Feng Min, Meng Fanxing, Jia Yuhan, Wang Yanlin, Ji Guozhen, Gao Chong, Luo Jing
Department of Rheumatology, the Second Hospital of Shanxi Medical University, Taiyuan, Shanxi, China.
Shanxi Medical University, Taiyuan, Shanxi, China.
Inflammation. 2024 Oct 16. doi: 10.1007/s10753-024-02157-5.
Patients with rheumatoid arthritis (RA) have increased mortality and morbidity rates owing to cardiovascular diseases (CVD). Timely detection of CVD in RA can greatly improve patient prognosis; however, this technique remains challenging. We aimed to investigate the risk factors for CVD incidence in patients with RA.
This retrospective study included RA patients without CVD risk factors (n = 402), RA with CVD risk factors (n = 394), and RA with CVD (n = 201). Their data on routine examination indicators, vascular endothelial growth factor (VEGF), and immune cells were obtained from medical records. The characteristic variables between each group were screened using univariate analysis, least absolute shrinkage and selection operator (LASSO), random forest (RF), and logistic regression (LR) models, and individualized nomograms were further established to more conveniently observe the likelihood of CVD in RA.
Univariate analysis revealed significantly elevated levels of white blood cells (WBC), blood urea nitrogen (BUN), creatinine, creatine kinase (CK), lactate dehydrogenase (LDH), VEGF, serum total cholesterol (TC), triglycerides (TG), low-density lipoprotein (LDL), apolipoprotein B100 (ApoB100), and apolipoprotein E (ApoE) in RA patients with CVD, whereas apolipoprotein A1 (ApoA1) and high-density lipoprotein/cholesterol (HDL/TC) were decreased. Furthermore, the ratio of regulatory T (Treg) cells exhibiting excellent separation performance in RA patients with CVD was significantly lower than that in other groups, whereas the ratios of Th1/Th2/NK and Treg cells were significantly elevated. The LASSO, RF, and LR models were also used to identify the risk factors for CVD in patients with RA. Through the final selected indicators screened using the three machine learning models and univariate analysis, a convenient nomogram was established to observe the likelihood of CVD in patients with RA.
Serum lipids, lipoproteins, and reduction of Treg cells have been identified as risk factors for CVD in patients with RA. Three nomograms combining various risk factors were constructed to predict CVD occurring in patients with RA (RA with/without CVD risk factors).
类风湿关节炎（RA）患者因心血管疾病（CVD）导致死亡率和发病率增加。及时检测RA患者的CVD可大大改善患者预后；然而，这项技术仍然具有挑战性。我们旨在研究RA患者发生CVD的危险因素。
这项回顾性研究纳入了无CVD危险因素的RA患者（n = 402）、有CVD危险因素的RA患者（n = 394）和患有CVD的RA患者（n = 201）。他们的常规检查指标、血管内皮生长因子（VEGF）和免疫细胞数据来自病历。使用单因素分析、最小绝对收缩和选择算子（LASSO）、随机森林（RF）和逻辑回归（LR）模型筛选每组之间的特征变量，并进一步建立个体化列线图，以更方便地观察RA患者发生CVD的可能性。
单因素分析显示，患有CVD的RA患者白细胞（WBC）、血尿素氮（BUN）、肌酐、肌酸激酶（CK）、乳酸脱氢酶（LDH）、VEGF、血清总胆固醇（TC）、甘油三酯（TG）、低密度脂蛋白（LDL）、载脂蛋白B100（ApoB100）和载脂蛋白E（ApoE）水平显著升高，而载脂蛋白A1（ApoA1）和高密度脂蛋白/胆固醇（HDL/TC）降低。此外，在患有CVD的RA患者中表现出优异区分性能的调节性T（Treg）细胞比例显著低于其他组，而Th1/Th2/NK和Treg细胞比例显著升高。LASSO、RF和LR模型也用于识别RA患者发生CVD的危险因素。通过使用三种机器学习模型和单因素分析筛选出的最终选定指标，建立了一个方便的列线图，以观察RA患者发生CVD的可能性。
血清脂质、脂蛋白和Treg细胞减少已被确定为RA患者发生CVD的危险因素。构建了三个结合各种危险因素的列线图，以预测RA患者（有/无CVD危险因素的RA）发生CVD的情况。
