Fifteen studies were included (n=1,309). These included three placebo-controlled RCTs (n=379) and 12 retrospective studies (n=930).
Resolution of abdominal pain (two RCTs and three retrospective studies): One RCT reported no significant difference between corticosteroids and placebo in the time to pain resolution. The other RCT found a significant reduction in the corticosteroid group (reduction 1.2 days, p=0.03). Significant heterogeneity was found for the retrospective studies (p=0.01). After excluding one study in which treatment was not started until after 21 days, a significant increase in the proportion of patients with pain resolution within 24 hours of treatment and no significant heterogeneity was found, OR 5.42 (95% CI: 1.60 to 18.29).
Surgical intervention for intussusception (one RCT and two retrospective studies): The RCT and meta-analysis of the two retrospective studies both found a non-significant reduction with corticosteroids, OR 0.16 (95% CI: 0.01 to 3.62) and OR 0.75 (95% CI: 0.13 to 4.46; no significant heterogeneity).
Henoch-Schonlein purpura recurrence (two RCTs and five retrospective studies): Meta-analysis of the two RCTs found a non-significant reduction in Henoch-Schonlein purpura recurrence with corticosteroids, OR 0.32 (95% CI: 0.07 to 1.49; no significant heterogeneity). Significant heterogeneity was found for analysis of the five retrospective studies (p<0.01); no evidence of publication bias was found, One retrospective study reported a significant reduction in Henoch-Schonlein purpura recurrence with corticosteroids. One study reported a significant reduction in Henoch-Schonlein purpura with routine care.
Cumulative renal abnormalities (three RCTs and five retrospective studies): Significant heterogeneity was found for the meta-analysis of the RCTs (p=0.03) and the retrospective studies (p<0.01). No evidence of publication bias was found for either type of study. One RCT and two retrospective studies reported a significant decrease in cumulative renal abnormalities with corticosteroid. One retrospective study reported a significant increase in cumulative renal abnormalities with corticosteroids.
Persistent renal abnormalities (three RCTs and one retrospective study): Meta-analysis of the three RCTs found a significant reduction in the risk of persistent renal disease in corticosteroid groups, OR 0.43 (95% CI: 0.19 to 0.96; no significant heterogeneity and no evidence of publication bias were found). The retrospective study found no significant effect of corticosteroids.
No evidence of a significant corticosteroid dose-response effect was found (p=0.07).