摘要
目的 调查分析急性心肌梗死(AMI)患者创伤后应激障碍(PTSD)的发生率和主要危险因素,构建可视化的风险列线图预测模型并进行效能验证,为临床实践提供可操作的量化工具。方法 采用单中心、回顾性队列观察的方法,选择2020年1月—2022年1月南通大学附属如皋医院诊断初发AMI患者224例为建模集,选择2022年4月—2023年4月初发AMI患者75例为验证集。出院后3个月将其分为PTSD组与无PTSD组,比较两组人口学资料和临床资料,采用多因素Logistic回归模型筛选PTSD的危险因素,并构建列线图模型。内部验证Bootstrap法重复抽样1 000次计算一致性指数,校准曲线和受试者工作特征(ROC)曲线分别进行内部和外部验证。结果 建模集PTSD发生率为29.5%(66/224),验证集PTSD发生率为33.3%(25/75),比较无统计学差异(χ^(2)=0.397,P=0.529)。建模集中PTSD组女性、吸烟和糖尿病史比例较无PTSD组明显增多,肌酸激酶同工酶(CK-MB)、失眠评分和对疾病进展的恐惧评分较无PTSD组显著升高,而有配偶和家庭人均月收入较无PTSD组减少,左心室射血分数(LVEF)较无PTSD组降低(P<0.05)。Logistic回归显示,女性(OR=2.815,95% CI=2.124~3.658,P<0.001)、糖尿病史(OR=1.385,95% CI=1.032~1.859,P=0.003)、CK-MB水平(OR=1.203,95% CI=1.009~1.756,P=0.006)、失眠评分(OR=2.125,95% CI=1.653~2.754,P<0.001)和对疾病进展的恐惧评分(OR=2.462,95% CI=1.856~2.936,P<0.001)均是影响AMI患者出院后3个月PTSD的独立危险因素。R软件建立列线图,总分220分。内部验证显示,建模集列线图的一致性指数为0.825。校准曲线显示,列线图预测PTSD的概率与建模集和验证集PTSD的实际发生率有较好的吻合度。ROC曲线分析显示,列线图预测PTSD的曲线下面积(AUC)在建模集和验证集分别为0.902(95% CI=0.823~0.936,P<0.001)和0.863(95% CI=0.812~0.923,P<0.001)。结论 AMI患者出院后3个月PTSD的发生率较高,女性、糖尿病史、CK-MB水平、失眠
Objective To investigate and analyze the incidence and main risk factors of post traumatic stress disorder(PTSD) in patients with acute myocardial infarction(AMI),construct a visualized risk nomogram predictive model,and validate its effectiveness,so as to provide an operable quantitative tool for clinical practice.Methods The single center,retrospective cohort observative study was conducted,224 firstly diagnosed of AMI patients in Rugao Hospital affiliated to Nantong University from January 2020 to January 2022 were selected as the model set,other 75 firstly diagnosed AMI patients from April 2022 to April 2023 were selected as the validation set.Three months after discharge,they were divided into the PTSD group and non-PTSD group,the demography data and clinical data between the groups were compared,then multivariate Logistic regression model was used to screen risk factors to PTSD,and construct a nomogram model.Internal validation was conducted using the Bootstrap method with 1 000repeated samples to calculate the consistency index.The calibration curve and receiver operating characteristic(ROC)curve were used for internal and external validation,respectively.Results The incidence of PTSD in the model set was29.5%(66/224),while in the validation set it was 33.3%(25/75).There was no significant difference in the incidence of PTSD(χ~2=0.397,P=0.529).Univariate comparison found that women,smoking and diabetes history in PTSD group of the model set were significantly more than those in non-PTSD group,creatine kinase isozyme(CK-MB),insomnia score and fear of disease progression score were significantly higher,while spouse and monthly household income were less,and left ventricular ejection fraction(LVEF) was lower,too(P<0.05).Logistic regression showed that women(OR=2.815,95% CI=2.124-3.658,P<0.001),history of diabetes(OR=1.385,95% CI=1.032-1.859,P=0.003),CK-MB level(OR=1.203,95% CI=1.009-1.756,P=0.006),insomnia score(OR=2.125,95% CI=1.653-2.754,P<0.001) and fear of disease progression score(OR=2.462,95% CI=1.856-
作者
陈培培
李雯
沙飞燕
CHEN Peipei;LI Wen;SHA Feiyan(Department of Cardiology,Rugao Hospital Affiliated to Nantong University,Rugao People's Hospital,Rugao Jiangsu 226500,China)
出处
《中国急救复苏与灾害医学杂志》
2024年第11期1413-1417,共5页
China Journal of Emergency Resuscitation and Disaster Medicine
基金
江苏省“六大人才高峰”高层次人才选拔培养资助项目(编号:WSN-062)。
