International Journal of Infectious Diseases 78 (2019) 34–38 Contents lists available at ScienceDirect International Journal of Infectious Diseases journal homepage: www.elsevier.com/locate/ijid Epidemiology and diagnosis of pleural tuberculosis in a low incidence country with high rate of immigrant population: A retrospective study Antonio Macíasa , Adrián Sánchez-Montalváa,b,* , Fernando Salvadora,b , Ana Villarc , Teresa Tórtolad, Nuria Saborita , Israel Molinaa a Infectious Diseases Department, Vall d’Hebron University Hospital, PROSICS Barcelona, Universitat Autònoma de Barcelona, Barcelona, Spain b Grupo de Estudio de Infecciones por Micobacterias, Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica, Madrid, Spain c Pneumology Department, Vall d’Hebron University Hospital, Universitat Autònoma de Barcelona, Barcelona, Spain d Microbiology Department, Vall d’Hebron University Hospital, Universitat Autònoma de Barcelona, Barcelona, Spain A R T I C L E I N F O A B S T R A C T Article history: Background: The confirmatory diagnosis of pleural tuberculosis (pTB) remains challenging. The aim of Received 3 August 2018 this study was to describe the clinical and epidemiological characteristics of pTB patients and assess the Received in revised form 6 September 2018 yield of different diagnostic procedures in a low burden country with a high rate of immigrant Accepted 8 October 2018 population. Corresponding Editor: Eskild Petersen, Aarhus, Denmark Methods: All adult patients with pTB between 2007 and 2014 were studied retrospectively. Results: One hundred and three out of 843 patients with tuberculosis had pTB. Fifty-three (54.1%) were male, and the median age was 45 years (range 18–87 years). Fifty-two (50.49%) patients were immigrants. Keywords: Pleural tuberculosis A confirmed diagnosis was reached in 16 patients (15.5%) by microbiological studies of pleural effusion. Immigrant population Lung involvement was demonstrated by sputum smear microscopy in 13/49 (26.5%), sputum GeneXpert Tuberculosis diagnostic tools MTB/RIF test in 13/20 (65%), and sputum culture in 16/37 (43.2%). High-resolution computed tomography (CT) showed lung involvement in 47.7% of the patients. The cure rate was 91.3% at the 1-year follow-up. Three patients died, all of them within the first month after diagnosis. Conclusions: The detection of lung involvement increased by two-fold when lung CT was used; this correlated with the likelihood of finding a positive microbiological result on sputum sample testing. Pleural microbiological studies had a low diagnostic yield, and sputum could have a complementary role. © 2018 The Author(s). Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). Introduction countries receiving immigrants, taking in over 280 000 of them (Migration and migrant population statistics, 2016). Due to this Tuberculosis (TB) affects around nine million people and persistent flow of immigrants and the increasing rates expected in accounts for 1.5 million deaths worldwide every year (World the coming years, TB will continue to be a major health issue in the Health Organization, 2014). In countries with a low incidence of TB, EU (European Centre for Diseases Prevention and Control/WHO the epidemiology and clinical presentation of the disease is Regional Office for Europe, 2013). changing, with an increasing number of cases of extrapulmonary After tuberculous lymphadenitis, pleural tuberculosis (pTB) is involvement and a higher proportion of cases in non-native the most common extrapulmonary presentation of TB disease. pTB patients (Baussano and Mercadante, 2013). According to the can appear as a complication of a primary disease or as a long-term European Union Statistical Agency, 3.4 million immigrants came to effect of dissemination occurring many years after the initial the European Union (EU) in 2013, and Spain was in the top five infection (Davies and Pai, 2007). Factors such as age, chronic diseases, and immunosuppressive conditions have been associated with an increased incidence of pleural involvement (Gopi et al., 2007). The incidence of TB in Barcelona in 2015 was 14.9 cases per * Corresponding author at: Departamento de Enfermedades Infecciosas, Hospital 100 000 population, with 10% of pleural involvement (Informea- Universitario Vall d’Hebron, Edificio General, 6a planta, Passeig Vall d’Hebron, 119- nual, 2015). 