Skip to content
Review Open access

Pathogen-associated heterogeneity in pulmonary sepsis: from inflammatory phenotypes and transcriptomic endotypes to precision medicine

Oct 2026 · Frontiers in Immunology · Vol 17 · 0 citations · 85 references
Medicine

Abstract

Sepsis is a heterogeneous syndrome, and pulmonary infection is its commonest source. Pulmonary sepsis carries a particularly high mortality, especially in patients who develop refractory septic shock, yet it is still managed as a single disease. Conventional diagnostic criteria, based on organ dysfunction, describe the consequences of the host response rather than its biological diversity, and this mismatch helps to explain why immunomodulatory trials have repeatedly failed in unselected patients. In this review we summarise how pathogen characteristics interact with the alveolar immune microenvironment to produce distinct host-response phenotypes, and how those phenotypes are captured by clinical classifiers, plasma protein biomarkers and blood transcriptomic endotypes — among them SRS1/2, Mars1–4, SENECA and USE-1/2. Three problems currently limit translation. The classification systems overlap only in part, and should not be used interchangeably. Endotypes change over time, so a single measurement at baseline can misclassify a patient. And most of the evidence linking particular pathogens to particular phenotypes is associative, and comes from ARDS, general sepsis, severe pneumonia or COVID-19 cohorts rather than from pulmonary sepsis itself. We then review the phenotype-dependent responses to fluid strategy, corticosteroids, statins, endotoxin removal, cytokine-directed therapy and checkpoint inhibition, and consider why several biomarker-guided trials produced pharmacodynamic signals without clinical benefit. Pathogen-informed, repeatedly measured host-response phenotypes, tested in adaptive trials with enrichment, offer the most realistic route to precision immunotherapy in pulmonary sepsis.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.