红细胞参数对高血压的影响及高血压发病风险预测模型研究

发布时间：2018-05-18 07:39

本文选题：高血压 + 红细胞计数　；参考：《山东大学》2017年硕士论文

【摘要】：高血压是心脑血管疾病最重要和最常见的危险因素,中国高血压的发病和患病形势严峻。识别高血压危险因素,构建高血压发病风险预测模型,评估高血压发病风险,发现高危人群,对高危人群进行干预可延缓甚至阻止高血压的发生。目前多个国家和地区建立了高血压发病风险预测模型,但以往的高血压发病风险预测模型通常采用传统的预测参数(年龄、收缩压、舒张压、体质指数、吸烟、饮酒和高血压家族史),缺乏新的预测因子,使得模型预测能力受限。近些年来多项研究发现红细胞参数(红细胞计数、血红蛋白含量、血细胞比容)可能是高血压的预测因子,有望助于提高模型预测能力。为此,本文基于队列探讨红细胞参数对高血压的影响,确定其是否可以作为高血压预测因子,如果可以,则在考虑红细胞参数的基础上构建高血压发病风险预测模型。资料与方法:本研究基于"山东多中心纵向健康管理队列",采用队列中2005年至2010年期间参加健康查体的体检者构建队列,体检三次及以上,排除首次体检时有高血压、心血管疾病、脑卒中、年龄小于18岁的体检者,最终有12497人(男7537人、女4960人)进入队列。描述性分析的基础上,控制其他影响因素,分性别采用Cox比例回归分析方法研究红细胞参数(红细胞计数、血红蛋白含量、血细胞比容)对高血压的影响;分性别纳入红细胞参数的基础上,构建高血压Cox风险回归预测模型,并用ROC曲线下面积AUC及O/E进行评价。结果:1.该健康管理队列12497人共随访了 38958人年,其中有2785人(男2021、女764人)发生高血压,高血压的发病密度为71.48/1000人年。2.将红细胞参数按照四分位数分为四类(Q1,Q2,Q3,Q4),则红细胞参数与其他基线变量之间的关系如下:无论男女,多数基线变量随红细胞参数的增大而增高,但有统计学意义的基线变量在不同红细胞参数中略有不同。Cochran-Armitage趋势性检验显示,对于男性,仅有血细胞比容与高血压发生率间存在趋势性(Z=-3.1628,P0.0001);而女性,三个红细胞参数均与高血压发生率间存在趋势性(红细胞计数,Z=-4.2950,P0.0001;血红蛋白含量,Z=-5.8120,P0.0001;血细胞比容,Z=-6.5504,P0.0001)。3.红细胞参数与高血压发生风险的Cox比例回归分析:对于男性,仅调整年龄时三个红细胞参数Q4的相对危险度(RR,以Q1为参照组)、红细胞参数四分类模型趋势性检验及红细胞参数每增加1个标准差的RR值有统计学意义,调整更多协变量时无统计学意义。对于女性,模型调整不同协变量时,Q3和Q4的RR值、红细胞参数四分类模型趋势性检验及红细胞参数每增加1个标准差的RR值均有统计学意义(P0.05);调整年龄、吸烟、饮酒、规律锻炼、体质指数、收缩压、空腹血糖、高密度脂蛋白后,红细胞计数Q2、Q3、Q4的RR值分别是1.140、1.285、1.240,血红蛋白Q2、Q3、Q4的RR值分别是1.069、1.309、1.311,血细胞比容Q2、Q3、Q4的 R值分别是 1.019、1.263、1.234。4.多因素Cox比例回归分析构建高血压发病风险预测模型:采用后退法进行变量筛选,经多因素Cox比例回归分析构建分性别的高血压发病风险预测模型,纳入男性模型的有年龄、体质指数、收缩压、舒张压、γ-谷氨酰转移酶、空腹血糖、饮酒、年龄与体质指数的交互项及年龄与舒张压的交互项。纳入女性模型的有年龄、体质指数、收缩压、舒张压、空腹血糖、血细胞比容、饮酒和吸烟。5.男性高血压发病风险预测模型的ROC曲线下面积AUC(95%CI)为0.761(0.752,0.771),十折交叉验证后 AUC(95%CI)为 0.760(0.751,0.770),O/E为0.9561。女性高血压发病风险预测模型的AUC(95%CI)为0.750(0.738,0.762),十折交叉验证后 AUC(95%CI)为 0.747(0.735,0.759),O/E 为 0.9707。结论:1.红细胞计数、血红蛋白含量、血细胞比容升高将增加高血压发病的风险,这种关联在女性尤为明显。2.血细胞比容最终纳入女性高血压发病风险预测模型,血细胞比容是女性高血压发生的预测因子。3.分性别构建的高血压发病风险预测模型判别能力和校准能力效果良好,可用于评估高血压的发病风险。
[Abstract]:Hypertension is the most important and most common risk factor for cardiovascular and cerebrovascular diseases. The incidence and incidence of hypertension in China are severe. Identifying the risk factors of hypertension, constructing the prediction model of hypertension risk, assessing the risk of hypertension, finding high-risk groups and intervening in high-risk groups can delay or even prevent the occurrence of hypertension. Many countries and regions have established a predictive model for the risk of hypertension, but the previous prediction models of hypertension risk usually adopt traditional predictive parameters (age, systolic pressure, diastolic pressure, body mass index, smoking, drinking and family history of hypertension), lack of new pretest factors and limited prediction ability. In recent years, many studies have been made. It is found that red blood cell parameters (red blood cell count, hemoglobin content, hematocyte specific volume) may be a predictor of hypertension and may help improve model prediction. Therefore, this paper is based on a cohort study to determine the effect of red cell parameters on hypertension and determine whether it can be used as a predictor of hypertension. If possible, the red blood cell is considered. Based on the parameters, a model for predicting the risk of hypertension was constructed. Data and methods: Based on the "Shandong multi center longitudinal health management queue", a cohort of health checkup participants from 2005 to 2010 in the cohort was constructed and examined for three times and above, excluding hypertension, cardiovascular disease, stroke, and the first physical examination. At the age of 18 years of age, 12497 people (7537 men and 4960 women) entered the cohort. On the basis of descriptive analysis, other factors were controlled and the Cox proportional regression analysis was used to study the effects of red cell