脑卒中患者静脉血栓栓塞症风险预测模型的构建及验证

王荣荣; 周乾宇; 郭园丽; 王盼盼; 何雯倩; 赵明扬; 张配嘉; 胡博; 吴田田; 要子慧; 王昱; 孙长青

doi:10.16462/j.cnki.zhjbkz.2023.10.008

脑卒中患者静脉血栓栓塞症风险预测模型的构建及验证

doi: 10.16462/j.cnki.zhjbkz.2023.10.008

1.
郑州大学护理与健康学院社区护理教研室，郑州 450001
2.
郑州大学公共卫生学院社会医学与卫生事业管理教研室，郑州 450001
3.
郑州大学第一附属医院神经内科，郑州 450052

基金项目:

国家社会科学基金项目 20BRK041

河南省科技攻关项目 212102310767

详细信息

通讯作者:
孙长青，E-mail：suncq@zzu.edu.cn

中图分类号: R12
计量
- 文章访问数: 337
- HTML全文浏览量: 476
- PDF下载量: 42
- 被引次数: 0
出版历程
- 收稿日期: 2022-10-21
- 修回日期: 2023-01-13
- 网络出版日期: 2023-10-23
- 刊出日期: 2023-10-10

Development and validation of a risk prediction model for venous thromboembolism in stroke patients

1.
Department of Community Care, School of Nursing and Health, Zhengzhou University, Zhengzhou 450001, China
2.
Department of Social Medicine and Health Management, School of Public Health, Zhengzhou University, Zhengzhou 450001, China
3.
Department of Neurology, the First Affiliated Hospital of Zhengzhou University, Zhengzhou 450052, China

Funds:

National Social Science Foundation of China 20BRK041

The Key Science and Technology Program of Henan Province 212102310767

More Information

Corresponding author: SUN Changqing, E-mail: suncq@zzu.edu.cn

摘要

摘要: 目的构建并验证脑卒中患者发生静脉血栓栓塞症(venous thromboembolism, VTE)的风险预测模型，为脑卒中患者VTE的预防控制提供科学依据。方法研究对象为河南省脑卒中队列的675例脑卒中患者，按7∶3随机划分为训练集(473例)和测试集(202例)。使用随机森林算法筛选变量、logistic回归模型分析方法构建模型，并绘制列线图。通过受试者工作特征曲线下面积(area under curve, AUC)、Hosmer-Lemeshow检验等评价模型的预测效能，使用决策曲线分析(decision curve analysis, DCA)评估模型的临床使用价值；并采用五折交叉验证对模型进行内部验证。结果最终纳入模型的预测因子为年龄、住院天数、日常生活活动能力(activity of daily living, ADL)、肌力、尿酸(uric acid, UA)、D-二聚体、纤维蛋白原(fibrinogen, Fib)和总胆固醇(total cholesterol, TC)。模型在训练集中的Hosmer-Lemeshow检验P=0.872，AUC=0.924(95% CI: 0.898~0.950)；测试集Hosmer-Lemeshow检验P=0.597，AUC=0.902(95% CI: 0.852~0.951)。DCA曲线表明，模型在训练集和测试集中均具有较高的临床净获益。五折交叉内部验证结果显示，模型在训练集和测试集中的平均AUC分别为0.913和0.929。结论该研究构建的脑卒中患者VTE风险预测模型能有效预测VTE的发生，可为高风险患者早期识别和预防性治疗提供参考。
- 脑卒中 /
- 静脉血栓栓塞症 /
- 预测模型 /
- 随机森林 /
- 列线图
Abstract: Objective To develop and validate the risk prediction model of venous thromboembolism (VTE) in stroke patients, so as to provide a scientific basis for the prevention and control of VTE. Methods A total of 675 stroke patients were enrolled from our stroke cohort of Henan Province. The data were randomly divided into a training (473 patients) and a testing dataset (202 patients) by a ratio of 7∶3. Then, we used a random forest algorithm for variable selection and logistic regression analysis to construct the model, and a nomogram was drawn. The prediction efficiency of the model was evaluated by the area under the receiver operating characteristic curve and Hosmer-Lemeshow test. Decision curve analysis (DCA) was used to evaluate the clinical application value of the model and the five-fold cross-validation was utilized to verify the model internally. Results The predictors that ultimately entered the prediction model were age, hospital stays, ADL, myodynamia, uric acid, D-dimer, fibrinogen, and total cholesterol. In the training dataset, the Hosmer-Lemeshow test yielded P=0.872 and the AUC was 0.924 (95% CI: 0.898-0.950). The testing dataset showed that the Hosmer-Lemeshow test yielded P=0.597 and the AUC was 0.902 (95% CI: 0.852-0.951). DCA curves indicated that the model had high clinical net benefits in both datasets. Internal verification presented that the average AUCs of the model in the training and testing datasets were 0.913 and 0.929, respectively. Conclusions The risk prediction model developed in this study can effectively predict VTE occurrence in stroke patients, offering a valuable tool for identifying high-risk individuals and implementing early preventive measures.
- Stroke /
- Venous thromboembolism /
- Prediction model /
- Random forest /
- Nomogram

