Prognostic Evaluation Of Different Critical Illness Scoring Systems In Children With Sepsis

Objective:To explore the predictive value of four scoring systems,Pediatric Clinical Illness Score（PCIS）,Pediatric Logistic Organ Dysfunction Score 2（PELOD-2）,Pediatric Risk of Mortality Ⅲ（PRISM Ⅲ）,and Pediatric Sequential Organ Failure Assessment（pSOFA）,in the severity of disease and prognosis in paediatric sepsis.Materials and methods:In strict accordance with the inclusion and exclusion criteria,the clinical data of 159 children with sepsis treated in the PICU of the First Hospital of Jilin University from August2015 to December 2020 were retrospectively collected.The children were divided into survival group（n = 97）and death group（n = 62）based on clinical outcomes during hospitalization and the differences between the two groups were compared.The children were stratified according to severity of sepsis into sepsis group（n=18）,severe sepsis group（n=49）and septic shock group（n=92）,comparing the differences between the death and survival subgroups of each group.The pSOFA,PRISM Ⅲ,PCIS,and PELOD-2 scores were calculated according to the worst value of each index within 24 h of admission.Using the ROC curve,we analyzed: 1.the predictive effectiveness of each critical illness scoring system for the development of severe sepsis or septic shock in children with sepsis;2.the predictive effectiveness of each critical illness scoring system for the development of poor prognosis in overall sepsis and each stratification group of sepsis.Results:1.This study collected 159 patients,male to female ratio was 1.01: 1;the median age was2（0.83,9）years;the total length of stay was 14（5,22）days,and the length of stay in PICU was 13（5,21）days.97 cases were in the survival group（61.01%）and 62 cases in the death group（38.99%）;There were 18 children with sepsis（11.32%）,49 children with severe sepsis（30.82%）and 92 children with septic shock（57.86%）.The most common infection site was respiratory tract infection（43.4%）which was followed by digestive tract infection（13.21%）,skin soft tissue infection（8.81%）,intracranial infection（5.03%）,bloodstream infection（5.03%）.For underlying diseases,34 cases（22.67%）had underlying diseases,including 11 cases of hematological malignancies（32.35%）,8 cases of rheumatic immune system diseases（23.53%）,7 cases of congenital heart disease（20.59%）,4 cases of nervous system diseases（11.76%）,3 cases of congenital malformations（8.82%）,1 case of congenital genetic metabolic diseases（2.94%）.2.There were statistically significant differences in the total length of stay,length of PICU stay,multiple organ dysfunction syndrome（MODS）incidence,underlying illness,PCIS,PELOD-2,PRISM Ⅲ and pSOFA between the nonsurvivors and survivors in all case groups（all P < 0.05）;The proportion of children with severe sepsis and septic shock was higher in the nonsurvivors（98.4% vs.82.5%）,with a significant difference（P<0.05）.3.Subgroup analysis was as follows:（1）In severe sepsis subgroup,compared with the survivors,the nonsurvivors had shorter PICU length of stay and total hospital stay,and lower PCIS score,PELOD-2,and PRISM Ⅲ scores,all of which were statistically different（all P <0.05）;Compared with survivors,nonsurvivors had higher pSOFA,which had not significant difference（P > 0.05）.（2）In septic shock subgroup,compared with survivors,nonsurvivors had shorter PICU length of stay,total hospital length of stay,and higher PRISM Ⅲ score,all of which had statistical significance（all P < 0.05）;compared with survivors,nonsurvivors had lower PCIS score,PELOD-2 and pSOFA score,all of which had no significant difference（P >0.05）.4.The predictive efficacy of each scoring system for the death of sepsis: Areas Under Curve（AUCs）of pSOFA,PELOD-2,PRISM Ⅲ and PCIS scores for predicting the prognosis of children with sepsis in PICU were 0.656,0.697,0.716 and 0.687,respectively（all P < 0.05）,indicating that compared with pSOFA,PELOD-2 and PCIS scores,PRISM Ⅲ had higher predictive value for the adverse prognosis of children with sepsis.5.The AUCs of PCIS,PELOD-2 and PRISM Ⅲ scores for predicting death in children with severe sepsis in the intensive care unit were 0.726,0.684 and 0.712,respectively（all p <0.05）,indicating PCIS and PRISM Ⅲ scores had predictive value（the best cut-off values were78 and 10.5,with sensitivities of 90.9% and specificities of 55.6% and 48.1%,respectively）,with the PCIS score having a higher predictive value;the pSOFA score had no predictive value for the occurrence of death in children with severe sepsis（AUC: 0.611,P=0.173 > 0.05）.6.The AUC of PRISM Ⅲ for predicting death in paediatric septic shock in the PICU was0.692（P < 0.05）,indicating that the score was not an adequate predictor of death in children with septic shock in the PICU.pSOFA,PELOD-2 and PCIS had no predictive value for death in children with septic shock（P values: 0.186,0.071,0.172,all P > 0.05）.7.The AUC of pSOFA,PRISM Ⅲ,PELOD-2 and PCIS scores for predicting severe sepsis or septic shock in children with sepsis were 0.933,0.672,0.875 and 0.869,respectively（all P < 0.05）,indicating that compared with PELOD-2,PCIS and PRISM Ⅲ.The pSOFA score had the greatest predictive value for severe sepsis or septic shock in children with sepsis（the optimal cut-off value was 5.5,the sensitivity and specificity were 90.8 % and 94.4 %,respectively）.Conclusion:Among the four severe scoring systems,the PRISM Ⅲ score had the highest value in assessing the risk of death in all septic patients in the pediatric intensive care unit;the PCIS and PRISM Ⅲ scores had a low predictive power for predicting the risk of death in children with severe sepsis,and the pSOFA scoring system had the highest predictive power for disease worsening in children with sepsis.