糖尿病酮症酸中毒合并重症胰腺炎预后列线图预测模型构建

邓双南; 帅毓梅

doi:10.13898/j.cnki.issn.2097-5252.2025.12.019

糖尿病酮症酸中毒合并重症胰腺炎预后列线图预测模型构建

Construction of a prognostic nomogram prediction model for diabetic ketoacidosis complicated with severe pancreatitis

摘要

摘要:
目的探讨糖尿病酮症酸中毒(DKA)合并重症急性胰腺炎(SAP)病人预后的影响因素，并构建列线图预测模型。
方法采用便利抽样法选择136例DKA合并SAP病人作为研究对象，根据病人预后将其分为存活组和死亡组，收集2组临床资料进行单因素分析和二元logistic回归分析，构建列线图预测模型，并验证模型的预测性能及临床效用。
结果 136例DKA合并SAP病人中，死亡66例，死亡率为48.53%。单因素分析结果显示，性别、感染性休克、循环衰竭、急性生理与慢性健康(APACHE Ⅱ)评分、序贯器官衰竭(SOFA)评分、C反应蛋白(CRP)、白细胞计数(WBC)、三酰甘油(TG)、血尿素氮(BUN)、Ca²⁺均为DKA合并SAP病人死亡的影响因素。二元logistic回归分析结果显示，感染性休克、APACHE Ⅱ评分、CRP、TG、Ca²⁺均为DKA合并SAP病人预后死亡的独立危险因素。构建列线图预测模型，总分>104分时，病人死亡风险为90%，模型验证结果显示，训练集预测预后死亡AUC=0.991(95%CI: 0.979~1.000)，测试集预测预后死亡AUC=0.961(95%CI: 0.883~1.000)，Hosmer-Lemeshow检验χ²=1.10，P=1.00，训练集和测试集MAE分别为0.021和0.171，训练集和测试集MSE分别为0.001和0.035，决策曲线分析和临床影响曲线分析均证明该预测模型具有良好的临床实用性。
结论 DKA合并SAP病人死亡风险较高，通过构建列线图预测模型强化病人预后评估，可为病人治疗方案的制定及优化提供参考依据。

Abstract:
Objective To investigate the influencing factors of prognosis of patients with diabetic ketoacidosis(DKA) combined with severe acute pancreatitis(SAP) and to construct a nomogram prediction model.
Methods A total of 136 patients with DKA combined with SAP were selected as the research subjects by convenience sampling method.According to the prognosis of patients, the patients were divided into the survival group and death group.The clinical data of two groups were collected for univariate analysis and binary logistic regression analysis to construct a nomogram prediction model, and the predictive performance and clinical utility of model were verified.
Results Among the 36 patients with DKA combined with SAP, 66 died, with a mortality rate of 48.53%.The results of univariate analysis showed that the gender, septic shock, circulatory failure, acute physiology and chronic health (APACHE Ⅱ) score, sequential organ failure assessment(SOFA) score, C reactive protein(CRP), white blood cell count(WBC), triglyceride(TG), blood urea nitrogen (BUN), and Ca²⁺ were the nfluencing factors of the death of patients with DKA combined with SAP.The results of binary logistic regression analysis showed that septic shock, APACHE Ⅱ score, CRP, TG, and Ca²⁺ were the independent risk factors of prognosis and death in patients with DKA combined with SAP.A nomogram prediction model was constructed.When the total score was above 104 points, the risk of patient death was 90%.The model validation results showed that the AUC for predicting prognostic death in the training set was 0.991(95%CI: 0.979-1.000), and that in the test set was 0.961 (95%CI: 0.883-1.000).Hosmer-Lemeshow test found χ²=1.10, P=1.00, the MAE of the training set and test set were 0.021 and 0.171, respectively, and the MSE of the training set and test set were 0.001 and 0.035, respectively.Both the decision curve analysis and clinical impact curve analysis proved that this prediction model had good clinical practicability.
Conclusions DKA combined with SAP patients has a higher risk of mortality.By contructing a column chart prediction model to enhance patient prognosis evaluation, it can provide reference for the formulation and optimization of patient treatment plants.

HTML全文

参考文献(27)

施引文献

资源附件(0)