129, CP 08035, Barcelona, Spain. E-mail address: adsanche@vhebron.net (A. Sánchez-Montalvá). https://doi.org/10.1016/j.ijid.2018.10.005 1201-9712/© 2018 The Author(s). Published by Elsevier Ltd on behalf of International Society for Infectious Diseases. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). A. Macías et al. / International Journal of Infectious Diseases 78 (2019) 34–38 35 pTB is a problematic disease due to the difficulty reaching a the drug susceptibility test was broadened to the following confirmed bacteriological diagnosis; hence many treatments are antibiotics: amikacin, capreomycin, streptomycin, ethionamide, started based on a presumed diagnosis (Light, 2010; Richter et al., moxifloxacin, and ofloxacin. Molecular biology testing was done 1994). The percentage of confirmed diagnosis varies widely (from with the GeneXpert MTB/RIF assay (Cepheid, Sunnyvale, CA, USA) 12% to 80%) depending on the study location and prevalence of the on pleural samples. disease (Porcel et al., 2015; Na, 2014). Although molecular biology methods are being generalized for samples other than sputum, the Imaging assessment diagnostic yield of molecular biology studies on pleural effusion is variable and usually lower than that for sputum (Trajman et al., A chest X-ray was performed for all patients. The need for a 2014; Porcel et al., 2013). high-resolution CT scan was at the discretion of the treating Pulmonary involvement in pTB has been described in many physician. Lung parenchyma findings suggestive of TB, such as series, and it is suggested to approach 50% in studies using high- nodules, infiltration, cavities, and pleural thickening, were resolution computed tomography (CT) (Kim et al., 2006). Induced recorded. sputum or even bronchoscopy may increase the yield of the culture diagnosis of pTB, which is extremely useful in settings with a high Statistical analysis burden of resistance to the first-line anti-TB drugs (Conde et al., 2003). Categorical variables are presented as the absolute number and The objective of this study was to describe the clinical and proportion, and continuous variables are expressed as the median epidemiological characteristics of patients with pTB in a low and range. The Chi-square test or Fisher’s exact test was used to incidence burden area and establish the yield of the different compare the distribution of categorical variables, as appropriate, microbiological techniques. and the Mann–Whitney U-test or Student t-test was used to assess continuous variables (following evaluation for a normal distribu- Materials and methods tion through the Kolmogorov–Smirnov test). Results were considered statistically significant if the two-tailed p-value was Study population <0.05. IBM SPSS Statistics version 21.0 (IBM Corp., Armonk, NY, USA) was used for the statistical analyses. This was a retrospective observational study performed at the Vall d’Hebron University Hospital, a tertiary reference Ethical considerations centre included in the International Health Program of the Catalan Health Institute (PROSICS), Barcelona, Spain. Patients The study protocol was approved by the Institutional Review older than 18 years of age with a diagnosis of pTB between Board of Vall d’Hebron University Hospital (Barcelona, Spain). An January 2007 and December 2014 were included. Epidemiologi- exemption from obtaining informed consent was granted. cal, clinical, and microbiological information was obtained from Procedures were performed in accordance with the ethical all patients through a review of the medical records. The standards laid down in the Declaration of Helsinki as revised in diagnostic approach and the treatment were at the discretion of 2012 in Fortaleza, Brazil (World Medical Association, 2016). the treating physician. Results Pleural tuberculosis definitions Of an overall 843 patients with TB, a total of 103 patients The diagnosis of pTB was classified as bacteriologically fulfilled the inclusion criteria for pTB. The median age of the study confirmed or probable. Bacteriologically confirmed pTB was population was 45 years (range 18–87 years) and 51.4% were male. defined as a positive Ziehl–Neelsen (ZN) smear, positive culture, Twenty (19.4%) of the patients had an immunosuppressant and/or positive PCR assay from a pleural sample. Probable pTB was condition, including seven patients with an HIV infection. More considered if a positive response to anti-TB drugs was found and data regarding patient characteristics and epidemiological