Analysis of risk factors and construction of risk warning model for late-onset Escherichia coli sepsis in newborns
Objective To explore risk factors of late-onset Escherichia coli sepsis in newborns and construct a column chart model for the late-onset Escherichia coli sepsis in newborns.Methods A total of 226 neonates who admitted to our hospital from January 2021 to July 2023 were selected as the model group,and another 100 neonates who admitted to our hospital over a period from August 2021 to October 2023 were selected as the verification group.Logistic regression was used to analyze risk factors for neonatal late-onset Escherichia coli sepsis,and R software was used to build a column chart model for late-onset Escherichia coli sepsis in newborns and the model was verified.Results There were significant differences in gestational age,mechanical ventilation,birth weight,parenteral nutrition,premature rupture of membranes and long-term antibiotic exposure between the late-onset Escherichia coli septicemia group and the non-late-onset Escherichia coli septicemia group(x2=11.492,11.605,10.748,11.494,6.089 and 7.222 respectively,all P＜0.05).That was,compared with the non-late-onset Escherichia coli septicemia group,the neonates in the late-onset Escherichia coli septicemia group had lower gestational age,higher proportion of mechanical ventilation,lower birth weight,higher proportion of parenteral nutrition,longer duration of premature rupture of membranes and longer antibiotic exposure.Logistic regression analysis showed that gestational age＜37 weeks(OR=4.031,95％CI:1.492-10.982),mechanical ventilation(OR=3.064,95％CI:1.306-7.188),low birth weight(OR=3.262,95％CI:1.200-8.873),parenteral nutrition(OR=2.852,95％CI:1.092-7.450),long time of premature rupture of membranes(OR=1.517,95％CI:1.275-1.806)and prolonged antibiotic exposure(OR=2.947,95％CI:1.216-7.140)were risk factors for late-onset Escherichia coli septicemia in newborns(all P＜0.05).The calibration curves in both the model group and the verification group showed that the predictive values were basically consistent with the actual values.The area under ROC curve of the model group was 0.856(95％CI:0.820-0.892),and that of the verification group was 0.846(95％CI:0.800-0.893).The decision curve analysis showed that when the threshold probability were at l％to 100％in the model group and 1％to 98％in the verification group,the net benefit values of the nomogram for predicting neonatal late-onset Escherichia coli sepsis were higher.Conclusion Gestational age＜37 weeks,mechanical ventilation,low birth weight,parenteral nutrition,long time of premature rupture of membranes and prolonged antibiotic exposure are risk factors for neonatal late-onset Escherichia coli sepsis.The nomogram model of neonatal late-onset Escherichia coli sepsis is of high accuracy and clinical application value.
万方数据
维普
