Impact of corticosteroids on initiation and half-year durability of humoral response in COVID-19 survivors

Background The impact of corticosteroids on humoral responses in coronavirus disease 2019 (COVID-19) survivors during the acute phase and subsequent 6-month period remains unknown. This study aimed to determine how the use of corticosteroids influences the initiation and duration of humoral responses in COVID-19 survivors 6 months after infection onset. Methods We used kinetic antibody data from the lopinavir–ritonavir trial conducted at Jin Yin-Tan Hospital in January 2020, which involved adults hospitalized with severe COVID-19 (LOTUS, ChiCTR2000029308). Antibody samples were collected from 192 patients during hospitalization, and kinetic antibodies were monitored at all available time points after recruitment. Additionally, plasma samples were collected from 101 COVID-19 survivors for comprehensive humoral immune measurement at the half-year follow-up visit. The main focus was comparing the humoral responses between patients treated with systemic corticosteroid therapy and the non-corticosteroid group. Results From illness onset to day 30, the median antibody titre areas under the receiver operating characteristic curve (AUCs) of nucleoprotein (N), spike protein (S), and receptor-binding domain (RBD) immunoglobulin G (IgG) were significantly lower in the corticosteroids group. The AUCs of N-, S-, and RBD-IgM as well as neutralizing antibodies (NAbs) were numerically lower in the corticosteroids group compared with the non-corticosteroid group. However, peak titres of N, S, RBD-IgM and -IgG and NAbs were not influenced by corticosteroids. During 6-month follow-up, we observed a delayed decline for most binding antibodies, except N-IgM (β −0.05, 95% CI [−0.10, 0.00]) in the corticosteroids group, though not reaching statistical significance. No significant difference was observed for NAbs. However, for the half-year seropositive rate, corticosteroids significantly accelerated the decay of IgA and IgM but made no difference to N-, S-, and RBD-IgG or NAbs. Additionally, corticosteroids group showed a trend towards delayed viral clearance compared with the non-corticosteroid group, but the results were not statistically significant (adjusted hazard ratio 0.71 95% CI 0.50–1.00; P = 0.0508). Conclusion Our findings suggested that corticosteroid therapy was associated with impaired initiation of the antibody response but this did not compromise the peak titres of binding and neutralizing antibodies. Throughout the decay phase, from the acute phase to the half-year follow-up visit, short-term and low-dose corticosteroids did not significantly affect humoral responses, except for accelerating the waning of short-lived antibodies.