Asian Cardiovasc Thorac Ann 2004;12:23-29
© 2004 Asia Publishing EXchange Ltd
Simple Method for Rapid Diagnosis of Tuberculosis Pleuritis: A Statistical Approach
Mostafa Ghanei, MD,
Jafar Aslani, MD,
Hossein Bahrami, MD1,
Hossein Adhami, MD
Baqiyatallah University of Medical Sciences
1 Trauma Research Center, Tehran, Iran
For reprint information contact: Mostafa Ghanei, MD Tel: 98 21 805 3770 Fax: 98 21 804 0106 Email: m.ghanei{at}bmsu.ac.ir Baqiyatallah University of Medical Sciences, Mollasadra Ave, 19945-546, Tehran, Iran.
 |
ABSTRACT
|
|---|
Among the extrapulmonary presentations of tuberculosis, pleural tuberculosis is the second most frequent. Failure to diagnose and treat pleural tuberculosis can result in progressive disease with involvement of other organs in as many as 65% of patients. Conventional methods such as direct examination of pleural fluid, pleural fluid culture and pleural biopsy have proven to be insufficient for diagnoses of pleural tuberculosis. In this study, we examined a statistical method by combining the diagnostic efficiency of adenosine deaminase activity, pleural fluid protein, lactate dehydrogenase and cellular components in patients with tuberculous pleural effusions. Eighty eight patients over 12 years of age presenting with pleural effusions were included. A positive result by either three of the methods was considered to be indicative of a positive diagnosis of pleural tuberculosis. The determination of adenosine deaminase activity, lactate dehydrogenase levels, and lymphocyte to neutrophil ratio in the pleural fluid yielded a sensitivity of 100% for pleural tuberculosis. A patient was considered positive if any of the three tests was positive, with a specificity of 100%. A positive diagnosis was made when all three tests were positive. Similarly, these different approaches to the combination of pleural adenosine deaminase and lactate dehydrogenase result in sensitivity and specificity of 91.4% and 100% respectively.
|
INTRODUCTION
|
|---|
Pulmonary tuberculosis (TB) is the most frequent cause of death by an infectious agent worldwide.1 Among the extrapulmonary presentations after tuberculous lymphadenitis, pleural TB is the second most frequent.2 Failure to diagnose and treat pleural TB can result in progressive disease with the involvement of other organs in as many as 65% of patients.3 Treatment based solely on clinical suspicion results in over treatment. Conventional methods have proven to be insufficient for diagnoses of pleural TB. Direct examination of pleural fluid is inefficient because sensitivity is about 1%.4,5 Pleural fluid (PF) culture is more sensitive than direct examination but Mycobacterium tuberculosis requires 4 to 6 weeks to grow.6 Although the sensitivity of pleural biopsy specimens is higher, this procedure requires greater expertise, and is more invasive.
Methods such as the quantification of interferon (IFN-gamma) and polymerase chain reaction (PCR) to detect a specific sequence of the Mycobacterium tuberculosis genome have shown higher sensitivity than culturing or direct examination of PF.7–9 However, these diagnostic aids have not being widely utilized for the routine laboratory diagnosis of pleural TB. Previous studies showed that in populations with a high prevalence of TB, the measurement of adenosine deaminase (ADA) activity in PF, has proven to be sensitive and specific for diagnosis of pleural TB.10–12 In this study, we examined a statistical method by combining the diagnostic efficiency ofADA activity, PF protein, lactate dehydrogenase (LDH) and cellular components in patients with tuberculous pleural effusions and correlated the results with clinical signs and symptoms in a population with a high prevalence of TB. Subsequently, we determined the positive predictive values (PPVs) and negative predictive values (NPVs) of individual and combined methods relative to the prevalence of pleural TB in patients with pleural effusion.
|
PATIENTS AND METHODS
|
|---|
Eighty eight patients (> 12 years of age) presenting with pleural effusions to our service between September 1999 and September 2001 were included in this study. Informed consent was obtained. Clinical signs and symptoms, demographic data, and radiological results were recorded. The response to intradermal purified protein derivative (PPD) was evaluated. An induration of > 10 mm was considered to be positive, and an induration between 5 and 10 mm considered a suspicious result.
