Establishment and evaluation of cardio-cerebrovascular disease prediction model in Henan
-
摘要:
目的 构建河南省心脑血管疾病发病预测模型,为河南省心脑血管疾病的防控提供科学依据。 方法 采用队列研究设计,选取2013—2014年参加中国慢性病及危险因素监测的18岁以上河南省常住居民,将基线无心脑血管疾病、无失访、无关键变量缺失的5 757人作为随访队列。 采用加权的Cox回归分析构建模型,以C统计量和校准曲线评估模型的区分度和校准度,并采用Bootstrap法进行内部验证。 结果 平均随访时间7.01±1.02年,心脑血管疾病累积发病率为9.23%。 纳入预测模型的变量有年龄[风险率(HR)= 1.05,95%置信区间(CI):1.04~1.06,P<0.001]、收缩压(HR =1.01,95%CI: 1.01~1.02,P=0.001)、吸烟(HR= 1.44,95%CI:1.19~1.74,P=0.002)、高血压服药(HR= 1.81,95%CI:1.32~2.50,P=0.002)和糖尿病(HR= 1.49,95%CI:1.11~2.00,P=0.014)。 模型的C统计量为0.76(0.74~0.78),校准曲线显示预测值与观察值有较好的一致性。 采用Bootstrap法验证结果显示,C统计量均值的偏差为0.000 16。 结论 该模型具有较好的可信度和准确性,可为河南省心脑血管疾病高危人群的筛查提供科学依据。 Abstract:Objective To develop a cardio-cerebrovascular disease prediction model and provide scientific evidence for the prevention and control of cardio-cerebrovascular disease in Henan province. Methods Using a cohort study design, 5 757 permanent residents aged ≥18 years without baseline cardio-erebrovascular diseases, loss to follow up and missing data in China chronic diseases and risk factor surveillance in Henan during 2013−2014 were selected as the follow-up cohort. Weighted Cox regression analysis was used to construct the model, and C-statistics and calibration curve were used to evaluate the differentiation and calibration of the model, and Bootstrap method was used for internal validation. Results The mean follow-up time was 7.01±1.02 years, and the cumulative incidence of cardio-cerebrovascular disease was 9.23%. Variables included in the prediction model were age [hazard rate (HR) : 1.05, 95% confidence interval (CI): 1.04–1.00, P<0.001], systolic blood pressure (HR: 1.01, 95%CI: 1.01–1.02, P=0.001), smoking (HR: 1.44, 95%CI: 1.19–1.74, P=0.002), anti-hypertension treatment (HR: 1.81, 95%CI: 1.32–2.50, P=0.002), and diabetes (HR: 1.49, 95%CI: 1.11–2.00, P=0.014). The C statistic of the model was 0.76 (0.74–0.78), and the calibration curve showed that the predicted value was highly consistent with the observed value. The results of Bootstrap method showed that the bias of the mean value of C statistic was 0.000 16. Conclusion The model can be used to identify individuals at high risk for cardio-cerebrovascular disease in Henan with good reliability and accuracy. -
Key words:
- Cardio-cerebrovascular disease /
- Risk factor /
- Risk assessment /
- Prediction model /
- Cohort study
-
表 1 2020年河南省居民心脑血管病未发病组和心脑血管病发病组人群的基线特征比较
Table 1. Baseline characteristics of population with or without cardio-cerebrovascular disease in Henan in 2020
变量 总体
(n=5 757,加权n=48 741 947)心脑血管病未发病组
