Tuberculosis Exposure With Risk of Behçet Disease Among Patients With Uveitis

Zhenyu Zhong; Guannan Su; Qingyun Zhou; Akira Meguro; Masaki Takeuchi; Nobuhisa Mizuki; Shigeaki Ohno; Weiting Liao; Xiaojie Feng; Jiadong Ding; Yingnan Gao; Wei Chi; Meifen Zhang; Xianbo Zuo; Peizeng Yang

doi:10.1001/jamaophthalmol.2020.6985

. 2021 Feb 18;139(4):1–8. doi: 10.1001/jamaophthalmol.2020.6985

Tuberculosis Exposure With Risk of Behçet Disease Among Patients With Uveitis

Zhenyu Zhong ¹, Guannan Su ¹, Qingyun Zhou ¹, Akira Meguro ², Masaki Takeuchi ², Nobuhisa Mizuki ², Shigeaki Ohno ³, Weiting Liao ¹, Xiaojie Feng ¹, Jiadong Ding ⁴, Yingnan Gao ⁴, Wei Chi ⁵, Meifen Zhang ⁶, Xianbo Zuo ⁷, Peizeng Yang ^1,^✉

¹The First Affiliated Hospital of Chongqing Medical University, Chongqing, China

²Department of Ophthalmology, Yokohama City University School of Medicine, Yokohama, Japan

³Department of Ophthalmology, Faculty of Medicine and Graduate School of Medicine, Hokkaido University, Sapporo, Japan

⁴The First Affiliated Hospital of Zhengzhou University, Henan Province Eye Hospital, Zhengzhou, China

⁵State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China

⁶Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China

⁷The First Affiliated Hospital of Anhui Medical University, Institute of Dermatology, Anhui Medical University, Hefei, China

Accepted for Publication: December 18, 2020.

Published Online: February 18, 2021. doi:10.1001/jamaophthalmol.2020.6985

^✉

Corresponding Author: Peizeng Yang, MD, PhD, The First Affiliated Hospital of Chongqing Medical University, Youyi Road 1, Chongqing 400016, China (peizengycmu@126.com).

Author Contributions: Drs Yang and Zhong had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Zhong and Su are co–first authors.

Concept and design: Zhong, Takeuchi, Yang.

Acquisition, analysis, or interpretation of data: Zhong, Su, Zhou, Meguro, Mizuki, Ohno, Liao, Feng, Ding, Gao, Chi, Zhang, Zuo.

Drafting of the manuscript: Zhong, Takeuchi.

Critical revision of the manuscript for important intellectual content: Zhong, Su, Zhou, Meguro, Mizuki, Ohno, Liao, Feng, Ding, Gao, Chi, Zhang, Zuo, Yang.

Statistical analysis: Zhong, Meguro, Takeuchi, Mizuki.

Obtained funding: Yang.

Administrative, technical, or material support: Zhong, Su, Meguro, Mizuki, Ohno, Yang.

Supervision: Takeuchi, Yang.

Conflict of Interest Disclosures: Dr Takeuchi reported support from Ministry of Health, Labour and Welfare, Japan (grant H29-Nanchi-Ippan-050) during the conduct of the study. Dr Mizuki reported grants from the Ministry of Health, Labour and Welfare, Japan during the conduct of the study. No other disclosures were reported.

Funding/Support: The work was supported by National Natural Science Foundation Key Program (grant 81930023), Natural Science Foundation Major International (Regional) Joint Research Project (grant 81720108009), Chongqing Outstanding Scientists Project (grant 2019), Chongqing Key Laboratory of Ophthalmology (CSTC; grant 2008CA5003), Chongqing Science & Technology Platform and Base Construction Program (grant cstc2014pt-sy10002), the Chongqing Chief Medical Scientist Project (2018), and the Ministry of Health, Labour and Welfare, Japan (grants H20-Nanchi-Ippan-006, H23-Nanchi-Ippan-006, H26-Nanchi-Ippan-054, and H29-Nanchi-Ippan-050).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

^✉

Corresponding author.

PMCID: PMC7893545 PMID: 33599689

Key Points

Question

What is the evidence for tuberculosis exposure in association with noninfectious uveitis in a human population?

Findings

In this mendelian randomization and observational cohort study involving 8349 people and a T-SPOT.TB analysis involving 237 participants with a history of uveitis, human genetic and biomarker data demonstrated that tuberculosis was a risk factor for Behçet disease.

Meaning

The present study provides evidence linking tuberculosis exposure with the causative mechanism of a noninfectious uveitis entity, Behçet disease, and may add to knowledge concerning the pathologic consequences of tuberculosis.

This mendelian randomization and cohort study evaluates associations of tuberculosis with several noninfectious uveitis conditions, particularly Behçet disease, in cohorts in China and Japan, including a subgroup given T-SPOT.TB assays.

Abstract

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.

