional review board, which decided that informed consent was not required for this retrospective study that only involved examination of patient records and images.
A diagnosis of bacterial pneumonia was made in patients with a new pulmonary infiltrate and production of mucopurulent sputum together with respiratory symptoms (fever, cough, dyspnoea and chest pain).19 PPE were defined as exudates occurring in the pleural space, adjacent to areas of bacterial pneumonia, and that improved with chest tube drainage and/or antibiotic treatment. PPE were divided into three categories based upon pathogenesis: (i) simple PPE occurring when vascular permeability increases due to the production of pro-inflammatory cytokines, and defined as non-purulent exudates that satisfied the criteria of negative fluid microbiological analysis, fluid pH >7.2 and fluid glucose >400 mg/L; (ii) complicated PPE occurring when neutrophil migration and activation of the coagulation cascade are accelerated by bacterial invasion, and defined as non-purulent exudates that satisfied at least one of the criteria of presence of microorganisms on Gram stain or culture, fluid pH <7.2 or fluid glucose <400 mg/L; and (iii) empyema, as characterized by fibrinopurulent collections caused by massive neutrophil phagocytosis and bacterial death.20
Tuberculous pleuritis was diagnosed when exudates satisfied at least one of the following criteria: demonstration of acid-fast bacilli or a positive result for culture of M. tuberculosis from fluid; the presence of caseous granulomas in pleural tissue; lymphocytic exudative effusions with pADA levels >50 IU/L;5 or lymphocytic exudative effusions that improved with anti-tuberculosis treatment. Cytology and pleural biopsy were both negative for malignancy, and other causes of pleural effusions were discarded with reasonable justification. Lymphocytic effusions were defined as fluid with a lymphocyte count >50% of the total white blood cell count, as conventionally defined.14,20,21 Pleural effusions were diagnosed as malignant when fluid cytology or pleural biopsy findings were positive for malignancy or when a primary or metastatic tumour had been diagnosed adjacent to the pleura with no other explanation for the effusion. All other exudative effusions were included in the miscellaneous group.
At the time of thoracentesis or chest drainage, fluid was collected from the pleural effusions. Pleural fluid levels of protein, LDH, Na+, Cl-, K+, amylase and glucose were measured on a biochemistry analysis system (JCA-BM1650, JEOL, Tokyo, Japan). Differential cell counts were measured on a haematology analysis system (XE-2100, Sysmex, Kobe, Japan). Pleural fluid pH was measured using an arterial blood gas analysis system (IL GEM Premier3000; IL Japan, Tokyo, Japan). Pleural fluid pH levels were not measured for patients with empyema because the membranes of the machine may be damaged by particulate matter in the fluid. Pleural ADA levels were analyzed using a commercially available colorimetric assay kit (Kyokuto Pharmaceutical Industrial Co. Ltd, Fukui, Japan).
Statistical analyses were performed using the StatView J 5.0 statistical program (Abacus Concepts Inc., Berkeley, CA, USA). Differences between two independent samples were assessed for statistical significance using the Mann–Whitney U-test. One-way factorial analysis of variance and multiple comparison tests were used to compare three or more independent parameters. Categorical data were analyzed using Fisher's exact probability test. Correlations between two independent parameters were assessed using Spearman's rank correlation test. Multiple linear regression analysis was performed to determine which clinical factors were associated with high levels of pADA. P values <0.05 were considered to indicate statistically significant differences.
- Top of page
Seventy-five patients were diagnosed as having PPE; in 24 cases because of the presence of microorganisms on Gram stain or culture of fluids, and in 51 cases because the exudates satisfied the criteria for both bacterial pneumonia and PPE. Tuberculous effusions were diagnosed in 21 patients; 13 of whom had lymphocytic effusions with pADA levels >50 IU/L, five patients had acid-fast bacilli or positive results for culture of M. tuberculosis from fluid, and three had lymphocytic effusions with pADA levels ≤50 IU/L that improved in response to anti-tuberculosis chemotherapy with no other explanation for the effusions. In 41 patients, the aetiology of the effusion was malignancy (lung cancer in 26 patients, metastatic lung tumour in nine, mesothelioma in three and lymphoma or leukaemia in three). Miscellaneous exudative effusions were diagnosed in 10 patients (five with collagen vascular disease, two with trauma, two with pancreatitis and one with parasitic infection).