(n=5 052,加权n=44 245 101)心脑血管病发病组
(n=705,加权n=4 496 846)t/χ2值 P值 性别 2.670 0.102 男性 24 072 555 (49.39) 21 641 648 (48.91) 2 430 907 (54.06) 女性 24 669 393 (50.61) 22 603 454 (51.09) 2 065 939 (45.94) 年龄(岁) 43.41±15.93 41.81±15.19 59.19±14.37 15.700 <0.001 文化程度 20.430 <0.001 初中及以下 36 703 548 (75.30) 32 810 673 (74.16) 3 892 875 (86.57) 高中及以上 12 038 400 (24.70) 11 434 429 (25.84) 603 971 (13.43) 婚姻状态 1.980 0.159 已婚/同居 42 369 870 (86.93) 38 583 740 (87.21) 3 786 130 (84.19) 未婚/离婚/丧偶/分居 6 372 077 (13.07) 5 661 361 (12.79) 710 716 (15.81) 城乡 6.170 0.013 城市 19 989 831 (41.01) 17 831 465 (40.30) 2 158 367 (48.00) 农村 28 752 116 (58.99) 26 413 637 (59.70) 2 338 479 (52.00) 吸烟 6.090 0.014 否 32 324 309 (66.32) 29 646 419 (67.01) 2 677 890 (59.55) 是 16 417 639 (33.68) 14 598 683 (32.99) 1 818 956 (40.45) 饮酒 6.690 0.010 否 30 882 929 (63.36) 27 698 463 (62.60) 3 184 466 (70.82) 是 17 859 019 (36.64) 16 546 638 (37.40) 1 312 380 (29.18) 蔬菜水果摄入不足 0.420 0.515 否 27 829 168(57.09) 25 179 215 (56.91) 2 649 953 (58.93) 是 20 912 779 (42.91) 19 065 887 (43.09) 1 846 893 (41.07) 红肉摄入过多 0.640 0.422 否 42 114 995 (86.40) 38 147 376 (86.22) 3 967 620 (88.23) 是 6 626 952 (13.60) 6 097 726 (13.78) 529 226 (11.77) 体力活动 6.430 0.011 高水平 20 926 808 (42.93) 19 351 643 (43.74) 1 575 165 (35.03) 中等水平 14 853 080 (30.47) 13 293 736 (30.05) 1 559 344 (34.68) 低水平 12 962 059 (26.60) 11 599 723 (26.21) 1 362 337 (30.29) 高血压服药 145.790 <0.001 否 43 939 720 (90.15) 40 872 632 (92.38) 3 067 089 (68.21) 是 4 802 227 (9.85) 3 372 470 (7.62) 1 429 757 (31.79) 糖尿病患病 41.790 <0.001 否 43 061 191 (88.35) 39 649 065 (89.61) 3 412 125 (75.88) 是 5 680 757 (11.65) 4 596 036 (10.39) 1 084 721 (24.12) BMI 24.52±3.72 24.48±3.69 24.86±3.92 1.560 0.120 WC 84.47±10.49 84.31±10.50 86.10±10.35 2.710 0.007 SBP 130.74±21.16 129.18±20.12 146.18±24.61 10.450 <0.001 DBP 76.69±11.85 76.19±11.58 81.64±13.24 6.520 <0.001 心率 75.85±10.88 75.82±10.88 76.02±10.88 0.310 0.759 TC 4.46±0.96 4.44±0.96 4.64±0.97 3.150 0.002 TG 1.43±1.25 1.41±1.27 1.60±1.10 2.890 0.004 HDL-C 1.28±0.34 1.28±0.34 1.24±0.35 −1.670 0.095 LDL-C 2.73±0.82 2.71±0.82 2.90±0.85 3.580 <0.001 注:表中数据均进行加权调整;表头括号中n是实际例数,加权n是加权例数。连续性变量采用加权均值±标准差描述,分类变量采用加权例数(构成比)描述;BMI. 体质指数;WC. 腰围;SBP. 收缩压;DBP. 舒张压;TC. 总胆固醇;TG. 甘油三酯;HDL-C. 高密度脂蛋白胆固醇;LDL-C. 低密度脂蛋白胆固醇 表 2 心脑血管疾病发病危险因素的单因素Cox回归分析
Table 2. Univariate Cox regression analysis on risk factors for cardio-cerebrovascular disease
