Researchers have identified four promising blood-based metabolites that may help distinguish individuals with latent tuberculosis infection (LTBI) from those without infection, according to a new study published in BIO Integration. The findings provide fresh insights into the metabolic changes associated with latent tuberculosis and could support the development of improved diagnostic approaches in the future.
Latent tuberculosis infection remains a major global health challenge because infected individuals do not show symptoms and current diagnostic methods have important limitations. The absence of a definitive gold-standard test makes accurate identification of LTBI particularly difficult.
In the study, researchers investigated plasma metabolic profiles using an untargeted metabolomics strategy. Individuals with LTBI were recruited from close contacts of tuberculosis patients and confirmed as QuantiFERON-TB Gold positive, while participants without LTBI were selected from prison detainees who tested negative for the same assay.
The research team analyzed plasma samples from 199 participants using ultra-high-performance liquid chromatography coupled with quadrupole time-of-flight tandem mass spectrometry. Advanced statistical methods were then applied to identify metabolites that differed significantly between the two groups.
The analysis revealed 43 metabolites with significant differences between LTBI and non-LTBI participants. Among these, four metabolites, leucylleucine, tryptophyl-phenylalanine, lysoPE(18:1(11Z)/0:0), and biliverdin, showed particularly strong potential for distinguishing latent tuberculosis infection.
These metabolites demonstrated high diagnostic discrimination in the study population, with area under the curve (AUC) values ranging from 0.975 to 0.981. Models combining selected metabolites produced even higher classification performance during internal validation, with some AUC values approaching 1.00.
Despite these encouraging results, the researchers emphasized that the findings should be considered preliminary. Since metabolite selection and model evaluation were conducted within the same study cohort, the reported diagnostic performance may be overestimated. Furthermore, no external validation was performed to confirm the results in independent populations.
The study also highlighted several limitations. Participants in the LTBI and non-LTBI groups originated from different source populations, which may have introduced selection bias and unmeasured confounding factors. Additionally, all identified metabolites were classified at Metabolomics Standards Initiative level 2 and were not confirmed using authentic reference standards. No targeted metabolomic validation was conducted.
Researchers concluded that the study provides exploratory evidence of metabolic differences associated with latent tuberculosis infection and identifies several candidate biomarkers worthy of further investigation. However, larger studies involving well-matched populations and targeted validation approaches will be necessary before these biomarkers can be considered for clinical application.
The findings contribute to ongoing efforts to develop more accurate and accessible tools for the detection of latent tuberculosis infection, a critical component of global tuberculosis control strategies.
