The aim of this study was to identify the risk factors for persistence of coronary artery abnormalities (CAAs) in Kawasaki disease and to compare the differences between complete (n=25) and incomplete (n=18) forms of the disease in relation to CAAs. The patients' demographic (age and sex), clinical (season of admission, duration of fever, form of Kawasaki disease), laboratory (complete blood count, eosinophil count, serum biochemistry, C-reactive protein level, erythrocyte sedimentation rate [ESR], urinary analysis), echocardiographic and therapeutic data were evaluated retrospectively from the medical records. The incidence of sterile pyuria was nearly twofold greater in patients with initial CAAs than in patients in which CAAs were not present (25% and 13%, respectively; p=0.33). In multivariate logistic regression analysis, ESR and pyuria were found to be associated with persistence of CAAs (p=0.035 and p=0.046, respectively). In addition, we found that duration of fever was significantly associated with persistence of CAAs (p=0.045). However, gender, age at presentation, peripheral blood eosinophilia, low albumin level, CRP, leukocytosis and anemia were not predictive for persistence of CAAs. There was no difference between the complete and incomplete form of the disease in regard to persistence of CAAs. As a result, we have determined that duration of fever, high levels of ESR and presence of sterile pyuria can be used to predict the persistence of CAAs in Kawasaki disease.