The effectiveness of treatment of patients with a combination of tuberculosis and HIV infection (TB/HIV) remains insufficient. According to the who global report on tuberculosis, while the success rate for new cases and relapses of tuberculosis was 85%, for similar TB/HIV patients it was only 75%. However, the reduction in the percentage of successful treatment of TB/HIV patients was primarily due to their higher mortality rate. There was also a large proportion of TB/HIV cases that were lost to follow-up (this international definition is similar to the Russian definition of the outcome of "interrupted chemotherapy"). Aim: to identify predictors of the death of patients with a combination of tuberculosis and HIV infection and to develop organizational measures to reduce their mortality. Materials and methods. Potential predictors of the lethal outcome of 1727 patients registered in 2016 in 13 regions of the Russian Federation were studied, of which 1003 successfully completed the treatment, and 724 died during the course of treatment. Criteria for inclusion in the study: the presence of co-infection, the successful completion of a course of chemotherapy or the death of a patient. Criterion for exclusion from the study: registration of other outcomes of the course of treatment. The influence of age, gender, place of residence (city or rural residence), the history of previous treatment, the appointment of antiretroviral therapy and co-trimoxazole, smear microscopy result, drug resistance of mycobacterium tuberculosis, and destructive changes in the lungs were studied. The analysis was done in two stages: first, we tested the hypothesis about the effect of potential predictors on the death rate using contingency tables, and then checked the predictor independence using logistic regression. Results. The probability of death was statistically significantly increased: the male gender of the patient, his rural residence, prior treatment default or a chronic history of tuberculosis, a positive sputum smear, multidrug resistance of mycobacterium tuberculosis, and the presence of lung tissue destruction. Antiretroviral drugs have been shown to be less likely to be fatal. Independent predictors were: the appointment of antiretroviral therapy, multidrug resistance of mycobacterium tuberculosis, massive bacterial excretion, and living in rural areas. Conclusion. To reduce the mortality of patients with co-infection, their full coverage with antiretroviral therapy should be ensured, measures to actively detection TB cases in people living with HIV, and the availability of treatment for patients in rural areas should be strengthened. It is advisable to conduct a detailed study of death predictors in patients with a combination of HIV and multidrug-resistant tuberculosis.