Sepsis: deciphering pathogenesis and transforming diagnosis and treatment through systems immunology

Abstract

Background. Sepsis is responsible for over 11 million deaths per year, constituting one of the most heterogeneous critical illness syndromes. Whilst the implementation of bundle–based care has been demonstrated to be effi cacious in reducing mortality, this eff ect has been shown to plateau in well–resourced settings due to the presence of biological heterogeneity. Systems immunology methods have been demonstrated to off er signifi cant potential for the development of precision diagnosis and treatment strategies. Objective. The objective of this study was to undertake a systematic review of the extant evidence on systems immunology in the context of sepsis diagnosis and treatment. In addition, the study will involve a critical analysis of the limitations of the leading conceptual frameworks in this area, and the delineation of future research priorities. Materials and methods. A systematic analysis and synthesis of 43 publications from PubMed/MEDLINE, Cochrane Library, and Frontiers (2016–2025) was conducted, with priority given to retrospective analyses of randomized controlled trials (RCTs), systematic reviews, and large–cohort studies (n >100). The following keywords were utilised: sepsis endotypes, phenotypes, precision medicine, and transcriptomics. The protocol was registered in the international registry of systematic reviews PROSPERO (CRD420261347322). Images were created using the Nano Banana 2.0 AI image editor (https://nano–banana2.co/), and the resulting data were reviewed, with changes and corrections made by the authors in accordance with the literature reviewed in this article. Results. Hancock et al. (2024) identifi ed fi ve distinct endotypes, each characterised by approximately 200 unique diff erentially expressed genes, in transcriptomic data from emergency department patients (before formal diagnosis). The five types of infl ammation are as follows: NPS (Neutrophilic–Suppressive): the most deadly type, and a potential target for dexamethasone; INF (Infl ammatory): characterised by hyperinfl ammation, opposed to NPS; IFN (Interferon): the predominance of interferon pathways; ADA (Adaptive): involving adaptive immunity; and IHD (Innate Host Defence): the activation of natural host defence. The SUBSPACE consortium (2025, n=3,380) validated four consensus endotypes and proposed the Hi–DEF two–dimensional framework (OR for severe outcomes with system–wide dysregulation=7.1, 95%CI 5.6–8.9). RCTs confi rmed heterogeneity of treatment eff ect: corticosteroids, simvastatin, and activated protein C demonstrated diff erential eff ects by subphenotype. Conclusions. Systems immunology is transforming the management of sepsis, moving away from universal bundle protocols towards precision strategies. Key future priorities include prospective endotype–enriched adaptive RCTs, the development of bedside–compatible diagnostics, and improvements in global accessibility.