| 子痫前期孕妇血清尿酸影响因素的预测模型构建与验证 |
| 投稿时间：2025-04-25 修订日期：2025-05-14 点此下载全文 |
| 引用本文：闫坤,陆进,周健红,闫华,田玲,邹青,门敏超.子痫前期孕妇血清尿酸影响因素的预测模型构建与验证[J].医学研究杂志,2025,54(9):98-102, 192 |
| DOI：
10.11969/j.issn.1673-548X.2025.09.017 |
| 摘要点击次数: 36 |
| 全文下载次数: 16 |
|
| 基金项目:安徽省教育厅自然科学研究重点项目(2023AH051929) |
|
| 中文摘要:目的 构建并验证子痫前期(preeclampsia,PE)孕妇血清尿酸(serum uric acid,SUA)影响因素的预测模型,为早期识别和干预PE孕妇提供科学依据。方法 本研究搜集2021年3月1日~2024年6月30日蚌埠市第三人民医院产科收治的171例PE孕妇的住院号、年龄、身高、体重、孕周、尿蛋白、SUA、口服葡萄糖耐量试验(oral glucose tolerance test,OGTT)、谷丙转氨酶(alanine aminotransferase,ALT)、血红蛋白(hemoglobin,HB)、白蛋白(albumin,ALB)、谷草转氨酶(aspartate aminotransferase,AST)、血清肌酐(serum creatinine,SCR)、甘油三酯(triglycerides,TG)、总胆固醇(total cholesterol,TC)、血清尿素(serum urea,SUR)、低密度脂蛋白胆固醇(low density lipoprotein cholesterol,LDL)、高密度脂蛋白胆固醇(high density lipoprotein cholesterol,HDL)以及D-二聚体(D-dimer,D2D)等19个变量。根据SUA的表达高低进行分组,采用最小绝对值收敛和选择算子算法(least absolute shrinkage and selection operator,LASSO)、多因素Logistic以及神经网络模型筛选影响其的变量,并构建列线图预测模型；通过受试者工作特征(receiver operating characteristic,ROC)曲线、决策分析曲线(decision analysis curves,DCA)以及校准曲线验证模型的预测能力及临床价值。结果 通过LASSO、多因素Logistic以及神经网络模型筛选影响PE孕妇SUA的3个变量ALB(OR=0.893,95%CI:0.808~0.986)、AST(OR=4.292,95%CI:2.043~9.016)和SCR(OR=4.2,95%CI:1.495~11.801)；并基于其成功构建SUA的列线图预测模型,且模型的ROC为0.7646,DCA的净获益区间为20%~100%,校准曲线与理想曲线高度吻合。结论ALB、AST和SCR是影响PE孕妇SUA重要变量,可为PE孕妇的早期预测和干预提供科学依据。 |
| 中文关键词:子痫前期 血清尿酸 神经网络 列线图 |
| |
| Construction and Validation of a Predictive Model for Influencing Factors of Serum Uric Acid in Preeclampsia |
|
|
| Abstract:Objective To construct and validate a predictive model for influencing factors of serum uric acid (SUA) in preeclampsia (PE), providing a scientific basis for early identification and intervention. Methods This study collects 19 variables from 171 PE pregnant women admitted to the Department of Obstetrics at Bengbu Third People′s Hospital between March 1,2021 to June 30,2024 including hospitalization number, age, height, weight, urine protein, gestational week, SUA, oral glucose tolerance test (OGTT), alanine aminotransferase (ALT), hemoglobin (HB), albumin (ALB), aspartate aminotransferase (AST), serum creatinine (SCR), triglycerides (TG), total cholesterol (TC), serum urea (SUR), low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), and D-dimer (D2D). According to the expression level of SUA, the important variables affecting SUA were screened by LASSO, multi-factor Logistic and neural network model, and the nomogram prediction model was constructed. The predictive performance and clinical utility of the model were validated using the receiver operating characteristic (ROC) curve, decision curve analysis (DCA), and calibration curve. Results Three variables-ALB(OR=0.893,95%CI:0.808-0.986), AST(OR=4.292,95%CI:2.043-9.016) and SCR(OR=4.2,95%CI:1.495-11.801)-were identified as significant influencing factors of SUA in PE pregnant women through LASSO, multivariate Logistic regression, and neural network models. A nomogram prediction model for SUA was successfully established, achieving an ROC value of 0.7646. The net benefit range of DCA was 20%-100%, and the calibration curve showed high consistency with the ideal curve. Conclusion ALB, AST, and SCR are critical variables influencing SUA in PE pregnant women, offering a scientific basis for early prediction and intervention. |
| keywords:Preeclampsia Serum uric acid Neural network Nomogram |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