When the cut-off value for pADA levels was set at 50 IU/L,5 the percentage of patients with tuberculous effusions, as well as serum protein, pleural fluid protein, LDH (pleural LDH (pLDH)) and K+ (pleural potassium (pK)) levels, were higher in patients with pADA levels >50 IU/L than in patients with pADA levels ≤50 IU/L. In contrast, the percentage of patients with malignant effusions, as well as serum Na+, pleural fluid Na+ and glucose levels, were lower in patients with pADA levels >50 IU/L than in patients with pADA levels ≤50 IU/L. There were no differences in mean age, gender, the percentage of patients with PPE or miscellaneous effusions, or the percentage of patients with comorbidities, between the two groups (Table 1). Multiple linear regression analysis showed that pLDH and pK were associated with high pADA levels (P < 0.0001) (Table 2).
Table 1. Characteristics and clinical data for patients with pleural ADA levels >50 IU/L and those with pleural ADA levels ≤50 IU/L
| ||Unit||Pleural ADA >50 IU/L (n = 44)||Pleural ADA ≤50 IU/L (n = 103)||P value|
|Age||years||72 ± 13||75 ± 13||0.2731|
|Males||n (%)||33 (75)||78 (76)||0.9999|
|Plural effusion aetiology|| || || || |
| PPE||n (%)||21 (48)||54 (52)||0.7189|
| Tuberculosis||n (%)||18 (41)||3 (3)||<0.0001|
| Malignant disease||n (%)||4 (9)||37 (36)||0.0006|
| Miscellaneous diseases||n (%)||1 (2)||9 (9)||0.2823|
|Comorbidities|| || || || |
| COPD||n (%)||2 (5)||11 (11)||0.3452|
| Previous tuberculosis||n (%)||1 (2)||2 (2)||0.9999|
| Cerebrovascular disease||n (%)||2 (5)||8 (8)||0.7236|
| Hypertension||n (%)||11 (25)||24 (23)||0.8348|
| Chronic hepatitis||n (%)||2 (5)||2 (2)||0.5833|
| Chronic renal disease||n (%)||8 (18)||13 (13)||0.4416|
| Diabetes mellitus||n (%)||8 (18)||12 (12)||0.3023|
| Malignant disease||n (%)||2 (5)||12 (12)||0.5866|
| Other diseases||n (%)||2 (5)||3 (3)||0.6357|
|Blood analyses|| || || || |
| WCC||×109/L||8.6 ± 4.7||9.2 ± 4.2||0.3653|
| CRP||mg/L||103 ± 84||96 ± 86||0.4810|
| Protein||g/L||67 ± 9||63 ± 9||0.0476|
| BUN||mg/L||200 ± 190||230 ± 190||0.2771|
| Creatinine||mg/L||13 ± 15||13 ± 16||0.3889|
| LDH||IU/L||208 ± 110||214 ± 86||0.2961|
| Na+||mEq/L||135 ± 5||138 ± 5||0.0041|
| K+||mEq/L||4.0 ± 0.7||4.1 ± 0.6||0.4443|
| Cl-||mEq/L||100 ± 6||102 ± 6||0.0773|
| Amylase||IU/L||76 ± 41||62 ± 36||0.0718|
| Glucose||mg/L||1460 ± 750||1270 ± 390||0.6281|
|Pleural effusion analyses|| || || || |
| Protein||g/L||43 ± 13||41 ± 10||0.0281|
| LDH||IU/L||5318 ± 7680||736 ± 1540||0.0001|
| pH||7.20 ± 0.29||7.33 ± 0.25||0.1281|| |
| Na+||mEq/L||134 ± 7||138 ± 5||0.0133|
| K+||mEq/L||5.3 ± 2.3||4.1 ± 0.9||0.0290|
| Cl-||mEq/L||101 ± 7||105 ± 5||0.1023|
| Amylase||IU/L||55 ± 36||60 ± 74||0.4325|
| Glucose||mg/L||650 ± 680||980 ± 470||0.0003|
| Lymphocytic effusions||n (%)||20 (45)||50 (49)||0.8571|
Table 2. Multivariate analysis of factors associated with increased pleural adenosine deaminase levels
| ||Coefficient estimate||Standard error||P value|
|Pleural fluid protein||−0.072||6.095||0.5199|
|Pleural fluid LDH||0.419||0.002||0.0006|
|Pleural fluid Na+||−0.073||1.369||0.6368|
|Pleural fluid K+||0.303||3.496||0.0044|
|Pleural fluid glucose||−0.145||0.097||0.1428|
There were linear correlations between pLDH levels and pADA levels in patients with PPE (ρ = 0.774, P < 0.0001), and tubercular (ρ = 0.403, P = 0.0717), malignant (ρ = 0.467, P = 0.0031) and miscellaneous effusions (ρ = 0.784, P = 0.0186). However, there was a significant linear correlation between pK levels and pADA levels in patients with PPE but not in those with the other three types of effusion (Fig. 1).
