Hui Xu1, Guoshuang Feng2, Siyu Cai3, Jinrong Liu1, Xiaolei Tang1, Hui Liu1, Haiming Yang1, Huiming Li1, Shunying Zhao1. 1. From the Department of Respiratory Medicine, Beijing Children's Hospital. 2. Big Data and Engineering Research Center, Beijing Children's Hospital. 3. Center for Clinical Epidemiology and Evidence-based Medicine, Beijing Children's Hospital, Capital Medical University, National Center for Children's Health, Beijing, China.
Abstract
BACKGROUND: Tuberculous pleural effusion (TPE) is often misdiagnosed as severe Mycoplasma pneumoniae pneumonic effusion (SMPPE) in children at early stage. The aim of this study was to develop a predictive model based on clinical and laboratory indices to make accurate differential diagnosis. METHODS: Patients included in this study were 167 children (83 patients with TPE and 84 with SMPPE), containing 117 patients for predictive model development and 50 patients for external validation. Multivariate logistic regression analysis was conducted to select potentially useful characteristics for discrimination of TPEs. External validation was performed for model evaluation. RESULTS: Multivariate analysis revealed that blood neutrophils and serum lactate dehydrogenase were significant independent factors to discriminate between TPEs and SMPPEs. The results indicated that blood neutrophils ≤69.6% and concentration of serum lactate dehydrogenase ≤297 U/L were the extremely important discrimination factors of TPEs. The area under the receiver operating characteristic curve of the model was 0.9839. The accuracy rate, sensitivity and specificity of the model were 94.02%, 98.28% and 89.83%, respectively. Meanwhile, the accuracy rate of the external validation from the 50 patients was 94.0%. CONCLUSIONS: Applying a predictive model with clinical and laboratory indices can facilitate the differential diagnosis of TPE from SMPPE in children, which seems helpful when a microbiologic or histologic diagnosis of pleural tuberculosis could not be established.
BACKGROUND:Tuberculous pleural effusion (TPE) is often misdiagnosed as severe Mycoplasma pneumoniae pneumonic effusion (SMPPE) in children at early stage. The aim of this study was to develop a predictive model based on clinical and laboratory indices to make accurate differential diagnosis. METHODS:Patients included in this study were 167 children (83 patients with TPE and 84 with SMPPE), containing 117 patients for predictive model development and 50 patients for external validation. Multivariate logistic regression analysis was conducted to select potentially useful characteristics for discrimination of TPEs. External validation was performed for model evaluation. RESULTS: Multivariate analysis revealed that blood neutrophils and serum lactate dehydrogenase were significant independent factors to discriminate between TPEs and SMPPEs. The results indicated that blood neutrophils ≤69.6% and concentration of serum lactate dehydrogenase ≤297 U/L were the extremely important discrimination factors of TPEs. The area under the receiver operating characteristic curve of the model was 0.9839. The accuracy rate, sensitivity and specificity of the model were 94.02%, 98.28% and 89.83%, respectively. Meanwhile, the accuracy rate of the external validation from the 50 patients was 94.0%. CONCLUSIONS: Applying a predictive model with clinical and laboratory indices can facilitate the differential diagnosis of TPE from SMPPE in children, which seems helpful when a microbiologic or histologic diagnosis of pleural tuberculosis could not be established.