Min Yang,1– 3,* Wenfei Wang,3,* Peize Zhang,4 Guizhen Liu,5 Hailin Lu,6 Mingjie He,5 Guofang Deng,2,3 Xiaoyou Chen1,7 

1Department of Pharmacology, Beijing Tuberculosis and Thoracic Tumor Research Institute, Beijing Chest Hospital, Capital Medical University, Beijing, People’s Republic of China; 2Second Department of Pulmonary Medicine and Tuberculosis, the Third People’s Hospital of Shenzhen, Southern University of Science and Technology, Shenzhen, Guangdong, People’s Republic of China; 3National Clinical Research Center for Infectious Disease, the Third People’s Hospital of Shenzhen, Southern University of Science and Technology, Shenzhen, Guangdong, People’s Republic of China; 4Fourth Department of Pulmonary Medicine and Tuberculosis, the Third People’s Hospital of Shenzhen, Southern University of Science and Technology, Shenzhen, Guangdong, People’s Republic of China; 5The First Affiliated Hospital of Gannan Medical University, Ganzhou, Jiangxi, People’s Republic of China; 6Gannan Medical University, Ganzhou, Jiangxi, People’s Republic of China; 7Infectious Diseases Department, Beijing Ditan Hospital, Capital Medical University, Beijing, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Xiaoyou Chen; Guofang Deng, Email chenxy1998@hotmail.com; jxxk1035@yeah.net

Objective: We aimed to investigate dysregulated metabolic pathways and identify diagnostic and therapeutic targets in patients with tuberculosis-diabetes (TB-DM).
Methods: In our prospective cohort study, plasma samples were collected from healthy individuals, diabetic (DM) patients, untreated TB-only (TB-0)/TB-DM patients (TB-DM-0), and cured TB (TB-6)/TB-DM patients (TB-DM-6) to measure the levels of amino acids, fatty acids, and other metabolites in plasma using high-throughput targeted quantification methods.
Results: Significantly different biological processes and biomarkers were identified in DM, TB-DM-0, and TB-DM-6 patients. Moreover, quinolinic acid (QA) showed excellent predictive accuracy for distinguishing between DM patients and TB-DM-0 patients, with an AUC of 1 (95% CI 1– 1). When differentiating between TB-DM-0 patients and TB-DM-6 patients, the AUC was 0.9297 (95% CI 0.8460– 1). Compared to those in DM patients, the QA levels were significantly elevated in TB-DM-0 patients and decreased significantly after antituberculosis treatment. We simultaneously compared healthy controls and untreated tuberculosis patients and detected an increase in the level of QA in the plasma of tuberculosis patients, which decreased following treatment.
Conclusion: These findings improve the current understanding of tuberculosis treatment in patients with diabetes. QA may serve as an ideal diagnostic biomarker for TB-DM patients and contribute to the development of more effective treatments.

Keywords: pulmonary tuberculosis, diabetes, metabolomics, quinolinic acid, biomarker