Differences in biomarker levels and proteomic survival prediction across two COVID-19 cohorts with distinct treatments

Prognostic biomarkers have been widely studied in COVID-19, but their levels may be influenced by treatment strategies. This study examined plasma biomarkers and proteomic survival prediction in two unvaccinated hospitalized COVID-19 cohorts receiving different treatments. In a derivation cohort (n = 126) from early 2020, we performed plasma proteomic profiling and evaluated innate and complement system immune markers. A proteomic model based on differentially expressed proteins predicted 30-day mortality with an area under the curve (AUC) of 0.81. The model was tested in a validation cohort (n = 80) from late 2020, where patients received remdesivir and dexamethasone, and performed with an AUC of 0.75. Biomarker levels varied considerably between cohorts, sometimes in opposite directions, highlighting the impact of treatment regimens on biomarker expression. These findings underscore the need to account for treatment effects when developing prognostic models, as treatment differences may limit their generalizability across populations.