Cytokine Profile in Different Post-Covid-19 Condition Phenotypes

ABSTRACT

Background: At least 10% of SARS-CoV-2 infected patients suffer from persistent symptoms for >12 weeks, known as post-COVID-19 condition (PCC) or Long Covid. Reported symptomatology is diverse with >200 physical and neurological debilitating symptoms. Here, we analyzed pro-inflammatory cytokine levels as a potential mechanism underlying persistent symptomatology. Method(s) Clinical data and samples used belong to the KING cohort extension, which includes clinically well characterized PCC (N=358, 59 persistent symptoms evaluated), COVID-19 recovered and uninfected subjects. We used Gower distances to calculate symptom's similarity between PCC and Ward's hierarchical clustering method to identify different symptom patterns among PCC patients. Cytokine levels of randomly selected PCC, recovered and uninfected subjects (N=193) were measured on plasma samples collected >6 months after acute infection using the 30-Plex Panel for Luminex. Mann- Whitney t-test was used to compare PCC vs recovered groups and Kruskal-Wallis t-test for >2 groups comparisons (PCC vs recovered vs Uninfected and within PCC clusters). FDR correction was applied for statistical significance (p-adj). Result(s) Hierarchical clustering identified 5 different PCC clusters according to their symptomatology, where PCC3 and PCC5 clusters showed higher prevalence of women ( >80%) and more persistent symptoms, while acute COVID-19 was mild in >80% of the patients. We selected 91 PCC (belonging to each cluster), 57 recovered and 45 uninfected subjects for cytokine profiling (Table 1). 13 soluble markers were significantly elevated (IL-1beta, Eotaxin, MIP-1beta, MCP-1, IL-15, IL-5, HGF, IFN-alpha, IL-1RA, IL-7, MIG, IL-4 and IL-8) in PCC and recovered groups compared to uninfected subjects (all p-adj< 0.04). In addition, PCC subjects tended towards higher levels of IL-1RA compared to recovered group (padj= 0.071). Within PCC clusters, FGF-basic and RANTES were elevated while IL-2 and MIG were decreased in PCC3 and PCC5 compared to the other PCC clusters (all p-adj< 0.04). TNF-alpha, IP-10, G-CSF and MIP-1alpha were decreased in PCC3 and PCC5 not reaching statistical significance (all p-adj=0.07). Conclusion(s) Some cytokines remained altered in all SARS-CoV-2 infected subjects independently of persistent symptoms after 6 months from acute infection. Differences between PCC and recovered individuals are limited after correction. Importantly, PCC cytokine profiles showed differences between clusters, which suggests different PCC subsyndromes with distinct etiology. Subjects Characteristics (Table Presented).