Aspirin is the basic treatment for Kawasaki disease, however its optimal dose is controversial. We investigated the therapeutic efficacy of high-dose (100 mg/kg/day, n = 30) versus low-dose (30 mg/kg/day, n = 30) aspirin. Duration of fever, transaminase, plasma thromboxane B2 (TxB2) and 6-keto-prostaglandin F1 alpha (PGF1 alpha) levels were compared before enrollment and on days 4, 7 and 14. In the high-dose group, duration of fever was significantly shorter than that of low-dose group (3.2 +/- 1.8 versus 5.4 +/- 4.3 days, p less than 0.05), however, serum glutamic pyruvic transaminase levels were elevated (157.4 +/- 187.7 versus 48.0 +/- 58.2I.U./liter, p less than 0.005). No differences in the incidence of coronary artery lesions were observed (5 of 30 versus 7 of 30). Plasma TxB2 production was completely blocked in both groups, plasma 6-keto-PGF1 alpha levels in the high-dose group on day 14 was lower than that in the low-dose group (39 +/- 26 versus 160 +/- 207 pg/ml, p less than 0.05). This latter observation suggest that high-dose therapy may be disadvantageous as anti-thrombotic treatment, and supports the notion that low dose therapy is safe in the acute stage of Kawasaki disease.