HTML全文

图 1 脑卒中患者VTE风险预测模型的列线图

VTE, 静脉血栓栓塞症; ADL, 日常行为活动能力; UA, 尿酸; Fib, 纤维蛋白原; TC, 总胆固醇。

Figure 1. A nomogram of a VTE risk prediction model for stroke patients

VTE, venous thromboembolism; ADL, activity of daily living; UA, uric acid; Fib, fibrinogen; TC, total cholesterol.

下载: 全尺寸图片幻灯片

表 1 训练集与测试集病例的基线特征

Table 1. Baseline characteristics of patients in training and testing datasets

变量 Variable	训练集 Training datasets (n=473)				测试集 Testing datasets (n=202)
变量 Variable	VTE组^① VTE group ^① (n=121, 25.6%)	非VTE组^① Non-VTE group ^① (n=352, 74.4%)	Z/χ²值 value	P值 value	VTE组^① VTE group ^① (n=43, 21.3%)	非VTE组^① Non-VTE group ^① (n=159, 78.7%)	Z/χ²值 value	P值 value
性别 Gender			8.476	0.004			2.709	0.100
男性 Man	61(21.0)	230(79.0)			24(17.9)	110(82.1)
女性 Woman	60(33.0)	122(67.0)			19(27.9)	49(72.1)
年龄/岁 Age/years	68(57, 74)	58(50, 68)	-5.033	< 0.001	67(57, 75)	58(48, 68)	-3.596	< 0.001
住院天数/d Length of stay/d	17(10, 25)	10(8, 15)	-6.805	< 0.001	15(10, 21)	10(7, 13)	-4.207	< 0.001
BMI/(kg·m^-2)			3.875	0.268			2.256	0.500
＜18.5	4(50.0)	4(50.0)			1(33.3)	2(66.7)
18.5~＜24.0	52(27.1)	140(72.9)			19(25.7)	55(74.3)
24.0~＜28.0	56(24.9)	169(75.1)			18(19.4)	75(80.6)
≥28.0	9(18.8)	39(81.2)			5(15.6)	27(84.4)
ADL			71.515	< 0.001			20.336	< 0.001
无须依赖 No dependence	2(2.2)	87(97.8)			3(6.5)	43(93.5)
轻度依赖 Mild dependence	26(15.4)	143(84.6)			9(13.2)	59(86.8)
中度依赖 Moderate dependence	35(38.9)	55(61.1)			9(28.1)	23(71.9)
重度依赖 Heavily dependence	58(46.4)	67(53.6)			22(39.3)	34(60.7)
卒中类型 Stroke subtypes			10.582	0.004			14.255	< 0.001
缺血性 Ischemic	93(23.1)	310(76.9)			31(17.5)	146(82.5)
出血性 Hemorrhagic	23(37.1)	39(62.9)			9(40.9)	13(59.1)
混合型 Mixed	5(62.5)	3(37.5)			3(100.0)	0(0.0)
吸烟史 Smoking history			4.543	0.103			0.850	0.714
从不吸烟 Never smoking	95(28.3)	241(71.7)			31(22.5)	107(77.5)
曾经吸烟 Ever smoking	9(20.9)	34(79.1)			5(23.8)	16(76.2)
现在吸烟 Current smoking	17(18.1)	77(81.9)			7(16.3)	36(83.7)
饮酒史 Drinking history			6.072	0.048			1.406	0.504