Figure 1. Correlations between pleural fluid levels of adenosine deaminase (ADA) and K+ in patients with parapneumonic effusions, tuberculous effusions, malignant effusions and miscellaneous exudative effusions.
Download figure to PowerPoint
In patients with PPE, pADA, pLDH and pK levels increased and pleural glucose levels decreased according to the following stages: simple PPE (n = 43), complicated PPE (n = 23) and empyema (n = 9). Plural pH was lower in complicated PPE than in simple PPE. Bacterial organisms were identified in fluids from 17 (74%) complicated PPE and from 7 (78%) patients with empyema (Table 3). The percentage of patients with pADA levels >50 IU/L was 5% for simple PPE, 43% for complicated PPE and 100% for empyema.
Table 3. Pleural fluid analyses in patients with simple PPE, complicated PPE or empyema
|Parameter||Unit||Simple PPE (n = 43)||Complicated PPE (n = 23)||Empyema (n = 9)|
|ADA||IU/L||25.3 ± 15.3||55.6 ± 34.8*||242.6 ± 137.9*†|
|Protein||g/L||39 ± 11||42 ± 8||31 ± 20†|
|LDH||IU/L||517 ± 330||3652 ± 4963*||17125 ± 7161*†|
|pH||—||7.40 ± 0.14||6.95 ± 0.12*||—‡|
|K+||mEq/L||4.0 ± 0.7||4.8 ± 1.6*||8.6 ± 2.0*†|
|Amylase||IU/L||37 ± 20||34 ± 20||71 ± 62*†|
|Glucose||mg/L||1050 ± 44||540 ± 68*||1 ± 1*†|
|Lymphocytic effusions||n (%)||12 (28)||3 (13)||0|
|Bacteria in fluids||n (%)||0||17 (74)||7 (78)|
| Streptococcus pneumoniae||n (%)||0||3 (13)||1 (11)|
| Streptococcus milleri||n (%)||0||3 (13)||1 (11)|
| Staphylococcus aureus||n (%)||0||7 (30)||2 (22)|
| Klebsiella pneumoniae||n (%)||0||0||1 (11)|
| Pseudomonas aeruginosa||n (%)||0||2 (9)||0|
| Enterobacter cloacae||n (%)||0||1 (4)||1 (11)|
| Other||n (%)||0||1 (4)||1 (11)|
There was no correlation between pK levels and pleural pH in the 66 patients with PPE (patients with empyema were excluded because pleural pH was not measured). In patients with complicated PPE, there was an inverse linear correlation between pK levels and pleural pH (n = 23, ρ = −0.489, P = 0.0201), whereas there was no correlation in patients with simple PPE (n = 43, ρ = 0.086, P = 0.538) (Fig. 2).
Figure 2. Correlations between pleural fluid levels of K+ and pleural pH in patients with parapneumonic effusions (PPE). Open circles indicate patients with simple PPE and closed circles indicate patients with complicated PPE. ○, simple PPE stage (n = 43); ●, complicated PPE stage (n = 23).
Download figure to PowerPoint
Comparing the characteristics and clinical data for the tuberculous effusion group with those for the non-tuberculous effusion group (patients with PPE, malignant and miscellaneous effusions were combined), there were no differences between the two groups, except for the percentage of patients with previous tuberculosis, and for pADA, pleural protein and pLDH levels, pleural pH, and the percentage of patients with lymphocytic effusions (Table 4). Pleural ADA levels were >50 IU/L in 18 patients (86%) in the tubercular effusion group and 43 patients (34%) in the non-tuberculous effusion group (28% of patients with PPE, 4% of those with malignant effusions and 10% of those with miscellaneous effusions). Setting the cut-off value for pK at 5.0 mEq/L, there was only one patient (5%) in the tubercular effusion group and 15 patients (12%) in the non-tubercular effusion group (all with PPE) with pADA levels >50 IU/L, as well as pK levels >5.0 mEq/L (Fig. 1).
