Abstract
BACKGROUND: Differentiating tuberculous pleural effusion (TPE) from non-tuberculous benign pleural effusion (non-TB BPE) can be challenging, especially in patients with low levels of pleural effusion adenosine deaminase (pADA) and negative etiological evidence. This study aimed to evaluate the diagnostic performance of the pADA to serum C-reactive protein ratio (pADA/sCRP) and to develop a scoring model for diagnosing TPE. METHODS: This retrospective study included 364 patients with pleural effusion, comprising 121 with TPE and 243 with non-TB BPE from Peking Union Medical College Hospital. Clinical, laboratory, and imaging data were collected, and comparisons were made between the two groups. The diagnostic performance of the pADA/sCRP ratio and other statistically significant variables was assessed. Six valuable factors were selected for multivariate regression analysis to establish a predictive model, which was displayed as a nomogram. RESULTS: The pADA/sCRP ratio demonstrated superior diagnostic performance compared to pADA alone, with an area under the curve (AUC), sensitivity, and specificity for identifying TPE of 0.68, 71%, and 64%, respectively. Six variables were selected to develop a nomogram, including night sweats, calcification on chest computed tomography, pleural effusion lymphocyte ratio, pADA/sCRP, hemoglobin, and neutrophil. With a cutoff value of 20 points, the AUC, sensitivity, and specificity for distinguishing TPE from non-TB BPE were 0.836, 83.4%, and 64.9%, respectively. The validation cohort confirmed the model with the AUC, sensitivity, and specificity of 0.815, 61.1%, and 82.4%, respectively. CONCLUSION: The pADA/sCRP ratio exhibited improved diagnostic performance compared to pADA alone. The novel scoring system based on a nomogram demonstrated good diagnostic efficacy in differentiating TPE from non-TB BPE.