影响早产儿脑损伤发生的危险因素分析及风险预测模型的构建与检验

Analysis of risk factors affecting brain injury in premature infants and construction and testing of risk prediction models

目的探讨影响早产儿脑损伤发生的危险因素,构建并检验风险模型的预测效能。方法收集2014年1月-2016年1月我院新生儿科收治的476例早产儿的临床资料,根据颅脑超声结果,将新生儿分为脑损伤组与未损伤组,分析可能影响脑损伤发生的相关危险因素;根据危险因素权重,构建早产儿脑损伤"危险评分-预测概率"模型。连续性收集2016年2月-2017年12月收治的486例早产儿资料作为验证组,验证模型的预测效能。结果 (1)单因素分析结果显示,分娩方式、机械通气、产前存在高危因素、低体重、低胎齡、1min Apgar评分、低血糖水平是影响早产儿发生脑损伤的危险因素,差异有统计学意义(P<0.05)。(2)当新生儿体质量<1 310g、胎龄(GA)<32周、1 min Apgar评分<3、pH <7.15、二氧化碳分压[P(CO_2)]≥75 mmHg、碱剩余(BE)<-10.0mmol/L、血糖(GLU)<3mmol/L时,早产儿脑损伤的发生率,差异有统计学意义(P<0.05)。(3)多因素Logistic分析结果显示,GA、pH、P(CO_2)、BE、GLU、机械通气、产前高危因素是影响影响早产儿脑损伤发生的独立性危险因素。(4)早产儿脑损伤危险分数总分为18分,其中,产前危险因素评分值最高,为5分;其次,各因素的危险分数依次为行机械通气4分,BE<-10mmol/L:3分,P(CO_2)≥65mmHg:2分,GA:2分,pH<7.25:1分,GLU<2mmol/L:1分。危险评分从0～18分对应的预测值分别为3.9%、5.5%、7.8%、10.8%、14.9%、19.8%、27.4%、34.5%、43.5%、51.8%、54.0%、66.7%、77.0%、83.3%、83.5%、90.0%、92.0%、94.0%、96.0%,发病风险呈显著的上升趋势。(5)利用ROC曲线对验证组486例早产儿脑损伤发生率进行检验发现,ROC曲线下面积为0.79,Hosmer-Lemeshow拟合优度检验(χ~2=1.52,P=0.93)。模型拟合效度好,预测价值高。结论影响早产儿发生脑损伤的高危因素较多,临床应进行及时有效评估。预测模型P预测=1/(1+exp[-(β0+Sc×βm)])有较好的评估效能,有一定的临床应用价值。

Objective To study the risk factors affecting the occurrence of brain injury in preterm infants,and to construct and test the predictive effectiveness of the risk model.Methods the clinical data of 476 premature infants who were treated in our hospital from January 2014 to January 2016 were collected,and the neonates were divided into brain injury group and non injury group,and the possible risk factors were analyzed.At the same time,the data of 486 preterm infants were collected to verify the predictive effectiveness of the model.Results(1)Delivery mode,mechanical ventilation,high-risk factors before delivery,low body weight,low fetal weight,1 min Apgar score and low GLU level were risk factors(P<0.05).(2)When the weight is less than 1 310 g,gestational age less than 32W,1 min Apgar score<3,pH<7.15.PCO 2=75 mmHg,BE<-10.0 mmol/L,GLU<3 mmol/L,There was significant difference in the incidence of brain injury in preterm infants(P<0.05).(3)The results of multiple factors analysis showed that GA,PH,PCO 2,BE,GLU,mechanical ventilation,and prenatal risk factors were independent risk factors.(4)The risk of premature brain damage scores was 18 points,the highest score of prenatal risk factors.Risk score is 5 points;secondly the factors in mechanical ventilation for 4 points,BE<-10mmol/L:3,PCO 2=65 mmHg:2,GA 2,pH<7.25:1,GLU<2 mmol/L:1 branch.(5)Using the ROC curve to test the incidence of 486 cases of neonatal brain injury in the validation group,it was found that the area under the ROC curve was 0.79,and Hosmer-Leme show goodness of fit test showed thatχ2 value=1.52 and P=0.93.The model has good fitting validity and high prediction value.Conclusion There are many high risk factors affecting the brain injury in preterm infants.The clinical evaluation should be carried out in time and effectively.Prediction model P predicts=1/(1+exp[-(β0+Sc×βm)]with good evaluation efficiency,and has certain clinical application value.