A single specimen of 50 to 100 mL of pleural fluid was submitted for Ziehl-Neelsen staining, ADA activity determination, measurement of protein sensitivity and LDH levels, and differentiation of leukocytes. Simultaneously, a pleural biopsy was performed and the specimen was submitted for histopathologic examination. A blood sample was also taken at the same time for biochemical evaluation. Pleural fluid samples were concentrated by configuration for 20 min at 10,000 g at 4°C. The pellet was resuspended in 1.5 mL DNAse free buffer (pH = 8.0). The sample was divided to perform direct smears (Ziehl-Neelsen staining) and cultures in Ogawa-Kudoh medium. ADA activity was determined in 1 mL pleural fluid using the colorimetric method described by Giusti and Galanti.13 A positive result was defined as a value > 47 U·L-1 based on the previous analysis of pleural fluid samples with proven TB and non-TB. A positive control sample and two negative control samples for which the ADA value was known were included in each group of clinical samples analyzed. Pleural LDH levels to serum LDH levels greater than 0.6 and pleural protein level to serum protein more than 0.5 were considered exudative. Lymphocyte to neutrophil ratios > 0.75 in pleural fluid and > 0.3 in serum were considered to be in favor of pleural TB.
The results of clinical evaluations, diagnostic tests, and ADA activity determinations were analyzed using computer software (SPSS, version 10.0). The sensitivity, specificity, PPV, NPV of ADA activity, pleural fluid LDH and protein levels and the lymphocyte to neutrophil ratio were evaluated. The results of positive acid fast bacilli staining (AFB) of pleural biopsy, growth of mycobacterium tuberculosis in biopsy specimen culture, or typical histopathologic findings, were used as the "gold standard". The sensitivity and specificity of the diagnostic methods were determined using the result of each of the diagnostic methods for each patient. Therefore, the unit of analysis was the patient, not the clinical specimen or test. Comparisons between the diagnostic methods were performed using a McNemar chi-square test ( p < 0.05). PPV and NPV were estimated using Bayesian analysis with a TB prevalence range of 1 to 50% for the patient population with pleural effusions.
The formula for Bayesian analysis is:
where Sn is sensitivity, Sp is specificity, and P equals to the prevalence.
|
RESULTS
|
|---|
The study population included 54 men and 34 women with a mean age of 56.7 years. According to the final diagnosis the study population (88 patients) was distributed into two groups as follows: 17 patients (19.3%) with pleural TB confirmed by compatible pleural biopsy and/or culture; and 71 patients (80.7%) with pleural effusion due to an etiology other than Mycobacterium tuberculosis. Within the group with another etiology, 6 patients had pleural effusions associated with chronic heart failure, 32 patients had malignancies, 8 bacterial pneumonia and 25 patients with other diseases including renal insufficiency, pancreatitis, or pulmonary thromboemboli (PTE). 82.4% of the patients with pleural TB (14/17) presented with dyspnea compared with 88.7% of patients with other diagnoses (63/71). Cough was seen in 64.7% (11/17) and 69.0% (49/71) of the patients with pleural TB and patients with other diagnoses, respectively. 23.5% of the patients with pleural TB had chest pain at their presentation while this symptom was seen in 32.4% of patients with other diagnoses. Productive cough with blood was present in 35.3% (6/17) of patients with pleural TB and 35.2% (25/71) of patients with other diagnoses. No statistically significant association was found between pleural TB and any of the symptoms mentioned above.
Measurement of ADA activity in pleural fluid was the single most specific method (97.1%) for the diagnosis of pleural TB (Table 1
). The sensitivity of ADA activity was 52.6%, a value that is inferior to that of pleural biopsy. Nevertheless, ADA activity had the highest PPV (83.3%) among single diagnostic methods, which remained high over a range of prevalence from 0.01 to 0.5 (Figure 1). ADA was positive in 2 of 71 patients with non-tuberculous pleural effusion. PPD was applied and read in all patients. In the patient group with confirmed pleural TB, 7.1% was PPD positive, 21.4% were suspicious, and 81.5% had negative PPD responses. In the group having a confirmed non-tuberculous effusion, 9.7% patients were PPD positive, 16.1% were suspicious and others (76.2 %) had negative PPD.
