INTRODUCTION: The increasing burden of diabetes mellitus in low- and middle-income countries negatively impacts tuberculosis control. To understand this dual burden in Eswatini, we describe the prevalence and predictors of elevated baseline blood glucose and unfavourable tuberculosis treatment outcomes. METHODS: We conducted a prospective cohort study at 11 health facilities in Eswatini and included adults ≥18 years commencing tuberculosis treatment. Blood glucose measurements were taken at baseline, months 2 and 5, and patients' sociodemographic and clinical data were extracted. We computed the prevalence of elevated blood glucose and used logistic regression to determine the predictors of elevated baseline blood glucose and unfavourable treatment outcomes. RESULTS: Of 369 consecutively enrolled patients, the mean age was 38.4 (SD 12.9) years, and 202 (54.7%) were males. The prevalence of elevated baseline blood glucose was 8.0% (95% CI: 5.5, 11.3)
8.9% in males (95% CI: 5.6, 13.9)
highest at ≥55 years (13.6%
95% CI: 6.2, 27.3) and in patients with reactive HIV at 9.5% (95% CI: 6.5, 13.7). A family history of diabetes mellitus (adjusted OR (AOR) 2.80
95% CI: 1.08, 7.32) and a reactive HIV status (AOR 4.62
95% CI: 1.06, 20.11) significantly predicted elevated baseline blood glucose. Three-quarters (n=276, 75.4%) had a favourable tuberculosis treatment outcome
more males (n=59, 66%) had an unfavourable treatment outcome (p=0.020), the most common unfavourable outcome being death (n=34, 9.2%). Hypertension (AOR 4.84
95% CI: 1.48, 15.7), unemployment (AOR 2.01
95% CI: 1.08, 3.71) and high school education (AOR 0.32
95% CI: 0.16, 0.64) were associated with unfavourable treatment outcome. CONCLUSION: Our study shows the need to optimise care for patients receiving treatment for tuberculosis by integrating screening for and treatment of diabetes and hypertension, prioritising males, those aged ≥55 years and those with a reactive HIV status to limit unfavourable outcomes and death.