类风湿性关节炎患者发生冠心病风险列线图预测模型的建立

Establishment of nomogram prediction model of coronary heart disease risk in patients with rheumatoid arthritis

目的类风湿性关节炎(RA)是动脉粥样硬化形成的危险因素。文中旨在探讨建立RA患者发生冠心病(CHD)的列线图预测模型。方法回顾性分析2012年3月至2018年9月常州市第二人民医院风湿免疫科收治的315例RA患者的临床资料。根据是否发生CHD分为RA+CHD组(n=32)和RA组(n=283),统计分析2组一般资料、实验室指标等;采用单因素分析以及多因素Logistic回归筛选RA患者发生CHD的危险因素,联合各独立因素构建RA患者发生CHD的列线图预测模型,并对模型的预测性和区分度进行内部验证。结果 Logistic回归分析结果显示尿酸(OR=1.006, 95%CI:1.001~1.011)、颈-股脉搏波传播速度(OR=1.437, 95%CI:1.032~2.000)、颈动脉斑块积分(OR=1.277, 95%CI:1.027~1.586)、心电图改变(OR=2.623, 95%CI:1.074~6.405)、RA病程(OR=1.142, 95%CI:1.025~1.272)、糖化血红蛋白(OR=1.588, 95%CI:1.020~2.473)以及高血压(OR=2.575, 95%CI:1.073~6.175)是RA患者发生CHD的危险因素(P<0.05)。利用上述指标构建列线图预测模型,对RA患者发生CHD的列线图模型进行内部验证,C-index指数为0.89(0.825~0.903),校正后ROC曲线下面积为0.867(95%CI:0.825~0.903)。结论 RA患者发生CHD的列线图预测模型预测能力和区分能力较好,对RA患者发生CHD具有较高的预测价值。

Objective Rheumatoid arthritis(RA) is a risk factor for the formation of atherosclerosis. This study aims to explore the establishment of a nomogram prediction model for coronary heart disease(CHD) in RA patients. Methods To retrospectively analyze the clinical data of 315 patients with RA admitted to the Department of Rheumatology and Immunology, Changzhou Second People’s Hospital from March 2012 to September 2018. The patients were divided into RA with the CHD group and RA group, respectively. General information, laboratory indicators, and inspection indicators were compared between the two groups, and the independent predictive factors for RA patients at risk of CHD were screened by single-factor analysis and binary multivariate logistic regression analysis. Single-factor analysis and multivariate logistic regression were used to screen risk factors for CHD in RA patients, combined with independent factors to construct a nomogram prediction model for CHD in RA patients, and internally verified the predictability and discrimination of the model. Results The results of Logistic regression analysis showed that uric acid(UA, OR=1.006, 95% CI: 1.001-1.011), carotid-femoral pulse wave velocity(cfPWV, OR=1.437, 95%CI: 1.032-2.000), Crouse plaque score at the cervical artery(OR=1.277, 95%CI:1.027-1.586), electro cardiographic changes(OR=2.623, 95%CI:1.074-6.405), the course of RA disease(OR=1.142, 95%CI: 1.025-1.272), glycosylated hemoglobin(HbA1 c, OR=1.588, 95%CI: 1.020-2.473) and hypertension(OR=2.575, 95%CI:1.073-6.175) were independent predictive factors of RA with CHD(P<0.05). The nomogram prediction model was constructed using the above indicators, and the nomogram model of CHD in RA patients was internally verified. The C-index was 0.89(0.825-0.903) and the ROC curve was drawn. The area under the ROC curve after the correction was 0.867(0.825-0.903). Conclusion The nomogram prediction model for the occurrence of CHD in RA patients has good predictability and discrimination and has high predictive value for the occurrence of CHD in RA patients.