Predictive Value Of First Blood Routine And Coagulation Function Parameters Within 24hours Of Admission For Critical Illness In Children

Objective: The parameters of blood routine and coagulation function tests were frequently used to predict disease severity and mortality outcomes in children and adults.At present,the Pediatric Critical Illness Score（PCIS）and simplified PCIS were commonly used in China to assess the severity of children’s illness.However,the simplified PCIS still needs to score 8 items,which is not simple enough.In this study,based on the PCIS the correlation between the parameters of the first blood routine and coagulation function within 24 hours of admission and critical illness in patients was explored,and a nomogram prediction model was established to evaluate its predictive value for critical illness in children.Methods: A retrospective study was conducted.From April 2015 to December2019,patients who were admitted to the pediatric intensive care unit of the First Affiliated Hospital of Bengbu Medical College and underwent blood routine and coagulation function tests within 24 hours of admission were enrolled.The patients were randomly divided into a training cohort and a validation cohort at a ratio of 7:3according to R language sample function.The patients of the two cohorts were divided into a critical group and a non-critical group according to the PCIS and simplified PCIS score,respectively.The correlation between the variables of the training cohort and critical illness was analyzed by univariate and multivariate logistic regression analysis,and a nomogram prediction model was established by the R language software.The concordance index and calibration curve were used to verify the prediction of critical illness in children in the training and validation cohort,respectively,and the decision curve and clinical impact curve were used to verify the clinical application value of the nomogram model for the prediction of critical illness in the training and validation cohort,respectively.The receiver operating characteristic curve was used to determine the optimal risk threshold probability of critical illness in children in the validation cohort,and the sensitivity,specificity,positive predictive value,and negative predictive value were calculated.Results: Of the 496 patients,there were 283 males（57%）and 213 females（43%）,with a median age of 2.00（0.57,5.88）years.There were 347 patients in the training cohort,including 113 in the critically ill group and 234 in the non-critically ill group,and 149 patients in the validation cohort,including 48 in the critically ill group and 101 in the non-critically ill group.There was no significant difference in the baseline data between the training and validation cohort（P>0.05）.Hemoglobin（HB）（OR=0.986,95%CI=0.973-1.000,P=0.049）,reticulocyte percentage（RET%）（OR=1.930,95%CI=1.358-2.745,P<0.001）,prothrombin time（PT）prolongation（OR=1.138,95%CI=1.032-1.256,P=0.010）,and D-dimer（DD）（OR=1.018,95%CI=1.001-1.035,P=0.040）were significantly associated with critical illness in children in the training cohort.The nomogram prediction model was constructed using HB,RET%,PT prolongation and DD,and the concordance index（0.711 for training cohort,95%CI=0.651-0.771;0.685 for validation cohort,95%CI=0.585-0.785）,calibration curve,decision curve and clinical impact curve suggested that this nomogram model can be used to predict critical illness in children.The receiver operating characteristic curve analysis showed that the area under the curve of HB + RET% + PT prolongation + DD combined index for predicting critical illness of children in the validation cohort was 0.650（95%CI=0.568-0.727,P=0.003）.At the optimal risk threshold probability of 0.3366,the sensitivity,specificity,positive predictive value,and negative predictive value of the combined index for predicting critical illness in children in the validation cohort was 50%（95%CI=35.2%-64.8%）,81.2%（95%CI=72.2%-88.3%）,55.8%（95%CI=43.5%-67.4%）,and 77.4%（95%CI=71.7%-82.2%）,respectively.Conclusions HB,RET%,PT prolongation,and DD within 24 hours of admission were independent predictors of critical illness in children,of which RET%,PT prolongation and DD were risk factors,and HB was a protective factor.The nomogram prediction model can evaluate the occurrence probability of critical illness by HB,RET%,PT prolongation,and DD,and has a certain accuracy,simplicity,and clinical application value.