Background: There are three scoring systems for the prediction of non-response to IVIG in Japan. Steroid combination therapy for initial treatment is recommended for high risk cases in the new guideline from the scientific committee the Japanese society of pediatric cardiology and cardiac surgery in 2012. Objective: to evaluate efficacy of steroid combination therapy for high risk cases which have high scores in more than two prediction scoring systems. Subjects and Methods: From January 2009 to 2014, a total of 54 hospitalized patients with high scores in more than two prediction scoring systems were enrolled. The patients were separated into two groups. Group S (n=21): who were administered steroid+IVIG, Group N (n=33): who were administered only IVIG. Clinical data, clinical course, and circumstances of additional therapies were compared. Results: In the patients who had high scores in two prediction scoring systems, there were fewer patients who received additional therapies in group S compared to that in group N (13% 1/8 vs. 38% 8/21). However, for patients who had high scores in all prediction scoring systems, those who received additional therapies had not decreased in group S compared to that in group N (50% 6/12 vs. 46% 6/13). Conclusions: Steroid combination therapy is beneficial for the patient who has high scores in two or less prediction scoring systems. On the other hand, in the most severe cases those who have high scores in all prediction scoring systems, Steroid combination therapy is less efficacious and other additional therapy should be considered in the early stage to improve prognosis.