Comparative risk of tuberculosis infection with different TNF-α inhibitors in immune-mediated inflammatory diseases: a systematic review and network meta-analysis

Background: Tumor necrosis factor-α inhibitors (TNFi) are established to increase the risk of tuberculosis (TB). However, the comparative risk across different TNFi agents remains poorly defined due to a lack of head-to-head comparative studies. This network meta-analysis (NMA) aimed to evaluate and compare the risk of TB infection associated with various TNFi therapies in patients with immune-mediated inflammatory diseases (IMIDs) based on real-world, long-term cohort studies.

Methods: We conducted a systematic search of PubMed, EMBASE, Cochrane Library, and Web of Science from inception to May 30, 2025, for cohort studies reporting TB events in patients with IMIDs treated with TNFi. Study selection, data extraction, and risk of bias assessment were performed by three independent reviewers using the Newcastle-Ottawa Scale. A Bayesian arm-based NMA with random-effects models was used to estimate log risk ratio (logRR) and 95% credible intervals (CrIs) for TB infection across different TNFi agents compared with TNFi-naive.

Results: A total of 19 cohort studies involving 396, 044 patients were included. Compared to TNFi-naive, infliximab (IFX) was associated with the highest risk of TB (logRR = 2.32, 95% CrI: 1.12-3.32), followed by adalimumab (ADA) (logRR = 1.72, 95% CI: 0.42-2.65) and etanercept (ETN) (logRR = 1.39, 95% CI: 0.33-2.42). Certolizumab pegol (CZP) was associated with the lowest risk among TNFi agents.

Conclusion: TNFi treatment in patients with IMIDs is associated with a significantly increased risk of TB infection. Among the TNFi agents, IFX was associated with the highest risk, while ETN and CZP demonstrated lower risks. These findings can inform clinical decision-making, suggesting that ETN or CZP may be preferable in patients with high TB risk, while emphasizing that vigilant TB monitoring remains paramount regardless of the chosen agent.

Systematic Review Registration: https://www.crd.york.ac.uk/prospero/, identifier CRD42022331674.

1 Introduction

Immune-mediated inflammatory diseases (IMIDs) encompass a heterogeneous group of prevalent conditions, including rheumatoid arthritis (RA), spondyloarthropathy (SpA), connective tissue disorders, inflammatory cutaneous conditions, and inflammatory bowel disease (IBD), etc (1). Inadequate disease control can lead to progressive disability, loss of work capacity, reduced quality of life, and substantial socioeconomic burdens (2). While the precise pathogenesis of IMIDs is not fully elucidated, TNF-α existing in both soluble (sTNF-α) and transmembrane (tmTNF-α) forms has been identified as a key pro-inflammatory cytokine driving disease pathology (3). Consequently, TNFi which block the inflammatory effects of TNF-α, are widely used in the treatment of IMIDs and have significantly improved clinical outcomes (4). Currently, five TNFi agents are approved for clinical use: etanercept (ETN), adalimumab (ADA), infliximab (IFX), golimumab (GOL), and certolizumab pegol (CZP). IFX, ADA, and GOL are full-length IgG1 monoclonal antibodies against TNF-α. ETN is a fusion protein consisting of the extracellular domain of the human TNF receptor 2 (TNFR2) linked to the Fc portion of human IgG1. CZP is a PEGylated Fab′ fragment of a humanized anti-TNF-α monoclonal antibody and lacks the Fc region (5). The widespread use of TNFi has been accompanied by growing concerns regarding associated infections, particularly tuberculosis (TB) (6). However, due to the absence of head-to-head comparative trials, direct evidence comparing the TB risk among different TNFi agents is limited. Previous systematic reviews and meta-analyses have reported that the risk of TB in patients treated with TNFi is 1.6 to 25.1 times higher than that in the general population (7–12). Subgroup analyses suggest that monoclonal antibody-based TNFi agents confer a higher risk of TB than soluble receptor analogs (7–9). Some randomized controlled trials (RCTs) have shown no significant difference in TB incidence between IFX and ETN in RA patients (13, 14), however, these studies were limited by small sample sizes and short follow-up durations, which may not accurately reflect real-world risk differences. Existing systematic reviews and meta-analyses are predominantly based on RCTs, which often involve homogeneous patient populations, fixed drug regimens, and relatively short treatment durations. In contrast, rheumatic diseases are chronic and complex, frequently requiring long-term treatment and individualized therapy adjustments based on disease activity factors that complicate the accurate assessment of TB risk. Furthermore, most available data pertain to RA, with limited information on other IMIDs.

To better reflect clinical practice, this meta-analysis incorporated cohort studies involving patients with various IMIDs treated with TNFi and followed for at least one year, thereby providing a more robust evaluation of TB risk. This network meta-analysis (NMA) therefore aims to synthesize long-term, real-world evidence from diverse global populations to provide a comprehensive ranking of TB risk among all five TNFi agents, ultimately offering higher-quality evidence to guide individualized clinical decision-making, particularly in TB-endemic areas.

Registration: This systematic review and meta-analysis was registered on PROSPERO (CRD42022331674, Last updated:2025.12.25).

