Jian Zhao,1,* Rushun Dai,2,3,* Yi Zhao,1,* Jiaping Tan,4,* Di Hao,1 Jie Ren,1 Xianwen Wang,1 Yanqing Chen,1 Hu Peng,1 Yugang Zhuang,1,5 Shuqin Zhou,1 Yuanzhuo Chen1 

1Department of Emergency, Shanghai 10th People’s Hospital, Tongji University School of Medicine, Shanghai, 200072, People’s Republic of China; 2Shanghai Institute of Thoracic Oncology, Shanghai Chest Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, 200030, People’s Republic of China; 3Department of Clinical Laboratory Medicine, Shanghai 10th People’s Hospital, Tongji University School of Medicine, Shanghai, 200072, People’s Republic of China; 4Department of Emergency, Chenzhou First People’s Hospital, Chenzhou, Hunan Province, 423000, People’s Republic of China; 5Department of Critical Care Medicine, Shanghai 10th People’s Hospital, Tongji University School of Medicine, Shanghai, 200072, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Shuqin Zhou; Yuanzhuo Chen, Email shuqin4344@126.com; chenyuanzhuo@tongji.edu.cn

Background: Sepsis is a heterogeneous clinical syndrome. Identifying distinct clinical phenotypes may enable more targeted therapeutic interventions and improve patient care.
Objective: This study aims to use clustering analysis techniques to identify different immune subtypes in sepsis patients and explore their clinical relevance and prognosis.
Methods: The study included 236 patients from the EICU at Shanghai Tenth People’s Hospital, who met the Sepsis 3.0 diagnostic criteria. Blood samples were collected to measure lymphocyte subsets and cytokine levels, along with demographic and clinical data. K-means clustering analysis was used to categorize patients into three groups based on immune and inflammatory markers.
Results: Three immune subtypes were identified: the high immune activation subtype (Cluster 1), characterized by high levels of CRP and WBC, high levels of T cells, NK cells, and B cells, and low levels of IL-6, IL-8, and IL-10; the moderate immune activation subtype (Cluster 2), characterized by moderate levels of CRP, WBC, T cells, NK cells, B cells, IL-6, IL-8, and IL-10; and the high inflammation and immune suppression subtype (Cluster 3), characterized by very high levels of IL-6, IL-8, and IL-10, low levels of T cells, NK cells, and B cells, and relatively lower CRP levels. Patients in Cluster 3 had a significantly increased 28-day mortality risk compared to those in Cluster 1 (HR = 21.65, 95% CI: 7.46– 62.87, p < 0.001). Kaplan-Meier survival curves showed the lowest survival rates for Cluster 3 and the highest for Cluster 1, with the differences among the three groups being highly statistically significant (p < 0.0001).
Conclusion: This study identified three immune subtypes of sepsis that are significantly associated with clinical outcomes. These findings provide evidence for personalized treatment strategies to improve patient outcomes.

Keywords: sepsis, immune subtypes, clustering analysis, prognosis, cytokines