Objectives: To evaluate the characteristics of HIV-infected patients treated in Mayotte, focusing on health care pathways. Methods: Observational study, with retrospective collection of data from medical charts according to a standardized questionnaire. All HIV-infected patients who consulted at the Centre Hospitaller de Mayotte from 1990 through 2010 were included. Patients were classified as lost to follow-up if they had not consulted for at least 6 months and if no information about their outcome was available. Variables associated with loss to follow-up were analyzed by bivariate analysis (Chi t ), and then multivariate analysis (logistic regression). Results: Overall, 236 patients were enrolled (134 females, 102 males), with a mean age of 30.7 +/- 12 years, and a median CD4 cell count of 390/mm(3) at diagnosis. HIV testing was most often ordered because of: i) HIV-related symptoms (n = 70); ii) risk factors (n = 67); iii) pregnancy (n = 51). The 52 patients (22%) lost to follow-up were less likely to have medical insurance (52% versus 67%, p = 0.05) or antiretroviral treatment (56% versus 71%, p = 0.04), and were more frequently aged younger than 15 years (15% versus 2%, p<0.001). On multivariate analysis, initial management at the Centre Hospitaller de Mayotte (OR 2.62 [1.09-6.29], p = 0.03, and the absence of antiretroviral treatment (OR 1.97 [1.00-3.89], p = 0.05) were significantly associated with loss to follow-up. Conclusion: The HIV epidemic in Mayotte follows an African pattern in terms of main transmission pathways, age and circumstances of diagnosis, but the median CD4 cell count at diagnosis is closer to the pattern in metropolitan France. The rate of patients lost to follow-up is high (22%), mostly linked to problems with access to care.