Abstract
BACKGROUND: Type 2 diabetes mellitus (DM) increases both the risk of acquiring latent tuberculosis (TB) infection (LTBI) and progression to active TB. The immunological mechanisms underlying the increased susceptibility remain poorly understood. This study aimed to elucidate the impact of DM on Mtb-specific cytokine and chemokine responses and identify potential biomarkers for LTBI regardless of DM status. METHODS: This cross-sectional study recruited 164 participants with LTBI-DM (n = 51), LTBI-only (n = 63), DM-only (n = 12) and healthy controls (n = 38) at Kiruddu National Referral Hospital. Cytokine and chemokine responses were measured in ESAT-6 and CFP-10 whole blood stimulated supernatants using the Luminex assay. Linear regression with age as a covariate and false discovery rate (p < 0.050) correction were performed on log(2)-transformed concentrations. For biomarker analysis, comparisons were performed for LTBI (LTBI-DM + LTBI-only) versus no LTBI (DM-only + healthy controls) groups, and receiver operating characteristic analyses were performed for both univariate and multivariate analyses. RESULTS: LTBI-DM was associated with decreased Mtb-specific IFN-γ (p(adjusted)=0.006), TNF (p(adjusted)=0.023), IL-12 (p(adjusted)=0.004), IP-10 (p(adjusted)=0.004), IL-8 (p(adjusted)=0.030) and IL-10 (p(adjusted)=0.043) responses compared to LTBI-only. Univariate analysis identified IFN-γ (AUC: 0.80, sensitivity: 81%, specificity: 78%), IL-2 (0.84, 75%, 88%), IL-12 (0.77, 65%, 82%) and IP-10 (0.73, 77%, 62%) as top-performing biomarkers for LTBI identification regardless of DM status. A combined four-marker biosignature yielded an AUC of 0.92 (CI: 0.88-0.97) with 76% sensitivity and 96% specificity. CONCLUSION: Diabetes mellitus impairs Th1, inflammatory, and regulatory cytokine and chemokine responses during LTBI. A multivariate biosignature of IFN-γ, IL-2, IL-12 and IP-10 outperforms single-marker assays (e.g., variability in IFN-γ due to immunosuppression in DM) and offers robust identification of LTBI regardless of DM status, highlighting the additive value of combining mediators that capture distinct immunological axes. The improved biosignature specificity may minimise false positives in DM populations, who often exhibit comorbidities that could mimic LTBI-associated inflammation. Our findings support validating combined biomarker approaches to enhance LTBI diagnosis and prognosis in the immunocompromised DM population.