Introduction

Uveitis is one of leading causes of visual acuity loss. Its prevalence was estimated to be 38 to 714 cases per 100 000 people.¹ Uveitis can be clinically categorized as infectious or noninfectious. Epidemiological data show that noninfectious uveitis can account for most cases of uveitis in some populations.^2,3 Most noninfectious uveitis conditions are presumed to be immune-mediated disease, driven by autoinflammatory or autoimmune responses, such as Vogt-Koyanagi-Harada disease, uveitis associated with ankylosing spondylitis, and Behçet disease.⁴ Although the hypothesis is supported by experimental studies that exposure of infectious agents may trigger an aberrant immune response and contribute to noninfectious uveitis,^5,6 the causal association of a definite pathogen with human noninfectious uveitis conditions has not been well established in a population.⁷

Tuberculosis is currently one of the most prevalent infectious diseases and is caused by Mycobacterium tuberculosis. Approximately 25% of the global population may have a latent tuberculosis infection, of which 5% to 10% could progress to an active disease in their lifetime.⁸ It has been suggested that M tuberculosis could induce inappropriate host responses to self-antigens, thus triggering the subsequent development of an autoimmune or autoinflammatory disorder.^9,10 However, the frequent use of glucocorticoids and immunosuppressive or biologic agents in the treatment of autoimmune or autoinflammatory diseases may also confer an increased risk of tuberculosis infection or reactivation.^11,12 The association between tuberculosis and autoimmune disease is thus not yet clear, since one seems to pose a risk of the other.

Currently, it remains unknown whether tuberculosis exposure may contribute to the development of noninfectious uveitis. Interpretations of observational analyses might be biased in assessing associations because of potential reverse causality and residual confounders. A conventional randomized clinical trial of tuberculosis infection cannot be ethical and feasible. In this study, we therefore sought to investigate the potential association based on the principle of mendelian randomization.¹³ This approach has been successfully applied to provide evidence that can guide public health interventions.^14,15,16 The role of host genetic factors in determining susceptibility to tuberculosis has been well established.^17,18 Genome-wide association studies (GWAS) have identified common genetic variants associated with tuberculosis.^17,18 These variants, randomly allocated at birth, could be a natural determinant of tuberculosis risk that may be used for inference on health outcomes.¹³ We aim to assess whether tuberculosis is a risk for several noninfectious uveitis conditions, including Behçet disease, Vogt-Koyanagi-Harada disease, uveitis associated with ankylosing spondylitis, and Fuchs uveitis syndrome, by using these genetic variants in the population.

Methods

Study Design

We performed a mendelian randomization in which genetic variants strongly associated with tuberculosis were used as natural indicators of risk or likelihood in tuberculosis infection, to test potential association with the risk of noninfectious uveitis conditions (Figure 1). In the mendelian randomization design, we selected genetic variants directly associated with tuberculosis, not associated with confounders and any alternative pathways that could affect the outcome. Under the assumption that the randomly inherited genotype of individuals determines an assignment for tuberculosis susceptibility at birth, the outcome of genetically determined tuberculosis susceptibility can be analogous to the intention-to-treat effect in a randomized clinical trial.¹³ This mendelian randomization study was designed in June 2019, and data were analyzed during July 2019 and January 2020 based on the Chinese and Japanese GWAS data sets. In the Chinese cohort, a total of 2584 patients with uveitis and 4417 healthy control participants of Chinese ancestry were enrolled at 5 centers in China from April 2008 through January 2018 for the genome-wide genotyping and association study (Peizeng Yang, MD, PhD; written communication; July 15, 2019). Among Chinese patients with uveitis, diagnosis of Behçet disease, Vogt-Koyanagi-Harada disease, and ankylosing spondylitis was made according to international standardized criteria.^19,20,21 A diagnosis of Fuchs uveitis syndrome was principally clinical, based on the description of classic literature.^22,23 In the Japanese cohort, 611 individuals with Behçet disease and 737 healthy control participants were enrolled at 8 centers in Japan between January 2002 and June 2009, after which the main GWAS results were published elsewhere.²⁴ Our primary analysis was based on the Chinese GWAS data set, and the Japanese data set was used as a replication. In addition, we performed a prospective observational study in which we examined the T-SPOT.TB positivity rate in individuals with Behçet disease compared with control participants. Consecutive participants with a history of uveitis were screened at the First Affiliated Hospital of Chongqing Medical University according to predefined exclusion criteria. Individuals were excluded if any of the following criteria were met: they were aged 50 years or older (this criterion was based on the consideration that the age of onset of Behçet disease is about 20 to 40 years old,²⁵ while tuberculosis infection can occur in older individuals; the observation may be associated with inversion of cause and effect, if elderly individuals are involved); were HIV positive; had serous effusion; or had used systemic corticosteroids of more than 20 mg per day in the last 3 months and had other serious medical conditions, such as hepatic failure, kidney failure, and heart failure. A total of 237 participants were enrolled for this investigation between July and November 2019.

Figure 1. — Mendelian randomization tests indicate an association by first establishing via linear regression that tuberculosis-associated genetic variants increase the risk of tuberculosis (β1). These variants are then tested for a linear association with the risk of uveitis conditions (β2). An assessment of association with the risk of uveitis conditions per change in the risk of tuberculosis can be obtained by β2/β1.

The GWAS study received ethical approval from the local institutional ethics committees (eMethods in the Supplement). All the participants provided written informed consent to use their genetic data for any investigation, and this investigation was designed in June 2019 and approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University. All the procedures adhered to the Declaration of Helsinki. There was no financial compensation for participants.

