Yin Hongqiang, Su Ruijuan, Liu Dongmei, Deng Yawen, Ma Ning
Heart Center, National Center for Children's Health, Beijing Children's Hospital, Capital Medical University, No. 56 Nanlishi Avenue, Xicheng District, Beijing, 100045, China.
Department of Ultrasound, Beijing-Shijitan Hospital, Capital Medical University, No. 10 Tieyi Road, Yangfangdian, Haidian District, Beijing, 100038, China.
Eur J Pediatr. 2025 May 22;184(6):355. doi: 10.1007/s00431-025-06181-x.
This study aimed to identify risk factors for the progression of coronary artery lesions (CALs) in children with Kawasaki disease (KD) and to establish a nomogram for predicting this risk. We retrospectively analyzed clinical and echocardiographic data from KD patients diagnosed at Beijing Children's Hospital from 1 January 2021 to 30 December 2023.The patients were categorized into the progression and non-progression groups on the basis of coronary artery Z-scores and diameters at the 1-month follow-up compared with baseline. Univariate logistic regression identified significant indicators, supplemented by factors from the literature. We used full permutation to examine potential combinations, followed by multivariate logistic regression to calculate the Akaike information criterion (AIC) and area under the curve (AUC) for each model. We selected the best values for establishing a prediction score and nomogram. Model performance was assessed using the AUC, calibration curves, and tenfold cross-validation. Among 1249 patients, 183 (14.7%) experienced progression of CALs, while 1066 (85.3%) showed improvement or stability. Eight independent factors were identified: the baseline maximum Z-score, age, percentage of neutrophils, hemoglobin concentrations, erythrocyte sedimentation rate, albumin, fibrinogen, and intravenous immunoglobulin resistance. The nomogram model showed an AUC of 0.788, with a mean AUC of 0.775 and an accuracy of 85.6% after tenfold cross-validation.
The baseline maximum Z-score, age, percentage of neutrophils, hemoglobin concentrations, erythrocyte sedimentation rate, albumin, fibrinogen, and intravenous immunoglobulin resistance are predictive factors for CALs progression in KD. The established nomogram shows high accuracy and reliability, aiding clinicians in decision-making.
• Since the introduction of IVIG therapy, most children with KD show CALs regression, yet a subset experience progressive CALs despite treatment. • CALs progression is associated with increased adverse cardiovascular events, yet predictors of this progression remain poorly characterized.
• The eight-factor predictive model developed in this study effectively identifies progression risks in CALs following treatment, providing a basis for personalized clinical management. • Echocardiography, the primary modality for assessing coronary arteries in children, demonstrates that early baseline Z-score evaluation serves as the strongest predictor for CALs progression, while non-coronary cardiac abnormalities show no significant association.
本研究旨在确定川崎病（KD）患儿冠状动脉病变（CALs）进展的危险因素，并建立预测该风险的列线图。我们回顾性分析了2021年1月1日至2023年12月31日在北京儿童医院确诊的KD患者的临床和超声心动图数据。根据1个月随访时与基线相比的冠状动脉Z评分和直径，将患者分为进展组和非进展组。单因素逻辑回归确定了显著指标，并辅以文献中的因素。我们使用全排列来检验潜在组合，然后进行多因素逻辑回归以计算每个模型的赤池信息准则（AIC）和曲线下面积（AUC）。我们选择最佳值来建立预测评分和列线图。使用AUC、校准曲线和十折交叉验证评估模型性能。在1249例患者中，183例（14.7%）出现CALs进展，而1066例（85.3%）病情改善或稳定。确定了八个独立因素：基线最大Z评分、年龄、中性粒细胞百分比、血红蛋白浓度、红细胞沉降率、白蛋白、纤维蛋白原和静脉注射免疫球蛋白抵抗。列线图模型的AUC为0.788，十折交叉验证后的平均AUC为0.775，准确率为85.6%。
基线最大Z评分、年龄、中性粒细胞百分比、血红蛋白浓度、红细胞沉降率、白蛋白、纤维蛋白原和静脉注射免疫球蛋白抵抗是KD患儿CALs进展的预测因素。所建立的列线图显示出高准确性和可靠性，有助于临床医生进行决策。
• 自引入静脉注射免疫球蛋白（IVIG）治疗以来，大多数KD患儿的CALs会消退，但仍有一部分患儿尽管接受了治疗，CALs仍会进展。• CALs进展与不良心血管事件增加有关，但这种进展的预测因素仍未得到充分描述。
• 本研究开发的八因素预测模型有效地识别了治疗后CALs的进展风险，为个性化临床管理提供了依据。• 超声心动图是评估儿童冠状动脉的主要方式，表明早期基线Z评分评估是CALs进展的最强预测因素，而非冠状动脉心脏异常无显著关联。
