中国组织工程研究

中国组织工程研究 ›› 2025, Vol. 29 ›› Issue (15): 3215-3226.doi: 10.12307/2025.802

• 骨与关节图像与影像Bone and joint imaging • 上一篇    下一篇

系统性红斑狼疮并发股骨头坏死危险因素列线图预测模型的建立和验证

徐文博1，汪利合2，李松伟2，3，史鹏博2   

  1. 1河南中医药大学，河南省郑州市   450046；2河南中医药大学第一附属医院，河南省郑州市   450000；3河南省中医院，河南省郑州市   450002
  • 收稿日期:2024-04-01 接受日期:2024-06-17 出版日期:2025-05-28 发布日期:2024-11-05
  • 通讯作者: 汪利合，主任医师，硕士生导师，河南中医药大学第一附属医院，河南省郑州市 450000
  • 作者简介:徐文博，男，1996年生，汉族，河南省洛阳市人，河南中医药大学在读硕士，主要从事中医药防治骨关节病方面的研究。
  • 基金资助:
    河南省中医药拔尖人才培养项目(豫中医科教[2018]35号)，项目负责人：汪利合；中国博士后科学基金面上项目 (2022M721066)，项目负责人：史鹏博；河南省重点研发与推广专项(232102310474)，项目负责人：史鹏博

Establishment and validation of a Nomogram prediction model for risk factors of osteonecrosis of the femoral head in systemic lupus erythematosus

Xu Wenbo1, Wang Lihe2, Li Songwei2, 3, Shi Pengbo2   

  1. 1Henan University of Chinese Medicine, Zhengzhou 450046, Henan Province, China; 2First Affiliated Hospital of Henan University of Chinese Medicine, Zhengzhou 450000, Henan Province, China; 3Henan Provincial Hospital of TCM, Zhengzhou 450002, Henan Province, China
  • Received:2024-04-01 Accepted:2024-06-17 Online:2025-05-28 Published:2024-11-05
  • Contact: Wang Lihe, Chief physician, Master’s supervisor, First Affiliated Hospital of Henan University of Chinese Medicine, Zhengzhou 450000, Henan Province, China
  • About author:Xu Wenbo, Master candidate, Henan University of Chinese Medicine, Zhengzhou 450046, Henan Province, China
  • Supported by:
    Henan Province Traditional Chinese Medicine Top Talent Training Project, No. Yu Traditional Chinese Medicine Science and Education [2018]35 (to WLH); General Program of China Postdoctoral Science Foundation, No. 2022M721066 (to SPB); Key Research & Development and Promotion Project in Henan Province, No. 232102310474 (to SPB) 

PDF

206

摘要：

文题释义：
股骨头坏死：作为骨科一种致残率很高的疾病，其特征是由于股骨头周围动静脉循环的中断，不足以营养骨细胞及骨髓成分，造成其部分死亡，从而引起骨组织坏死及修复，使股骨头结构改变及塌陷，最终引起髋关节疼痛和功能障碍。股骨头坏死作为系统性红斑狼疮患者常见的并发症之一，严重影响患者的生活质量。
列线图预测模型：列线图又称诺莫图，可以将复杂的回归方程呈现为简单易懂的可视化图形，目前广泛应用于疾病危险因素和疾病诊断的研究。其基于多因素回归分析结果模型，在模型中根据每个危险因素对结局变量的影响进行量化赋分，因此可以方便地在列线图上计算每个危险因素的得分和总分，进而计算出该个体结局事件的预测值。

摘要
背景：股骨头坏死是系统性红斑狼疮患者常见的并发症，若能早期对其发生风险进行预测与验证，将有助于避免或延缓股骨头坏死的发展。
目的：分析系统性红斑狼疮患者并发股骨头坏死的危险因素，构建系统性红斑狼疮患者并发股骨头坏死的列线图预测模型并进行验证。
方法：回顾性分析2013年1月至2022年12月首次就诊于河南中医药大学第一附属医院的914例系统性红斑狼疮患者的病历资料，根据是否发生股骨头坏死分为发生股骨头坏死组(n=100)和未发生股骨头坏死组(n=814)。采用单因素、LASSO回归和多因素Logistic回归分析筛选和确定系统性红斑狼疮并发股骨头坏死的危险因素。同时将数据集按照7∶3的比例随机分为训练集和测试集，并基于多因素Logistic回归分析结果，构建系统性红斑狼疮并发股骨头坏死的列线图预测模型。同时，使用受试者工作特征曲线、Hosmer-Lemeshow校准曲线和决策曲线对列线图的性能进行评估。
结果与结论：①股骨头坏死组与未发生股骨头坏死组患者在系统性红斑狼疮病程、系统性红斑狼疮疾病活动度评分、狼疮性肾炎、呼吸系统受累、胃肠道受累、干燥综合征、骨质疏松、抗核糖核蛋白抗体阳性、补体C3降低、环磷酰胺、吗替麦考酚酯、生物抑制剂、糖皮质激素最大日剂量、糖皮质激素冲击治疗方面差异有显著性意义(P < 0.05)；②采用LASSO回归分析方法筛选出10个与系统性红斑狼疮并发股骨头坏死风险相关的预测变量，将其纳入多因素Logistic回归分析，结果显示系统性红斑狼疮病程、呼吸系统受累、干燥综合征、骨质疏松、抗核糖核蛋白抗体阳性、环磷酰胺、吗替麦考酚酯、生物抑制剂、糖皮质激素最大日剂量是系统性红斑狼疮患者发生股骨头坏死的独立危险因素(P < 0.05)；③训练集中预测发生风险的受试者工作特征曲线下面积为0.802(95%CI=0.742-0.862)，测试集预测发生股骨头坏死风险受试者工作特征曲线下面积为0.811(95%CI=0.745-0.876)；Hosmer-Lemeshow校准曲线拟合度较好(训练集，P=0.447；验证集，P=0.870)；决策曲线显示使用列线图预测模型预测系统性红斑狼疮患者发生股骨头坏死的风险是有益的；④月经异常为女性系统性红斑狼疮患者并发股骨头坏死的危险因素之一；⑤此次研究结果提示，系统性红斑狼疮并发股骨头坏死的危险因素是多因素的，同时建立了一个包含9个危险因素的列线图预测模型，可将其用于预测系统性红斑狼疮患者发生股骨头坏死的风险；此外，首次报道了月经异常为女性系统性红斑狼疮并发股骨头坏死的危险因素之一。



