In France, the entire population theoretically has access to health coverage, but in fact a section of the poorest population does not. Institutions have therefore been set up to provide medical care for the destitute. The objectives of this study were to describe the social characteristics of the HIV-positive destitute population attending an out-patient clinic providing free health care for the destitute in a Paris University Hospital, to compare their clinical-epidemiological characteristics with those of non-destitute HIV-positive patients, and to evaluate the quality of their care. We performed a historical prospective study wherein a cohort of 115 HIV-positive destitute patients (defined as having no health coverage at their first consultation) was compared with a control cohort of 183 HIV-positive non-destitute patients attending the same clinic. Ninety-five per cent of the destitute patients had no stable employment, 32% had no source of income, 75% had no permanent residence and 27% were i.v. drug abusers. Fifty-nine per cent were foreigners, most of whom had legal residence papers and had been in France for more than 3 years. When comparing the control and the destitute groups, the latter had a three times greater risk of developing tuberculosis (RH = 3.2, CI 95% = [1.1-9.4]). Medical compliance, access to antiretroviral treatment and hospitalization were identical in both groups. No difference was observed in terms of occurrence of a new AIDS-related disease during follow-up when full-blown AIDS before entry, CD4 count at entry and transmission group were taken into account in multivariate analysis. From the moment that destitute patients attended this adapted medico-social facility, their access to care was the same as, if not better than, that of the other patients. The development of out-patient medico-social facilities for HIV-positive destitute patients must be a public health priority even for those countries theoretically providing generalized health coverage.