GLP-1 Receptor Agonists Linked to Lower Tuberculosis Risk in Type 2 Diabetes
核心洞察
A large retrospective cohort study in Nature Communications found GLP-1 receptor agonist therapy was associated with a significantly lower incidence of tuberculosis (搜索) in patients with type 2 diabetes (搜索).
The analysis compared GLP-1RAs against sulfonylureas, metformin, DPP-4 inhibitors (搜索), and SGLT2 inhibitors (搜索) using propensity score-matched data from the TriNetX (搜索) network of roughly 153 million individuals.
Lower pulmonary TB risk was observed across all four comparisons, while extrapulmonary TB risk was significantly lower only versus sulfonylureas and DPP-4 inhibitors (搜索).
A new study published in Nature Communications places a widely used class of diabetes medicines under the microscope, asking whether glucagon-like peptide-1 receptor agonists (搜索) (GLP-1RAs) influence the risk of tuberculosis (搜索) in people living with type 2 diabetes (搜索). The research, led by KM Liao, JY Wu, and CC Lai, sits at the intersection of two major global health challenges: metabolic disease, which affects hundreds of millions of people worldwide, and tuberculosis, an infectious disease that continues to cause substantial illness and death despite the availability of effective antibiotics.
The investigation is especially timely because GLP-1 receptor agonists are moving rapidly from specialized diabetes care into mainstream medical practice. Drugs in this class are now widely discussed not only for their effects on glycated hemoglobin and body weight, but also for their potential cardiovascular and kidney benefits. As exposure increases across diverse populations, uncommon or indirect effects may become easier to detect.
Study Design and Methods
The association between GLP-1RA use and tuberculosis (搜索) risk was assessed in a large retrospective cohort study utilizing TriNetX (搜索), a global health research network containing electronic health records from approximately 153 million individuals across 143 healthcare organizations. Incidence of tuberculosis was compared between individuals with type 2 diabetes (搜索) initiating GLP-1RAs and those starting other commonly prescribed glucose-lowering medications.
Eligibility criteria required participants to be new users of their assigned antidiabetic agents, excluding those with prior exposure to these medications. Treatment groups were determined by the first antidiabetic drug class initiated during the study period (intention-to-treat), and follow-up began one day after the initial prescription. Propensity score matching was performed to enhance comparability, incorporating demographic, clinical, comorbidity, diabetes complication, socioeconomic, lifestyle, and baseline medication variables.
The primary outcome was tuberculosis (搜索) incidence, with secondary outcomes of pulmonary and extrapulmonary tuberculosis. Follow-up continued until tuberculosis diagnosis, last recorded clinical encounter, death, or five years after the index date.
Key Findings
Within the matched cohorts, individuals receiving GLP-1RAs were compared to those receiving sulfonylureas, metformin, DPP-4 inhibitors (搜索), and SGLT2 inhibitors (搜索). Baseline characteristics such as age, sex, body mass index, and comorbidities were balanced between groups using propensity score matching.
The analysis revealed that GLP-1RA therapy was associated with a significantly lower incidence of tuberculosis (搜索) among patients with type 2 diabetes (搜索) compared with all four comparator drug classes. Lower pulmonary tuberculosis risk was observed in all four comparisons, while extrapulmonary tuberculosis risk was significantly lower only compared with sulfonylureas and DPP-4 inhibitors (搜索). Compared with SGLT2 inhibitors (搜索), pulmonary tuberculosis risk was significantly lower, whereas extrapulmonary tuberculosis risk did not differ significantly between groups.
Subgroup analyses comparing sulfonylureas revealed that the protective association was strongest among younger patients and those with lower body mass index (BMI). However, subgroup patterns varied across the other comparator groups.
Sensitivity analyses generally supported these findings across different patient populations and analytic approaches. However, in the U.S.-restricted analysis, the association with lower tuberculosis (搜索) risk was attenuated and was not statistically significant compared with SGLT2 inhibitors (搜索). No association was found between GLP-1RA use and the unrelated negative-control outcome of skin cancer, providing some reassurance against residual confounding unrelated to tuberculosis.
Biological and Clinical Context
Tuberculosis (搜索) is caused primarily by Mycobacterium tuberculosis (搜索), a bacterium that usually enters the body through inhaled airborne particles and most often affects the lungs. After infection, the immune system may contain the bacteria in a latent state; if immune control weakens, latent infection can reactivate and develop into active tuberculosis.
Type 2 diabetes (搜索) is already recognized as an important risk factor for tuberculosis (搜索). Persistently elevated blood glucose can interfere with several layers of host defense, including the function of macrophages, neutrophils, and lymphocytes. Diabetes may also impair inflammatory signaling and alter the structure and function of lung tissue, creating conditions in which tuberculosis bacteria are more likely to survive or escape containment.
The biological interpretation of the findings is complex. If an association were observed, it would not automatically prove that GLP-1 receptor agonists directly cause or prevent tuberculosis (搜索). Improved metabolic control could alter risk in one direction, while changes in inflammatory signaling could push it in another. The timing of treatment, baseline immune status, and the presence of complications such as chronic kidney disease could also shape outcomes.
Limitations and Future Directions
Because this was an observational electronic health record study, the findings cannot establish that GLP-1RAs prevent tuberculosis (搜索). The database lacked information on tuberculosis contact history, latent infection, radiographic findings, disease severity, medication dosage and adherence, and direct socioeconomic measures. The authors also noted possible geographic and time-related confounding and acknowledged that one exposure-restricted sensitivity analysis introduced immortal time bias because patients were classified using treatment information obtained after follow-up had begun.
To substantiate and clarify these observations, future research should focus on elucidating underlying mechanisms, conducting prospective clinical studies, and systematically reviewing tuberculosis (搜索) cases reported in randomized controlled trials. The findings may also encourage further laboratory research into GLP-1 signaling, macrophage function, and the mechanisms that determine whether M. tuberculosis remains latent or becomes active.