Analysis of severity in COVID-19 patients by using longitudinal immune profiles

COVID-19 severity varies due to immune heterogeneity, complicating treatment; therefore, we modeled oxygen saturation, CRP, and blood urea as severity markers for pulmonary, inflammatory, and renal functions. We analyzed cytokine, T cell receptor (TCR), and B cell receptor (BCR) expressions in 106 patients (NIH, Korea) across three stages (T1-T3) and applied a joint modeling framework, using linear mixed-effects models to identify immune features associated with severity. Models incorporating correlations among severity variables outperformed models assuming independence. IL-6 was significantly associated with all markers (negatively with oxygen saturation and positively with CRP and blood urea), whereas IL-8 was linked to oxygen saturation, and IGLV1-44 was consistently associated with CRP. Overall, accounting for correlations improved model performance and highlighted IL-6, IL-8, and IGLV1-44 as candidate immune features that may help refine severity assessment in heterogeneous COVID-19 cases.