Dose–response relationship and predictive value of soluble B7‐DC in bronchoalveolar lavage fluid and risk of refractory Mycoplasma pneumoniae pneumonia in children

Abstract

AbstractThis study was to investigate the clinical significance of soluble B7‐dendritic cell (sB7‐DC) concentration in bronchoalveolar lavage fluid (BALF) of children with Refractory Mycoplasma pneumoniae pneumonia (RMPP). A total of 298 patients with Mycoplasma pneumoniae pneumonia (MPP) were enrolled. Patients were divided into general MPP (GMPP) (n = 213) and RMPP groups (n = 85). Detection of sB7‐DC and serum inflammatory factors in BALF was performed by ELISA. The relationship between sB7‐DC and the risk of RMPP was assessed using restricted cubic spline (RCS) model. A base model for predicting RMPP was constructed using logistic regression analysis, and a compound model was created with the addition of sB7‐DC in the base model. ROC curves were plotted to evaluate the predictive value of the model. Column line plots were plotted to assess the contribution of each variable to the outcome event. Calibration curves were plotted and the Hosmer–Lemeshow test (HL test) was performed to assess the calibration performance of the model. Decision curve analysis (DCA) plots were plotted to assess determine whether sB7‐DC has clinical value. There was no statistical difference between sB7‐H3 and sB7‐H4 in the two groups (both p > 0.05). sB7‐DC levels were higher in the RMPP group than in the GMPP group (91.66 [77.36, 122.5] pg/ml vs. 64.87 [47.07, 86.46] pg/ml, p < 0.001). RCS analysis showed that the risk of RMPP gradually increased with the increase of sB7‐DC when sB7‐DC > 76.505 pg/ml. Both the base model and the compound model constructed with independent correlates of RMPP had some predictive value, and the models were well‐fitted. The column line graphs showed that the models had discriminative ability. Notably, the compound model had a higher predictive value, with a higher AUC value than the base model: 0.76 (0.65–0.87) versus 0.68 (0.54–0.81). The highest net benefit was close to 0.15 (only 0.1 in the base model). When the net benefit was >0, the high‐risk threshold took on a wide range of values. sB7‐DC in children with RMPP is an independent predictor of RMPP. sB7‐DC helps to improve quantitative prediction of RMPP risk and accurately guide medical decisions.