Abstract: Objective　To explore the high-risk factors for the occurrence and development of severe Kawasaki disease（SKD） and the evaluation of its clinical prediction value. Methods　The children with Kawasaki disease were retrospectively analyzed，who were diagnosed in the Heart Center of the Women and Children's Hospital Affiliated to Qingdao University from November 2016 to July 2021. They were divided into the SKD group（18 cases） and the ordinary Kawasaki disease（OKD） group（639 cases） according to whether there was serious organ system damage or unstable vital signs. After matching compare the differences between the two groups in clinical manifestations，auxiliary examinations，response to intravenous immunoglobulin，coronary artery lesion，and length of hospital stay. Stepwise logistic regression was used to explore the high-risk factors for SKD，The ROC curves of single factor and joint prediction were drawn. Results　After propensity score matching，there were 18 cases in the SKD group and 36 cases in the OKD group. There were significant differences between the two groups in terms of multiple serous cavity effusions，non-response to intravenous immunoglobulin and length of hospital stay （P<0.05）. There were significant differences in hemoglobin，C-reactive protein，procalcitonin，serum sodium，serum potassium，total bilirubin，direct bilirubin，aspartate aminotransferase，albumin，globulin，atrial natriuretic peptide precursor and proteinuria between the two groups （P<0.05）. Stepwise logistic regression analysis and ROC curve analysis showed that the increase in C-reactive protein and procalcitonin and the decrease of serum potassium were the risk factors and predictive indicators for the occurrence of SKD （P<0.05）. The sensitivity and specificity of the combination of the three methods in predicting the occurrence of SKD were 94.44% and 80.00%，respectively. The prediction threshold was 3.7 mmol/L for serum potassium，104.7 mg/L for C-reactive protein，and 1.11 ng/mL for procalcitonin，respectively. The predictive value of the three indicators combined was higher than that of C-reactive protein or procalcitonin （P<0.05） alone. Conclusion　The clinical manifestations of SKD are diverse. Multiple serous cavity effusions and non-response to intravenous immunoglobulin often occur. The increase in C-reactive protein and procalcitonin and the decrease in blood potassium are the high-risk factors for SKD.

Key words: Kawasaki disease, severe, child, high-risk factor

摘要： 目的　探讨重症川崎病发生发展的高危因素及其临床预测价值评价。 方法　回顾性分析2016年11月至2021年7月于青岛大学附属妇女儿童医院心脏中心诊断为川崎病的患儿657例，根据是否出现严重器官系统损害或生命体征不平稳分为重症组18例，普通组639例。匹配后对比两组患儿临床表现、辅助检查、丙种球蛋白应答情况、冠状动脉病变及住院时长等资料的差异，采用逐步法进行多因素Logistic回归探讨发生重症川崎病的高危因素，并绘制单因素及联合预测ROC曲线。结果　采用倾向性评分匹配后，重症组患儿18例，普通组患儿36例，两组多浆膜腔积液、丙种球蛋白无应答及住院时间比较，差异有统计学意义（P<0.05）；两组患儿血红蛋白、C反应蛋白、降钙素原、血清钠、血清钾、总胆红素、直接胆红素、天冬氨酸转氨酶、白蛋白、球蛋白、心房钠尿肽前体及蛋白尿比较，差异有统计学意义（P<0.05）。多因素Logistic回归分析及ROC曲线分析显示，C反应蛋白、降钙素原升高及血钾降低是发生重症川崎病的危险因素及预测指标（P<0.05）；三者联合应用预测发生重症川崎病的灵敏度和特异度分别为94.44%和80.00%，预测界值分别血清钾3.7 mmol/L、C反应蛋白 104.7 mg/L及降钙素原 1.11 μg/L；三项联合指标预测价值大于C反应蛋白或降钙素原单个指标（P<0.05）。结论　重症川崎病临床表现多样，常发生多浆膜腔积液及丙种球蛋白无应答；C反应蛋白、降钙素原升高及血钾降低是发生重症川崎病的高危因素。

关键词: 川崎病, 重症, 儿童, 高危因素