基于新生儿糖尿病低血糖风险预测模型构建血糖分级管理方案的效果评价  

Evaluation of the Effect of Blood Glucose Grading Management Scheme Based on Hypoglycemia Risk Prediction Model of Neonatal Diabetes

作　　者：弓雪茹 古建平 GONG Xueru;GU Jianping(Children's Hospital Affiliated to Zhengzhou University,Henan Children's Hospital,Zhengzhou Children's Hospital,Henan Zhengzhou 450000)

机构地区：[1]郑州大学附属儿童医院,河南省儿童医院,郑州儿童医院,河南郑州450000

出　　处：《深圳中西医结合杂志》2024年第5期6-10,共5页Shenzhen Journal of Integrated Traditional Chinese and Western Medicine

基　　金：河南省医学科技攻关计划联合共建项目(LHGJ20190900)。

摘　　要：目的:构建新生儿糖尿病患儿低血糖风险预测模型,并分析基于该模型的血糖分级管理方案的应用效果。方法:选取2021年6月至2023年6月郑州大学附属儿童医院收治的新生儿糖尿病患儿96例,记录其低血糖发生情况,将发生低血糖患儿纳入低血糖组(26例),未发生低血糖患儿纳入对照组(70例),采用logistic回归模型分析新生儿糖尿病患儿发生低血糖的影响因素,建立风险预测模型,并基于该模型构建的血糖分级管理方案,连续干预新生儿糖尿病患儿6个月,分析干预前后的低血糖发生率变化。结果:logistic回归分析结果显示,出生体质量、早产、低体温、喂养不佳、胰岛素过度治疗、妊娠期糖尿病是新生儿糖尿病低血糖发生的独立危险因素(P <0.05);风险预测模型:风险评分=[e^(Z)/(e^(Z)+1)]×100%;受试者工作特征(ROC)曲线分析结果显示,该模型的曲线下面积(AUC)值为0.812,95%CI为(0.719,0.884),截断值为0.11,灵敏度为76.92%,特异度为78.57%,约登指数为0.555;干预期间,新生儿糖尿病患儿低血糖发生率呈持续下降趋势,且干预1个月后各时间点的低血糖发生率均较入组时显著降低,差异均具有统计学意义(P <0.05)。结论:基于新生儿糖尿病低血糖风险预测模型构建的血糖分级管理方案可有效降低新生儿糖尿病患儿低血糖的发生风险。Objective To construct a hypoglycemia risk prediction model for neonatal diabetes mellitus patients,and to analyze the application effect of a graded blood glucose management scheme based on the model.Methods A total of 96 patients with neonatal diabetes mellitus admitted to Children's Hospital Affiliated to Zhengzhou University from June 2021 to June 2023 were selected,and the occurrence of hypoglycemia was recorded.The children with hypoglycemia were included in the hypoglycemia group(26 cases),and the children without hypoglycemia were included in the control group(70 cases).Logistic regression model was used to analyze the risk factors of hypoglycemia in neonates with diabetes mellitus,and a risk prediction model was established.A graded blood glucose management scheme was constructed based on the model.After continuous intervention for 6 months,the incidence of hypoglycemia before and after intervention was analyzed.Results logistic regression analysis showed that birth weight,premature birth,hypothermia,poor feeding,excessive insulin treatment,and gestational diabetes mellitus were independent risk factors for hypoglycemia in neonates with diabetes mellitus(P < 0.05).Risk prediction model:risk score = [e~Z/(e~Z + 1)] × 100 %.The results of receiver operating characteristic(ROC) curve analysis showed that the area under the curve(AUC) value of the curve was 0.812,95% CI was(0.719,0.884),the cut-off value was 0.11,the sensitivity was 76.92 %,the specificity was 78.57 %,and the Youden index was 0.555.During the intervention period,the incidence of hypoglycemia in patients with neonatal diabetes mellitus showed a continuous downward trend,and the incidence of hypoglycemia at each time point after 1 month of intervention was significantly lower than that at the time of enrollment,and the differences were statistically significant(P < 0.05).Conclusion The graded blood glucose management scheme based on the hypoglycemia risk prediction model for neonatal diabetes mellitus can effectively reduce the risk of hypogl

关 键 词：新生儿糖尿病 低血糖 风险预测模型 血糖分级管理 

分 类 号：R587.1[医药卫生—内分泌]