Introduction: Tuberculosis is the leading cause of death among people living with HIV/AIDS (PLHIV) in sub-Saharan Africa. In PLHIV, Smear-Negative Pulmonary Tuberculosis (SNPTB) and Extrapulmonary Tuberculosis (EPTB) are predominant. Presumptive anti-tuberculosis (anti-TB) treatment is often delayed leading to a high mortality rate.Objectives: To investigate the clinical outcomes of presumptive anti-TB treatment in HIV patients suspected of having TB and to determine the factors associated with patients’ death.Methods: We conducted a retrospective descriptive study from 1 January 2007 to 31 December 2008 in the Department of Internal Medicine of the Hospital Yalgado Ouédraogo on patients infected with HIV who received a presumptive anti-TB treatment. Defining patients with SNPTB or EPTB was based on the 2007 WHO's diagnostic algorithm of SNPTB and EPTB.Results: One hundred and sixteen patients of the 383 (30.2%) HIV patients hospitalized in this period were suspected of having TB. The average CD4 count was 86.1 cells/µl (SD = 42.3). A SNPTB was diagnosed in 67 patients (57.8%) and a EPTB in 49 patients (42.2%). The median length of hospitalization duration was 23.5 days. The average time of initiation of anti-TB treatment after admission was 22 days (SD = 9.2 days). Evolution during the hospital stay was favorable for 65 patients (56.0%), unfavorable for 48 patients (41.4% or 12.5% of all hospitalized patients), and 3 patients (2.6%) were treatment defaulters. In a multivariate analysis, hospitalization duration longer than 15 days and a delay of anti-TB treatment initiation of more than 30 days are independent factors associated with patients’ deaths.Conclusion: An urgent access to TB-diagnostic tools and a revision of the International algorithm for the diagnosis and treatment of SNPTB and EPTB in the context of HIV could help to reduce the delay of anti-TB treatment initiation and the mortality rate of PLHIV in sub-Saharan Africa.