健康体检人群发生正常高值血压的风险预测列线图模型构建及验证

Construction and Verification of Risk Prediction Nomogram Model for High-Normal Blood Pressure in Health Exmaination Population

目的构建并验证健康体检人群发生正常高值血压的风险预测列线图模型。方法采用便利抽样法,前瞻性选择2022年12月—2023年5月在广西中医药大学第一附属医院进行健康体检的人员为调查对象。采用风险因素调查表收集受试者的一般资料、生活方式及实验室检查指标,采用《中医体质分类与判定表》、健康促进生活方式量表Ⅱ(HPLP-Ⅱ)对受试者进行调查。建模集健康体检者发生正常高值血压的影响因素分析采用多因素Logistic回归分析,并根据多因素Logistic回归分析结果构建健康体检人群发生正常高值血压的风险预测列线图模型;采用ROC曲线、校准曲线分别评估列线图模型的区分度、准确性。结果本研究共发放调查问卷376份,回收有效问卷340份,有效回收率为90.43%。340例健康体检者正常高值血压发生率为40.59%(138/340)。按照7∶3的比例将健康体检者随机分为建模集(n=239)和验证集(n=101)。其中建模集中正常血压者142例,正常高值血压者97例。建模集正常血压者与正常高值血压者性别、年龄、BMI、腰围、吸烟者占比、TG、HDL-C、尿酸、尿素氮、γ-谷氨酰转移酶(GGT)、同型半胱氨酸(Hcy)、中医体质、HPLP-Ⅱ评分比较,差异有统计学意义(P<0.05)。多因素Logistic回归分析结果显示,年龄、尿酸、Hcy、湿热质、阴虚质是建模集健康体检者发生正常高值血压的独立影响因素(P<0.05)。基于多因素Logistic回归分析结果构建健康体检人群发生正常高值血压的风险预测列线图模型。ROC曲线分析结果显示,列线图模型预测建模集、验证集健康体检者发生正常高值血压的AUC分别为0.936〔95%CI(0.907~0.964)〕、0.940〔95%CI(0.898~0.983)〕。校准曲线分析结果显示,在建模集、验证集中,列线图模型的准确性均较高(Brier评分分别为0.095、0.071)。结论年龄、尿酸、Hcy、湿热质、阴虚质是健康体检人群发生正常高值血压的独立影响因素,而基于上述影响因素构建的健康体检人群发生正常高值血压的风险预测列线图模型具有良好的区分度和准确性。

Objective To construct and verify the risk prediction nomogram model for high-normal blood pressure in health exmaination population.Methods People undergoing health examination in the First Affiliated Hospital of Guangxi University of Chinese Medicine from December 2022 to May 2023 were used to prospectively selected as the survey subjects by convenience sampling method.The Risk Factor Questionnaire was used to collect the general information,lifestyle,and laboratory examination indicators of the subjects.The subjects were investigated with the TCM Constitution Classification and Judgment Table and the Health-Promoting Lifestyle ProfileⅡ(HPLP-Ⅱ).Multivariate Logistic regression analysis was used to analyze the influencing factors of high-normal blood pressure in the health examination population in the modeling set,and the risk prediction nomogram model for high-normal blood pressure in health examination population was constructed according to its results.The ROC curve was used to analyze the discrimination of the nomogram model,and the calibration curve was used to evaluate the accuracy of the nomogram model.Results A total of 376 questionnaires were distributed in this study,and 340 valid questionnaires were recovered,with an effective recovery rate of 90.43%.The incidence of high-normal blood pressure of 340 health examination population was 40.59%(138/340).The health exmaination population were randomly divided into the modeling set(n=239)and the validation set(n=101)in a 7∶3 ratio.There were 142 cases of normal blood pressure and 97 cases of high-normal blood pressure in the modeling set.There were significant differences in gender,age,BMI,waist circumference,proportion of smokers,TG,HDL-C,uric acid,urea nitrogen,γ-glutamyl transferase(GGT),homocysteine(Hcy),TCM constitution,and HPLP-Ⅱscore between subjects with normal blood pressure and subjects with high-normal blood pressure in the modeling set(P<0.05).Multivariate Logistic regression analysis showed that age,uric acid,Hcy,damp-heat constitution,and Yin deficiency constitution were independent influencing factors of high-normal blood pressure in health examination population in the modeling set(P<0.05).The risk prediction nomogram model for high-normal blood pressure in health exmaination population was constructed based on the results of multivariate Logistic regression analysis.ROC curve analysis showed that the AUC of the nomogram model in predicting high-normal blood pressure in health exmaination population in the modeling set and the validation set was 0.936[95%CI(0.907-0.964)]and 0.940[95%CI(0.898-0.983)],respectively.The calibration curve analysis showed that the accuracy of the nomogram model was high in the modeling set and the validation set(Brier scores were 0.095 and 0.071,respectively).Conclusion Age,uric acid,Hcy,damp-heat constitution,and Yin deficiency constitution are independent influencing factors of high-normal blood pressure in health examination population.The risk prediction nomogram model for high-normal blood pressure in health examination population constructed based on the above influencing factors has good discrimination and accuracy.