川崎病患儿静脉注射丙种球蛋白抵抗的影响因素分析及Nomogram预测模型构建  

Analysis of factors influencing intravenous immunoglobulin resistance in children with Kawasaki disease and construction of a Nomogram prediction model

作　　者：刘清彪[1] 赵武陵 钟小敏 梁建昌 LIU Qingbiao;ZHAO Wuling;ZHONG Xiaomin;LIANG Jianchang(Department of Pediatrics,Dongguan Tungwah Hospital,Dongguan 523000,Guangdong Province,China)

机构地区：[1]东莞东华医院儿科,广东东莞523000

出　　处：《新乡医学院学报》2025年第4期304-309,共6页Journal of Xinxiang Medical University

摘　　要：目的探讨川崎病患儿静脉注射丙种球蛋白(IVIG)抵抗的影响因素并构建Nomogram预测模型。方法选择2021年1月至2024年1月于东莞东华医院接受IVIG治疗的川崎病患儿220例为研究对象。根据是否发生IVIG抵抗将患者分为抵抗组(n=42)和非抵抗组(n=178)。收集所有患者的临床资料,采用单因素和Lasso-logistic回归分析IVIG抵抗的影响因素。根据影响因素构建不含血清细胞因子的常规Nomogram预测模型和含血清细胞因子的Nomogram预测模型,绘制受试者操作特征(ROC)曲线评估Nomogram预测模型的预测效能。结果川崎病患儿IVIG抵抗发生率为19.09%(42/220)。单因素分析结果显示,抵抗组患者发热持续时间、发生皮疹患者占比、颈部淋巴结肿大患者占比、冠状动脉扩张患者占比及血清C反应蛋白(CRP)、肿瘤坏死因子-α(TNF-α)、干扰素-γ(IFN-γ)水平显著高于非抵抗组,血钠、低密度脂蛋白胆固醇(LDL-C)、白蛋白(ALB)、白细胞介素-10水平显著低于非抵抗组(P<0.05)。Lasso-logistic回归分析结果显示,发热持续时间、血钠、颈部淋巴结肿大、皮疹、LDL-C、ALB、CRP、TNF-α、IFN-γ为川崎病患儿发生IVIG抵抗的影响因素(P<0.05)。常规Nomogram模型预测IVIG抵抗的曲线下面积(AUC)为0.848[95%置信区间(CI):0.790~0.907],含血清细胞因子的Nomogram模型预测IVIG抵抗的AUC为0.893(95%CI:0.845~0.941)。与常规Nomogram模型比较,含血清细胞因子的Nomogram预测模型的AUC明显增大(Z=2.014,P<0.05);含血清细胞因子Nomogram预测模型的净重新分类指数为0.017(95%CI:0.012~0.026,P<0.05)、综合判别改善指数为0.031(95%CI:0.024~0.050,P<0.05)。结论血清细胞因子CRP、TNF-α、IFN-γ是预测川崎病患儿发生IVIG抵抗的有效指标,基于以上血清细胞因子构建的Nomogram预测模型能更准确指导临床医生对川崎病患儿IVIG抵抗发生风险进行个体化预测。Objective To explore the factors influencing intravenous immunoglobulin(IVIG)resistance in children with Kawasaki disease(KD)and construct its Nomogram prediction model.Methods A total of 220 KD children treated with IVIG at Dongguan Tungwah Hospital from January 2021 to January 2024 were selected and then divided into a resistance group(n=42)and non-resistance group(n=178)according to whether IVIG resistance occurred.The clinical data of the patients were collected,and the factors influencing IVIG resistance were analyzed through univariate and Lasso-logistic regression method.Two conventional Nomogram prediction models were constructed based on the influencing factors,one including serum cytokines while the other excluding them.The receiver operating characteristic(ROC)curve was used to assess the performance of Nomogram prediction models.Results The incidence of IVIG resistance in KD children was 19.09%(42/220).The univariate analysis showed that the duration of fever,incidence of rash,incidence of cervical lymph node enlargement,incidence of coronary artery dilation,serum C-reactive protein(CRP)levels,tumor necrosis factor-α(TNF-α)levels,interferon-γ(IFN-γ)levels in the resistance group were significantly higher than those in the non-resistance group(P<0.05).The levels of serum sodium,low density lipoprotein cholesterol(LDL-C),albumin(ALB),and interleukin-10(IL-10)levels in the resistance group were significantly higher than those in the non-resistance group(P<0.05).Lasso-logistic regression analysis showed that the duration of fever,serum sodium,cervical lymph node enlargement,rash,LDL-C,ALB,CRP,TNF-α,and IFN-γwere factors influencing the development of IVIG resistance in KD children(P<0.05).The area under the curve(AUC)of the conventional Nomogram model for predicting IVIG resistance was 0.848[95%confidence interval(CI):0.790-0.907],while that of the Nomogram prediction model including serum cytokines for IVIG resistance prediction was 0.893(95%CI:0.845-0.941).Compared with the conventional Nomogram p

关 键 词：川崎病 儿童 丙种球蛋白抵抗 细胞因子 Nomogram预测模型 

分 类 号：R725.4[医药卫生—儿科]