变量 β值 HR值(95%CI) t值 P值 性别 男性 1.00 女性 −0.16 0.85 (0.72~1.01) −2.11 0.061 年龄 0.06 1.06 (1.06~1.07) 18.38 <0.001 文化程度 初中及以下 1.00 高中及以上 −0.73 0.48 (0.37~0.62) −6.35 <0.001 婚姻状态 已婚/同居 1.00 未婚/离婚/丧偶/分居 0.24 1.27 (1.03~1.56) 2.53 0.030 城乡 城市 1.00 农村 −0.42 0.66 (0.38~1.15) −1.66 0.128 吸烟 否 1.00 是 0.30 1.35 (1.13~1.61) 3.81 0.003 饮酒 否 1.00 是 −0.36 0.70 (0.58~0.85) −4.13 0.002 蔬菜水果摄入不足 否 1.00 是 −0.11 0.90 (0.67~1.21) −0.82 0.430 红肉摄入过多 否 1.00 是 −0.16 0.85 (0.40~1.85) −0.46 0.657 体力活动 高水平 1.00 中等水平 0.44 1.55 (1.13~2.13) 3.05 0.012 低水平 0.36 1.44 (1.17~1.77) 3.86 0.003 高血压服药 否 1.00 是 1.60 4.97 (3.75~6.58) 12.67 <0.001 糖尿病 否 1.00 是 0.95 2.60 (1.95~3.46) 7.37 <0.001 BMI (kg/m2) 0.03 1.03 (1.01~1.05) 3.32 0.008 WC (cm) 0.02 1.02 (1.00~1.03) 2.51 0.031 SBP (mmHg) 0.03 1.03 (1.02~1.04) 7.03 <0.001 DBP (mmHg) 0.03 1.04 (1.02~1.05) 5.05 <0.001 心率 (次/分) 0.00 1.00 (0.99~1.02) 0.47 0.650 TC (mmol/L) 0.16 1.17 (0.99~1.39) 2.05 0.068 TG (mmol/L) 0.10 1.10 (1.02~1.19) 2.84 0.018 HDL-C (mmol/L) −0.42 0.66 (0.37~1.17) −1.61 0.138 LDL-C (mmol/L) 0.22 1.25 (1.04~1.50) 2.68 0.023 注:BMI. 体质指数;WC. 腰围;SBP. 收缩压;DBP. 舒张压;TC. 总胆固醇;TG. 甘油三酯;HDL-C. 高密度脂蛋白胆固醇;LDL-C. 低密度脂蛋白胆固醇;HR, 风险比;CI, 置信区间 表 3 心脑血管疾病预测模型纳入变量分析
Table 3. Analysis on variables included in prediction model of cardio-cerebrovascular disease
新纳入危险因素 危险因素系数P值 C统计量 (95%CI) P值a NRI (95%CI) 模型1 − 0.76 (0.74~0.78) − − 模型1+性别 0.941 0.76 (0.74~0.78) 0.525 −0.03 (−0.09~0.13) 模型1+文化程度 0.158 0.76 (0.74~0.78) 0.960 0.20 (−0.13~0.28) 模型1+婚姻状态 0.283 0.76 (0.74~0.78) 0.020 −0.24 (−0.31~0.33) 模型1+城乡 0.113 0.76 (0.74~0.77) 0.080 0.10 (0.01~0.18) 模型1+饮酒 0.844 0.76 (0.74~0.78) 0.836 0.08 (−0.06~0.15) 模型1+体力活动 0.003 0.76 (0.74~0.78) 0.821 0.11 (0.02~0.17) 模型1+BMI 0.588 0.76 (0.74~0.78) 0.769 0.05 (−0.04~0.14) 模型1+WC 0.386 0.76 (0.74~0.78) 0.184 0.05 (−0.05~0.13) 模型1+DBP 0.556 0.76 (0.74~0.78) 0.986 0.19 (0.03~0.32) 模型1+TC 0.764 0.76 (0.74~0.78) 0.619 0.13 (−0.16~0.28) 模型1+TG 0.154 0.76 (0.74~0.78) 0.058 0.19 (−0.21~0.31) 模型1+HDL-C 0.212 0.76 (0.74~0.78) 0.315 0.09 (−0.06~0.18) 模型1+LDL-C 0.135 0.76 (0.74~0.78) 0.865 0.19 (−0.10~0.27) 注:a. 与模型1的C统计量进行比较的P值;BMI. 体质指数;WC. 腰围;TC. 总胆固醇;TG. 甘油三酯;HDL-C. 高密度脂蛋白胆固醇;LDL-C. 低密度脂蛋白胆固醇;NRI. 净重分类指数;CI. 置信区间;−表示无数据 表 4 心脑血管疾病发病风险预测模型参数
Table 4. Parameters of risk prediction model of cardio-cerebrovascular disease
变量 β值 $S\bar x$ HR (95%CI) t值 P值 年龄 0.05 0.00 1.05 (1.04~1.06) 12.67 <0.001 收缩压 0.01 0.00 1.01 (1.01~1.02) 4.51 0.001 吸烟(1=是,0=否) 0.37 0.09 1.44 (1.19~1.74) 4.27 0.002 高血压服药(1=是,0=否) 0.60 0.14 1.81 (1.32~2.50) 4.14 0.002 糖尿病(1=是,0=否) 0.40 0.13 1.49 (1.11~2.00) 2.98 0.014 注:HR. 风险比;CI. 置信区间 -