infor- one of the following conditions was present: (1) biochemical mation are shown in Table 1. Figure 1 depicts the number of total diagnosis: pleural exudates with lymphocyte predominance and TB and pTB cases per year. adenosine deaminase (ADA) level higher than 35 IU/L; (2) Regarding symptoms, the chief complaints were fever (n = 83, histopathological sample showing necrotizing granuloma and/or 80.5%), pleuritic chest pain (n = 70, 67.9%), cough (n = 56, 54.3%), caseous necrosis; (3) pleural effusion in a patient with confirmed general weakness (n = 41, 39.8%), dyspnoea (n = 39, 37.8%), diapho- TB at any other location. A positive response to treatment was resis (n = 37, 35.9%), and weight loss (n = 33, 33.9%). The median defined as symptom resolution and pleural effusion improvement duration of symptoms was 21 days (range 3–365 days). Four by at least 90% of the initial assessment. Pleural biopsy was patients (3.8%) had a personal history of previous TB (two performed at the discretion of the attending physician. The main pulmonary TB and two pTB). More clinical and epidemiological reasons for performing a biopsy were a high risk of multidrug- information is detailed in Table 2. resistant TB, high suspicion of an alternative diagnosis, and a poor When analyzing the biochemical characteristics of the pleural clinical course. effusion, 14 of 97 cases (14.4%) did not have a lymphocytic predominance in pleural effusion and five of 101 cases (4.95%) had Microbiological studies an ADA concentration below the suggested limit to consider the diagnosis. All of these cases had a confirmed TB diagnosis by Pleural samples were examined microscopically using ZN stain. culture or a pleural biopsy showing granulomas. Table 3 shows the All samples were cultured and incubated in a BACTEC MGIT 960 diagnostic yield according to biochemical characteristics of the (Becton Dickinson Diagnostic Systems, Baltimore, MD, USA). Drug pleural effusion. susceptibility testing of Mycobacterium tuberculosis was done using A confirmed diagnosis was made in 16 (15.5%) patients: 14 by the BACTEC MGIT 960 SIRE Kit for the first-line drugs: streptomy- microbiological studies of pleural effusion and six by microbiolog- cin, isoniazid, rifampicin, ethambutol, and pyrazinamide (Kent and ical studies of pleural biopsy, four of them with negative pleural Kubica, 1985). When resistance to any first-line drug was detected, effusion microbiological studies. The diagnosis for 87 patients 36 A. Macías et al. / International Journal of Infectious Diseases 78 (2019) 34–38 Table 1 Demographic characteristics of the patients (N = 103). Variable Number (%) Sex (male) 53 (51.5%) Age (years), median (range) 45 (18–87) Foreign born 52 (51.5%) Latin America 20 (38.4%) Africa 14 (26.9%) Asia 11 (21.2%) Eastern Europe 7 (13.4%) Duration of stay in Spain prior to diagnosis (years), median (range) 3 (0–25) Immunosuppressive conditions 20 (19.4%) HIV infection 7 (6.7%) Autoimmune diseases 4 (3.9%) Steroid therapy 4 (3.9%) Anti-TNF inhibitor therapy 2 (1.9%) Liver transplantation 1 (0.9%) Kidney replacement therapy 1 (0.9%) TNF, tumour necrosis factor. Figure 1. Number of pleural TB and total cases of TB at Vall d’Hebron University Hospital. Barcelona, Spain 2007–2014. (84.5%) was probable TB. Regarding the accuracy of diagnostic The great majority of therapy regimens (96%) included the techniques, the yield was superior in sputum and pleural biopsy standard four-drug treatment including isoniazid, rifampicin, samples than in pleural effusion (Table 4). pyrazinamide, and ethambutol. Only four patients were treated Twenty-four (23.3%) patients had lung involvement as seen on with a fluoroquinolone during follow-up, with side effects being chest X-ray. High-resolution CT was done for 57 (55.34%) patients. the cause of the switch in all cases. The median duration of the Twenty-seven out of 57 (47.4%) showed lung involvement on high- treatment was 6 months (confidence interval interquartile range resolution CT scan, with 16 patients having any tomographic 5.5–8.5 months). finding of tuberculous lung involvement not visualized in the X- After 1 year of follow-up, 94 (91.3%) patients were cured, five ray, thereby increasing the diagnostic yield of having lung were lost to follow-up, and four had died. Three of the deaths compromise by 2.3 times. Moreover, patients with lung compro- occurred in the hospital within the first month after