Genetic Variants

Five genetic variants (rs4240897, rs2269497, rs41553512, rs12437118, and rs6114027) associated with tuberculosis at genome-wide significance (P < 5.0 × 10⁻⁸) in the Chinese population were used as the instrumental variable for tuberculosis exposure (eTable 1 in the Supplement).^17,18 We obtained the summary estimates of associations between genetic variants and tuberculosis from 2 previously published GWAS, which finally included 4310 patients with tuberculosis vs 6386 healthy control participants¹⁷ and 2949 patients with tuberculosis vs 5090 control participants.¹⁸ The exposure allele for each single-nucleotide variant (SNV) was the allele associated with higher risk of tuberculosis. To our knowledge, there has been only 1 GWAS and replication study²⁶ reported on tuberculosis in the Japanese population, involving 393 young Japanese and Thai individuals with tuberculosis and 1255 healthy control participants, in which 1 SNV reached the significance at a P value less than 5.0 × 10^-8. To obtain sufficient power by multiple SNVs, we used a significance threshold of P < 5.0 × 10⁻⁶ to select genetic variants and obtained 3 eligible SNVs as proxies for tuberculosis in the Japanese population: rs6071980, rs4604310, and rs10792563. In the Chinese cohort, genome-wide genotyping was performed using the HumanOmniZhongHua-8 BeadChip (Illumina), and association analyses of SNV genotype with uveitis conditions were performed using logistic regression adjusted for sex, age, and testing additive, per-allele genetic outcomes. To perform 2-sample mendelian randomization analyses, we obtained summary statistics of these genetic variants from the Chinese and Japanese GWAS results. There was no sample overlap between each data set that we used. All SNVs in respective data sets are located on different chromosomes and are not in pairwise linkage disequilibrium. We evaluated each SNV and its relevant proxy SNVs with linkage disequilibrium (r² >0.80) for pleiotropic associations with potential confounders or other traits, except tuberculosis, by searching the PhenoScanner database (University of Cambridge; http://www.phenoscanner.medschl.cam.ac.uk/), which contained more than 65 billion genotype-phenotype association results from large-scale, published GWAS.²⁷ The genetic variant associated with other traits or having a proxy SNV associated with other traits at genome-wide significance (P < 5.0 × 10⁻⁸) was likely to be pleiotropic and was excluded in sensitivity analyses. All computations of pairwise linkage disequilibrium were based on the East Asia population in 1000 Genomes Project.²⁸

T-SPOT.TB Test

The T-SPOT.TB test is a diagnostic assay for past or present tuberculosis infection, by measuring effector T cells reactive to stimulation with M tuberculosis antigens ESAT-6 and CFP-10.²⁹ Samples were assayed using the T-SPOT.TB kit (Oxford Immunotec Ltd) according to the manufacturer’s instructions. The results were defined as positive when there was a positive response to ESAT-6 and/or CFP-10 antigens.

Statistical Analyses

The F statistics were calculated to evaluate the association strength.³⁰ For individual SNVs, mendelian randomization estimates of the association with outcomes were determined by the Wald-type ratio.³¹ We combined the associations of individual SNVs to produce an inverse variance–weighted mean of the associations in multiplicative random-effects and fixed-effects models.³¹ To avoid the potential heterogeneity bias, the estimates obtained by multiplicative random-effects inverse variance weighting were considered the primary outcome. All estimates were described as odds ratios (ORs) with 95% CIs per 2-fold increase in tuberculosis incidence.³² Heterogeneity was tested using the Cochran Q and I² statistics. An MR Egger regression was performed, whereby an intercept statistically different from null (0) could demonstrate the presence of directional pleiotropy (P < .05).³³ We performed a leave-1-out sensitivity analysis by excluding each SNV in turn to examine the outcome influenced by each genetic variant. Sensitivity analyses based on maximum likelihood, MR Egger regression, and weighted median and contamination mixture models were conducted under different assumptions about potential bias and pleiotropy.^31,33,34,35 Comparison of the results of sensitivity analyses with the primary analysis was performed using Fisher z scores.³⁶ We used logistic regression analysis to estimate the OR of T-SPOT.TB positivity for individuals with Behçet disease, whereby ORs were further adjusted for sex, age, body mass index, smoking history, drinking history, (urban or rural) residence, and education level. For the primary mendelian randomization estimate in the Chinese cohort, a 2-sided P < .0125 (by Bonferroni correction, [P < .05] / 4 entity outcomes) was considered to indicate statistical significance. This study did not plan for multiple-comparison adjustments for sensitivity analyses, so the results are reported with point estimates and 95% CIs only. The statistical analyses were performed using Stata version 15.0 (StataCorp) and R version 3.5.0 (R Foundation for Statistical Computing).

Results

Study Population and Genetic Variants

In the Chinese cohort, 999 were diagnosed with Behçet disease, 608 had Vogt-Koyanagi-Harada disease, 477 had uveitis associated with ankylosing spondylitis, and 500 had Fuchs uveitis syndrome; 4417 were control individuals. A Japanese cohort of 611 individuals with Behçet disease and 737 control participants was included as a replication; 509 of 611 patients (83.3%) in this cohort had uveitis.²⁴ 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 individuals (41.8%) were female and the mean (SD) age was 39.1 (12.7) years. All 5 genetic variants had sufficient association strength with tuberculosis in the Chinese population, as reflected by F statistics greater than 10.0 (rs4240897: F_10 694, 97.12; rs2269497: F_10 694, 62.39; rs41553512: F_10 694, 91.68; rs12437118: F₈₀₃₇, 57.83; rs6114027: F₈₀₃₇, 55.35; eTable 1 and eTable 2 in the Supplement). Additionally, a total of 116 individuals with Behçet disease and 121 control participants underwent T-SPOT.TB tests and were included in the observational analysis.

Primary Outcome

Estimates for the associations of individual variants with the risk of Behçet disease as well as other 4 conditions are shown in eTable 3 in the Supplement. Across all 5 tuberculosis-associated SNVs, a given genetic increase in the risk of tuberculosis was correlated with a concomitant increase in the risk of Behçet disease (rs4240897: OR, 1.27 [95% CI, 1.20-1.33]; P = 1.41 × 10⁻¹¹; rs2269497: OR, 1.51 [95% CI, 1.35-1.68]; P = 3.37 × 10⁻⁸; rs41553512: OR, 2.14 [95% CI, 1.78-2.57]; P = 7.93 × 10⁻¹¹; rs12437118: OR, 1.28 [95% CI, 1.19-1.37]; P = 1.72 × 10⁻¹¹; rs6114027: OR, 1.34 [95% CI, 1.23-1.47]; P = 2.37 × 10⁻¹¹; Figure 2; eFigure 1 and eTable 1 in the Supplement). The genetically determined tuberculosis susceptibility was associated with an increased risk of Behçet disease (Table 1); in the Chinese cohort, the odds ratio of Behçet disease per 2-fold increase in tuberculosis incidence was 1.26 (95% CI, 1.12-1.43; P = 1.47 × 10⁻⁴). No associations were detected for Vogt-Koyanagi-Harada disease (OR, 0.94 [95% CI, 0.65-1.38]; P = .77), uveitis associated with ankylosing spondylitis (OR, 0.99 [95% CI, 0.81-1.22]; P = .95), and Fuchs uveitis syndrome (OR, 1.03 [95% CI, 0.86-1.24]; P = .72).

Figure 2. — Per-allele association (log-transformed odds ratio) with the risk of Behçet disease is shown against the per-allele association with the risk of tuberculosis infection. Error bars show 95% CIs for each SNV. The regression line represents the outcome estimated by the inverse variance weighting.

Table 1. Estimated Association of Genetically Determined Tuberculosis Susceptibility With the Risk of Uveitis Conditions.

Outcome	Random-effects inverse-variance weighting^a		Fixed-effects inverse-variance weighting
Outcome	Odds ratio (95% CI)	P value	Odds ratio (95% CI)	P value
Behçet disease	1.26 (1.12-1.43)	1.47 × 10⁻⁴	1.26 (1.11-1.44)	5.37 × 10⁻⁴
Vogt-Koyanagi-Harada disease	0.94 (0.65-1.38)	.77	0.94 (0.79-1.13)	.54
Uveitis associated with ankylosing spondylitis	0.99 (0.81-1.22)	.95	0.99 (0.82-1.20)	.95
Fuchs uveitis syndrome	1.03 (0.86-1.24)	.72	1.03 (0.86-1.25)	.72

Open in a new tab

^{^a}

A random-effects inverse-variance weighted estimate was considered the primary outcome in this study.

No heterogeneity was detected in the inverse variance weighted estimate for Behçet disease (I² = 0%; P = .51; eTable 4 in the Supplement). Nevertheless, there was a certain degree of heterogeneity detected in estimates for Vogt-Koyanagi-Harada disease (I² = 88.8%), uveitis associated with ankylosing spondylitis (I² = 57.0%), and Fuchs uveitis syndrome (I² = 47.1%; eTable 4 in the Supplement). In addition, the MR Egger intercept test demonstrated no evidence of pleiotropy in the estimates for Behçet disease as well as the 3 other conditions (eTable 5 in the Supplement).

Sensitivity Analyses

Searching the 5 SNVs in the PhenoScanner for associations with other traits except tuberculosis indicated that SNV rs4240897 or its relevant SNVs that were in strong linkage disequilibrium were likely to have a pleiotropy (P = 1.91 × 10⁻³⁴ for association with plateletcrit; P = 3.77 × 10⁻²² for association with platelet count; eTable 6 in the Supplement). Nevertheless, excluding the SNV rs4240897 in the sensitivity analysis showed a similar result with that of the primary analysis (OR, 1.25 [95% CI, 1.07-1.46]; eFigure 2 in the Supplement). In addition, the leave-1-out sensitivity analysis did not suggest that a single SNV was driving the association (eFigure 2 in the Supplement). A concordant association of increased tuberculosis susceptibility with higher risk of Behçet disease was observed in analyses using maximum likelihood (OR, 1.27 [95% CI, 1.11-1.46]), MR Egger (OR, 1.60 [95% CI, 1.09-2.34]), weighted median (OR, 1.27 [95% CI, 1.06-1.51]), and contamination mixture (OR, 1.37 [95% CI, 1.08-1.75]) models (eFigure 3 and eFigure 4 in the Supplement).

