|
| 预测模型分析妊娠期糖尿病产妇不良妊娠结局影响因素的应用价值 |
| Application value of the prediction model in analyzing the influencing factors of adverse pregnancy outcomes in puerperae with gestational diabetes mellitus |
| 投稿时间：2024-04-11 |
| DOI：10.3969/j.issn.1000-0399.2025.05.008 |
| 中文关键词: 妊娠期糖尿病|不良妊娠|影响因素|决策树模型|回归分析 |
| 英文关键词: Gestational diabetes mellitus|Adverse pregnancy outcome|Influencing factor|Decision tree model|Logistic regression |
| 基金项目:广西卫健委自筹经费科研课题(编号:Z20211554)~~ |
|
| 摘要点击次数: 21 |
| 全文下载次数: 20 |
| 中文摘要: |
| 目的 探究不同预测模型分析妊娠期糖尿病（GDM）产妇发生不良妊娠结局的应用价值。方法 回顾性分析2021年4月到2022年4月在梧州市人民医院确诊GDM的348例产妇的临床资料，观察并记录GDM患者不良妊娠结局，根据GDM产妇是否发生不良妊娠结局分成不良妊娠结局组（n=197）和正常妊娠结局组（n=151）。采用单因素分析比较不良妊娠结局和正常妊娠结局患者的一般资料，logistic回归分析影响GDM患者不良妊娠结局的因素，构建预测模型，Hosmer-Lemesho检验评估模型的拟合度。受试者工作特征（ROC）曲线评估logistic模型预测GDM产妇发生不良妊娠结局的效能。采用χ2自动交互检测法（CHAID）模型进一步筛选危险因素。结果 GDM患者主要不良妊娠结局为胎膜早破（12.07%）、早产（6.90%）、羊水过少（6.32%）、胎儿窘迫（6.03%）。不良妊娠结局组和正常妊娠结局组患者在产妇年龄、孕前身体质量指数（BMI）、合并多囊卵巢综合征、高血压、高脂血症等基础疾病，口服葡萄糖耐量试验（OGTT）异常项、家族糖尿病史和不良孕产史等方面比较，差异有统计学意义（P<0.05）;logistic回归分析结果显示，年龄（OR=1.904,95%CI:1.064～3.408,P=0.031）、孕前BMI（OR=1.879,95%CI:1.034～3.417,P=0.039）、OTGG异常项（OR=1.879,95%CI:1.050～3.364,P=0.034）、不良孕产史（OR=1.816,95%CI:1.023～3.383,P=0.042）、合并高脂血症（OR=1.842,95%CI:1.015～3.343,P=0.045）均为影响GDM产妇发生不良妊娠结局的危险因素（P<0.05），构建风险预测模型，得到P=1/[1+e~((-6.958+0.644×（年龄）+0.631（孕前BMI）+0.631（OTGG异常项）+0.621（不良孕产史）+0.611（合并高脂血症）)]，Hosmer-Lemeshowχ2=7.592,P>0.05。CHAID模型分析显示，年龄≥30岁的产妇不良妊娠结局发生率为71.3%，高于年龄<30岁产妇的44.5%（P<0.05）；在年龄<30岁的产妇中，孕前BMI不是影响产妇不良妊娠结局发生的危险因素（P>0.05），在年龄≥30岁的产妇中，孕前BMI≥24 kg/m2的产妇不良妊娠结局发生率为84.7%，高于孕前BMI<24 kg/m2产妇的63.3%（P<0.05）；在年龄≥30岁、孕前BMI<24 kg/m2的产妇中，确诊孕周≥25周的产妇发生不良妊娠结局的概率为77.4%，高于确诊孕周<25周的53.3%（P<0.05）。ROC曲线显示，logistic回归模型预测GDM产妇不良妊娠结局的曲线下面积为0.898（95%CI:0.862～0.928,P<0.05）。CHAID模型筛选出的影响因素为年龄、孕前BMI和确诊孕周 。结论 年龄是GDM患者发生不良妊娠结局最主要影响因素，决策树模型和logistic回归模型可互补分析影响GDM患者不良妊娠结局的影响因素。 |
| 英文摘要: |
| Objective To explore the application value of different prediction models in analyzing adverse pregnancy outcomes of puerperae with gestational diabetes mellitus （GDM）. Methods A retrospective analysis was performed on the clinical data of 348 puerperae with GDM confirmed in the Wuzhou People’s Hospital between April 2021 and April 2022, and the adverse pregnancy outcomes of patients were observed and recorded.According to pregnancy outcomes, patients were divided into the adverse pregnancy outcome group （n=197） and the normal pregnancy outcome group （n=151）.The differences in general data of the two groups of patients were compared by univariate analysis.Logistic regression was conducted to analyze the factors affecting adverse pregnancy outcomes in GDM patients, and the prediction model was constructed.The fit of model was evaluated by Hosmer-Lemesho test.The efficacy of logistic model was evaluated by receiver operating characteristic （ROC） curves.Chi-square automatic interactive detection （CHAID） model was used to further screen risk factors. Results The main adverse pregnancy outcomes in GDM patients were premature rupture of membranes （12.07%）, premature delivery （6.90%）, oligohydramnios（6.32%） and fetal distress （6.03%）.The maternal age, pre-pregnancy body mass index （BMI）, combined polycystic ovary syndrome, hypertension, hyperlipidemia and other underlying diseases, abnormal items of oral glucose tolerance test （OGTT）, family history of diabetes and history of adverse pregnancy demonstrated statistically significant differences in two groups of patients （P＜0.05）.Logistic regression analysis showed that age （OR=1.904, 95%CI:1.064～3.408, P=0.031）, progestational BMI （OR=1.879, 95%CI;1.034～3.417, P=0.039）, abnormal OGTT items（OR=1.879, 95%CI=:1.050～3.364, P=0.034）, adverse pregnancy history （OR=1.816, 95%CI;1.023～3.383, P=0.042） and hyperlipidemia （OR=1.842, 95%CI:1.015～3.343, P=0.045） were all risks factor for adverse pregnancy outcomes in GDM puerperae.The probability by risk prediction model was as follows: P=1/[1+e（-6.958+0.644×（age）+0.631（progestational BMI）+0.631（abnormal OTGG items）+0.621（adverse pregnancy history）+0.611（hyperlipidemia）], Hosmer-Lemeshow: χ2=7.592, P＞0.05.CHAID model analysis showed that incidence of adverse pregnancy outcomes in puerperae aged≥30 years was significantly higher than that＜30 years （71.3%vs 44.5%, P＜0.05）.Among puerperae aged＜30 years, progestational BMI was not a risk factor for the adverse pregnancy outcomes （P＞0.05）.Among puerperae aged ≥30 years, incidence of adverse pregnancy outcomes in those with puerperae BMI≥24 kg/m~2 was significantly higher than that ＜ 24 kg/m2（84.7%vs 63.3%, P＜0.05）.Among puerperae with age ≥30 years and progestational BMI＜24 kg/m~2, incidence of adverse pregnancy outcomes in those with gestational weeks≥25 weeks was significantly higher than that ＜25 weeks（77.4%vs 53.3%, P＜0.05）.ROC curves analysis showed that the AUC of the logistic regression model for predicting adverse pregnancy outcomes in GDM mothers was 0.898, with a 95%CI:0.862～0.928, and P＜0.05.The influencing factors screened by CHAID model were age, progestational BMI and confirmed gestational week. Conclusion Age is the most important influencing factor of adverse pregnancy outcomes in GDM patients.Decision tree model and Logistic regression model can complementarily analyze the influencing factors of adverse pregnancy outcomes in GDM patients. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