View this table:
[in this window]
[in a new window]
|
Table 1. Sensitivity, specificity, accuracy, and predictive values of individual diagnostic methods for pleural TB
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Figure 1. Predictive values of individual diagnostic methods according to prevalence of pleural TB among patients with pleural effusion. Top: PPV. Bottom: NPV.
|
|
The combination of two methods, with a positive result by either two of the methods considered to be indicative of a positive diagnosis of pleural TB, increased the diagnostic sensitivity over all individual methods (Table 2, Figure 2). The measurement of pleural fluid lymphocyte to neutrophil ratio combined with the determination of ADA activity yielded a sensitivity of 91.7%, whereas combination of pleural fluid protein levels with ADA activity presented the highest specificity (82.1%). The combination of two methods requiring both to be positive for diagnosis to be made increased the specificity of the methods, at the expense of sensitivity. The determination of ADA activity in pleural fluid plus the measurement of each pleural fluid LDH, protein, lymphocyte to neutrophil ratio in pleural fluid, or this ratio in serum was 100% specific for pleural TB.
View larger version (28K):
[in this window]
[in a new window]
|
Figure 2. Predictive values of combined methods [either/or] according to prevalence of pleural TB among patients with pleural effusion. Top: PPV. Bottom: NPV.
|
|
In the combination of three methods, with a positive result by either three of the methods considered to be indicative of a positive diagnosis of pleural TB, ADA activity plus LDH levels plus lymphocyte to neutrophil ratio in the pleural fluid yielded a sensitivity of 100%. This combination requiring all three methods to be positive for diagnosis to be made was 100% specific for pleural TB.
In the McNemar’s test, the difference in positive diagnostic proportion between combination of pleural fluid LDH and ADA level, pleural lymphocyte to neutrophil ratio (L/N) and pleural fluid LDH, and finally pleural ADA and L/N were statistically significant ( p values of < 0.01, 0.03, and < 0.01, respectively). Also, the pleural biopsy showed a statistically significant difference to each of pleural ADA, LDH, and L/N ( p values of < 0.01, < 0.01, and 0.01, respectively). Other paired binary responses between two tests did not show any significant difference.
|
DISCUSSION
|
|---|
Pleural effusions can present a challenging diagnostic problem. The differential diagnosis is diverse, but most common causes include congestive heart failure (CHF), malignancy, and pneumonia related effusion. Studies on pleural fluid yield a definitive or presumptive diagnosis in about 74% of cases.14 In the present investigation, conventional and alternative diagnostic methods were evaluated individually and in combination within the context of the case mix encountered in clinical practice. We have explored combinations of methods the complementary strengths of which offer diagnostic options that can be utilized advantageously in countries with high TB prevalence and lower technology. The understanding of the strengths and weaknesses of each diagnostic method, used individually or in combination, and the consideration of the results together with the symptoms and signs associated with a determined etiology of pleural effusion propitiates the rational and optimal election of methods in different diagnostic scenarios. The results of this study contribute a comprehensive view of the comportment of the repertoire of methods currently in use or available in the differential diagnosis of pleural effusion. These findings support the use of new and conventional diagnostic strategies in the management of pleural TB.
The evaluation of the efficiency of alternative diagnostic tests is conducted in reference to the best available standard, which is taken as the "gold standard." In the case of pleural TB and other paucibacillary forms of Mycobacterium tuberculosis infection, the "gold standards," i.e. culture and biopsy, present significant limitations in sensitivity and in the time and clinical expertise required to determine etiology. Under these circumstances, positive results of alternative diagnostic methods having greater sensitivity may be classified as false positives. However, doubt arises as to whether some of these results may in fact be true positives. In order to obtain the maximum utility of alternative diagnostic methods, it is important to consider this issue and to interpret results within the context of all laboratory and clinical findings.
The high sensitivity and specificity of ADA in the differential diagnosis of pleural TB have been substantiated in studies conducted in other areas of high TB prevalence, such as Mexico, Spain, and Brazil.9,11,15 This theoretical projection encourages the evaluation of ADA as a diagnostic aid considering previous conventional tests for the differential diagnosis of pleural TB in populations with moderate to high prevalence of TB. Nevertheless, the sensitivity and robustness of ADA activity in the diagnosis of pleural TB, together with its simplicity, speed, and low cost, urge the widespread implementation and routine utilization of the method in populations with a high prevalence of TB and its evaluation in areas of low prevalence.