parameters (red blood cell count, hemoglobin content, blood cell specific volume) on hypertension. On the basis of the number, the Cox risk regression model of hypertension was constructed, and the area AUC and O/E under the ROC curve were evaluated. Results: 1. the 12497 people of the health management queue were followed up for 38958 years, of which 2785 people (2021 men and 764 women) had hypertension, and the density of hypertension was 71.48/1000 person year.2. and the red blood cell parameters were according to four points. The number is divided into four categories (Q1, Q2, Q3, Q4), and the relationship between red blood cell parameters and other baseline variables is as follows: the majority of baseline variables increase with the increase of red cell parameters in both men and women, but a statistically significant baseline variable has a slightly different.Cochran-Armitage trend test in different red cell parameters. For men, only blood is thin. There was a tendency (Z=-3.1628, P0.0001) between the cell specific volume and the incidence of hypertension, while in women, the three red blood cell parameters were all with the incidence of hypertension (red blood cell count, Z=-4.2950, P0.0001; hemoglobin content, Z=-5.8120, P0.0001; blood cell specific volume, Z=-6.5504, P0.0001) the Cox ratio of the.3. red blood cell parameters to the risk of hypertension Regression analysis: for men, the relative risk degree of three red blood cell parameters (RR, Q1 as reference group) was adjusted only for age (RR, Q1 as reference group). The trend test of red cell parameter four classification model and 1 standard deviation of erythrocyte parameters were statistically significant. There was no statistical significance in adjusting more covariant quantity. For women, the model adjustment was different. When covariate, the RR value of Q3 and Q4, the trend test of the red cell parameter four classification model and the RR value of the red blood cell parameters every 1 standard deviations were statistically significant (P0.05); the adjustment of age, smoking, drinking, regular exercise, body mass index, systolic blood pressure, fasting blood glucose, high density lipoprotein, Q2, Q3, Q4 were 1.140,1.28, RR value of Q4, respectively 1.140,1.28. 5,1.240, the RR values of hemoglobin Q2, Q3, and Q4 were 1.069,1.309,1.311, the R values of blood cell specific volume Q2, Q3, and Q4 were 1.019,1.263,1.234.4. multifactor Cox proportional regression analysis to predict the risk of hypertension. Risk prediction model, including age, body mass index, systolic pressure, diastolic pressure, gamma glutamyl transferase, fasting blood glucose, drinking, interaction between age and body mass index, age and diastolic pressure, including age, body mass index, systolic blood pressure, diastolic pressure, fasting blood glucose, blood cell specific volume, drinking and smoking.5 The area AUC (95%CI) under the ROC curve of the risk prediction model for male hypertension was 0.761 (0.752,0.771), AUC (95%CI) was 0.760 (0.751,0.770) after ten fold cross validation, O/E was the AUC (95%CI) of the 0.9561. female hypertension risk prediction model (0.738,0.762), 0.747 after ten fold cross validation, 0, 0. .9707. conclusion: 1. red cell count, hemoglobin content and increased blood cell specific volume will increase the risk of hypertension. This association is particularly evident in women's.2. blood cell specific volume eventually incorporated into the prediction model of the risk of hypertension in women. Blood cell specific volume is a predictor of female hypertension.3., a gender based hypertension. The risk prediction model is effective in discriminating ability and calibrating ability, and can be used to assess the risk of hypertension.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R544.1

【相似文献】