Importance: Although experimental studies support the hypothesis that exposure of infectious agents may trigger an aberrant immune response and contribute to noninfectious uveitis, the association of a definite pathogen with human noninfectious uveitis conditions appears not to have been well established in a population.
Objective: To evaluate associations of tuberculosis infection with risk of several noninfectious uveitis conditions.
Design, setting, and participants: These mendelian randomization and observational analyses were conducted with the genetic data of a Chinese cohort enrolled between April 2008 and January 2018 and a Japanese cohort enrolled between January 2002 and June 2009. We recruited participants for T-SPOT.TB (Oxford Immunotec) assays between July and November 2019. The Chinese cohort included patients with uveitis associated with Behçet disease or other uveitis conditions and control participants. The Japanese cohort and the group given T-SPOT.TB assays included individuals with Behçet disease and control participants. Data analyses for this study were completed from July 2019 to January 2020.
Exposures: Genetic variants associated with tuberculosis as natural proxies for tuberculosis exposure.
Main outcomes and measures: The primary outcome was the odds ratio (OR) for Behçet disease, estimated by an inverse variance weighted mean of associations with genetically determined tuberculosis susceptibility. The T-SPOT.TB positivity rate was examined in individuals with Behçet disease and compared with that of control participants.
Results: The Chinese cohort included 999 patients with uveitis associated with Behçet disease, 1585 with other uveitis conditions, and 4417 control participants. The Japanese cohort included 611 individuals with Behçet disease and 737 control participants. The group given T-SPOT.TB assays included 116 individuals with Behçet disease and 121 control participants. Of the Chinese individuals with Behçet disease and control participants, 2257 (41.7%) were female and the mean (SD) age was 35.4 (12.5) years. In the Japanese cohort, 564 (41.8%) were female and the mean (SD) age was 39.1 (12.7) years. Genetically determined tuberculosis susceptibility was associated with an increased risk for Behçet disease. The OR for Behçet disease per 2-fold increase in tuberculosis incidence was 1.26 (95% CI, 1.12-1.43; P = 1.47 × 10-4). Replication using the Japanese cohort yielded similar results (OR, 1.16 [95% CI, 1.08-1.26]). In T-SPOT.TB assays, having a positive result, indicating a history of tuberculosis infection, was found to be an independent risk factor for Behçet disease (OR, 2.26 [95% CI, 1.11-4.60]).
Conclusions and relevance: These human genetic and biomarker data demonstrated that tuberculosis exposure was a risk factor for Behçet disease. This study provides novel evidence linking an infectious agent to the risk of a noninfectious uveitis condition.