Abstract

AIMS: Diabetes mellitus (DM) is associated with worse tuberculosis (TB) treatment outcomes, especially among those with poor glycemic control. We examined whether a structured clinical algorithm could improve glycemic control in TB patients with DM. METHODS: In an open label randomized trial, TB-DM patients were randomized to scheduled counselling, glucose monitoring, and adjustment of medication using a structured clinical algorithm (intervention arm) or routine DM management (control arm), with glycated hemoglobin (HbA1c) at month 6 as the primary end point. RESULTS: We randomized 150 pulmonary TB-DM patients (92% culture positive, 51.3% male, mean age 53 years). Baseline mean HbA1c was 11.0% in the intervention arm (n=76) and 11.6% in the control arm (n=74). At 6 months, HbA1c had decreased more in the intervention arm compared with the control arm (a difference of 1.82% HbA1c, 95% CI 0.82-2.83, p<0.001). Five patients were hospitalized in the intervention arm and seven in the control arm. There was more hypoglycemia (35.0% vs 11.8%; p=0.002) in the intervention arm. Two deaths occurred in the intervention arm, one due to cardiorespiratory failure and one because of suspected septic shock and multiorgan failure. CONCLUSION: Regular monitoring and algorithmic adjustment of DM treatment led to improved glycemic control.