Replication and Observational Analysis

In the Japanese cohort (eTable 7 in the Supplement), genetically determined tuberculosis susceptibility scaled per 2-fold increase in tuberculosis incidence conferred a 1.16-fold change in the odds ratio for Behçet disease (95% CI, 1.08-1.26; P = 9.13 × 10⁻⁵ in the random-effects model). No heterogeneity and pleiotropy were detected in the estimate (eTable 4 and eTable 5 in the Supplement). In the observational analysis, a positive T-SPOT.TB test result, indicating past or present tuberculosis infection, was observed in 38 individuals with Behçet disease (32.8%) and 19 control participants (15.7%) (eFigure 5 in the Supplement). After adjustment for covariables, T-SPOT.TB–positive status was an independent risk factor for having Behçet disease among patients with uveitis (OR, 2.26 [95% CI, 1.11-4.60]; P = .03; Table 2).

Table 2. Multivariable Logistic Regression Analysis of Factors Associated With Behçet Disease.

Variables	Participants, No. (%)		Unadjusted		Multivariable adjusted
Variables	With Behçet disease (n = 116)	Control (n = 121)	Odds ratio (95% CI)	P value	Odds ratio (95% CI)	P value
T-SPOT.TB test positivity	38 (32.8)	19 (15.7)	2.62 (1.40-4.88)	.003	2.26 (1.11-4.60)	.03
Male	99 (85.3)	66 (54.5)	4.85 (2.59-9.08)	<.001	4.66 (2.25-9.67)	<.001
Age, y
<20	8 (6.9)	21 (17.4)	0.27 (0.10-0.72)	.009	0.73 (0.22-2.37)	.60
20-29	31 (26.7)	22 (18.2)	1 [Reference]	NA	1 [Reference]	NA
30-39	50 (43.1)	31 (25.6)	1.15 (0.57-2.32)	.71	1.42 (0.64-3.17)	.40
40-49	27 (23.3)	47 (38.8)	0.41 (0.20-0.84)	.02	0.55 (0.24-1.28)	.17
BMI
<18.5 (Underweight)	11 (9.5)	16 (13.2)	0.58 (0.25-1.34)	.20	0.63 (0.23-1.72)	.36
18.5-23.9 (Normal weight)	68 (58.6)	57 (47.1)	1 [Reference]	NA	1 [Reference]	NA
≥24 (Overweight and obesity)	37 (31.9)	48 (39.7)	0.65 (0.37-1.13)	.12	0.70 (0.36-1.38)	.30
Smoking history
Never	76 (65.5)	75 (62.0)	1 [Reference]	NA	1 [Reference]	NA
Formerly	9 (7.8)	9 (7.4)	0.99 (0.37-2.62)	.98	1.11 (0.38-3.29)	.85
Currently occasionally	6 (5.2)	1 (0.8)	5.92 (0.70-50.37)	.10	5.12 (0.52-51.02)	.16
Currently regularly	25 (21.6)	36 (29.8)	0.69 (0.38-1.25)	.22	0.66 (0.33-1.34)	.25
Drinking history
Never	77 (66.4)	98 (81.0)	1 [Reference]	NA	1 [Reference]	NA
Formerly	28 (24.1)	11 (9.1)	3.24 (1.52-6.92)	.002	1.91 (0.78-4.66)	.16
Currently occasionally	8 (6.9)	7 (5.8)	1.46 (0.51-4.19)	.49	0.85 (0.25-2.86)	.79
Currently regularly	3 (2.6)	5 (4.1)	0.76 (0.18-3.30)	.72	0.38 (0.07-2.01)	.25
Urban resident	77 (66.4)	88 (72.7)	0.74 (0.43-1.29)	.29	0.66 (0.33-1.31)	.23
Education level
Primary school and less	7 (6.0)	21 (17.4)	1 [Reference]	NA	1 [Reference]	NA
Middle school	36 (31.0)	30 (24.8)	3.60 (1.35-9.62)	.01	2.69 (0.85-8.50)	.09
High school	21 (18.1)	32 (26.4)	1.97 (0.71-5.44)	.19	1.89 (0.58-6.20)	.29
College and higher	52 (44.8)	38 (31.4)	4.11 (1.58-10.64)	.004	4.33 (1.36-13.73)	.01

Open in a new tab

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NA, not applicable.

Discussion

This mendelian randomization study showed that an increased risk or likelihood of tuberculosis infection was associated with a higher risk of Behçet disease among individuals with a history of uveitis. Clinical biomarker data also suggested that tuberculosis infection was an independent risk factor for developing Behçet disease. The present study provides evidence in people linking infectious agents to the risk of noninfectious uveitis and may also add to our knowledge concerning the pathologic consequences of tuberculosis. The study did not provide sufficient evidence showing the association of tuberculosis with the risk of Vogt-Koyanagi-Harada disease, uveitis associated with ankylosing spondylitis, and Fuchs uveitis syndrome, which warrants further investigations.

