IP10 levels refine severity prognostication in COVID-19 and implicate causes of clinical deterioration: lessons for pandemic preparedness

Abstract

SummaryBackgroundEmerging pandemics place immense strains on healthcare systems that may be ameliorated by rapid development of biomarkers whose measurements may predict disease severity and additionally inform about disease causation. Conspicuously, such routine measures rarely include immunological cytokines or chemokines, despite their contributions to host protection and immunopathology.MethodsMultiplex bead-array and ELISA-based serum cytokine and chemokine measurements, routinely employed clinical laboratory measures, and clinical outcomes were collectively fed into predictive model development for prognostication of COVID-19 severity in an unvaccinated UK cohort (Discovery; early-to-mid 2020), with subsequent external validation among a pauci-vaccinated UK cohort (early 2021) and part-vaccinated India cohort (early 2022 to early 2023). Correlates of disease severity were assessed by high-content spectral flow cytometry.FindingsIncorporating a practical test for the chemokine IP10 (a.k.a. CXCL10) alongside routine clinical laboratory assays increased at-admission test accuracy to prognosticate intensive care requirements or in-hospital mortality at 30 days in the Discovery and Validation cohorts. In the India cohort, high IP10 levels predicted terminal deterioration among unvaccinated persons. High-resolution immune-profiling within subsets of the Discovery and India cohorts associated a T cell-centric signature with disease severity and with high IP10 levels, thereby identifying candidate drivers of COVID-19 deterioration.ConclusionsIP10 levels measured at or around hospital admission offer a practical biomarker enhancing COVID-19 patient outcome prognostication, particluarly in unvaccinated individuals, and offer mechanistic insights into pathogenesis. Thus, prompt application of systems immune-profiling in future pandemics might rapidly identify prognostic and mechanistic biomarkers of patient deterioration, aiding clinical decision-making at a time of severe healthcare strain.FundingMedical Research Council grant, CARDINNATE.