Understanding why some sepsis patients spiral into prolonged immune failure while others recover has frustrated clinicians for decades. A new synthesis of single-cell genomics, spatial omics, and functional immunology studies offers the most integrated picture yet of how the immune system remodels itself during sepsis — and what blood-based markers might finally help predict who is at greatest risk of irreversible immune collapse.
This narrative review centers on peripheral blood mononuclear cells (PBMCs) — primarily monocytes and lymphocyte subsets — as a dynamic, serially accessible lens into sepsis immunopathology. The authors synthesize evidence showing that monocytes undergo state transitions from classical inflammatory phenotypes toward dysfunctional, immunosuppressive states, accompanied by pronounced downregulation of HLA-DR, a surface molecule critical for antigen presentation. Simultaneously, lymphocyte populations exhibit accelerated apoptosis and exhaustion, while immunometabolic reprogramming shifts cellular energy utilization in ways that impair effector function. Together, these changes explain a paradox well known to intensivists: patients can be hyperinflamed and immunosuppressed at the same time, with the balance shifting unpredictably over the course of illness. Transcriptomic signatures, monocyte phenotyping assays, and exhaustion marker panels are identified as candidate biomarkers for immune risk stratification.
The clinical implications here are real but still contingent. Sepsis immunology has long promised precision medicine but delivered little at bedside. The barriers catalogued in this review — cohort heterogeneity, timing-dependent immune states, assay standardization gaps, and sparse prospective validation — are not new complaints; they have stalled immune-targeted sepsis trials for over a decade. What is incrementally new is the resolution that single-cell and spatial omics now bring to mapping monocyte state transitions, potentially allowing researchers to define actionable immune phenotypes rather than broad, heterogeneous patient categories. Whether transcriptomic stratification can be operationalized in real ICU timelines remains the central unsolved problem. This is a high-quality synthesis that advances the conceptual framework considerably, but translational progress will require large, prospectively designed cohort studies with standardized assay protocols.