Molecular Subtypes in Acute Respiratory Illness and Sepsis
Philip A. Efron, Brendon P. SciclunaSepsis and acute respiratory illness are clinical syndromes that encompass profound biological heterogeneity. The resulting heterogeneity of treatment effect is now widely understood as the explanation to why immunomodulatory trials in critical care have not me their study endpoints, thereby misclassifying effective therapies as futile while concealing harm in biologically ineligible subgroups. High-dimensional molecular profiling offers a route to resolve this heterogeneity into tractable subgroups or subtypes. Here we review molecular subtypes across adult and paediatric sepsis, pneumonia, and acute respiratory distress syndrome (ARDS), spanning transcriptomic, proteomic, metabolomic, and lipidomic platforms. Blood transcriptomic frameworks derived independently in multiple countries, and subsequently harmonized by international consensus efforts, converge on recurring immune archetypes: a myeloid inflammatory state dominated by neutrophil activation; a mixed myeloid and lymphoid dysregulated state carrying the highest mortality; and a relatively immunocompetent state with preserved lymphocyte function and the most favourable clinical trajectory. Plasma proteomic and metabolomic profiling recapitulate this architecture at orthogonal molecular layers, arguing that these archetypes are platform-independent rather than artefacts of any single technology. Cross-syndrome convergence is asymmetric, which plausibly reflects the methodological differences between top-down latent class analysis in ARDS and bottom-up unsupervised clustering in sepsis. Secondary analyses of randomized controlled trials consistently show subtype-dependent responses to treatment strategies, with a coherent directionality with potential benefit in hyperinflammatory states, and harm in immunosuppressed or adaptive immune-prevalent states. Subtype membership is not static; trajectory, not baseline assignment, most strongly encodes prognosis. Realising precision critical care now requires prospective, subtype-stratified trials of which none has yet been completed.