Fig. 1. Relationship between IL-6 levels and ICU duration for COVID-19 ICU patients who had at least one plasma IL-6 measurement after administration of either corticosteroids or other classes of drugs.

a Among the COVID19 patients that were admitted to ICU at some point during their hospital stay and had undergone an IL6 test, the patients that were on corticosteroids (n = 43) experienced longer durations on higher levels of respiratory support compared to the patients that were not on corticosteroids (n = 56). b Relationship between ICU duration and IL-6 levels following the administration of non-topical corticosteroids for n = 21 ICU patients who received corticosteroids. c Relationship between ICU duration and IL-6 levels following the administration of one or more of the following drugs: tocilizumab, azithromycin, hydroxychloroquine, or antivirals (n = 64 patients); these patients did not receive corticosteroids. d Summarizing the number of patients in each of the four quadrants (Q1, Q2, Q3, and Q4) in (b) and (c) above. A Fisher exact test of proportions suggests that a higher proportion of patients who receive corticosteroids and have an IL-6 value over 10 pg/mL have poor outcomes (ICU stay ≥15 days or death) when compared to the equivalent proportion of those treated with non-corticosteroid drugs including antivirals, tocilizumab, azithromycin, and hydroxychloroquine (p = 0.0006).