妊娠合并自身免疫性疾病不良妊娠结局预测模型的构建及验证

Construction and Verification of Adverse Pregnancy Outcome Prediction Model for Pregnancy Complicated with Autoimmune Diseases

目的:构建及验证妊娠合并自身免疫性疾病患者发生不良妊娠结局的个体化预测模型。方法:回顾性分析2016年01月至2022年12月入住青岛大学附属医院产科且入院诊断为妊娠合并明确自身免疫性疾病的患者的资料。对患者的临床资料进行应用单因素分析、多因素Logistic回归分析、LASSO筛选患者发生不良妊娠结局的独立危险因素,建立相应的nomogram预测模型。采用测试集评估预测效能,利用ROC曲线、Bootstrop自抽样法对模型进行验证,进一步评估模型的区分能力和校准性。结果:1) 经过单因素、多因素及LASSO回归分析,筛选出预测妊娠合并自身免疫性疾病发生不良妊娠结局的最优预测变量:高血压、其它并发症数量、血红蛋白、N/L比值、D-二聚体。2) 应用上述最优预测变量建立妊娠合并自身免疫性疾病不良妊娠结局的预测模型的性能较好,灵敏度为74.3%,特异度为82.5%,AUC为0.843 (95% CI: 0.789~0.898),验证集模型的灵敏度为51.3%,特异度为93.1%,AUC为0.75 (95% CI: 0.656~0.845)。3) 通过Bootstrop重复自抽样法获得校准曲线,模拟曲线与实际曲线具有较强一致性。结论:本研究构建预测妊娠合并自身免疫性疾病不良妊娠结局的预测模型,可以作为预测妊娠合并自身免疫性疾病结局的较为简单的工具。

Objective: To construct and verify an individualized prediction model for adverse pregnancy outcomes in pregnant patients with autoimmune diseases. Methods: The data of patients admitted to the Department of Obstetrics, Affiliated Hospital of Qingdao University from January 2016 to December 2022 who were diagnosed with pregnancy combined with definite autoimmune diseases were retrospectively analyzed. Using univariate analysis, multivariate Logistic regression analysis and LASSO to screen independent risk factors for adverse pregnancy outcomes, a nomogram prediction model was established. Test set was used to evaluate the prediction efficiency, ROC curve and Bootstrop self-sampling method were used to verify the model, and the distinguishing ability and calibration of the model were further evaluated. Results: 1) Through univariate, multivariate and LASSO regression analysis, the optimal predictors of adverse pregnancy outcomes with autoimmune diseases were selected: hypertension, number of other complications, hemoglobin, N/L ratio, D-dimer. 2) The application of the above optimal predictors to establish a predictive model for adverse pregnancy outcomes of pregnancy combined with autoimmune diseases has a good performance, with a sensitivity of 74.3%, specificity of 82.5%, and AUC of 0.843 (95% CI: The sensitivity and specificity of the validation set model were 51.3%, 93.1%, and AUC was 0.75 (95% CI: 0.656~0.845). 3) The calibration curve was obtained by Bootstrop repeated self-sampling method, and the simulated curve had a strong consistency with the actual curve. Conclusion: In this study, a prediction model for adverse pregnancy outcomes of pregnancy with autoimmune diseases was established, which can be used as a relatively simple tool to predict the outcome of pregnancy with autoimmune diseases.