Introduction Kawasaki disease is a systemic pediatric vasculitis with a wide range of clinical presentations. Although often viewed as a single disease entity, emerging evidence indicates considerable phenotypic diversity. However, population-specific studies are still limited, especially in high-incidence areas like Japan. This study aimed to identify different clinical subgroups among patients with Kawasaki disease in Japan using cluster analysis. Methods We retrospectively reviewed 321 patients diagnosed with Kawasaki disease from 2020 to 2025 at a single center in Japan. Eighteen clinical and laboratory variables were used for principal component analysis and hierarchical clustering. The clusters were compared based on clinical features, treatment responses, and coronary outcomes. Temporal and seasonal trends were also examined. Results Three phenotypic clusters were identified: Cluster 1 included younger children with classical Kawasaki disease’s features and low treatment resistance; Cluster 2 exhibited hepatobiliary involvement, high intravenous immunoglobulin resistance prediction scores, and frequent need for additional therapy; and Cluster 3 comprised older children with elevated inflammation markers, cervical lymphadenopathy, and delayed diagnosis. Postpandemic, Cluster 3 increased, and seasonal peaks were observed in winter (Cluster 3) and spring (Cluster 2). Conclusion Cluster analysis identified distinct Kawasaki disease subtypes in a high-incidence Japanese population, each with unique clinical and temporal features. These findings support using phenotypic clustering to improve risk stratification and guide personalized treatment approaches. <Table Float="No" ID="Taba"> <tgroup cols="2"> <colspec align="left" colname="c1" colnum="1" /> <colspec align="left" colname="c2" colnum="2" /> <tbody> <row> <entry align="left" nameend="c2" namest="c1"> Key Points • Three distinct Kawasaki disease phenotypes were identified in Japanese children, each with unique clinical features, biochemical profiles, and treatment responses. • Temporal and seasonal patterns suggest an interplay between environmental triggers and host genetic susceptibility in the expression of KD among Japanese children. </entry> </row> </tbody> </tgroup> </Table>

Subgroups of Kawasaki disease in Japanese patients and temporal change: a retrospective data-driven cluster analysis

期刊论文
发表日期：

2026年1月9日

查看全文

Shintaro Fujiwara,
Yousuke Higuchi,
Junya Shimizu,
Ryosuke Matsuki,
Takahiro Namba

摘要

Introduction

Kawasaki disease is a systemic pediatric vasculitis with a wide range of clinical presentations. Although often viewed as a single disease entity, emerging evidence indicates considerable phenotypic diversity. However, population-specific studies are still limited, especially in high-incidence areas like Japan. This study aimed to identify different clinical subgroups among patients with Kawasaki disease in Japan using cluster analysis.

Methods

We retrospectively reviewed 321 patients diagnosed with Kawasaki disease from 2020 to 2025 at a single center in Japan. Eighteen clinical and laboratory variables were used for principal component analysis and hierarchical clustering. The clusters were compared based on clinical features, treatment responses, and coronary outcomes. Temporal and seasonal trends were also examined.

Results

Three phenotypic clusters were identified: Cluster 1 included younger children with classical Kawasaki disease’s features and low treatment resistance; Cluster 2 exhibited hepatobiliary involvement, high intravenous immunoglobulin resistance prediction scores, and frequent need for additional therapy; and Cluster 3 comprised older children with elevated inflammation markers, cervical lymphadenopathy, and delayed diagnosis. Postpandemic, Cluster 3 increased, and seasonal peaks were observed in winter (Cluster 3) and spring (Cluster 2).

Conclusion

Cluster analysis identified distinct Kawasaki disease subtypes in a high-incidence Japanese population, each with unique clinical and temporal features. These findings support using phenotypic clustering to improve risk stratification and guide personalized treatment approaches.

Key Points

• Three distinct Kawasaki disease phenotypes were identified in Japanese children, each with unique clinical features, biochemical profiles, and treatment responses.

• Temporal and seasonal patterns suggest an interplay between environmental triggers and host genetic susceptibility in the expression of KD among Japanese children.