基于LASSO回归建立妊娠高血压疾病发生的ADAM10、PLGF、RDW-CV相关风险列线图模型

Established of the risk nomogram model of ADAM10,PLGF and RDW-CV related to hypertensive disorder complicating pregnancy based on LASSO regression

目的建立基于LASSO回归的妊娠高血压疾病发生的解整合素样金属蛋白酶10(ADAM10)、胎盘生长因子(PLGF)、红细胞分布宽度变异系数(RDW-CV)相关风险预测模型。方法采取前瞻性研究,选择2022年3月至2023年3月于南充市中心医院接受产检的55例合并妊娠高血压(GH)产妇作为研究对象,另选取同期55例未发生GH产妇作为研究对象。产妇均接受血清ADAM10、PLGF及RDW-CV水平检测,统计并比较两组基线资料;采用LASSO回归筛选可能影响产妇GH发生的因素并进行二元Logistic回归分析,根据回归分析结果构建列线图模型并进行内部验证;采用受试者操作特征(ROC)曲线分析列线图模型预测产妇GH发生的价值。结果GH组体质量指数≥24kg/m^(2)、有高血压家族史患者占比高于非GH组,血清ADAM10及RDW-CV高于非GH组,血清PLGF低于非GH组(P<0.05);LASSO回归模型,最终筛选出5个潜在的相关因素,分别为体质量指数、高血压家族史及ADAM10、PLGF、RDW-CV水平;行二元Logistic回归分析结果显示,有高血压家族史、血清ADAM10及RDWCV高表达是产妇发生GH的危险因素(OR>1,P<0.05);血清PLGF高表达是产妇发生GH的保护因素(OR<1,P<0.05);使用Bootstrap内部验证法对列线图模型,C-index值0.971,表明模型具有良好的区分度,绘制受试者操作特征(ROC)曲线对列线图模型进行内部验证,结果显示,列线图模型预测孕妇发生GH的AUC 0.971(P<0.05)有一定预测价值。结论高血压家族及血清ADAM10、PLGF、RDW-CV水平与产妇GH的发生关系密切,以此构建模型对于产妇GH的发生预测性能较好。

Objective To establish a risk prediction model of disintegrin-like metalloproteinase 10(ADAM10),pla-cental growth factor(PLGF)and red blood cell distribution width coefficient of variation(RDW-CV)in hypertensive disorder complicating pregnancy based on LASSO regression.Methods A prospective study was conducted to select 55 pregnant women with GH who underwent prenatal examination in Sichuan Nanchong Central Hospital from March 2022 to March 2023 as the study subjects,and 55 pregnant women without GH during the same period were selected as the study subjects.Women who did not experience GH during the same period as the study subjects.All postpartum women underwent serum ADAM10,PLGF,and RDW-CV levels testing,and baseline data were collected and compared between the two groups.LASSO regres-sion was used to screen the factors that may affect the occurrence of maternal GH and binary logistic regression analysis was performed.According to the results of regression analysis,a nomogram model was constructed and verified internally.Re-ceiver operator characteristic(ROC)curve was used to analyze the value of nomogram model in predicting the occurrence of maternal GH.Results The proportion of patients with body mass index≥24 kg/m^(2) and family history of hypertension in GH group was higher than that in non-GH group,serum ADAM10 and RDW-CV were higher than those in non-GH group,and serum PLGF was lower than that in non-GH group(P<0.05).LASSO regression model was used to screen out 5 potential related factors,including body mass index,family history of hypertension and levels of ADAM10,PLGF and RDW-CV.The results of binary logis-tic regression analysis showed that having a family history of hypertension,high expression of ADAM10 and RDWCV in serum were the risk factors for GH in pregnant women(OR>1,P<0.05).High expression of serum PLGF is a protective factor for postpar-tum GH(OR<1,P<0.05).The Bootstrap internal validation method was used to verify the nomogram model.The C-index value was 0.971,indicating that the model had good discrimination.The ROC curve was drawn to verify the nomogram model internally.The results showed that the nomogram model predicted GH in pregnant women.AUC 0.971(P<0.05)has certain predictive value.Conclusion Hypertension family and serum ADAM10,PLGF,RDW-CV levels are closely related to the occurrence of maternal GH,and the model constructed based on this has a good predictive performance for the occurrence of maternal GH.