【摘要】 目的:通过常规孕期保健中孕12～19⁺⁶周时血常规及肝肾功能检测结果构建妊娠期高血压疾病（HDP）的早期预测模型。方法:纳入2017年1月至2019年12月在合肥市妇幼保健院行孕期保健并住院分娩孕妇858例，其中孕晚期发生妊娠期高血压（GH）或子痫前期（PE）70例为HDP组，其余788例为非HDP组。对两组血常规和肝肾功能进行单因素比较，采用多因素logistic回归分析HDP的危险因素并建立预测模型，应用R语言软件中回归建模策略程序包建立诊断的列线图模型，并从多维度评价模型效能。结果:平均血红蛋白浓度、血小板、平均血小板体积、谷草转氨酶、谷氨酰胺基转移酶、尿酸及葡萄糖是HDP的独立危险因素，平均血红蛋白含量是HDP的独立保护因素。基于logistic回归建立HDP的早期预测模型，最终公式：Log（Pi）=-21.353+0.045×平均血红蛋白浓度（g/L）+0.009×血小板（10⁹/L）-0.270×平均血红蛋白含量（pg）+0.353×平均血小板体积（fL）+0.011×谷草转氨酶（IU/L）+0.023×谷氨酰胺基转移酶（IU/L）+0.007×尿酸（mol/L）+0.096×葡萄糖（mmol/dL）。该模型的受试者工作特征（ROC）曲线下面积（AUC）为0.762（95%CI为0.703～0.820）,最佳阈值为0.096,此时特异度为0.793,敏感度为0.571。Bootstrap重抽样1000次进行内部验证后，AUC为0.741,布里尔分数为0.068,校准曲线的预测概率与实际概率相近。结论:孕12～19⁺⁶周时血常规和肝肾功能的检测结果中多个指标是HDP的风险因素，本研究构建的HDP早期列线图预测模型具有良好的预测效能和临床实用价值。更多还原

【Abstract】 Objective:To construct an early prediction model for hypertensive disorders in pregnancy（HDP） based on the results of blood routine and liver and kidney functions at 12～19⁺⁶ weeks of pregnancy.Method:858 pregnant women who underwent prenatal care and delivered at Hefei Maternal and Child Health Hospital from January 2017 to December 2019 were included in the study.Among them, 70 cases of gestational hypertension（GH） or preeclampsia（PE） were included in the HDP group, while the remaining 788 cases were included in the non-HDP group.Single factor comparison was conducted between two groups of blood routine and liver and kidney functions.Multivariate logistic regression was used to analyze the risk factors of HDP and establish a predictive model.The regression modeling strategy package in R language software was used to establish a nomogram, and the model effectiveness was evaluated from multiple dimensions.Result:MCHC,PLT,MPV,AST,GGT,UA and Glu were independent risk factors for HDP,MHC was an independent protective factor for HDP.Based on the risk factors screened by logistic regression, a predictive model for HDP was established, and the final formula was: Log（Pi）=-21.353+0.045×MCHC（g/L）+0.009×PLT（10⁹/L）-0.270×MHC（pg）+0.353×MPV（fL）+0.011×AST（IU/L）+0.023×GGT（IU/L）+0.007×UA（μmol/L）+0.096×Glu（mmol/dL）.The area under the receiver operating characteristic（ROC） curve（AUC） of this model was 0.762（95%CI:0.703～0.820）,with an optimal threshold of 0.096.At this point, the specificity was 0.793,and the sensitivity was 0.571.After internal validation using Bootstrap sampling, the AUC was 0.741,the Brier score was 0.068,and the predicted probability of the calibration curve was similar to the actual probability.Conclusion:Multiple indexes in the blood routine and liver and kidney functions at 12～19⁺⁶ weeks of pregnancy are risk factors for HDP.The early nomogram prediction model for HDP constructed in this study has good predictive performance and clinical practical value.更多还原