关键词: 系统性红斑狼疮, 股骨头坏死, 危险因素, 列线图, 预测模型, 月经异常, LASSO回归, 多因素Logistic回归

Abstract: BACKGROUND: Osteonecrosis of the femoral head is a common complication in patients with systemic lupus erythematosus. The prediction and validation of the risk in advance will help to avoid or delay the progression of osteonecrosis of the femoral head.
OBJECTIVE: To analyze risk factors for the occurrence of osteonecrosis of the femoral head in systemic lupus erythematosus patients and to construct and validate a nomogram prediction model of systemic lupus erythematosus patients with osteonecrosis of the femoral head. 
METHODS: A retrospective study was conducted to analyze the medical records of 914 systemic lupus erythematosus patients who first visited First Affiliated Hospital of Henan University of Chinese Medicine between January 2013 and December 2022. All patients were divided into osteonecrosis of the femoral head (n=100) and non-osteonecrosis of the femoral head (n=814) groups according to whether they had suffered from osteonecrosis of the femoral head or not. Univariate, LASSO regression, and multifactorial logistic regression analyses were used to screen and identify the risk factors for systemic lupus erythematosus complicating osteonecrosis of the femoral head. The dataset was also randomly divided into training and test sets in a ratio of 7:3. A nomogram prediction model of the risk of systemic lupus erythematosus complicating osteonecrosis of the femoral head was constructed based on the results of the multifactorial logistic regression analysis. The performance of the nomogram was evaluated using the receiver operating characteristic curve, Hosmer-Lemeshow calibration curve, and decision curve analysis.
RESULTS AND CONCLUSION: (1) There were significant differences in disease duration of systemic lupus erythematosus, systemic lupus erythematosus disease activity, lupus nephritis, respiratory involvement, gastrointestinal involvement, Sjögren's syndrome, osteoporosis, anti-ribonucleoprotein, complement C3 decrease, cyclophosphamide, mycophenolate mofetil, biologics, maximum daily dose of glucocorticosteroids, and pulses of intravenous methylprednisolone between the osteonecrosis of the femoral head and non-osteonecrosis of the femoral head groups (P < 0.05). (2) Ten predictor variables related to the risk of osteonecrosis of the femoral head in patients with systemic lupus erythematosus were screened using LASSO regression analysis. Multivariate logistic regression analysis further confirmed disease duration of systemic lupus erythematosus, respiratory involvement, Sjögren's syndrome, osteoporosis, anti-ribonucleoprotein, cyclophosphamide, mycophenolate mofetil, biologics, and maximum daily dose of glucocorticosteroids were independent risk factors for osteonecrosis of the femoral head in systemic lupus erythematosus patients (P < 0.05). (3) The area under the receiver operating characteristic curve for predicting the risk of occurrence of osteonecrosis of the femoral head in systemic lupus erythematosus patients was 0.802 (95%CI=0.742-0.862) in the training set and 0.811 (95%CI=0.745-0.876) in the testing set. The Hosmer-Lemeshow calibration curve fit was well (P=0.447 in raining set; P=0.870 in testing set). Decision curve analysis showed that it was beneficial in predicting the risk of osteonecrosis of the femoral head in systemic lupus erythematosus patients using the nomogram prediction model. (4) Menstrual abnormalities were one of the risk factors for osteonecrosis of the femoral head in female systemic lupus erythematosus patients. (5) The results suggest that the risk factors for systemic lupus erythematosus complicating osteonecrosis of the femoral head are multi-factorial, and a nomogram prediction model containing nine risk factors was also developed, which could be used to predict the risk of osteonecrosis of the femoral head in systemic lupus erythematosus patients. In addition, we reported for the first time that menstrual abnormalities were one of the risk factors for systemic lupus erythematosus complicating osteonecrosis of the femoral head in female.


Key words: systemic lupus erythematosus, osteonecrosis of the femoral head, risk factor, nomogram, prediction model, menstrual abnormality, LASSO regression, multivariate logistic regression

中图分类号: