空腹血糖受损预测模型与风险评分的建立  

Establishment of prediction model and risk score of impaired fasting glucose

作　　者：刘晓磊 刘相佟[1,2] 陶丽新 刘佳[3] 陈硕 康晓平[5] 张静波[4] 郭秀花 LIU Xiao-lei;LIU Xiang-tong;TAO Li-xin;LIU Jia;CHEN Shuo;KANG Xiao-ping;ZHANG Jing-bo;GUO Xiu-hua(School of Public Health,Capital Medical University,Beijing 100069,China;Beijing Key Laboratory of Clinical Epidemiology,Beijing 100069,China;Capital Medical University Yangjing Medical College,Beijing 101321,China;Beijing Physical Examination Center,Beijing 100077,China;Beijing Xiaotangshan Hospital,Beijing 10221l,China)

机构地区：[1]首都医科大学公共卫生学院,北京100069 [2]临床流行病学北京市重点实验室,北京100069 [3]首都医科大学燕京学院,北京101321 [4]北京市体检中心,北京100077 [5]北京市小汤山医院,北京102211

出　　处：《实用预防医学》2022年第9期1039-1044,共6页Practical Preventive Medicine

基　　金：国家自然科学基金(81872708)。

摘　　要：目的构建北京市成年人10年内发生空腹血糖受损的预测模型,并建立风险评分可视化列线图。方法本研究利用2009—2018年北京市健康管理队列研究的数据。以2009—2013年第一次数据为基线数据,采用Cox比例风险回归来构建空腹血糖受损发生的预测模型,建立空腹血糖受损发生的可视化列线图,通过C指数来对列线图进行评估。结果共纳入18-89岁居民5574例,其中训练集4459例,平均随访(6.78±1.78)年,共有544例(12.20%)发生空腹血糖受损;验证集1115例,平均随访(6.83±1.82)年,共128例(11.48%)发生空腹血糖受损。多因素Cox回归分析结果显示高龄、体质指数大、腰臀比大、高血压、甘油三酯异常、谷草转氨酶异常、有父母糖尿病史的人群更容易发生空腹血糖受损。验证结果:训练集C指数=0.74(95%CI:0.72-0.76),验证集C指数=0.73(95%CI:0.68-0.77)。结论本研究所构建的10年内成年人空腹血糖受损发生的列线图效能良好,对可能进展为空腹血糖受损的高风险人群有预测价值。Objective To construct a prediction model of fasting glucose impairment within 10 years in adults in Beijing,and to establish a visual nomogram of risk score.Methods This study used the data about Beijing health management cohort study from 2009 to 2018.Taking the first data from 2009 to 2013 as the baseline data,the Cox proportional hazards regression was used to construct the prediction model of fasting glucose impairment.We established the visual nomogram of fasting glucose impairment,and evaluated the nomogram through the C index.Results A total of 5,574 residents aged 18-89 years were included in this study,including 4,459 in the training set.The average follow-up was(6.78±1.78)years.A total of 544(12.20%)cases had impaired fasting blood glucose.In the validation set,1,115 cases were followed up for an average of(6.83±1.82)years.A total of 128(11.48%)cases had impaired fasting blood glucose.Multivariate Cox regression analysis showed that people with advanced age,high body mass index,high waist-to-hip ratio,hypertension,triglyceride abnormality,aspartate aminotransferase abnormality,and parental diabetes history were more likely to suffer impaired fasting blood glucose.The validation results revealed that the C index of the training set was 0.74(95%CI:0.72-0.76),and the C index of the validation set 0.73(95%CI:0.68-0.77).Conclusion The nomogram of adult fasting glucose impairment within 10 years constructed by this study is effective and has predictive value for high-risk persons who may progress to fasting glucose impairment.

关 键 词：空腹血糖受损 预测模型 列线图 纵向研究 

分 类 号：R587.1[医药卫生—内分泌]