ABSTRACT
During 2½ year period, 378 patients diagnosed with tuberculosis and admitted to a general hospital for care of the poor in Salvador, Bahia, were tested serologically for HIV-1, HTLV-I, and HTLV-II. The patients' mean age was 41.8 (range 14-89); they were hospitalized for a mean of 62 ñ 43 days; 70 percent were being treated for the first time; most of the remainder were being retreated after non-compliance with previously recommended anti-tuberculosis medication and a few required second-line therapy for relapsed disease. None had had previous serologic testing for retroviruses. Among the study population, 59 (16 percent) were found to be positive for retroviral infection. The distribution was as follows: 18 (4.8 percent) had HIV-1, 32 (8.5 percent) had HTLV-I, 2 of these had both HTLV-I and HTLV-II, 9 (2.4 percent) had both HIV-1 and HTLV-I. The rates of positive serologic tests for retroviral infection in this Salvador is 0.2 percent for HIV-1 and 1.0 percent for HTLV-I. Thus, there is a higher than expected frequency of retroviral infections among patients hospitalized for treatment of tuberculosis. The prognosis for treated patients was determined by recording the cause of death and the mortality rate. In the 319 patients with negative serologic testing for retroviruses the were 25 death (8 percent). In 32 patients with HTLV-I infection there were 8 death (25 prcent), and in 18 patients with HIV-1 infection there were 6 deaths (33 percent). In 9 patients with both HIV-1 and HTLV-I there were 5 deaths (56 percent). The causes of death in each serological group were primarily related to progression of tuberculosis rather than complications of rapid progression of the retroviral infection. We conclude that co-infection and disease due to either HIV-1 or HTLV-I/II infection and tuberculosis is common, that the ocurrence of HTLV-I in this population is higher than previously recognized, and that prognosis associated with the management of tuberculosis is adversely affected by the presence of either retroviral infection. In a few patients with both retroviral infections, mortality was very high. All patients with tuberculosis should be tested for retroviral infection because of the prognostic and therapeutic implications.