Efficacy and Safety of Different Doses of Systemic Corticosteroids in COPD Exacerbation

BACKGROUND: Although systemic corticosteroids (SCS) have long been used to treat patients with COPD exacerbation, the recommended dose remains controversial. We aimed to perform a meta-analysis and an indirect treatment comparison to investigate the efficacy and safety of different doses of SCS in subjects with COPD exacerbation. METHODS: Studies were identified by searching different databases for randomized controlled trials that investigated the efficacy and safety of SCS with placebo in subjects with exacerbation of COPD. The different doses of SCS were assigned to low-dose (ie, initial dose <= 40 mg prednisone equivalent/d [PE/d]), medium-dose (initial dose = 40-100 mg PE/d, and high-dose (initial dose > 100 mg PE/d) groups. The indirect treatment comparison was performed between low-, medium-, and high-dose SCS groups. RESULTS: Twelve trials with 1,375 participates were included. Compared to placebo, the risk of treatment failure was lower in the low-dose SCS groups (risk ratio 0.61 [95% CI 0.43-0.88], P = .007) and high-dose SCS groups (risk ratio 0.64 [95% CI 0.48-0.85], P = .002); the FEV1 was significantly improved in low-dose (mean difference 0.09 [95% CI 0.06-0.12], P < .001), medium-dose (mean difference 0.23 [95% CI 0.02-0.44], P = .036), and high-dose SCS groups (mean difference 0.09, [95% CI 0.03-0.15], P < .001, respectively). Regarding safety, the incidence of hyperglycemia was higher in high-dose SCS groups versus placebo (risk ratio 2.52 [95% CI 1.13-5.62], P = .02). The indirect comparison between low-, medium-, and high-dose SCS found that the risk of treatment failure and changes in FEV1 were similar between these doses of SCS. CONCLUSIONS: This meta-analysis indicates that low-dose SCS (initial dose <= 40 mg PE/d) was sufficient and safer for treating subjects with COPD exacerbation, and it was noninferior to higher doses of SCS (initial dose > 40 mg PE/d) in improving FEV1 and reducing the risk of treatment failure. However, our findings need to be verified in head-to-head randomized controlled trials.