As shown, the determination of ADA activity plus LDH levels plus lymphocyte to neutrophil ratio in the pleural fluid yielded a sensitivity of 100% for pleural TB, while a patient is considered positive if any of three tests are positive with a specificity of 100%, and a positive diagnosis when all three tests are positive. Thus, by measuring these parameters we can reach a diagnosis of pleural TB with an acceptable degree of confidence. Even in the absence of L/N ratio, these different approaches to the combination of pleural ADA and LDH result in a sensitivity and specificity of 91.4% and 100%, respectively. This seems to be an acceptable diagnostic method for pleural TB.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 3. Predictive values of combined methods [both positive] according to prevalence of pleural TB among patients with pleural effusion. Top: PPV. Bottom: NPV.
|
|
|
REFERENCES
|
|---|
- Raviglione M, Luelmo F. Update on the global epidemiology of tuberculosis. Curr Issues Public Health
1996;2:192–7.[Medline]
- Mehta JB, Dutt A, Harvill L, Mathews KM. Epidemiology of extrapulmonary tuberculosis. A comparative analysis with pre-AIDS era. Chest
1991;99:1134–8.[Abstract/Free Full Text]
- Roper WH, Waring JJ. Primary serofibrinous pleural effusion in military personnel. Am Rev Tuberc
1955;71:616–34.[Medline]
- Yew WW, Chan CY, Kwan SY, Cheung SW, French GL. Diagnosis of tuberculous pleural effusion by the detection of tuberculoestearic acid in pleural aspirates. Chest
1991;100:1261–3.[Abstract/Free Full Text]
- Escudero Bueno C, Garcia Clemente M, Cuesta Castro B, Molinos Martin L, Rodriguez Ramos S, Gonzalez Panizo A, et al. Cytologic and bacteriologic analyzes of fluid and pleural biopsy with Cope’s needle. Arch Intern Med
1990;150:1190–4.[Abstract/Free Full Text]
- De Wit D, Maartens G, Steyn L. A comparative study of the polymerase chain reaction and conventional procedures for the diagnosis of tuberculous pleural effusion. Tuber Lung Dis
1992;73:262–7.[Medline]
- Aoki Y, Katoh O, Nakanishi Y, Kuroki S, Yamada H. A comparison study of IFN-gamma, ADA and CA125 as the diagnostic parameters in tuberculous pleuritis. Respir Med
1994;88:139–43.[Medline]
- Kim YC, Park KO, Bom HS, Lim SC, Park HK, Na HJ, et al. Combining ADA, protein and IFN–gamma best allows discrimination between tuberculous and malignant pleural effusion. Korean J Intern Med
1997;12:225–31.[Medline]
- Salazar-Lezama M, Quiroz-Rosales H, Banales-Mendez JL, Sanchez-Guzman M, Villarreal-Velarde H, Baez-Saldana R et al. Diagnostic methods of primary tuberculous pleural effusion in a region with high prevalence of tuberculosis. A study in Mexican population. Rev Invest Clin
1997;49:453–6.[Medline]
- Burgess LI, Maritz FJ, Le Roux I, Taljaard JJ. Use of adenosine deaminase as a diagnostic tool for tuberculous pleurisy. Thorax
1995;50:672–4.[Abstract/Free Full Text]
- Chalhoub M, Cruz AA, Mareilio C, Netto MB. Value of determining the activity of adenosine deaminase (ADA) in the differential diagnosis of pleural effusions. Rev Assoc Med Bras
1996;42:139–46.
- Villena V, Navarro-Gonzalvez JA, Garcia-Benayas C, Manzanos JA, Echave J, Lopez-Encuentra A et al. Rapid automated determination of adenosine deaminase and lysozyme for differentiating tuberculous and nontuberculous pleural effusions. Clin Chem
1996;42:218–21.[Abstract/Free Full Text]
- Giusti G, Galanti B. Adenosine deaminase. Bergemeyer HV, et al. Methods of enzyme analysis. New York, NY. Academic Press.1984.
- Collins TR, Sahn SA. Thoracocentesis. Clinical value, complications, technical problems and patient experience. Chest
1987;91:817–22.[Abstract/Free Full Text]
- Valdes L, Alvarez D, San Jose E, Penela P, Valle JM, Garcia-Pazos JM, et al. Tuberculous pleurisy: a study of 254 patients. Arch Intern Med
1998;158:2017–21.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