从不饮酒 Never drinking	99(28.4)	249(71.6)			34(23.4)	111(76.6)
曾经饮酒 Ever drinking	6(22.2)	21(77.8)			1(9.1)	10(90.9)
现在饮酒 Current drinking	16(16.3)	82(83.7)			8(17.4)	38(82.6)
输血史 Transfusion history	17(41.5)	24(58.5)	5.948	0.015	9(45.0)	11(55.0)	5.961	0.015
卒中史 Stroke history	34(26.4)	95(73.6)	0.056	0.813	25(34.2)	48(65.8)	11.458	0.001
并发症 Complications
糖尿病 Diabetes	36(23.7)	116(76.3)	0.423	0.515	16(23.9)	51(76.1)	0.402	0.526
高血压 Hypertension	93(26.8)	254(73.2)	1.018	0.313	30(21.7)	108(78.3)	0.053	0.818
高脂血症 Hyperlipidaemia	14(26.4)	39(73.6)	0.022	0.883	3(15.0)	17(85.0)	0.190	0.663
房颤 Atrial fibrillation	17(60.7)	11(39.3)	19.297	< 0.001	6(75.0)	2(25.0)	11.199	0.001
感染 Infection	65(51.6)	61(48.4)	61.016	< 0.001	21(48.8)	22(51.2)	24.746	< 0.001
恶性肿瘤 Malignant tumor	8(80.0)	2(20.0)	13.106	< 0.001	4(100.0)	0(0.0)	10.677	< 0.001
脱水类药物 Hyperosmotic drugs	56(48.3)	60(51.7)	41.580	< 0.001	17(39.5)	26(60.5)	10.856	0.001
抗血小板药物 Antiplatelet drugs	85(23.1)	283(76.9)	5.371	0.020	29(17.9)	133(82.1)	5.597	0.018
激素类药物 Hormone drugs	6(50.0)	6(50.0)	2.653	0.103	3(42.9)	4(57.1)	0.901	0.343
溶栓 Thrombolysis	17(51.5)	16(48.5)	12.533	< 0.001	4(25.0)	12(75.0)	0.004	0.952
深静脉置管 Deep vein catheterization			31.648	< 0.001			14.289	< 0.001
无 No	96(22.2)	337(77.8)			33(17.8)	152(82.2)
中心静脉 Central venous	22(61.1)	14(38.9)			9(64.3)	5(35.7)
股静脉 Femoral vein	3(75.0)	1(25.0)			1(33.3)	2(66.7)
PT/s	10.9(10.2~12.0)	10.5(9.9~11.1)	-3.961	< 0.001	11.1(10.4~12.2)	10.6(10.0~11.0)	-2.747	0.006
PTA/%	102.0(91.7~113.5)	108.0(99.3~116.8)	-3.746	< 0.001	100.0(87.0~112.0)	106.0(99.0~116.0)	-2.516	0.012
APTT/s	27.6(24.7~29.7)	28.7(26.3~31.1)	-2.988	0.003	27.6(24.5~32.1)	28.9(26.7~31.4)	-1.179	0.238
Fib/(g·L^-1)	3.67(2.95~4.61)	2.93(2.55~3.44)	-6.623	< 0.001	3.83(2.78~4.53)	2.88(2.45~3.39)	-3.898	< 0.001
TT/s	15.3(14.0~16.8)	15.3(14.3~16.9)	-0.780	0.435	15.1(13.7~16.7)	15.7(14.4~17.1)	-1.775	0.076
D-二聚体/(mg·L^-1) D-dimer/(mg·L^-1)	1.02(0.37~3.06)	0.15(0.08~0.27)	-11.546	< 0.001	0.83(0.29~2.11)	0.16(0.08~0.24)	-6.685	< 0.001
