F C Abrao1, I R L Bruno de Abreu1, D H Miyake, D H Miyaki1, M A M Busico2, R N Younes3. 1. Department of Thoracic Surgery, Hospital Santa Marcelina, São Paulo, Brazil. 2. Department of Pneumology, Instituto Clemente Ferreira, São Paulo, Brazil. 3. Department of Surgery, Hospital Sao Jose, University of São Paulo, São Paulo, São Paulo, Brazil.
Abstract
OBJECTIVE: 1) To determine factors affecting adenosine deaminase (ADA) levels in pleural fluid (PF), and 2) to establish the optimal ADA cut-off level for a Brazilian population. DESIGN: ADA levels in PF of 309 patients were analysed to investigate pleural effusion. All patients were evaluated for age, sex and presence of tuberculosis (TB) based on a positive pleural biopsy. Differences in ADA levels between groups were analysed using Kruskal-Wallis one-way analysis of variance. Logistic regression analysis was also carried out to predict the occurrence of TB. ADA cut-off levels were selected using the receiver operating characteristic (ROC) curve. RESULTS: The mean PF ADA level was significantly higher in the tuberculous pleural group than in non-tuberculous pleural patients (63.3 ± 29 IU/l vs. 19 ± 31 IU/l, P < 0.001). There was a significant correlation between PF ADA levels and age: for patients aged ⩾45 years, the ROC curve for ADA had an area under the curve of 0.91. An ADA level of 29 IU/l resulted in a sensitivity of 88.6% and specificity of 91.5%. CONCLUSIONS: There is a significant negative correlation between PF ADA level and age. The use of a lower ADA cut-off reduces the number of false-negative results.
OBJECTIVE: 1) To determine factors affecting adenosine deaminase (ADA) levels in pleural fluid (PF), and 2) to establish the optimal ADA cut-off level for a Brazilian population. DESIGN: ADA levels in PF of 309 patients were analysed to investigate pleural effusion. All patients were evaluated for age, sex and presence of tuberculosis (TB) based on a positive pleural biopsy. Differences in ADA levels between groups were analysed using Kruskal-Wallis one-way analysis of variance. Logistic regression analysis was also carried out to predict the occurrence of TB. ADA cut-off levels were selected using the receiver operating characteristic (ROC) curve. RESULTS: The mean PF ADA level was significantly higher in the tuberculous pleural group than in non-tuberculous pleuralpatients (63.3 ± 29 IU/l vs. 19 ± 31 IU/l, P < 0.001). There was a significant correlation between PF ADA levels and age: for patients aged ⩾45 years, the ROC curve for ADA had an area under the curve of 0.91. An ADA level of 29 IU/l resulted in a sensitivity of 88.6% and specificity of 91.5%. CONCLUSIONS: There is a significant negative correlation between PF ADA level and age. The use of a lower ADA cut-off reduces the number of false-negative results.