To analyze the effects of adjuvant glucocorticoid therapy on severe patients with COVID-19 in order to optimize the treatment for severe COVID-19.

195 patients confirmed as severe COVID-19 in non-ICU wards of one Hospital hospitalizing on Feb 28, 2020, were enrolled and divided into glucocorticoid treatment group(GC group) and non-glucocorticoid treatment group(non-GC group) by medication history. We retrospectively analyzed all patients′data including general information, glucocorticoid therapy information, temperature and best value of SpO₂ in one day when recorded, some results of complete blood count, values of C reaction protein(CRP) and procalcitonin(PCT), glucocorticoid associated adverse reaction, pharyngeal swab SARS-CoV-2 nucleic acid negative time and final outcomes.

Patients in both groups were characterized by male, similar average age(63.01±11.93 vs. 62.20±14.26, years) and with no significant difference on clinical characters, comorbidity, clinical course, antiviral therapy and temperature. Value of SPO₂ in GC group was lower than which in non-GC group[96(92-98)vs. 97(95-98), (P<0.05)]. Values of WBC and absolute neutrophil count (ANC) in GC group were normal but significantly higher than which in non-GC group. Compared with non-GC group, value of absolute lymphocyte count in GC group was significantly lower [0.90(0.68-1.10)vs. 1.29(0.93-1.62), P<0.05] and value of CRP in the same group was significantly higher [34.56(9.69-62.27)vs. 7.60(2.19-37.28), P<0.05]. After treating, value of SpO₂ in GC group was significantly higher than which in non-GC group [99(98-99)vs. 98(97-99), P<0.05)] admitted into hospital one week later. Mean while, values of WBC and ANC in GC group rose respectively, compared to the baseline level, and were significantly higher than which in non-GC group but normal. And value of absolute lymphocyte count in GC group rose significantly and was similar to which in non-GC group [1.39(1.08-2.18)vs.1.36(1.05-1.82), P>0.05], after treating, but with no difference. GC group′s patients′value of CRP significantly fell down and was similar to which in non-GC group admitted into hospital one week later. Analyses of subgroups divided by the accumulated dose of glucocorticoid showed no significant difference. Pharyngeal swab SARS-CoV-2 nucleic acid negative median time in both groups were similar with no statistical difference. Six cases of hyperglycemia, including blood pressure significantly higher than baseline level in two of the above, were observed in GC group, characterized by male and elder than 70 years. No case with adverse reaction was observed in non-GC group. All patients in both groups were discharged from hospital after recovery with no deterioration and death. Values of PCT in both groups were not analysed in order to avoiding bias for data missing.

The appropriate dose and course of glucocorticoid with other intensive treating may improve oxygenation of severe COIVD-19 patients and help the patients obtain more clinical benefits with controlled risk and favorable outcomes, for the absence of high grade level evidence-based medical recommendation for antiviral therapy. However, there are some limitations in our study and it is necessary to research the adjuvant glucocorticoid therapy for COVID-19 by a randomized controlled trial.