[1] Liu SW, Li YC, Zeng XY, et al. Burden of cardiovascular diseases in China, 1990–2016: findings from the 2016 global burden of disease study[J]. JAMA Cardiol, 2019, 4(4): 342–352. DOI: 10.1001/jamacardio.2019.0295. [2] Institute for Health Metrics and Evaluation (IHME). GBD compare data visualization[M]. Seattle, WA: IHME, University of Washington, 2018. [3] 中国心血管健康与疾病报告编写组. 中国心血管健康与疾病报告2021概要[J]. 中国循环杂志,2022,37(6):553–578. DOI:10.3969/j.issn.1000−3614.2022.06.001.The Writing Committee of the Report on Cardiovascular Health and Diseases in China. Report on cardiovascular health and diseases in China 2021: an updated summary[J]. Chin Circulat J, 2022, 37(6): 553–578. DOI: 10.3969/j.issn.1000−3614.2022.06.001. [4] Zhou MG, Wang HD, Zeng XY, et al. Mortality, morbidity, and risk factors in China and its provinces, 1990–2017: a systematic analysis for the Global Burden of disease study 2017[J]. Lancet, 2019, 394(10204): 1145–1158. DOI: 10.1016/S0140−6736(19)30427−1. [5] Jousilahti P, Laatikainen T, Peltonen M, et al. Primary prevention and risk factor reduction in coronary heart disease mortality among working aged men and women in eastern Finland over 40 years: population based observational study[J]. BMJ, 2016, 352: i721. DOI: 10.1136/bmj.i721. [6] D'Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham heart study[J]. Circulation, 2008, 117(6): 743–753. DOI: 10.1161/CIRCULATIONAHA.107.699579. [7] Conroy RM, Pyörälä K, Fitzgerald AP, et al. Estimation of ten-year risk of fatal cardiovascular disease in Europe: the SCORE project[J]. Eur Heart J, 2003, 24(11): 987–1003. DOI: 10.1016/s0195−668x(03)00114−3. [8] Hippisley-Cox J, Coupland C, Vinogradova Y, et al. Derivation and validation of QRISK, a new cardiovascular disease risk score for the United Kingdom: prospective open cohort study[J]. BMJ, 2007, 335(7611): 136. DOI: 10.1136/bmj.39261.471806.55. [9] Yang XL, Li JX, Hu DS, et al. Predicting the 10-year risks of atherosclerotic cardiovascular disease in Chinese population: the China-PAR project (prediction for ASCVD risk in China)[J]. Circulation, 2016, 134(19): 1430–1440. DOI: 10.1161/CIRCULATIONAHA.116.022367. [10] Yang SC, Han YT, Yu CQ, et al. Development of a model to predict 10-Year risk of ischemic and hemorrhagic stroke and ischemic heart disease using the China Kadoorie Biobank[J]. Neurology, 2022, 98(23): e2307–2317. DOI: 10.1212/WNL.0000000000200139. [11] 王丽敏, 张梅, 李镒冲, 等. 2013年中国慢性病及其危险因素监测总体方案[J]. 中华预防医学杂志,2018,52(2):191–194. DOI:10.3760/cma.j.issn.0253−9624.2018.02.015.Wang LM, Zhang M, Li YC, et al. Scheme of the Chinese chronic non-communicable disease and risk factor surveillance[J]. Chin J Prev Med, 2018, 52(2): 191–194. DOI: 10.3760/cma.j.issn.0253−9624.2018.02.015. [12] 赵振平, 王丽敏, 李镒冲, 等. 2013年中国慢性病及其危险因素监测系统省级代表性评价[J]. 