Diagnostic performance of D-dimer in predicting pulmonary embolism in tuberculous pleural effusion patients

Abstract

Abstract Background Tuberculous pleural effusion (TPE) patients usually have elevated D-dimer levels. The diagnostic performance of D-dimer in predicting pulmonary embolism (PE) in the TPE population is unclear. This study aimed to assess the diagnostic performance of D-dimer for PE in the TPE population and explore its potential mechanism. Methods We retrospectively analysed patients who were admitted to Xinhua Hospital and Weifang Respiratory Disease Hospital with confirmed TPE between March 2014 and January 2020. D-dimer levels were compared between patients with and without PE. To test the diagnostic performance of D-dimer in predicting PE, receiver operating characteristic curve analysis was performed. Positive predictive value (PPV) and negative predictive value (NPV) were also reported. To explore the potential mechanism of PE in TPE, inflammatory biomarkers were compared between PE and non-PE patients. Results This study included 248 patients (170 males and 78 females) aged 43 ± 20.6 years. Elevated D-dimer levels (≥ 0.5 mg/L) were detected in 186/248 (75%) patients. Of the 150 patients who underwent computed tomography pulmonary angiography, 29 were diagnosed with PE. Among the TPE population, the PE patients had significantly higher D-dimer levels than the non-PE patients (median, 1.06 mg/L vs. 0.84 mg/L, P < 0.05). The optimal cut-off value for D-dimer in predicting PE in TPE was 1.18 mg/L, with a sensitivity of 89.7% and a specificity of 77.8% (area under curve, 0.893; 95% confidence interval 0.839–0.947; P < 0.01). The PPV was 49.1%, while the NPV was 96.9% at a D-dimer cut-off of 1.18 mg/L for PE. PE patients had lower median WBC and interleukin (IL)-8 values (5.14 × 109/L vs. 6.1 × 109/L, P < 0.05; 30.2 pg/ml vs. 89.7 pg/ml, P < 0.05) but a higher median IL-2 receptor value (1964.8 pg/ml vs. 961.2 pg/ml, P < 0.01) than those in the non-PE patients. Conclusions D-dimer is an objective biomarker for predicting PE in patients with TPE. A D-dimer cut-off of 1.18 mg/L in the TPE population may reduce unnecessary radiological tests due to its excellent sensitivity, specificity, and NPV for PE. The imbalance of prothrombotic and antithrombotic cytokines may partly be attributed to the formation of pulmonary emboli in patients with TPE.