CRP/(mg·L^-1)	17.25(2.98~69.12)	2.42(1.01~7.71)	-7.203	< 0.001	20.28(5.01~102.95)	2.94(1.29~7.48)	-5.458	< 0.001
血红蛋白/(g·L^-1) Hemoglobin/(g·L^-1)	128.8(115.9~139.7)	137.0(125.8~147.0)	-4.347	< 0.001	125.0(117.2~144.0)	140.3(126.9~149.0)	-2.592	0.010
PLT/(10⁹·L^-1)	223.0(180.5~272.5)	215.5(179.3~258.8)	-0.947	0.344	223.0(169.0~273.0)	217.0(186.0~251.0)	-0.109	0.913
UA/(μmol·L^-1)	222(182~288)	273(222~335)	-4.992	< 0.001	252(208~328)	274(227~336)	-0.953	0.913
TC/(mmol·L^-1)	3.88(3.23~4.64)	3.71(3.06~4.47)	-2.084	0.037	3.81(3.09~4.57)	3.83(3.25~4.76)	-0.260	0.795
TG/(mmol·L^-1)	1.26(0.93~1.95)	1.21(0.87~1.62)	-1.407	0.159	1.32(0.95~1.61)	1.31(0.99~1.84)	-0.838	0.402
LDL/HDL	2.26(1.73~2.77)	2.09(1.51~2.73)	-1.421	0.155	2.38(1.76~3.29)	2.28(1.72~2.95)	-0.644	0.520
Hcy/(μmol·L^-1)	13.57(10.87~17.64)	14.11(11.11~17.92)	-0.684	0.494	13.22(10.35~16.57)	14.38(12.61~19.61)	-2.145	0.032
GCS	15(9~15)	15(15~15)	-7.608	< 0.001	15(9~15)	15(15~15)	-4.559	< 0.001
肌力/级 Muscle strength/grade	3(2~4)	5(4~5)	-8.527	< 0.001	3.8(2~5)	5(4~5)	-3.232	0.001
NIHSS评分 NIHSS score	5(3~8)	3(2~6)	-4.937	< 0.001	4(2~10)	3(1~5)	-2.978	0.003
mRs评分 mRs score	0(0~2)	0(0~1)	-2.370	0.018	1(0~2)	0(0~1)	-1.233	0.217
吞咽功能/级 Swallowing function/grade	1(1~2)	1(1~1)	-4.282	< 0.001	1(1~2)	1(1~1)	-2.651	0.008
注：PT, 凝血酶原时间; PTA, 凝血酶原时间活动度; APTT, 活化部分凝血酶原时间; TT, 凝血酶时间; CRP, C-反应蛋白; PLT, 血小板计数; UA, 尿酸; TC, 总胆固醇; TG, 甘油三酯; LDL/HDL, 低密度脂蛋白/高密度脂蛋白; Hcy, 同型半胱氨酸; NIHSS, 美国国立卫生院神经功能缺损。 ①以[人数(占比/%)]或[M(P₂₅, P₇₅)]表示。 Note: PT, prothrombin time; PTA, prothrombin time activity; APTT, activated partial thromboplastin time; TT, thrombin time; CRP, C-reactive protein; PLT, blood platelet count; UA, uric acid; TC, total cholesterol; TG, triglyceride; LDL/HDL, low density lipoprotein/high-density lipoprotein; Hcy, homocysteine; NIHSS, national institution of health stroke scale. ① [Number of people (proportion/%)] or [M(P₂₅, P₇₅)].

下载: 导出CSV

表 2 脑卒中患者VTE的多因素logistic回归模型分析

Table 2. Multivariate logistic regression model analysis of the risk of VTE in stroke patients

变量 Variable	β	s	Wald值 value	OR值 value (95% CI)	P值 value
常量 Constant	-3.730	1.077	-3.463	0.024(0.002~0.174)	0.001
年龄＞63岁 Age >63 years	0.858	0.336	2.554	2.359(1.227~4.609)	0.011
ADL(参照：无须依赖) ADL(reference: no dependence)
轻度依赖 Mild dependence	0.907	0.822	1.103	2.476(0.591~17.234)	0.270
中度依赖 Moderate dependence	1.909	0.837	2.281	6.745(1.559~47.997)	0.023
重度依赖 Heavily dependence	0.321	0.915	0.351	1.378(0.260~10.854)	0.726
住院天数＞14 d length of stay >14 days	1.015	0.336	3.019	2.761(1.432~5.383)	0.003
肌力＞3级 Muscle strength >3 grades	-0.890	0.370	-2.403	0.411(0.197~0.846)	0.016
D-二聚体＞0.53/(mg·L^-1) D-dimer >0.53/(mg·L^-1)	2.521	0.363	6.945	12.435(6.208~25.896)	＜0.001
Fib＞3.95/(g·L^-1)	1.054	0.402	2.620	2.869(1.305~6.355)	0.009
UA＞227/(μmol·L^-1)	-0.811	0.329	-2.463	0.445(0.231~0.845)	0.014
TC＞3.37/(mmol·L^-1)	1.333	0.385	3.460	3.793(1.831~8.344)	0.001
注：VTE, 静脉血栓栓塞症; ADL, 日常行为活动能力; Fib, 纤维蛋白原; UA, 尿酸; TC, 总胆固醇。 Note: VTE, venous thromboembolism; ADL, activity of daily living; Fib, fibrinogen; UA, uric acid; TC, total cholesterol.

下载: 导出CSV

参考文献(17)

[1]	王深明. 静脉血栓栓塞症诊治现状与展望[J]. 中国实用外科杂志, 2010, 30(12): 992-994. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGWK201012004.htm Wang SM. The present situation and forecast of venous thromboembolism[J]. Chin J Pract Surg, 2010, 30(12): 992-994. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGWK201012004.htm
[2]	Liu XC, Chen XW, Li ZL, et al. Anatomical distribution of lower-extremity deep venous thrombosis in patients with acute stroke[J]. J Stroke Cerebrovasc Dis, 2020, 29(7): 104866. DOI: 10.1016/j.jstrokecerebrovasdis.2020.104866.
[3]	Khan F, Tritschler T, Kahn SR, et al. Venous thromboembolism[J]. Lancet, 2021, 398(10294): 64-77. DOI: 10.1016/S0140-6736(20)32658-1.
[4]	Ding DL, Sekar P, Moomaw CJ, et al. Venous thromboembolism in patients with spontaneous intracerebral hemorrhage: a multicenter study[J]. Neurosurgery, 2018, 84(6): E304-E310. DOI: 10.1093/neuros/nyy333.
[5]	Kahn SR, Morrison DR, Diendere G, et al. Interventions for implementation of thromboprophylaxis in hospitalized patients at risk for venous thromboembolism[J]. Cochrane Database Syst Rev, 2018, 4(4): D8201. DOI: 10.1002/14651858.CD008201.pub3.
[6]	谢长清. 脑卒中急性期并发深静脉血栓形成不同风险评估工具的比较研究[D]. 苏州: 苏州大学, 2020. Xie CQ. A comparative study of different risk assessment tools for acute stroke complicated with deep vein thrombosis[D]. Suzhou: Suzhou university, 2020.
[7]	Yang YQ, Wang X, Huang Y, et al. Ontology-based venous thromboembolism risk assessment model developing from medical records[J]. BMC Med Inform Decis Mak, 2019, 19(Suppl 4): 151. DOI: 10.1186/s12911-019-0856-2.
[8]	Pan X, Wang Z, Chen QM, et al. Development and validation of a nomogram for lower extremity deep venous thrombosis in patients after acute stroke[J]. J Stroke Cerebrovasc Dis, 2021, 30(5): 105683. DOI: 10.1016/j.jstrokecerebrovasdis.2021.105683.
[9]	陈新宇, 周军良, 李婷婷, 等. 基于随机森林算法和Logistic回归分析的老年肌肉衰减症影响因素[J]. 中华疾病控制杂志, 2022, 26(3): 357-361. DOI: 10.16462/j.cnki.zhjbkz.2022.03.020. Chen XY, Zhou JL, Li TT, et al. Influential factors of sarcopenia in older adults based on random forest and Logistic regression[J]. Chin J Dis Control Prev, 2022, 26(3): 357-361. DOI: 10.16462/j.cnki.zhjbkz.2022.03.020.
[10]	中华医学会神经病学分会, 中华医学会神经病学分会脑血管病学组. 中国各类主要脑血管病诊断要点2019[J]. 中华神经科杂志, 2019, 52(9): 710-715. DOI: 10.3760/cma.j.issn.1006-7876.2019.09.003. Chinese Society of Neurology; Chinese Stroke Society. Main diagnostic points of cerebro-vascular diseases in China 2019[J]. Chin J Neurol, 2019, 52(9): 710-715. DOI: 10.3760/cma.j.issn.1006-7876.2019.09.003.
[11]	Wang Y, Shi Y, Dong Y, et al. Clinical risk factors of asymptomatic deep venous thrombosis in patients with acute stroke[J]. Clin Appl Thromb Hemost, 2019, 25: 1076029619868534. DOI: 10.1177/1076029619868534.
[12]	Cheng HR, Huang GQ, Wu ZQ, et al. Individualized predictions of early isolated distal deep vein thrombosis in patients with acute ischemic stroke: a retrospective study[J]. BMC Geriatr, 2021, 21(1): 140. DOI: 10.1186/s12877-021-02088-y.
[13]	Riley RD, Ensor J, Snell KIE, et al. Calculating the sample size required for developing a clinical prediction model[J]. BMJ, 2020, 368: m441. DOI: 10.1136/bmj.m441.
[14]	张丹丹, 丁唱, 欧阳朝威, 等. 重型创伤性脑损伤患者发生静脉血栓栓塞的危险因素分析[J]. 中华创伤杂志, 2022, 38(4): 346-353. DOI: 10.3760/cma.j.cn501098-20211202-00631. Zhang DD, Ding C, Ou Yang CW, et al. Risk factors of venous thromboembolism in patients with severe traumatic brain injury[J]. Chin J Trauma, 2022, 38(4): 346-353. DOI: 10.3760/cma.j.cn501098-20211202-00631.
[15]	Spasić I, Ubavić M, Sumarac Z, et al. Influence of lipid metabolism disorders on venous thrombosis risk[J]. J Med Biochem, 2021, 40(3): 245-251. DOI: 10.5937/jomb0-27106.
[16]	Ren WW, Huang QQ, Fan KL, et al. Uric acid played a role in the association between gender and deep vein thrombosis in patients with stroke[J]. Nutr Metab Cardiovasc Dis, 2021, 31(9): 2700-2706. DOI: 10.1016/j.numecd.2021.05.032.
[17]	Vila E, Sole M, Masip N, et al. Uric acid treatment after stroke modulates the Kruppel-like factor 2-VEGF-A axis to protect brain endothelial cell functions: Impact of hypertension[J]. Biochem Pharmacol, 2019, 164: 115-128. DOI: 10.1016/j.bcp.2019.04.002.