Risk factors affecting the mortality of HIV-infected patients with pulmonary tuberculosis in the cART era: a retrospective cohort study in China

Tuberculosis infection still places a great burden on HIV-infected individuals in China and other developing countries. Knowledge of the survival of HIV-infected patients with pulmonary tuberculosis (PTB) would provide important insights for the clinical management of this population, which remains to be well described in current China. HIV-infected patients with PTB admitted to Shanghai Public Health Clinical Center from January 2011 to December 2015 were retrospectively enrolled. In this cohort, the survival prognosis was estimated by the Kaplan-Meier method, while univariate and multivariate Cox proportional hazards models were used to determine the risk factors affecting mortality. After reviewing 4914 admitted patients with HIV infection, 359 PTB cases were identified. At the time of PTB diagnosis, the patients' median CD4 T cell count was 51 /mm (IQR: 23-116), and 27.30% of patients (98/359) were on combination antiretroviral therapy (cART). For the 333 cases included in the survival analysis, the overall mortality was 15.92% (53/333) during a median 27-month follow-up. The risk factors, including age older than 60 years (HR: 3.18; 95% CI: 1.66-6.10), complication with bacterial pneumonia (HR: 2.64; 95% CI: 1.30-5.35), diagnosis delay (HR: 2.60; 95% CI: 1.42-4.78), CD4 T cell count less than 50/mm (HR: 2.38; 95% CI: 1.27-4.43) and pulmonary atelectasis (HR: 2.20; 95% CI: 1.05-4.60), might independently contribute to poor survival. Among patients without cART before anti-TB treatment, the later initiation of cART (more than 8 weeks after starting anti-TB treatment) was found to increase the mortality rate (OR: 4.33; 95% CI: 1.22-15.36), while the initiation of cART within 4-8 weeks after starting anti-TB treatment was associated with the fewest deaths (0/14). The subjects in this study conducted in the cART era were still characterized by depressed immunological competence and low rates of cART administration, revealing possible intervention targets for preventing TB reactivation in HIV-infected individuals under current circumstances. Furthermore, our study indicated that the timely diagnosis of PTB, prevention of secondary bacterial pneumonia by prophylactic management and optimization of the timing of cART initiation could have significant impacts on decreasing mortality among HIV/PTB co-infected populations. These findings deserve further prospective investigations to optimize the management of HIV/PTB-co-infected patients. NCT01344148 , Registere