原发性高血压患者发生尿微量白蛋白列线图的构建和评价

目的:研究未治疗的原发性高血压患者发生尿微量白蛋白的危险因素,构建列线图预测模型并验证其有效性。方法:本研究连续纳入2018年1月—2019年12月在我院门诊和病房未治疗的原发性高血压共456例患者。随机分为建模组(318例)与验证组(138例)。以上述指标构建列线图模型,在建模组中,应用LASSO回归和非条件二分类Logistic回归分析确定其独立危险因素,然后采用易侕软件构建列线图预测模型;通过验证组数据对所构建的列线图模型进行验证。应用ROC曲线、Hosmer-Lemeshow拟合优度检验和校准曲线对列线图模型的区分度和校准度进行评价,最后采用临床决策曲线评估模型的临床实用性。结果:多因素Logistic回归模型显示尿酸、空腹血糖、静息心率和心血管风险水平分层是未治疗的原发性高血压患者发生尿微量白蛋白的独立预测因子。在建模组ROC曲线AUC值为0.813,95%CI:0.768~0.859,对发生尿微量白蛋白具有较好的预测能力。验证组ROC曲线AUC值为0.800,95%CI:0.719~0.871,均提示该模型具有较好的区分度和判别能力。校正曲线和Hosmer-Lemeshow拟合优度检验(P=0.688)均提示在预测值和实际值之间具备较好的一致性。结论:尿酸、空腹血糖、静息心率和心血管风险水平分层是未治疗的原发性高血压患者发生尿微量白蛋白的独立预测因子。构建列线图模型可较为直观、准确地预测未治疗的原发性高血压患者发生尿微量白蛋白的风险。

Construction and evaluation of urinary microalbumin nomogram in essential hypertension

Objective:To study the risk factors of urinary microalbumin in untreated patients with essential hypertension and develop and validate a nomogram model.Methods:A total of 456 patients with untreated essential hypertension in outpatient and ward of our hospital between 2018 and 2019 were recruited.They were randomly divided into two groups,the development group(n=318)and the validation group(n=138).the independent risk factors were determined by LASSO regression and non-conditional binary logistic regression analysis in the development group.A nomogram to predict the probability of microalbumin was constructed with the Empower Statsversion 2.17.8.The model is validated by validation group data.AUC value and calibration curve were employed to evaluate the performance of the nomogram.Finally,a clinical decision curve was used to evaluate the clinical practicality of the nomogram.Results:Using logistic regression analysis,we found that uric acid,fasting glucose,office heart rate,and cardiovascular risk levels were independent predictors of microalbumin.The nomogram was then developed using these four variables.The nomogram showed an AUC of 0.813(95%CI:0.768-0.859).In addition,in the validation group,the AUC was 0.800(95%CI:0.719-0.871).These data showed that the nomogram had a good discrimination ability.The calibration plot indicated that the nomogram-predicted probabilities compared very well with the actual probabilities(Hosmer-Lemeshow test,P=0.688).Conclusion:Uric acid,fasting glucose,office heart rate,and cardiovascular risk levels are independent predictors of microalbumin in untreated essential hypertension.The nomogram is a predictive tool for the risk of microalbumin.