中华预防医学杂志,2018,52(2):165–169. DOI:10.3760/cma.j.issn.0253−9624.2018.02.009.Zhao ZP, Wang LM, Li YC, et al. Provincial representativeness assessment of China Non-communicable and Chronic disease risk factor surveillance system in 2013[J]. Chin J Prev Med, 2018, 52(2): 165–169. DOI: 10.3760/cma.j.issn.0253−9624.2018.02.009. [13] Zhang M, Liu SW, Yang L, et al. Prevalence of smoking and knowledge about the hazards of smoking among 170 000 Chinese adults, 2013-2014[J]. Nicotine Tob Res, 2019, 21(12): 1644–1651. DOI: 10.1093/ntr/ntz020. [14] Xiao D, Pengpeng Y, Yichong L, et al. Prevalence of drink-driving among adults in China: A nationally representative survey in 2010[J]. Traffic Inj Prev, 2017, 18(8): 795–800. DOI: 10.1080/15389588.2016.1161760. [15] World Health Organization. Diet, nutrition and the prevention of chronic diseases: report of a Joint WHO/FAO Expert Consultation[R]. Geneva: World Health Organization, 2002. [16] Vingeliene S, Chan DSM, Vieira AR, et al. An update of the WCRF/AICR systematic literature review and meta-analysis on dietary and anthropometric factors and esophageal cancer risk[J]. Ann Oncol, 2017, 28(10): 2409–2419. DOI: 10.1093/annonc/mdx338. [17] Armstrong T, Bull F. Development of the world health organization global physical activity questionnaire (GPAQ)[J]. J Public Health, 2006, 14(2): 66–70. DOI: 10.1007/s10389−006−0024−x. [18] American Diabetes Association. 2. Classification and diagnosis of diabetes: standards of medical care in diabetes-2021[J]. Diabetes Care, 2021, 44 Suppl 1: S15–33. DOI: 10.2337/dc21−S002. [19] Collins GS, Reitsma JB, Altman DG, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement[J]. BMJ, 2015, 350: g7594. DOI: 10.1136/bmj.g7594. [20] 杨淞淳, 孙至佳, 吕筠, 等. 心血管疾病风险预测模型研究进展[J]. 中华心血管病杂志,2022,50(12):1243–1251. DOI:10.3760/cma.j.cn112148−20220324−00202.Yang SC, Sun ZJ, Lyu J, et al. Research progress on risk prediction models of cardiovascular disease[J]. Chin J Cardiol, 2022, 50(12): 1243–1251. DOI: 10.3760/cma.j.cn112148−20220324−00202. [21] Zhiting G, Jiaying T, Haiying H, et al. Cardiovascular disease risk prediction models in the Chinese population- a systematic review and meta-analysis[J]. BMC Public Health, 2022, 22(1): 1608. DOI: 10.1186/s12889−022−13995−z. [22] 邢小龙. 基于传统和遗传危险因素的脑卒中个体化发病风险预测研究[D]. 