Objective: To analysis the relationship between glucocorticosteroids (GCS) usage and side effects in the treatment of severe acute respiratory syndrome (SARS). Methods: All clinical records of probable SARS patients in Beijing were collected and input into an Epi6 database, in which 1 291 patients had entire information and met the clinical criteria of SARS. The usage of GCS and GCS associated side effects were analyzed retrospectively. Results: Patients accepted GCS therapy were 83.96% (n= 1 084), whereas 16.04%(n=207) did not take GCS. The average dosage of GCS was 160 mg/d in the first week, and then reduced to 80 mg/d and 40 mg/d in the second and the third weeks, respectively. Initial blood glucose, systolic pressure (SBP), and diastolic pressure (DBP) were no significant difference between GCS group and non-GCS group. The highest blood glucose during the treatment in GCS group was markedly higher than that in non-GCS group [(8.68? 4.80 ) mmol/L vs (6.39?3.71) mmol/L, P05). After GCS administration, SBP and DBP were increased gradually, and reached their peaks in the fourth week [SBP (117.2?14.0) mm Hg and DBP (72.5?9.1) mm Hg vs SBP (120.0?12.5) mm Hg and DBP (74.5?8.7) mm Hg, P