Kawasaki disease is an acute, febrile systemic vasculitis occurring mostly in young children. It is a leading cause of acquired heart disease, and may lead to long term coronary artery damage, particularly aneurysm formation. First line treatment includes high-dose intravenous immunoglobulin (IVIG) and aspirin. Adjunctive corticosteroids are recommended for patients with fever and inflammation that is unresponsive to initial IVIG treatment (Pediatrics 2004 Dec;114(6):1708). However, evidence for the efficacy of steroids as part of primary treatment for Kawasaki disease has been inconsistent (J Pediatr 2006 Sep;149(3):336, N Engl J Med 2007 Feb 15;356(7):663). The recently published RAISE trial lends support to the use of primary corticosteroids for children with severe Kawasaki disease who are at high risk of non-response to IVIG.

A total of 248 Japanese children with severe Kawasaki disease diagnosed within 9 days of fever onset were randomized without blinding to prednisolone IV vs. placebo. All children had been assessed to have high risk for unresponsiveness to IVIG and for coronary artery abnormalities using the Kobayashi score (Circulation 2006 Jun 6;113(22):2606). Initial prednisolone dosing was 2 mg/kg/day to maximum of 60 mg/day in 3 divided doses by IV injection for 5 days. When C-reactive protein was ≤ 5 mg/L, prednisolone was tapered to 0.5 mg/kg/day over 15 days in 5-day steps. Children whose fever was resolved at 5 days had oral prednisolone. All children received IVIG 2 g/kg over 24 hours and aspirin 30 mg/kg/day until afebrile, followed by aspirin 3-5 mg/kg/day for ≥ 28 days after fever onset.

The addition of prednisolone was associated with a reduction in fever duration after enrollment (median 1 day vs. 2 days, p < 0.0001) and with reduced need for additional treatment due to lack of response to first treatment (5% vs. 30%, p < 0.0001, NNT 4) (level 2 [mid-level] evidence). The prednisolone group had lower rates of coronary abnormalities at 1, 2, or 4 weeks (3% vs. 23%, p < 0.0001, NNT 5). Serious adverse events occurred in about 2% of children in each group (not significant), and there were no significant differences in relapse rates (11% vs. 12%). (Lancet 2012 Mar 7 early online).

It should be noted that while Kobayashi score has been shown to accurately predict risk in Japanese children, it had low sensitivity when applied to a cohort of North American children with Kawasaki disease (J Pediatr 2011 May;158(5):831). More generally applicable risk prediction tools may be needed to determine when children will benefit from steroid treatment.

For more information, see the Kawasaki disease topic in DynaMed.