the diagnosis. mise had a higher yield on diagnosis in the sputum sample, In brief, one patient suffered a haemophagocytic syndrome compared with those who did not have lung compromise (72.2% vs. associated to the TB, another patient had multifactorial encepha- 25%; p = 0.002). lopathy (age, renal replacement therapy, TB infection, and anti-TB In the drug susceptibility test assessment, no resistance to anti- treatment) that led to death, and one cirrhotic patient developed a TB drugs was found. Thirty-nine (37.8%) patients developed any fulminant hepato-renal syndrome concomitant with the TB adverse effect related to the anti-TB drugs. The most frequent were infection. The remaining patient died of reasons unrelated to TB nausea (n = 8, 7.7%), mild hepatic cytolysis (n = 7, 6.7%), and during follow-up. There was no case of relapse during the study cholestasis (n = 5, 4.8%). period. A. Macías et al. / International Journal of Infectious Diseases 78 (2019) 34–38 37 Table 2 Epidemiological and clinical data of patients with pleural TB (2007–2014)a . Overall Spanish (n = 51) Immigrant (n = 52) p-Value Sex, male 53 (51.4%) 26 (50.9%) 27 (51.9%) 0.92 Age (years) 45 (18–87) 46.9 (18–87) 34.3 (18–77) <0.001 Time to consultation (days) 21 (3–365) 28 (4–365) 20 (3–180) 0.31 Extrapleural involvement 31 (30%) 25 (49%) 24 (46.2%) 0.85 HIV co-infection 7 (6.7%) 5 (9.8%) 2 (3.8%) 0.27 Immunosuppressionb 10 (9.7%) 6 (11.7%) 4 (7.6%) 0.52 Positive TST 45 (83.3%) 19 (73.1%) 26 (96.3%) Empiric anti-TB treatment 87 (84.4%) 42 (82.3%) 45 (86.5%) 0.37 Confirmed pleural TB 16 (15.5%) 9 (17.6%) 7 (13.5%) 0.55 Pleural biopsy 17 (16.5%) 10 (19.6%) 7 (13.4%) 0.28 TB, tuberculosis; TST, tuberculin skin test. a Data are presented as the number (percentage) or the median (range). b Immunosuppression includes autoimmune diseases and/or immunosuppressant agents, corticoid therapy, biological therapy, and transplantation. Table 3 Proportion of diagnosis according to characteristics of the pleural effusion. Samples/methods Total Confirmed Probable p-Value ADA-positivea in pleural effusion 96/101 (95%) 14/16 (87.5%) 82/85 (96.5%) 0.177 Pleural exudates 98/100 (98%) 15/16 (93.8%) 83/84 (98.8%) 0.296 Lymphocytic pleural effusion 83/97 (85.6%) 11/16 (68.8%) 72/81 (88.9%) 0.052 ADA-positivea + lymphocytic effusion + pleural exudate 77/96 (80.2%) 10/16 (62.5%) 67/80 (83.8%) 0.081 ADA, adenosine deaminase. a ADA > 35 IU/IU/L. Table 4 Diagnostic accuracy of the different techniques employed. Pleural effusion Sputum Pleural biopsy Ziehl–Neelsen Culture PCR Ziehl–Neelsen Culture PCR Any diagnostic finding Total 2/79 10/60 8/52 13/49 16/37 13/20 4/17a 7/17b Diagnostic 2.5% 16.6% 15.38% 26.5% 43.2% 65% 23.53%a accuracy 41.2%b a Positive smear, positive culture, or positive molecular biology test. b Granuloma or caseous necrosis. predominance > 50%, and ADA > 35 mg/dl, IU/L the three condi- Discussion tions were present in 80.2% (77/96 cases). The yield of PCR on pleural effusion was lower (seven positive In this study, the clinical and epidemiological characteristics of cases out of 51, 13.7%) than that observed for PCR on sputum (13 pTB in a tertiary hospital of a low burden country, over eight positive cases out of 20, 65%). According to other publications, the consecutive years, are described. In Vall d’Hebron University sensitivity of the GeneXpert assay on pleural effusion is usually low Hospital, pTB accounts for 12.2% of the total number of TB cases. Of (around 25%), with high specificity (around 95%) (Lusiba et al., note, half of the cohort was foreign-born, mainly from Latin 2014). The reasons for this are not clear, but could probably be America and Africa (in concordance with the immigrant popula- attributed to the presence of PCR inhibitors in pleural fluid and a tion in Barcelona). The pTB diagnosis was mainly based on clinical low bacillary load (Du et al., 2015). information, biochemical characteristics of the pleural effusion, Regarding pleural biopsy, despite being considered the most and microbiological results obtained for another location, with useful technique to confirm a suspected diagnosis, with rates of only 22 (21.3%) having a microbiological confirmatory result from a positivity between 60% and 95%, the indication for this in pTB is not pleural sample. Sputum samples showed a good yield, especially in clear (Kirsch et al., 1997). Performing a pleural biopsy is not patients with lung involvement