Previous studies have shown that the 65-kD heat shock protein (HSP65) from Mycobacterium is highly homologous to human 60-kD heat shock protein, which might provide a role for antigenic mimicry in the development of Behçet disease.³⁷ Several reports documented that tuberculosis was complicated with Behçet disease and that antituberculosis chemotherapy could relieve the symptoms of Behçet disease.^38,39 Nevertheless, these studies were case reports or case series with small sample sizes. The major strength of this study was the use of naturally randomly inherited variants to indicate the tuberculosis exposure from the perspective of the time of birth, which might minimize the possibility of reverse causality and provide nearly randomized evidence in this population. Our mendelian randomization estimates were consistent with the T-SPOT.TB observations as well as previous experimental and clinical reports, implying that tuberculosis could be a risk factor for Behçet disease. Elucidation of the precise mechanisms involved, however, requires further studies.

Our study may provide several implications for clinical practice. First, there might be a need to carry out screening of Behçet disease for tuberculosis. We report here that the T-SPOT.TB positivity rate was 32.8% in individuals with Behçet disease, which is similar with a previous study.⁴⁰ Second, in terms of the frequent use of glucocorticoids and immunosuppressive or biologic agents, especially tumor necrosis factor inhibitors, for the treatment of Behçet disease, special attention should be paid to the risk of reactivation or progression of tuberculosis, especially in patients with positive T-SPOT.TB assays. Third, understanding the role of infectious agents in the development of Behçet disease may be helpful for the prevention and treatment of the disease. Together with existing clinical observations,^38,39 our results support the need for randomized clinical trials to evaluate the benefits of antituberculosis intervention for patients with Behçet disease associated with tuberculosis infection.

Limitations

This study has some limitations. First, although the SNVs were strongly associated with tuberculosis in the Chinese population, because of the lack of data, the actual tuberculosis infection status and the prevalence of tuberculosis could not be assessed in our population. Therefore, our mendelian randomization estimates only suggested that an increased risk or likelihood of tuberculosis infection was associated with a higher risk of Behçet disease, rather than that tuberculosis exposure directly increased the risk of Behçet disease. To address this issue, we examined the T-SPOT.TB positivity rate in another group of individuals with Behçet disease compared with control participants, which provided further evidence on the association between the history of tuberculosis and the risk of Behçet disease. Second, our study was based on a 2-sample mendelian randomization. This approach may lead to reduced precision in calculating effect sizes compared with 1-sample mendelian randomization analysis by using individual-level data. Because of power concerns, we did not perform a secondary 1-sample mendelian randomization test using the individual-level data of T-SPOT.TB cohort. Thus, these results may just indicate a positive association between the risk of tuberculosis infection and that of Behçet disease, but the estimated association could not be directly translated into the effect size in the real world. Third, the strength of the association between genetic variants and tuberculosis in the Japanese population was not as strong, and the analysis was liable to weak instrument bias.⁴¹ Therefore, the results from the Japanese population may not be reproducible and require further validation. Moreover, although much effort was made to address potential bias in the T-SPOT.TB analysis, because of the observational design, some unmeasured confounders may not have been completely excluded. Fourth, we did not detect a significant association for Vogt-Koyanagi-Harada disease, uveitis associated with ankylosing spondylitis, and Fuchs uveitis syndrome, which may be because of the small sample size and insufficient statistical power. In addition, we strictly followed international standardized criteria for the diagnosis of Behçet disease.²¹ Nevertheless, these criteria principally rely on clinical manifestations but not laboratory tests. To overcome this, we have introduced a multidisciplinary approach and specialist experience, especially the recommendations by rheumatologists, in the diagnosis of the disease.

Conclusions

In conclusion, human genetic and biomarker data demonstrated that an increased risk of tuberculosis infection was associated with a higher risk of Behçet disease among individuals with a history of uveitis. Our study provides novel evidence linking an infectious agent to the risk for noninfectious uveitis.

Supplement.

eMethods.

eTable 1. Characteristics of genetic variants associated with the risk of tuberculosis

eTable 2. F-statistics showing the association strength by five single nucleotide polymorphisms

eTable 3. Associations of variants related to tuberculosis with the risk of uveitis entities

eTable 4. Heterogeneity statistics in Mendelian randomization estimates

eTable 5. Directional pleiotropy estimated by MR Egger regression

eTable 6. Associations of SNPs with disease and trait in PhenoScanner database

eTable 7. Associations of SNPs with tuberculosis and Behçet’s disease in the Japanese cohort

eFigure 1. Inverse variance weighted estimates on the effect of tuberculosis exposure on the risk of uveitis entities

eFigure 2. Estimates for the effect of tuberculosis exposure on the risk of Behçet’s disease using leave-one-out sensitivity analysis

eFigure 3. Scatter plots of genetic associations with Behçet’s disease over associations with tuberculosis

eFigure 4. Mendelian randomization estimates of the effect of tuberculosis exposure on the risk of Behçet’s disease using maximum likelihood, weighted median, MR Egger and contamination mixture models

