Hyperglycemia during tuberculosis treatment increases morbidity and mortality in a contemporary cohort of HIV-infected patients in Rio de Janeiro, Brazil

Abstract

Background: Hyperglycemia occurs in tuberculosis (TB), but the long-term impact is unknown. We estimated the prevalence of hyperglycemia and compared the TB treatment outcomes and 1-year mortality rate according to the glycemic status noted during TB treatment.

Methods: We conducted a retrospective cohort analysis of adult patients who had TB and HIV coinfection and started receiving TB treatment at the Instituto Nacional de Infectologia Evandro Chagas, Brazil, between 2010-2015. Diabetes Mellitus (DM) and hyperglycemia were defined according to the American Diabetes Association. After excluding for known DM at baseline, the proportion of participants who developed new-onset DM after TB treatment was assessed. TB outcome was classified as successful or adverse (i.e., treatment failure, abandonment, and death). Kaplan-Meier survival curves were compared by the log-rank test based on the glycemic status of patients. Multivariate Cox regression models were used to assess the association between hyperglycemia and 1-year mortality. Two-sided p values <0.05 were considered statistically significant.

Results: We identified 414 euglycemic patients (87.5%), 49 hyperglycemic patients (10.3%), and 10 patients with known DM (2.1%). Diabetic patients were older compared to the euglycemic and hyperglycemic patients (47.9 vs. 37 vs. 39.7 years, respectively, p=0.001). Diabetic patients frequently had cavitation on chest image compared to hyperglycemic and euglycemic patients (50% vs. 23.4% vs. 15.3%, p=0.007, respectively). Hyperglycemic patients had more new-onset DM at follow-up compared to euglycemic (22 vs. 1; p<0001). Hyperglycemia was associated with adverse outcomes (71.4% vs. 24.6%, p<0.0001) compared to euglycemia. Crude 1-year mortality was significantly higher in patients with hyperglycemia compared with euglycemia (48.9% vs. 7.9%; unadjusted HR: 5.79 (3.74-8.96)). In the adjusted Cox models, hyperglycemia remained a significant factor for increased 1-year mortality (adjusted HR: 3.72 (2.17-6.38)].

Conclusions: Hyperglycemia frequently occurs in HIV-infected patients who commence TB treatment, and it increases the risks of adverse TB outcomes and 1-year mortality. Glucose testing during TB treatment detects patients at risk of adverse outcomes.

Keywords: Brazil; Diabetes mellitus; Hyperglycemia; Mortality; Tuberculosis.