Abstract:

Objective　To explore the risk factors of neonatal hyperbilirubinemia (NHB) complicated with brain injury and establish a Nomogram model for its risk prediction. Methods　A total of 133 children with NHB admitted to the First People's Hospital of Shangqiu City from February 2021 to February 2023 were selected as the research objects for a prospective study, including 77 boys and 56 girls, with a gestational age of (39.07±1.24) weeks and an age of (4.55±0.73) days old. They were divided into a brain injury group (25 cases) and a non-brain injury group (108 cases) based on whether they were complicated with brain injury or not. The clinical data of the two groups were compared. Univariate analysis (independent sample t test and χ² test) and multivariate logistic regression analysis were used to screen the influencing factors for brain injury in children with NHB, and R software was used to further construct the Nomogram model. Receiver operating characteristic curve (ROC) and calibration curve were used to verify the predictive efficacy of the model, and decision curve analysis (DCA) was used to evaluate the clinical net benefit of the model. Results　Among the 133 cases of NHB, 25 cases were complicated with brain injury, with an incidence of 18.80% (25/133). The results of multivariate logistic regression analysis showed that preterm birth, duration of NHB, serum unconnect bilirubin (UCB) level, bilirubin/albumin (B/A) level, bilirubine induced neurological dysfunction (BIND) score, and glucose-6-phosphate dehydrogenase (G-6-PD) deficiency were all independent risk factors for brain injury in NHB children (OR=2.798, 3.525, 4.563, 5.496, 3.800, and 3.963; all P<0.05). A Nomogram model was established based on the results of multivariate logistic regression analysis. The concordance index (C-index) of the verified prediction model was 0.875, the calibration curve was well fitted to the ideal curve, the area under ROC curve (AUC) was 0.835, the sensitivity was 88.00%, and the specificity was 73.15%. DCA evaluated the clinical application value of the model, and the results showed that the model had a high net benefit value when the threshold probability was in the range of 3% to 100%. Conclusions　Premature birth, duration of NHB, serum UCB level, serum B/A level, BIND score, and G-6-PD deficiency are independent risk factors for brain injury in children with NHB. The Nomogram model constructed on this basis has good predictive power and clinical application value.

Key words:

Hyperbilirubinemia, Brain injury, Newborns, Risk prediction, Nomegram model

摘要：

目的　探究新生儿高胆红素血症（NHB）并发脑损伤的危险因素，并建立其风险预测Nomogram模型。方法　选取2021年2月至2023年2月商丘市第一人民医院收治的133例NHB患儿作为研究对象进行前瞻性研究，其中男77例，女56例，胎龄（39.07±1.24）周，日龄（4.55±0.73）d。根据是否并发脑损伤分为脑损伤组（25例）和非脑损伤组（108例），比较两组患儿的临床资料。采用单因素分析（独立样本t检验、χ²检验）和多因素logistic回归分析筛选NHB患儿并发脑损伤的影响因素，利用R软件进一步构建Nomogram模型。采用受试者工作特征曲线（ROC）、校准曲线对模型预测效能进行验证，采用决策曲线分析（DCA）评估模型的临床净收益。结果　133例NHB患儿中有25例并发脑损伤，发病率为18.80%（25/133）；多因素logistic回归分析结果显示，早产、NHB持续时间、血清游离胆红素（UCB）水平、胆红素/白蛋白（B/A）水平、胆红素所致神经功能障碍（BIND）评分、合并葡萄糖-6-磷酸脱氢酶（G-6-PD）缺乏症均是NHB患儿并发脑损伤的独立危险因素（OR=2.798、3.525、4.563、5.496、3.800、3.963，均P<0.05）；根据多因素logistic回归分析结果建立Nomogram模型，经验证预测模型的一致性指数（C-index）为0.875，校准曲线与理想曲线拟合反映良好，ROC曲线下面积（AUC）为0.835，灵敏度为88.00%，特异度为73.15%；DCA评估模型的临床应用价值，结果显示当阈值概率在3%～100%范围内时，模型具有高净获益值。结论　早产、NHB持续时间、血清UCB水平、血清B/A水平、BIND评分、合并G-6-PD缺乏症是NHB患儿并发脑损伤的独立危险因素，在此基础上构建的Nomogram模型具有良好的预测效能和临床应用价值。

关键词:

高胆红素血症, 脑损伤, 新生儿, 风险预测, Nomogram模型