北京: 北京协和医学院, 2021. DOI: 10.27648/d.cnki.gzxhu.2021.000183.Xing XL. Prediction for personalized stroke risk with traditional and genetic risk factors backgroud and objective[D]. Beijing: Peking Union Medical College, 2021. DOI: 10.27648/d.cnki.gzxhu.2021.000183. [23] Uno H, Cai TX, Pencina MJ, et al. On the C-statistics for evaluating overall adequacy of risk prediction procedures with censored survival data[J]. Stat Med, 2011, 30(10): 1105–1117. DOI: 10.1002/sim.4154. [24] Demler OV, Paynter NP, Cook NR. Tests of calibration and goodness-of-fit in the survival setting[J]. Stat Med, 2015, 34(10): 1659–1680. DOI: 10.1002/sim.6428. [25] Capanu M, Gönen M. Building a nomogram for survey-weighted cox models using R[J]. J Stat Softw, 2015, 64(1): 1–17. DOI: 10.18637/jss.v064.c01. [26] Kannel WB, McGee D, Gordon T. A general cardiovascular risk profile: the Framingham Study[J]. Am J Cardiol, 1976, 38(1): 46–51. DOI: 10.1016/0002−9149(76)90061−8. [27] 唐迅, 张杜丹, 何柳, 等. China-PAR模型在北方农村人群中预测动脉粥样硬化性心血管疾病发病风险的应用[J]. 北京大学学报(医学版),2017,49(3):439–445. DOI:10.3969/j.issn.1671−167X.2017.03.011.Tang X, Zhang DD, He L, et al. Application of the China-PAR risk prediction model for atherosclerotic cardiovascular disease in a rural northern Chinese population[J]. J Peking Univ (Health Sci), 2017, 49(3): 439–445. DOI: 10.3969/j.issn.1671−167X.2017.03.011. [28] 陈暐烨, 刘晓非, 沈鹏, 等. China-PAR与WHO模型预测10年心血管病风险的准确性[J]. 中华流行病学杂志,2022,43(8):1275–1281. DOI:10.3760/cma.j.cn112338−20211206−00952.Chen WY, Liu XF, Shen P, et al. Accuracy of the China-PAR and WHO risk models in predicting the ten-year risks of cardiovascular disease in the Chinese population[J]. Chin J Epidemiol, 2022, 43(8): 1275–1281. DOI: 10.3760/cma.j.cn112338−20211206−00952. [29] 程水华, 朱建军, 王文, 等. 汇集队列风险方程与China-PAR模型在体检人群ASCVD风险预测中的应用[J]. 中国循证心血管医学杂志,2020,12(2):131–134. DOI:10.3969/j.issn.1674−4055.2020.02.02.Cheng SH, Zhu JJ, Wang W, et al. Application of pooled cohort risk equations and China-PAR model in ASCVD risk prediction among people with physical examination[J]. Chin J Evid Bases Cardiovasc Med, 2020, 12(2): 131–134. DOI: 10.3969/j.issn.1674−4055.2020.02.02. [30] Wu YF, Liu XQ, Li X, et al. Estimation of 10-year risk of fatal and nonfatal ischemic cardiovascular diseases in Chinese adults[J]. Circulation, 2006, 114(21): 2217–2225. DOI: 10.1161/circulationaha.105.607499. [31] Pylypchuk R, Wells S, Kerr A, et al. Cardiovascular disease risk prediction equations in 400 000 primary care patients in New Zealand: a derivation and validation study[J]. Lancet, 2018, 391(10133): 1897–1907. DOI: 10.1016/S0140−6736(18)30664−0. [32] Huang YY, Tian WB, Jiang CQ, et al. A simple model for predicting 10-year cardiovascular risk in middle-aged to older Chinese: Guangzhou biobank cohort study[J]. J Cardiovasc Transl Res, 2022, 15(2): 416–426. DOI: 10.1007/s12265−021−10163−3. -