Table 4. Characteristics and clinical data for patients with tuberculous and non-tuberculous effusions
| ||Unit||Tuberculous effusions (n = 21)||Non-tuberculous effusions (n = 126)||P value|
|Age||years||72 ± 16||74 ± 12||0.9978|
|Males||n (%)||16 (76)||95 (75)||0.754|
|Comorbidities|| || || || |
| COPD||n (%)||1 (5)||12 (10)||0.6935|
| Previous tuberculosis||n (%)||3 (14)||0||0.0026|
| Cerebrovascular disease||n (%)||0||10 (8)||0.3581|
| Hypertension||n (%)||7 (33)||28 (22)||0.2768|
| Chronic hepatitis||n (%)||0||4 (3)||>0.9999|
| Chronic renal disease||n (%)||3 (14)||18 (14)||>0.9999|
| Diabetes mellitus||n (%)||5 (24)||15 (12)||0.1671|
| Malignant disease||n (%)||0||14 (11)||0.4712|
| Other diseases||n (%)||0||5 (4)||>0.9999|
|Blood analyses|| || || || |
| WCC||×109/L||7.3 ± 2.2||9.3 ± 4.6||0.1942|
| CRP||mg/L||68 ± 49||104 ± 89||0.1707|
| Protein||g/L||67 ± 10||64 ± 9||0.1491|
| BUN||mg/L||240 ± 250||220 ± 180||0.6311|
| Creatinine||mg/L||15 ± 17||13 ± 15||0.3141|
| LDH||IU/L||206 ± 63||214 ± 98||0.9978|
| Na+||mEq/L||137 ± 5||137 ± 6||0.6914|
| K+||mEq/L||4.2 ± 0.8||4.1 ± 0.6||0.5649|
| Cl-||mEq/L||102 ± 5||101 ± 11||0.7419|
| Amylase||IU/L||87 ± 47||73 ± 35||0.34|
| Glucose||mg/L||1560 ± 810||1290 ± 470||0.582|
|Pleural effusion analyses|| || || || |
| ADA||IU/L||81 ± 37||46 ± 69||<0.0001|
| Protein||g/L||49 ± 8||40 ± 11||0.0001|
| LDH||IU/L||433 ± 293||2386 ± 5181||0.03|
| pH|| ||7.44 ± 0.14||7.26 ± 0.28||0.0381|
| Na+||mEq/L||135 ± 6||137 ± 6||0.4521|
| K+||mEq/L||4.0 ± 0.7||4.5 ± 1.6||0.1976|
| Cl-||mEq/L||104 ± 5||104 ± 6||0.9214|
| Amylase||IU/L||60 ± 28||59 ± 69||0.102|
| Glucose||mg/L||1020 ± 650||860 ± 540||0.5479|
| Lymphocytic effusions||n (%)||20 (95)||50 (40)||<0.0001|
- Top of page
In this study, we set out to determine whether, in exudative pleural effusions with pK levels >5.0 mEq/L, high pADA levels did not indicate the presence of tuberculous pleuritis.
To the best of our knowledge, measurement of pK levels in patients with exudative pleural effusions has not been reported. Measurement of pleural pH is useful for the diagnosis and management of patients with pleural effusions. It has been recommended that pleural pH should be measured using a blood gas analyzer.22 By coincidence, because the blood gas analyzer automatically measured electrolyte levels, we observed that pK levels were increased in some patients with high pADA levels. In the present study, there was a linear correlation between pLDH and pADA in patients with tuberculous pleuritis, as well as in patients with PPE, whereas there was a linear correlation between pK and pADA only in patients with PPE. For this reason, pK was used as a predictive factor, although multiple linear regression showed that pLDH was also associated with high pADA levels. When pK levels were >5.0 mEq/L, one of 16 patients with pADA levels >50 IU/L received a diagnosis of tuberculous pleuritis, whereas the other 15 patients received a diagnosis of PPE. On the other hand, pK levels were high in one patient (5%) with tuberculous effusion, three patients (7%) with malignant effusions and one (10%) with a miscellaneous effusion (Fig. 1); however, these high levels of pK were due to increased serum K+ levels.
Increased pK levels may be explained by the development of acidosis in pleural fluid due to anaerobic utilization of glucose by neutrophils and bacteria, and the fact that acidosis (low pH) in the pleural cavity causes an extracellular shift of potassium (high pK levels). In patients with PPE, pK and pLDH levels were increased, and pleural glucose levels and pH decreased according to the exudative stage (Table 3). There was no inverse linear correlation between pK levels and pleural pH in patients with simple PPE, whereas in patients with complicated PPE, there was an inverse correlation (data for patients with empyema was not included in this analysis) (Fig. 2). An inverse linear correlation between pK levels and pH in the pleural cavity may occur when neutrophil migration is accelerated by bacterial invasion, but not when vascular permeability increases due to the production of pro-inflammatory cytokines.
This was a retrospective study. However, additional measurement of pK during routine pleural fluid analysis is not disadvantageous to patients with pleural effusions. We suggest that measurement of pK in exudates with high pADA levels would be useful for distinguishing PPE from tuberculous pleuritis without the necessity for a pleural fluid differential cell count. In conclusion, high pADA levels do not necessarily indicate the presence of tuberculous pleuritis when pK levels exceed 5.0 mEq/L.