高血压病患者合并心房颤动的临床诊断模型的建立与验证

Establishmentand verificationof clinical diagnosis model in patients with hypertension complicated by atrial fibrillation

目的分析高血压病患者发生心房颤动(房颤)的影响因素,构建列线图模型,并分析该预测模型的准确性和临床有效性。方法分析2018年1月至2021年4月于广安门医院收治的748例高血压病患者的临床资料。按照是否合并患有房颤分为房颤组(n=165)和非房颤组(n=619)。应用Logistic回归分析高血压病患者发生房颤的影响因素,进一步构建列线图模型,通过Bootstrap方法对构建的模型进行内部验证,评价该模型的准确性。结果多因素Logistic回归分析显示,年龄(OR=1.040,95%CI:1.016~1.064,P<0.05)、女性(OR=1.720,95%CI:1.066~2.777,P<0.05)、三酰甘油(OR=0.717,95%CI:0.517~0.996,P<0.05)、胆固醇(OR=2.387,95%CI:1.105~5.159,P<0.05)、左心室舒张功能障碍(OR=0.121,95%CI:0.073~0.201,P<0.001)、左心房直径(OR=1.110,95%CI:1.072~1.149,P<0.001)、β受体阻滞剂的使用(OR=1.881,95%CI:1.177~3.007,P<0.05)和血清胆碱酯酶含量(OR=0.723,95%CI:0.639~0.819,P<0.001)是高血压患者发生房颤的独立危险因素。列线图模型显示,基于临床数据构建的列线图模型在预测高血压病患者发生房颤的风险能力较强。通过Bootstrap方法验证列线图模型发现校准曲线与理想曲线贴合良好。结论基于临床常见数据构建的风险预测模型可以较为准确的判断高血压病患者发生房颤的风险,有助于临床医师判断患者的情况,并提供及时有效的干预措施。

Objective To analyze the factors affecting atrial fibrillation(AF)in patients with hypertension,establish nomogram model,and analyze the accuracy and clinical efficacy of the model.Methods The clinical materials of hypertensive patients(n=748)were analyzed in Guang`anmen Hospital from Jan.2018 to Apr.2021.The patients were divided,according to AF occurrence,into AF group(n=165)and non-AF group(n=619).The influence factors of AF were analyzed by using Logistic regression analysis.A nomogrammodel was established and verified internally for reviewing accuracy by using Bootstrap method.Results The results of multi-factor Logistic regression analysis showed that age(OR=1.040,95%CI:1.016~1.064,P<0.05),female(OR=1.720,95%CI:1.066~2.777,P<0.05),triglyceride(OR=0.717,95%CI:0.517~0.996,P<0.05),cholesterol(OR=2.387,95%CI:1.105~5.159,P<0.05),left ventricular diastolic dysfunction(OR=0.121,95%CI:0.073~0.201,P<0.001)left atrial diameter(OR=1.110,95%CI:1.072~1.149,P<0.001),β-receptor blocker administration(OR=1.881,95%CI:1.177~3.007,P<0.05)and serum cholinesterase content(OR=0.723,95%CI:0.639~0.819,P<0.001)were independent risk factors of AF occurrence in hypertensive patients.The results of nomogrammodel analysis showed that the nomogrammodel based on clinical data had a stronger capability in predicting AF occurrence in hypertensive patients.Bootstrapvalidation confirmed that the model’s calibration curve closely aligned with ideal curve.Conclusion The risk predictive model based on common clinical data can precisely estimate AF risk in hypertensive patients,which is beneficial for clinicians in assessing patient’s conditions and implementingtimely and effective interventions.