BACKGROUND: We investigated the characteristics and causes of various uveitis subtypes in patients presenting to the Regional Eye Centre at the Royal Alexandra Hospital, University of Alberta, Edmonton, Alta., and estimated the incidence of anterior uveitis in northern Alberta. METHODS: A retrospective study was conducted of all patients presenting with uveitis to a single, full-time ophthalmologist at the Regional Eye Centre from September 2004 to June 2005. Uveitis was classified according to onset, severity, anatomical subtype, etiology, recurrence rate, and response to treatment. Statistical analysis was used to compare patients referred by ophthalmologists with those referred by non-ophthalmologists. RESULTS: Two hundred and nine eyes of 171 patients were included in the study. Ophthalmologist referrals consisted of 67.4% anterior, 14.0% intermediate, and 18.6% panuveitis, and non-ophthalmological referrals were 92.8% anterior, 5.4% intermediate, and 1.8% panuveitis. Referrals from ophthalmologists were significantly more likely to be chronic, recurrent, and (or) less responsive to treatment than referrals from other sources. INTERPRETATION: Referral bias strongly affects the proportions of uveitis subtypes seen. Human leukocyte antigen-B27-associated diseases (especially ankylosing spondylitis), sarcoidosis, and herpes infections should be considered among the most likely causes of uveitis to be diagnosed in this patient population.