摘要目的 探讨人体测量学指标与老年人群高血压发病的关联,比较新的人体测量学指标与传统指标对高血压发病风险的预测效力,建立老年人群高血压发病预测模型.方法 于2008年招募27 009名东风汽车公司退休职工进行基线调查,2013年进行了第1次随访,排除<60岁,以及患有高血压、冠心病、卒中和恶性肿瘤者,最后将资料完整的6 478名随访对象纳入本研究.采用多因素logistic回归模型分析传统人体测量学指标[体重指数(BMI)、腰围和腰高比],以及新的人体测量学指标[内脏脂肪指数(VAI)、身体形态指数(ABSI)和身体圆润指数(BRI)]与高血压的关联.采用多因素logistic回归模型构建高血压风险预测模型,采用受试者工作特征(ROC)曲线比较不同模型的预测效力.结果 共检出新发高血压患者1 787例,发病率为27.59%.调整了混杂因素之后,BMI、腰围、腰高比、VAI和BRI水平升高会增加老年人群高血压发生风险(P值均<0.05);与其最低三分位数相比,最高分位数人体测量学指标对应的高血压发病风险OR(95% CI)值在男性中分别为:VAI:1.45(1.22~1.73),BRI:1.86(1.55~2.23);在女性中分别为:VAI:1.55(1.22~1.96),BRI:1.60(1.27~2.01);ABSI与高血压发病的关联在男性[OR(95%CI):1.07(0.90~1.28)]和女性[OR(95%CI):1.03(0.82~1.29)]中均无统计学意义(P值均>0.05).高血压风险预测基础模型包括年龄、饮酒(仅男性)、教育水平(仅男性)、收缩压、舒张压和空腹血糖水平.与基础模型相比,分别纳入BMI[男性:ROC曲线下面积(AUC)=0.697]和BRI(男性:AUC=0.696)的扩展模型预测效力略有提高(P<0.05);纳入BMI后的模型对男性(AUC=0.697)和女性(AUC=0.685)高血压发病风险的预测效力最好.结论 在老年人群中, BMI、腰围、腰高比、VAI和BRI水平增加会提高高血压发病风险,纳入BMI模型预测高血压发病风险效力最好.

abstractsObjective To explore the relationship between anthropometric indices and the incident risk of hypertension, compare novel anthropometric indices with traditional indices in hypertension prediction and establish hypertension prediction models among elderly Chinese. Methods A total of 27 009 retirees from the Dongfeng Motor Corporation were recruited at baseline in 2008 and the first follow?up was conducted in 2013. After the exclusion of participants less than 60 years old, participants with hypertension, coronary heart disease, stroke, cancer, and those with missing data, a total of 6 784 elderly participants were enrolled in this study. A multivariate logistic regression model was used to analyze the relationship between traditional anthropometric indices, body mass index (BMI), waist circumference (WC), waist?to?height ratio (WHtR), novel anthropometric indices, visceral adiposity index (VAI), body shape index (ABSI), body roundness index (BRI), and the incident risk of hypertension. Meanwhile, a multivariate logistic regression model was used to establish hypertension prediction models. Receiver operating characteristic (ROC) curve was applied to compare the prediction ability of different models. Results A total of 1 787 incident cases of hypertension were identified, with the incidence of hypertension about 27.59%. Significant positive associations were detected between BMI, WC, WHtR, VAI, BRI and the incident risk of hypertension after adjusting for potential confounders (all P values<0.05). In men, the OR (95%CI ) (upper tertile vs lower tertile) was 1.45 (1.22-1.73) for VAI, and 1.86 (1.55-2.23) for BRI. In women, the OR (95% CI) (upper tertile vs lower tertile) was 1.55 (1.22-1.96) for VAI, and 1.60 (1.27-2.01) for BRI. For ABSI, no significant association was observed in either men (OR (95%CI): 1.07 (0.90-1.28)) or women (OR (95%CI): 1.03 (0.82-1.29), both P values >0.05). The basic hypertension prediction model included age, drinking (only in men), education status (only in men), systolic blood pressure (SBP), diastolic blood pressure (DBP), and fasting blood glucose levels. Based on the basic prediction model, BMI (in men: AUC=0.697) and BRI (only in men: AUC=0.696) improved area under ROC curve (AUC) significantly (P<0.05). BMI was the strongest predictor in both men (AUC=0.697) and women (AUC=0.685) in the extended model. Conclusion Significant positive associations were detected between BMI, WC, WHtR, VAI, BRI and the incident risk of hypertension among elderly Chinese. BMI was the strongest predictor in hypertension prediction model compared with other anthropometric indices.