Efficacy and Safety of 4-Month Rifapentine-Based Tuberculosis Treatments in Persons with Diabetes

Abstract

A previous study demonstrated noninferior efficacy of 4-month rifapentine/moxifloxacin regimen for tuberculosis (TB) treatment compared with the standard regimen. We explored results among study participants with diabetes. Among 2,516 randomized participants, 181 (7.2%) had diabetes. Of 166 participants with diabetes in the microbiologically eligible analysis group, 26.3% (15/57) had unfavorable outcomes in the control regimen, 13.8% (8/58) in the rifapentine/moxifloxacin regimen, and 29.4% (15/51) in the rifapentine regimen. The difference in proportion of unfavorable outcomes between the control and rifapentine/moxifloxacin arms in the microbiologically eligible analysis group was -12.5% (95% CI -27.0% to 1.9%); the difference between the control and rifapentine arms was 3.1% (95% CI -13.8% to 20.0%). Safety outcomes were similar in the rifapentine/moxifloxacin regimen and control arms. Among participants with TB and diabetes, the rifapentine/moxifloxacin arm had fewest unfavorable outcomes and was safe. Our findings indicate that the rifapentine/moxifloxacin regimen can be used in persons with TB and diabetes.

Keywords: United States; antimicrobial resistance; bacteria; diabetes; moxifloxacin; phase 3 clinical trial; respiratory infections; rifapentine; tuberculosis and other mycobacteria.

Figure 1

Unadjusted differences in unfavorable outcomes in each analysis population among participants with diabetes in a study assessing efficacy and safety of 4-month rifapentine-based tuberculosis treatments in persons with diabetes at sites in 12 countries (Brazil, Haiti, India, Kenya, Malawi, Peru, South Africa, Thailand, Uganda, United States, Vietnam, and Zimbabwe), January 2016–October 2018. Results of the efficacy results in all 5 analysis populations are shown: rifapentine/moxifloxacin regimen versus control regimen (A) and rifapentine regimen versus control regimen (B). Solid dots indicate primary results, open dots indicate secondary results, and error bars indicate 95% CIs. Dashed vertical line indicates the noninferiority margin of 6.6% for overall results in the randomized trial (18).

Figure 2

Analysis of time to sputum culture conversion (number of weeks from randomization) in liquid (A) and solid media (B) among participants with diabetes, by tuberculosis drug regimen, in the microbiologically eligible analysis population in a study assessing efficacy and safety of 4-month rifapentine-based tuberculosis treatments in persons with diabetes at sites in 12 countries (Brazil, Haiti, India, Kenya, Malawi, Peru, South Africa, Thailand, Uganda, United States, Vietnam, and Zimbabwe), January 2016–October 2018. Because scheduled study visits did not necessarily occur exactly at 8 weeks, the proportion of participants with culture conversion at 8 weeks is estimated from the Kaplan-Meier estimator at t = 10 weeks. Differences were not statistically significant for any comparisons.