found on imaging. without risks, although direct injury to the adjacent organs (liver, From a clinical perspective, these patients were mostly young kidneys, spleen) is very rare. The incidence of pneumothorax with adults, with no important co-morbid predisposing conditions, and a closed needle is variable and depends on the operator, but could with the classical signs and symptoms of fever, pleuritic chest pain, be between 8% and 18% (Gouda et al., 2006). In the present study, and cough. The duration of symptoms varied from 3 days to 1 year. pleural biopsy was performed at the discretion of the attending The yield of biochemical pleural effusion characteristics was physician and the yield was 64.7% (11/17 cases). The procedure was high, with an exudative effusion being the most common finding. performed in selected patients, in whom the diagnosis was not In a recent study of 548 cases performed by Sahn et al. (2013) the reached with other procedures. For this reason, the different yields diagnosis had 100% specificity if the pleural effusion met the of the diagnostic techniques should be interpreted with caution. following criteria: protein >5 g/dl, lymphocyte count > 80%, and Interestingly, in this study, the sputum yield remained high ADA > 45 mg/dl IU/L (sensitivity of 35%). In the patients included in even in the absence of involvement seen on chest X-ray. In a the present study, considering protein > 3.5 g/dl, lymphocyte previous report Conde et al. (2003), found a yield in induced 38 A. Macías et al. / International Journal of Infectious Diseases 78 (2019) 34–38 sputum of 55% in a cohort of Brazilian patients with presumptive References pTB and no apparent lung involvement on chest X-ray; thus sputum induction in patients with pleural effusion provides an Baussano I, Mercadante S. High rates of mycobacterium tuberculosis among socially marginalized immigrants in low-incidence area, 1991–2010, Italy. Emerg Infect invaluable opportunity to confirm a probable diagnosis and to Dis 2013;19(9):1437–45. perform a susceptibility study to guide treatment. Conde MB, Loivos AC, Rezende VM, Soares SL, Mello FC, Reingold AL, et al. Yield of From our perspective, we consider it justified to start anti-TB sputum induction in the diagnosis of pleural tuberculosis. Am J Respir Crit Care Med 2003;167(5):723–5. treatment in the case of an immigrant from a high incidence TB Davies PDO, Pai M. The diagnosis and misdiagnosis of tuberculosis. Int J Tuberc Lung country presenting with a febrile process accompanied by an Dis 2007;12(11):1226–34. exudative pleural effusion with a high number of lymphocytes and a Du J, Huang Z, Luo Q, Xiong G, Xu X, Li W, et al. Rapid diagnosis of pleural tuberculosis by Xpert MTB/RIF assay using pleural biopsy and pleural fluid high ADA level, unless there is a high risk of multidrug-resistant TB or specimens. J Res Med Sci 2015;20(1):26–31. an alternative diagnosis is highly probable. In this situation, a pleural Efsen AM, Schultze A, Post F, Panteleev A, Furrer H, Miller R, et al. Major challenges biopsy should be considered. In fact, the diagnosis was probable in 87 in clinical management of TB/HIV co-infected patients in Eastern Europe compared with Western Europe and Latin America. J Int AIDS Soc 2014;17(4 (84.5%) patients in the study cohort and treatment was started on Suppl 3):19505. this assumption, leading to a 100% therapeutic response. European Centre for Diseases Prevention and Control/WHO Regional Office for In the study cohort, the prognosis of the disease was favourable Europe. Tuberculosis surveillance and monitoring in Europe 2013. Stockholm: in the majority of patients and the cure rate was >90%, which is in European Centre for Disease Prevention and Control; 2013 https://ecdc.europa. eu/sites/portal/files/media/en/publications/Publications/Tuberculosis-surveil- concordance with the global plan to end TB (the STOP TB lance-monitoring-2013.pdf. [Accessed October 2017]. programme) (The paradigm Shift, 2015). The deaths that occurred Gopi A, Madhavan S, Sharma SK, Sahn SA. Diagnosis and treatment of tuberculous were not associated with the pleural involvement itself and were pleural effusion in 2006. Chest 2007;131:880–9. Gouda AM, Dalati TA, Al-Shareef NS. A comparison between Cope and Abrams related to multi-organ compromise, co-morbid conditions, and needle in the diagnosis of pleural effusion. Ann Thorac Med 2006;1:12–5. immunosuppression. In other