eFigure 5. Results of T-SPOT.TB tests

Click here for additional data file.^{(943.1KB, pdf)}

References

1.Tsirouki T, Dastiridou A, Symeonidis C, et al. A focus on the epidemiology of uveitis. Ocul Immunol Inflamm. 2018;26(1):2-16. doi: 10.1080/09273948.2016.1196713 [DOI] [PubMed] [Google Scholar]
2.Barisani-Asenbauer T, Maca SM, Mejdoubi L, Emminger W, Machold K, Auer H. Uveitis—a rare disease often associated with systemic diseases and infections—a systematic review of 2619 patients. Orphanet J Rare Dis. 2012;7:57-57. doi: 10.1186/1750-1172-7-57 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Yang P, Zhong Z, Du L, et al. Prevalence and clinical features of systemic diseases in Chinese patients with uveitis. Br J Ophthalmol. 2021;105(1):75-82. doi: 10.1136/bjophthalmol-2020-315960 [DOI] [PubMed] [Google Scholar]
4.Zhong Z, Su G, Kijlstra A, Yang P. Activation of the interleukin-23/interleukin-17 signalling pathway in autoinflammatory and autoimmune uveitis. Prog Retin Eye Res. 2020;16(100866):100866. doi: 10.1016/j.preteyeres.2020.100866 [DOI] [PubMed] [Google Scholar]
5.Horai R, Zárate-Bladés CR, Dillenburg-Pilla P, et al. Microbiota-dependent activation of an autoreactive t cell receptor provokes autoimmunity in an immunologically privileged site. Immunity. 2015;43(2):343-353. doi: 10.1016/j.immuni.2015.07.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Ye Z, Zhang N, Wu C, et al. A metagenomic study of the gut microbiome in Behcet’s disease. Microbiome. 2018;6(1):135. doi: 10.1186/s40168-018-0520-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Forrester JV, Kuffova L, Dick AD. Autoimmunity, Autoinflammation, and Infection in Uveitis. Am J Ophthalmol. 2018;189:77-85. doi: 10.1016/j.ajo.2018.02.019 [DOI] [PubMed] [Google Scholar]
8.Kahwati LC, Feltner C, Halpern M, et al. Primary care screening and treatment for latent tuberculosis infection in adults: evidence report and systematic review for the US Preventive Services Task Force. JAMA. 2016;316(9):970-983. doi: 10.1001/jama.2016.10357 [DOI] [PubMed] [Google Scholar]
9.Elkington P, Tebruegge M, Mansour S. Tuberculosis: an infection-initiated autoimmune disease? Trends Immunol. 2016;37(12):815-818. doi: 10.1016/j.it.2016.09.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Campisi L, Barbet G, Ding Y, Esplugues E, Flavell RA, Blander JM. Apoptosis in response to microbial infection induces autoreactive TH17 cells. Nat Immunol. 2016;17(9):1084-1092. doi: 10.1038/ni.3512 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Shobha V, Chandrashekara S, Rao V, et al. Biologics and risk of tuberculosis in autoimmune rheumatic diseases: a real-world clinical experience from India. Int J Rheum Dis. 2019;22(2):280-287. doi: 10.1111/1756-185X.13376 [DOI] [PubMed] [Google Scholar]
12.Jick SS, Lieberman ES, Rahman MU, Choi HK. Glucocorticoid use, other associated factors, and the risk of tuberculosis. Arthritis Rheum. 2006;55(1):19-26. doi: 10.1002/art.21705 [DOI] [PubMed] [Google Scholar]
13.Burgess S, Foley CN, Zuber V. Inferring causal relationships between risk factors and outcomes from genome-wide association study data. Annu Rev Genomics Hum Genet. 2018;19:303-327. doi: 10.1146/annurev-genom-083117-021731 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Smith JG, Luk K, Schulz C-A, et al. ; Cohorts for Heart and Aging Research in Genetic Epidemiology (CHARGE) Extracoronary Calcium Working Group . Association of low-density lipoprotein cholesterol-related genetic variants with aortic valve calcium and incident aortic stenosis. JAMA. 2014;312(17):1764-1771. doi: 10.1001/jama.2014.13959 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Holmes MV, Dale CE, Zuccolo L, et al. ; InterAct Consortium . Association between alcohol and cardiovascular disease: mendelian randomisation analysis based on individual participant data. BMJ. 2014;349:g4164. doi: 10.1136/bmj.g4164 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Millwood IY, Walters RG, Mei XW, et al. ; China Kadoorie Biobank Collaborative Group . Conventional and genetic evidence on alcohol and vascular disease aetiology: a prospective study of 500 000 men and women in China. Lancet. 2019;393(10183):1831-1842. doi: 10.1016/S0140-6736(18)31772-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Zheng R, Li Z, He F, et al. Genome-wide association study identifies two risk loci for tuberculosis in Han Chinese. Nat Commun. 2018;9(1):4072-4072. doi: 10.1038/s41467-018-06539-w [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Qi H, Zhang YB, Sun L, et al. Discovery of susceptibility loci associated with tuberculosis in Han Chinese. Hum Mol Genet. 2017;26(23):4752-4763. doi: 10.1093/hmg/ddx365 [DOI] [PubMed] [Google Scholar]
