急性冠脉综合征患者住院期间新发房颤预测模型的建立与验证

李军政; 许祥; 张志辉; 邬娜; 袁志权; 贾潇岳; 李成英; 吴龙; 向颖; 钟理; 李亚斐

doi:10.16462/j.cnki.zhjbkz.2021.02.016

急性冠脉综合征患者住院期间新发房颤预测模型的建立与验证

doi: 10.16462/j.cnki.zhjbkz.2021.02.016

1.
400038 重庆，陆军军医大学军事预防医学系军队流行病学教研室
2.
400038 重庆，陆军军医大学第一附属医院心内科
3.
401120 重庆，重庆医科大学附属第三医院心血管疾病中心

基金项目:

国家自然科学基金青年科学基金项目 81502883

详细信息

通讯作者:
LI Ya-fei, E-mail: liyafei2008@hotmail.com

中图分类号: R181.23
计量
- 文章访问数: 685
- HTML全文浏览量: 210
- PDF下载量: 85
- 被引次数: 0
出版历程
- 收稿日期: 2020-08-24
- 修回日期: 2021-01-02
- 刊出日期: 2021-02-10

Establishment and validation of a predictive model for new onset atrial fibrillation in patients with acute coronary syndrome during hospitalization

1.
Department of Epidemiology, College of Preventive Medicine, Army Medical University, Chongqing 400038, China
2.
Department of Cardiology, First Affiliated Hospital, Army Medical University, Chongqing 400038, China
3.
Cardiovascular Disease Center, Third Affiliated Hospital of Chongqing Medical University, Chongqing 401120, China

Funds:

Youth Science Foundation Project of National Natural Science Foundation of China 81502883

摘要

摘要: 目的建立急性冠脉综合征患者入院期间新发房颤预测模型，为入院早期发现高风险人群、及时采取干预措施提供科学依据。方法回顾性收集2010年1月至2019年12月在某三甲医院心内科住院、符合纳入排除标准的急性冠脉综合征患者资料，共1 915例患者，随机分为建模组和验证组，在建模组中运用多元Logistic回归分析模型和筛选出院内新发房颤的独立危险因素，建立回归预测模型，构建风险评分列线图，并在验证组中进行验证。采用受试者工作特征曲线(receiver operating characteristic curve, ROC)和曲线下面积(area under curve, AUC)以及Hosmer-lemoshow拟合优度检验和校准曲线分别评价模型的区分能力和校准度。结果建模组共纳入急性冠脉综合征患者958例，其中新发房颤62例；验证组纳入957例，其中新发房颤65例。建模组中共纳入7个指标，分别为：年龄、入院时心率、心衰Killip分级、N末端脑钠肽前体(N-terminal pro-brain natriuretic peptide, NT-proBNP)水平、左心房横径、右心房横径和中性粒细胞计数，以此建立回归预测模型。建模组的AUC为0.91(95% CI: 0.88~0.94)；验证组的AUC为0.86(95% CI: 0.81~0.91)。校准曲线图和拟合优度检验(建模组和验证组均有P>0.05)显示模型具有较好的校准度。结论本研究成功构建了急性冠脉综合征患者院内新发房颤的预测模型，具有较好区分度和校准度，利用列线图可方便直观地预测患者的房颤风险，为临床早期防治和改善预后提供依据。
- 急性冠脉综合征 /
- 新发房颤 /
- 预测模型 /
- 模型验证 /
- 列线图
Abstract: Objective To establish a predictive model of new onset atrial fibrillation in patients with acute coronary syndrome (ACS) during admission, so as to provide scientific basis for early detection of high-risk patients and timely intervention measures. Methods A retrospective cohort study was conducted among 1 915 patients with ACS who were admitted to the Department of Cardiology of a large general hospital between January 2010 and December 2019. Patients were randomly divided into two groups: model group and validation group. In the model group, a multivariate Logistic regression analysis model was used to screen the independent factors associated with new onset atrial fibrillation. Regression prediction model and nomogram were established, and validated in the validation group. Area under curve (AUC) of receiver operating characteristic curve (ROC) and Hosmer-lemoshow test were used to evaluate the discrimination and calibration of the model, respectively. Results There were 958 cases in the model group comprising 62 new onset atrial fibrillation cases, and 957 cases in the validation group comprising 65 new onset atrial fibrillation cases. In the model group, seven indicators were independently associated with atrial fibrillation, including age, heart rate at admission, Killip classification of heart failure, N-terminal pro-brain natriuretic peptide (NT-proBNP) level, left atrial diameter, right atrial diameter and neutrophil count. In the modeling group, the AUC was 0.91 (95% CI: 0.88-0.94) in model group and 0.86 (95% CI: 0.81-0.91) in validation group. Calibration plot and goodness of fit test (in modeling group and validation group, P>0.05) indicated that the prediction model had a good calibration ability. Conclusions In this study, the prediction model of new onset atrial fibrillation in patients with acute coronary syndrome was successfully constructed, which has a good discrimination and calibration, the nomogram could conveniently be used and intuitively predict the risk of atrial fibrillation, thus provides foundation for early intervention and improvement of prognosis in clinical practice.
- Acute coronary syndrome /
- New onset atrial fibrillation /
- Prediction model /
- Model validation /
- Nomogram

HTML全文

图 1 急性冠脉综合征患者院内新发房颤风险评分列线图

Figure 1. Nomogram of risk score of new onset atrial fibrillation in patients with acute coronary syndrome

下载: 全尺寸图片幻灯片

图 2 建模组(左)与验证组(右)ROC曲线

Figure 2. ROC curve of modeling group (left) and validation group (right)

下载: 全尺寸图片幻灯片

图 3 建模组(左)与验证组(右)Hosmer-lemoshow校准曲线

Figure 3. Hosmer-lemoshow calibration curve of modeling group (left) and validation group (right)

下载: 全尺寸图片幻灯片

表 1 建模组与验证组病例的基线特征[M(P₂₅，P₇₅)]

Table 1. Baseline characteristics of patients in modeling and validation group [M(P₂₅, P₇₅)]

变量	建模组				验证组
变量	未发生房颤	发生房颤	Z/χ²值	P值	未发生房颤	发生房颤	Z/χ²值	P值
人口学特征
年龄(岁)	67.0(59.0, 74.0)	74.5(70.0, 80.0)	-5.787^a	0.000	67.0(59.0, 74.0)	78.0(71.0, 81.0)	-7.126^a	0.000
性别[n(%)]			0.079^b	0.780			1.533^b	0.220
男性	606(67.6)	43(69.4)			563(63.1)	46(70.8)
女性	290(32.4)	19(30.6)			329(36.9)	19(29.2)
BMI(kg/m²)	24.4(22.2, 26.7)	24.5(21.9, 27.1)	-0.049^a	0.960	24.4(22.4, 26.5)	24.6(22.9, 26.7)	-0.501^a	0.620
吸烟[n(%)]	473(52.8)	34(54.8)	0.098^b	0.750	440(49.3)	34(52.3)	0.215^b	0.640
饮酒[n(%)]	360(40.2)	27(43.5)	0.274^b	0.600	334(37.4)	25(38.5)	0.268^b	0.870
入院检查
入院心率(次/min)	77.5(70.0, 86.5)	90.0(78.0, 102.0)	-4.951^a	0.000	78.0(71.0, 88.0)	87.0(78.0, 104.0)	-4.629^a	0.000
SPB(mm Hg)	130.0(118.0, 144.5)	130.0(114.0, 149.0)	0.034^a	0.970	130.0(118.0, 145.0)	130.0(110.0, 140.0)	1.520^a	0.130
DPB(mm Hg)	78.0(70.0, 87.0)	77.0(66.0, 84.0)	0.925^a	0.360	78.0(70.0, 85.0)	72.0(66.0, 80.0)	2.181^a	0.030
心衰分级[n(%)]			32.357^b	0.000			12.476^b	0.000
< 2级	630(70.3)	22(35.5)			615(68.9)	31(47.7)
≥2级	266(29.7)	40(64.5)			277(31.1)	34(52.3)
合并疾病[n(%)]
高血压	559(62.4)	44(71.0)	1.830^b	0.180	574(64.3)	43(66.2)	0.086^b	0.770
糖尿病	327(36.5)	28(45.2)	1.867^b	0.170	331(37.1)	26(40.0)	0.217^b	0.640
高血脂	213(23.8)	11(17.7)	1.177^b	0.280	203(22.8)	6(9.2)	6.495^b	0.010
心肌病	20(2.2)	3(4.8)	0.753^c	0.386	23(2.6)	2(3.1)	< 0.01^c	1.000
陈旧性脑梗死	68(7.6)	11(17.7)	7.900^b	0.010	67(7.5)	13(20.0)	12.335^b	0.000
慢性阻塞性肺疾病	22(2.5)	1(1.6)	< 0.01^c	1.000	15(1.7)	3(4.6)	1.459^c	0.227
慢性肾功能障碍	22(2.5)	6(9.7)	8.266^c	0.004	18(2.0)	7(10.8)	14.960^c	0.000
药物治疗[n(%)]
他汀类药物	808(90.2)	50(80.6)	5.637^b	0.020	814(91.3)	51(78.5)	11.413^b	0.001
阿司匹林	677(75.6)	43(69.4)	1.195^b	0.270	703(78.8)	41(63.1)	8.669^b	0.003
氯吡格雷	649(72.4)	43(69.4)	0.274^b	0.600	666(74.7)	51(78.5)	0.465^b	0.500
β受体抑制剂	556(62.1)	35(56.5)	0.770^b	0.380	527(59.1)	33(50.8)	1.724^b	0.190
血管紧张素转化酶抑制剂	357(39.8)	28(45.2)	0.682^b	0.410	363(40.7)	23(35.4)	0.710^b	0.400
实验室检查
肌钙蛋白阳性[n(%)]	325(36.3)	24(38.7)	0.149^b	0.700	287(32.2)	26(40.0)	1.686^b	0.190
NT-proBNP(pg/ml)	154.4(65.1, 441.6)	1 456.0(770.2, 2 730.0)	-9.139^a	0.000	177.2(80.1, 516.8)	1 100.0(467.3, 2 340.0)	-7.440^a	0.000
超敏C反应蛋白(mg/L)	2.8(1.0, 9.3)	6.5(1.9, 13.2)	-3.564^a	0.000	2.8(1.1, 9.8)	4.7(1.6, 11.2)	-2.071^a	0.040
D-二聚体(mg/L)	0.3(0.2, 0.9)	0.7(0.3, 2.2)	-3.614^a	0.000	0.3(0.2, 0.9)	1.0(0.4, 11.3)	-5.248^a	0.000
纤维蛋白原(g/L)	2.8(2.4, 3.4)	3.1(2.5, 3.9)	-1.966^a	0.050	2.9(2.4, 3.4)	2.9(2.5, 3.5)	-0.472^a	0.640
白细胞(10⁹/L)	6.7(5.4, 8.2)	7.7(6.0, 9.7)	-2.986^a	0.003	6.5(5.3, 8.3)	7.3(5.7, 10.0)	-2.618^a	0.009
中性粒细胞(10⁹/L)	4.3(3.2, 5.8)	5.1(3.8, 7.5)	-2.850^a	0.004	4.1(3.2, 5.8)	5.3(3.7, 7.6)	-2.878^a	0.004
淋巴细胞(10⁹/L)	1.5(1.1, 1.9)	1.5(1.2, 1.9)	-0.067^a	0.950	1.5(1.2, 1.9)	1.4(1.1, 1.8)	1.444^a	0.150
单核细胞(10⁹/L)	0.4(0.3, 0.6)	0.5(0.4, 0.8)	-2.642^a	0.008	0.4(0.3, 0.6)	0.6(0.4, 0.8)	-3.202^a	0.001
血小板(10⁹/L)	182.0(148.0, 223.0)	160.5(138.0, 201.0)	2.356^a	0.020	177.0(145.0, 216.0)	153.0(125.0, 192.0)	2.807^a	0.005
血红蛋白(g/L)	133.0(120.0, 144.0)	129.5(120.0, 139.0)	1.199^a	0.230	132.0(120.0, 143.5)	132.0(119.0, 142.0)	0.557^a	0.580
甘油三酯(mmol/L)	1.4(1.0, 2.0)	1.3(0.9, 1.9)	0.867^a	0.390	1.4(1.0, 2.0)	1.1(0.9, 1.8)	2.826^a	0.005
总胆固醇(mmol/L)	4.3(3.6, 5.1)	3.7(3.3, 4.6)	3.094^a	0.002	4.4(3.6, 5.2)	3.8(3.0, 4.4)	4.056^a	0.000
总蛋白(g/L)	65.9(61.8, 69.4)	64.8(61.9, 68.9)	0.821^a	0.410	65.7(61.7, 70.1)	64.7(60.1, 69.0)	1.440^a	0.150
白蛋白(g/L)	38.6(36.3, 40.8)	37.8(33.9, 39.6)	2.770^a	0.006	38.7(36.5, 41.0)	37.6(34.4, 39.5)	3.010^a	0.003
球蛋白(g/L)	27.2(24.7, 29.8)	28.2(25.6, 30.5)	-1.813^a	0.070	27.0(24.3, 30.0)	27.0(24.0, 29.9)	-0.032^a	0.9700
肾小球滤过率 (ml·min^-1·l^-1)	84.4(63.0, 95.2)	64.1(50.2, 84.0)	4.338^a	0.000	82.6(62.7, 94.0)	56.5(40.9, 76.5)	6.004^a	0.000
血肌酐(μmol/l)	74.0(63.0, 88.1)	86.9(75.0, 110.0)	-4.280^a	0.000	73.7(62.7, 89.0)	92.2(72.7, 124.5)	-5.019^a	0.000
超声心动图检查
左心房横径(mm)	36.0(33.0, 40.0)	40.0(38.0, 44.0)	-7.027^a	0.000	36.5(33.0, 40.0)	42.0(39.0, 47.0)	-7.313^a	0.000
右心房横径(mm)	34.0(31.0, 37.0)	38.0(35.0, 43.0)	-6.542^a	0.000	34.0(31.0, 37.0)	37.0(34.0, 43.0)	-5.288^a	0.000
左心室射血分数(%)	60.0(56.0, 65.0)	58.0(50.0, 65.0)	1.868^a	0.060	61.0(55.0, 66.0)	56.0(45.0, 63.0)	3.316^a	0.001
PCI治疗^d [n(%)]	336(37.5)	19(30.6)	1.168^b	0.280	310(34.8)	16(24.6)	2.772^b	0.100
注：^a：Wilcoxon秩和检验；^b：Pearson卡方检验；^c：连续校正的Pearson卡方检验；^dPCI治疗：经皮冠状动脉介入(percutaneous coronary intervention)治疗。

下载: 导出CSV

表 2 纳入模型的预测因子及统计检验结果

Table 2. Prediction factors and statistical results included in the model

截距和变量	β值	OR (95% CI)值	P值
截距	-16.545		< 0.001
年龄(岁)	0.043	1.044(1.011~1.078)	0.008
左心房横径(mm)	0.071	1.074(1.018~1.132)	0.008
右心房横径(mm)	0.098	1.102(1.040~1.169)	0.001
心衰分级
< 2		1.000
≥2	0.678	1.970(1.049~3.701)	0.035
入院时心率(次/min)
< 85		1.000
≥85	0.995	2.705(1.434~5.103)	0.002
NT-proBNP(log10转换)	1.202	3.328(1.903~5.821)	< 0.001
中性粒细胞计数	0.098	1.103(1.003~1.214)	0.044

下载: 导出CSV

参考文献(19)

[1]	段文涛, 张峰, 史东, 等.心房颤动相关危险因素及药物治疗研究新进展[J].武汉大学学报(医学版), 2020, 41(1):164-168. DOI: 10.14188/j.1671-8852.2018.0771. Duan WT, Zhang F, Shi D, et al. Recent advances in risk factors and drug therapy of atrial fibrillation[J]. Med J Wuhan University, 2020, 41(1):164-168. DOI: 10.14188/j.1671-8852.2018.0771.
[2]	Smith JN, Negrelli JM, Manek MB, et al. Diagnosis and management of acute coronary syndrome: an evidence-based update[J]. J Am Board Fam Med, 2015, 28(2):283-293. DOI: 10.3122/jabfm.2015.02.140189.
[3]	González-Pacheco H, Márquez MF, Arias-Mendoza A, et al. Clinical features and in-hospital mortality associated with different types of atrial fibrillation in patients with acute coronary syndrome with and without ST elevation[J]. J Cardiol, 2015, 66(2):148-154. DOI: 10.1016/j.jjcc.2014.11.001.
[4]	Wang CL, Chen PC, Juang HT, et al. Adverse outcomes associated with pre-existing and new-onset atrial fibrillation in patients with acute coronary syndrome: a retrospective cohort study[J]. Cardiol Ther, 2019, 8(1):117-127. DOI: 10.1007/s40119-019-0136-3.
[5]	Galvão Braga C, Ramos V, Vieira C, et al. New-onset atrial fibrillation during acute coronary syndromes: predictors and prognosis[J]. Rev Port Cardiol, 2014, 33(5):281-287. DOI: 10.1016/j.repc.2013.10.017.
[6]	Almendro-Delia M, Valle-Caballero MJ, Garcia-Rubira JC, et al. Prognostic impact of atrial fibrillation in acute coronary syndromes: results from the ARIAM registry[J]. Eur Heart Acute Cardiovasc Care, 2014, 3(2):141-148. DOI: 10.1177/2048872613517370.
[7]	Nagai M, Itoh T, Ishida M, et al. New-onset atrial fibrillation in patients with acute coronary syndrome may be associated with worse prognosis and future heart failure[J]. J Arrhythm, 2019, 35(2):182-189. DOI: 10.1002/joa3.12154.
[8]	Thygesen K, Alpert JS, Jaffe AS, et al. Third universal definition of myocardial infarction[J]. J Am Col Cardiol, 2012, 60(16):1581-1598. DOI: 10.1016/j.jacc.2012.08.001.
[9]	Camm AJ, Kirchhof P, Lip GY, et al. Guidelines for the management of atrial fibrillation: the task force for the management of atrial fibrillation of the European Society of Cardiology (ESC)[J]. Eur Heart J, 2010, 31(19):2369-2429. DOI: 10.1093/eurheartj/ehq278.
[10]	Dorje T, Wang X, Shao M, et al. Plasma N-terminal pro-brain natriuretic peptide levels predict new-onset atrial fibrillation in patients with acute myocardial infarction[J]. Int J Cardiol, 2013, 168(3):3135-3137. DOI: 10.1016/j.ijcard.2013.04.032.
[11]	Zhang H, Dong P, Yang X, et al. Prognostic indicators of new onset atrial fibrillation in patients with acute coronary syndrome[J]. Clin Cardiol, 2020, 43(6):647-651. DOI: 10.1002/clc.23363.
[12]	Ulus T, Isgandarov K, Yilmaz AS, et al. Predictors of new-onset atrial fibrillation in elderly patients with acute coronary syndrome undergoing percutaneous coronary intervention[J]. Aging Clin Exp Res, 2018, 30(12):1475-1482. DOI: 10.1007/s40520-018-0926-9.
[13]	Holl MJ, Van Den Bos EJ, Van Domburg RT, et al. NT-proBNP is associated with mortality and adverse cardiac events in patients with atrial fibrillation presenting to the emergency department[J]. Clini Cardiol, 2018, 41(3):400-405. DOI: 10.1002/clc.22883.
[14]	Mazzone A, Scalese M, Paradossi U, et al. Development and validation of a risk stratification score for new-onset atrial fibrillation in STEMI patients undergoing primary percutaneous coronary intervention[J]. Int J Clini Prac, 2018, 72(4):e13087. DOI: 10.1111/ijcp.13087.
[15]	Jensen MT, Pereira M, Araujo C, et al. Heart rate at admission is a predictor of in-hospital mortality in patients with acute coronary syndromes: results from 58 European hospitals: The European Hospital Benchmarking by Outcomes in acute coronary syndrome Processes study[J]. Eur Heart J Acute Cardiovasc Care, 2018, 7(2):49-57. DOI: 10.1177/2048872616672077.
[16]	He J, Yang Y, Zhang G, et al. Clinical risk factors for new-onset atrial fibrillation in acute myocardial infarction: a systematic review and meta-analysis[J]. Med, 2019, 98(26):e15960. DOI: 10.1097/md.0000000000015960.
[17]	Sepehri Shamloo A, Bollmann A, Dagres N, et al. Natriuretic peptides: biomarkers for atrial fibrillation management[J]. Clin Res in Cardiol, 2020, 109(8):957-966. DOI: 10.1007/s00392-020-01608-x.
[18]	Moons KG, Altman DG, Reitsma JB, et al. Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration[J]. Ann Int med, 2015, 162(1):W1-W73. DOI: 10.7326/m14-0698.
[19]	Karabaǧ Y, Rencuzogullari I, Çaǧdaš M, et al. Association between BNP levels and new-onset atrial fibrillation: a propensity score approach[J]. Herz, 2018, 43(6):548-554. DOI: 10.1007/s00059-017-4598-6.