Liver injury is a common but nonspecific feature of acute Kawasaki disease (KD), often obscuring the diagnosis and delaying treatment. This study aimed to develop and validate a nomogram to predict KD risk in febrile children with liver injury. We conducted a retrospective study of children admitted with fever and liver injury between January 2019 and December 2024. A nomogram was developed using multivariable logistic regression in a training cohort, internally validated, and further evaluated in an independent temporal validation cohort (January–September 2025). Among 1,031 children, KD was the second most common specific diagnosis (13.4%). Nine independent predictors were identified: duration of fever, absolute neutrophil count, absolute lymphocyte count, C-reactive protein, albumin, gamma-glutamyl transferase, total bilirubin, high-density lipoprotein, and serum potassium. The nomogram demonstrated excellent discrimination, with AUCs of 0.983 (training test), 0.962 (internal test), and 0.996 (temporal validation). Calibration and decision-curve analyses confirmed clinical utility.Conclusion: We constructed and validated a robust nomogram using accessible clinical variables. As a supplementary risk-assessment tool, it is designed to raise clinical suspicion of KD in febrile children with liver injury, thereby facilitating earlier intervention and optimized management. <Table Float="No" ID="Taba"> <tgroup cols="1"> <colspec align="left" colname="c1" colnum="1" /> <tbody> <row> <entry align="left" colname="c1"> What is Known: • Liver injury, a common complication of Kawasaki disease (KD), closely correlates with greater inflammation, intravenous immunoglobulin (IVIG) resistance, and higher coronary risk. • No specific predictive tool exists to identify KD among febrile children presenting with liver injury. </entry> </row> <row> <entry align="left" colname="c1"> What is New: • This study developed and validated the first nomogram to predict KD risk in febrile children with liver injury, using nine clinical parameters. • The model demonstrates excellent performance in both internal and prospective temporal validation. • It highlights low high-density lipoprotein (HDL) as a novel biomarker, useful not only for early diagnosis but also for monitoring acute inflammation in KD. </entry> </row> </tbody> </tgroup> </Table>

A nomogram for predicting Kawasaki Disease in acutely febrile children with liver injury

期刊论文
发表日期：

2026年3月3日

查看全文

Pingping Yang,
Guina Wang,
Jing Guo,
Wenlong Yan,
Lingfen Xu,
Lingzhe Meng

摘要

Liver injury is a common but nonspecific feature of acute Kawasaki disease (KD), often obscuring the diagnosis and delaying treatment. This study aimed to develop and validate a nomogram to predict KD risk in febrile children with liver injury. We conducted a retrospective study of children admitted with fever and liver injury between January 2019 and December 2024. A nomogram was developed using multivariable logistic regression in a training cohort, internally validated, and further evaluated in an independent temporal validation cohort (January–September 2025). Among 1,031 children, KD was the second most common specific diagnosis (13.4%). Nine independent predictors were identified: duration of fever, absolute neutrophil count, absolute lymphocyte count, C-reactive protein, albumin, gamma-glutamyl transferase, total bilirubin, high-density lipoprotein, and serum potassium. The nomogram demonstrated excellent discrimination, with AUCs of 0.983 (training test), 0.962 (internal test), and 0.996 (temporal validation). Calibration and decision-curve analyses confirmed clinical utility.

Conclusion: We constructed and validated a robust nomogram using accessible clinical variables. As a supplementary risk-assessment tool, it is designed to raise clinical suspicion of KD in febrile children with liver injury, thereby facilitating earlier intervention and optimized management.

What is Known:

• Liver injury, a common complication of Kawasaki disease (KD), closely correlates with greater inflammation, intravenous immunoglobulin (IVIG) resistance, and higher coronary risk.

• No specific predictive tool exists to identify KD among febrile children presenting with liver injury.

What is New:

• This study developed and validated the first nomogram to predict KD risk in febrile children with liver injury, using nine clinical parameters.

• The model demonstrates excellent performance in both internal and prospective temporal validation.

• It highlights low high-density lipoprotein (HDL) as a novel biomarker, useful not only for early diagnosis but also for monitoring acute inflammation in KD.