Little is known about the frequency and significance of clinically unapparent or occult hemorrhage in ITP. Therefore, we prospectively explored the sites and frequency of occult bleeding in children with severe ITP at diagnosis or upon symptomatic relapse in a prospective, single-institution cohort study of patients 18 years of age and a platelet count 10,000/mm 3 . Data collected included bleeding severity assessment, urinalysis, fecal occult blood testing, and non-contrast brain MRI. Stool and urine samples were tested within 7 days of diagnosis or symptomatic relapse. Three months after diagnosis or relapse a noncontrast brain MRI evaluated hemosiderin deposits resulting from prior localized hemorrhage. Fifty-two ITP patients were enrolled with a mean platelet count of 4,000/mm 3 . A significant occurrence of occult hemorrhage was identified in the urine (27%) compared with clinically overt hematuria (0.91%, P < 0.0005). CNS microbleeding in the superficial cortex of the left frontal lobe was identified in one child with occult bleeding in the urinary tract. There was no relationship between occult hemorrhage and bleeding manifestations on physical examination. Occult hemorrhage was not a harbinger of subsequent bleeding. Our findings suggest that occult hemorrhage occurs with greater frequency than overt bleeding in children with severe ITP. CNS microbleeding is a potential risk in this patient population. Assessment of brain microbleeds and microscopic hematuria in this patient population require additional study.