series, as in this study, mortality has Informe anual 2015 Situació Epidemiològica i Tendència de l`endèmia tuberculosa a usually been associated with advanced age and co-morbid Catalunya. Generalitat de Catalunya. Agència de Salut Pública de Catalunya. http://canalsalut.gencat.cat/web/.content/contingut_responsiu/salutAZ/T/ conditions not related to TB (Efsen et al., 2014). tuberculosi/documents_prof/arxius/informe_anual_tuberculosi_2015.pdf. The limitations of this study include those inherent to the [Accessed October 2017]. retrospective design. It should be noted that there was variability Kent PT, Kubica GP. Public health mycobacteriology: a guide for the Level III laboratory. Atlanta, GA: Centers for Disease Control and Prevention; 1985. among physician criteria and a lack of molecular biology tests at Kim HJ, Lee HJ, Kwon SY, Yoon HI, Chung HS, Lee CT, et al. The prevalence of the beginning of the study period. Moreover, pTB is a disease that is pulmonary parenchymal tuberculosis in patients with tuberculous pleuritis. challenging to confirm with a difficult-to-establish gold standard Chest 2006;129:1253–8. Kirsch CM, Kroe DM, Azzi RL, Jensen WA, Kagawa FT, Wehner JH. The optimal diagnosis, hindering the assessment of diagnostic methods. number of pleural biopsy specimens for a diagnosis of tuberculous pleurisy. In conclusion, pTB represents 12% of all TB diagnoses at Vall Chest 1997;112(3):702–6. d’Hebron University Hospital, with half of the cases diagnosed in Light RW. Update on tuberculous pleural effusion. Respirology 2010;15:451–8. immigrant patients. A confirmed diagnosis was achieved in 21.3% Lusiba JK, Nakiyingi L, Kirenga BJ, Kiragga A, Lukande R, Nsereko M, et al. Evaluation of Cepheid’s Xpert MTB/Rif test on pleural fluid in the diagnosis of pleural of cases using pleural samples. The detection of lung involvement tuberculosis in a high prevalence HIV/TB setting. PLoS One 2014;9(7). increases by two-fold when lung CT is used, which correlated with Migration and migrant population statistics. Eurostat. http://ec.europa.eu/eurostat/ the likelihood of finding a positive microbiological result in the statisticsexplained/index.php/Migration_and_migrant_population_statistics. [Accessed August 2016]. sputum sample. Pleural microbiological studies had low diagnostic Na MJ. Diagnostic tools of pleural effusion. Tuberc Respir Dis 2014;76(May (5)): yield, and sputum samples, even in the absence of X-ray 199–210. abnormalities, may support the diagnosis and provide information Porcel JM, Palma R, Valdés L, Bielsa S, San-José Esquerda EA. Xpert1 MTB/RIF in pleural fluid for the diagnosis of tuberculosis. Int J Tuberc Lung Dis 2013;17 on drug susceptibility of the bacillus. (9):1217–9. Porcel JM, Azzopardi M, Koegelenberg CF, Maldonado F, Rahman NM, Lee YC. The Consent for publication diagnosis of pleural effusions. Expert Rev Respir Med 2015;8:1–15. Richter C, Perenboom R, Swai AB, Kitinya J, Mtoni I, Chande H, et al. Diagnosis of tuberculosis in patients with pleural effusion in an area of HIV infection and Not applicable. limited diagnostic facilities. Trop Geogr Med 1994;46(5):293–7. Sahn SA, Huggins JT, San José ME, Álvarez-Dobaño JM, Valdés L. Can tuberculous pleural effusions be diagnosed by pleural fluid analysis alone?. Int J Tuberc Lung Availability of data and material Dis 2013;17(6):787–93. The paradigm shift. Global plan to end TB. 2016-2020. STOP TB partnership. United The datasets generated and/or analysed during the current Nations. http://www.stoptb.org/assets/documents/global/plan/GlobalPlan- study are not publicly available due to individual privacy reasons, ToEndTB_TheParadigmShift_2016-2020_StopTBPartnership.pdf. [Accessed Oc- tober 2015]. but are available from the corresponding author on reasonable Trajman A, da Silva Santos Kleiz de Oliveira EF, Bastos ML, Belo Neto E, Silva EM, da request. Silva Lourenço MC, et al. Accuracy of polimerase chain reaction for the diagnosis of pleural tuberculosis. Respir Med 2014;108(6):918–23. World Health Organization. Global tuberculosis report. World Health Organization; Funding 2014. World medical association Declaration of Helsinki ethical principles for medical The authors declare that they received no funding for the study. research involving human subjects. http://www.wma.net/es/30publications/ 10policies/b3/. [Accessed July 2016]. Conflict of interest The authors declare that they have no competing interests.