We investigated exposure to house dust mites and freshly cleaned carpets in patients with Kawasaki syndrome (KS) who resided in Denver, CO. House dust samples were analyzed for mites (Dermatophagoides pteronyssinus and Dermatophagoides farinae) and convalescent sera were assayed for immunoglobulin G antibody to mite antigen. House dust samples from case and neighborhood control homes showed no difference in the prevalence of mites. Fewer than 10% of dust samples from either group revealed evidence of live or dead mites. Nine of 18 (50%) children with KS had a prior history of exposure to freshly cleaned rugs within the 45 days of the onset of KS compared with 3 of 11 (27%) neighborhood controls. There was no difference in mean anti-mite antibody concentrations between convalescent sera from KS cases and sera from pediatric hospitalized controls. We conclude that in Denver KS can occur in the absence of significant mite exposure as defined by environmental studies or serologic means.