19.van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis: a proposal for modification of the New York criteria. Arthritis Rheum. 1984;27(4):361-368. doi: 10.1002/art.1780270401 [DOI] [PubMed] [Google Scholar]
20.Read RW, Holland GN, Rao NA, et al. Revised diagnostic criteria for Vogt-Koyanagi-Harada disease: report of an international committee on nomenclature. Am J Ophthalmol. 2001;131(5):647-652. doi: 10.1016/S0002-9394(01)00925-4 [DOI] [PubMed] [Google Scholar]
21.International Study Group for Behçet’s Disease . Criteria for diagnosis of Behçet’s disease. Lancet. 1990;335(8697):1078-1080. [PubMed] [Google Scholar]
22.Kimura SJ, Hogan MJ, Thygeson P. Fuchs’ syndrome of heterochromic cyclitis. AMA Arch Ophthalmol. 1955;54(2):179-186. doi: 10.1001/archopht.1955.00930020181003 [DOI] [PubMed] [Google Scholar]
23.La Hey E, Baarsma GS, De Vries J, Kijlstra A. Clinical analysis of Fuchs’ heterochromic cyclitis. Doc Ophthalmol. 1991;78(3-4):225-235. doi: 10.1007/BF00165685 [DOI] [PubMed] [Google Scholar]
24.Mizuki N, Meguro A, Ota M, et al. Genome-wide association studies identify IL23R-IL12RB2 and IL10 as Behçet’s disease susceptibility loci. Nat Genet. 2010;42(8):703-706. doi: 10.1038/ng.624 [DOI] [PubMed] [Google Scholar]
25.Davatchi F, Chams-Davatchi C, Shams H, et al. Behcet’s disease: epidemiology, clinical manifestations, and diagnosis. Expert Rev Clin Immunol. 2017;13(1):57-65. doi: 10.1080/1744666X.2016.1205486 [DOI] [PubMed] [Google Scholar]
26.Mahasirimongkol S, Yanai H, Mushiroda T, et al. Genome-wide association studies of tuberculosis in Asians identify distinct at-risk locus for young tuberculosis. J Hum Genet. 2012;57(6):363-367. doi: 10.1038/jhg.2012.35 [DOI] [PubMed] [Google Scholar]
27.Staley JR, Blackshaw J, Kamat MA, et al. PhenoScanner: a database of human genotype-phenotype associations. Bioinformatics. 2016;32(20):3207-3209. doi: 10.1093/bioinformatics/btw373 [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Auton A, Brooks LD, Durbin RM, et al. ; 1000 Genomes Project Consortium . A global reference for human genetic variation. Nature. 2015;526(7571):68-74. doi: 10.1038/nature15393 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Meier T, Eulenbruch HP, Wrighton-Smith P, Enders G, Regnath T. Sensitivity of a new commercial enzyme-linked immunospot assay (T SPOT-TB) for diagnosis of tuberculosis in clinical practice. Eur J Clin Microbiol Infect Dis. 2005;24(8):529-536. doi: 10.1007/s10096-005-1377-8 [DOI] [PubMed] [Google Scholar]
30.Palmer TM, Lawlor DA, Harbord RM, et al. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res. 2012;21(3):223-242. doi: 10.1177/0962280210394459 [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol. 2013;37(7):658-665. doi: 10.1002/gepi.21758 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Burgess S, Labrecque JA. Mendelian randomization with a binary exposure variable: interpretation and presentation of causal estimates. Eur J Epidemiol. 2018;33(10):947-952. doi: 10.1007/s10654-018-0424-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512-525. doi: 10.1093/ije/dyv080 [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol. 2016;40(4):304-314. doi: 10.1002/gepi.21965 [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Burgess S, Foley CN, Allara E, Staley JR, Howson JMM. A robust and efficient method for Mendelian randomization with hundreds of genetic variants. Nat Commun. 2020;11(1):376-376. doi: 10.1038/s41467-019-14156-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ. 2003;326(7382):219-219. doi: 10.1136/bmj.326.7382.219 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Gül A. Behçet’s disease: an update on the pathogenesis. Clin Exp Rheumatol. 2001;19(5)(suppl 24):S6-S12. [PubMed] [Google Scholar]
38.Hamill M, Remedios D, Kapembwa M. Orogenital ulceration with overlapping tuberculosis: epiphenomenon or expanding spectrum of Behçet disease? J Low Genit Tract Dis. 2006;10(4):219-222. doi: 10.1097/01.lgt.0000225897.29466.e4 [DOI] [PubMed] [Google Scholar]
39.Efthimiou J, Hay PE, Spiro SG, Lane DJ. Pulmonary tuberculosis in Behçet’s syndrome. Br J Dis Chest. 1988;82(3):300-304. doi: 10.1016/0007-0971(88)90073-3 [DOI] [PubMed] [Google Scholar]
40.Wu X, Chen P, Wei W, et al. Diagnostic value of the interferon-γ release assay for tuberculosis infection in patients with Behçet’s disease. BMC Infect Dis. 2019;19(1):323-323. doi: 10.1186/s12879-019-3954-y [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Burgess S, Small DS, Thompson SG. A review of instrumental variable estimators for Mendelian randomization. Stat Methods Med Res. 2017;26(5):2333-2355. doi: 10.1177/0962280215597579 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials