Construction of a prediction model for poor prognosis of children with RMPP based on serum SAA/CRP and RBP levels
CHANG Zhenzhen
GAO Wanqing
TONG Renxiang
CHEN Qian
LIU Nana
Abstract：Objective To construct a prediction model for poor grognosis of children with refractory Mycoplasma pneumoniae pneumonia(RMPP)based on serum levels of serum amyloid A to C-reactive protein ratio(SAA/CRP)and ret-inol-binding protein(RBP).Methods A total of 198 children with RMPP were enrolled,and their serum levels of SAA/CRP and RBP were measured.Based on the prognosis at 28 days after admission,the children were divided into the poor prognosis group and good prognosis group.Relevant clinical data were collected.Multivariate Logistic regression analysis was used to identify factors influencing poor prognosis and to construct a prediction model.The receiver operating charac-teristic(ROC)curve analysis was employed to evaluate the predictive value of individual indicators and the prediction mod-el for poor prognosis in RMPP children.Internal validation was performed using the Bootstrap method.Results At 28 days after admission,the incidence of poor prognosis among the 198 RMPP children was 30.30%(60/198).Compared with the good prognosis group,the poor prognosis group had a higher proportion of severe cases,longer duration of fever and oxygen therapy,elevated serum SAA/CRP levels,and decreased RBP levels(all P<0.05).Multivariate regression analysis revealed that the duration of fever,severe RMPP,long duration of oxygen therapy,and high SAA/CRP were inde-pendent risk factors for poor prognosis in RMPP children,while high RBP was an independent protective factor(all P<0.05).Based on these independent factors,a prediction model was constructed,and the regression equation was:Logit(P)=-10.707+0.347×duration of fever+1.811×disease severity+0.765×duration of oxygen therapy+1.085×SAA/CRP-0.258×RBP.The area under the ROC curve(AUC)for the model was 0.920,which was higher than the AUCs for duration of fever(0.674),disease severity(0.696),duration of oxygen therapy(0.682),SAA/CRP(0.755),and RBP(0.747)alone(all P<0.05).The Hosmer-Lemeshow test showed χ²=6.320,P=0.611,indicat-ing good model fit.Internal validation using 1000 Bootstrap resamples showed that the prediction model had a C-index of 0.920,indicating high discrimination and stability.Conclusions Elevated serum SAA/CRP levels and decreased RBP levels are associated with poor prognosis in children with RMPP.The prediction model developed based on these factors demonstrates high predictive efficacy.
Keywords：refractory Mycoplasma pneumoniae pneumoniaserum amyloid AC-reactive proteinretinol-binding proteinpredictive model
Publication Date：2026-01-25
Online Publishing Date：2026-03-05(First online date of this platform, not the publication date of the document)
Pages：5( 74-78 )