Objectives: To verify the prevalence of tuberculosis (TB)/HIV co-infection and the ability of the clinical history to detect the HIV infection in TB inpatients. Setting: Eduardo de Menezes Hospital, reference for both TB and AIDS. Patients and methods: All patients admitted with TB in a pneumology ward were evaluated prospectively from 1/1/1997 to 1/31/1998. The clinical history was directed to the presence of risk factors for AIDS or TB, previous treatments or abandoned treatments for TB, and TB clinical forms. Patients with AIDS defining illnesses, except for TB, and with previous anti-HIV tests were excluded. All patients had an ELISA anti-HIV serology, and when positive, a Western-Blot test was performed to confirm the previous result. X-square test and Fisher test were used for statistical analysis. Results: Sixty-five patients were divided into group I (positive serology for HIV, n = 6) and group II (negative serology for HIV, n = 59). There were no statistical differences between the groups comparing the risk factors for AIDS and TB, previous treatments or abandoned treatments for TB, or TB clinical forms. Conclusions: The high prevalence of TB/HIV co-infection (9.2%) reinforces that: 1) the clinical history was not able to detect a significant number of patients with TB/HIV co-infection and that: 2) anti-HIV serology should be performed in all patients with active TB forms.

Keywords: tuberculosis, acquired immundeficiency syndrome, prevalence