2 Materials and methods

2.1 Protocol

This systematic review and NMA were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (http://www.prisma-statement.org/) and registered in the PROSPERO database (CRD42022331674, Last updated:2025.12.25). The PRISMA 2020 Checklist was provided in Supplementary file 2.

2.2 Search strategy

A systematic literature search was conducted independently by two investigators (Q.Z. and Y.L.¹) in PubMed, EMBASE, the Cochrane Library, and Web of Science for English-language cohort studies published from inception to May 30, 2025. The search strategy combined Medical Subject Headings (MeSH) and keywords related to TNFi agents including etanercept, adalimumab, infliximab, golimumab, certolizumab pegol, TNF-α antagonist. The detailed search strategy is provided in Supplementary Table S1.

2.3 Eligibility criteria

Studies were included if they met the following criteria: (1) cohort studies including at least two cohorts of patients with IMIDs aged ≥18 years treated with different TNFi agents; (2) reported TB incidence; and (3) had a follow-up period of ≥1 year. When multiple publications from the same study population existed, only the most comprehensive or extended report was included. Studies with overlapping data were excluded.

2.4 Study selection and data extraction

Two reviewers (X.L. and Y.L.¹²³) independently screened titles and abstracts, reviewed full texts, and extracted data using a standardized form. The extracted information included first author, publication year, country, study period, sample size, percentage of female participants, and TB incidence. Any discrepancies were resolved through discussion or by consulting a third reviewer (S.C.).

2.5 Risk of bias assessment and quality of evidence

Publication bias was evaluated using funnel plots and Egger’s test (15). Two investigators (Y.L.¹ and X.L.) independently assessed the methodological quality of the included studies using the Newcastle-Ottawa Scale (NOS) for cohort studies. The NOS evaluated three domains: selection of study groups, comparability of groups, and outcome assessment. Studies were awarded a maximum of 9 stars, with higher scores indicating higher quality. Disagreements were resolved by consensus or by involving a third reviewer (S.C.).

2.6 Statistical analysis

We performed a Bayesian arm-based NMA using the “BUGSnet” package (16) in R (version 4.1.0; R Foundation for Statistical Computing). Markov chain Monte Carlo (MCMC) sampling was implemented using JAGS (17). Both fixed-effect and random-effects models were fitted, and model fit was compared using the deviance information criterion (DIC). Models with lower DIC values were preferred. Heterogeneity was assessed using the I² statistic and chi-square tests, where I² > 50% or p < 0.10 indicated substantial heterogeneity (18). Sensitivity analyses determined the effect of individual studies by sequential exclusion. Treatment rankings were estimated using the surface under the cumulative ranking curve (SUCRA). LogRR and 95% CrIs for TB infection across different TNFi agents compared with TNFi-naive were estimated. Subgroup analyses were conducted based on follow-up duration and the use of prophylactic anti-TB therapy. Consistency between direct and indirect evidence was assessed by comparing consistency and inconsistency models.

3 Results

3.1 Search results and study characteristics

The initial search identified 12, 009 articles. After screening, 19 cohort studies (19–37) involving 396, 044 patients were included in the analysis (Figure 1). The studies were published between 2004 and 2024 and conducted in 15 countries. The characteristics of the included studies were summarized in Table 1. All included studies had NOS scores ≥ 6, indicating moderate to high quality (Table 2). For the study by Lee et al. (2021) (34), data were extracted specifically for the patient cohort receiving only one TNF inhibitor to maintain consistency in the exposure definition across studies. Patients who received multiple biologics were excluded from our quantitative synthesis.

Figure 1

Figure 1. Flow diagram of study identification, screening, eligibility assessment, and inclusion. This study ultimately included 19 cohort studies that met the criteria.

Table 1

Table 1. The characteristics of the cohort studies included in the meta-analysis.

Table 2

Table 2. The Newcastle-Ottawa scale served to assess the risk of bias in the included studies.

3.2 Network geometry

The network plot of treatment comparisons was shown in Figure 2. The size of each node corresponds to the number of patients receiving that treatment, and the thickness of the edges represents the number of studies comparing connected treatments. Closed loops indicated the presence of direct comparisons involving more than two treatments. The network included multiple closed loops, with the most frequent direct comparison being between IFX and ADA.

Figure 2

Figure 2. The network diagram of this meta-analysis. Closed loops were detected between different treatment groups. Closed loops referred to direct comparisons including more than 2 comparators. The size of the nodes was proportional to the number of comparisons involving that treatment node, while the thickness of the edges indicated the number of studies that included the 2 connected treatments. ETN, Etanercept; ADA, Adalimumab; IFX, Infliximab; GOL, Golimumab; CZP, Certolizumab pegol.

3.3 Model selection and consistency

The random-effects model demonstrated a better fit (lower DIC) than the fixed-effect model (Figure 3) and was therefore selected for the primary analysis. Heterogeneity across the studies was significant (I² > 50%, p < 0.1, see Supplementary Figures S2), which was anticipated given the inclusion of diverse IMIDs and varying background TB risk across geographical regions. In terms of risk of bias, most studies were rated with some concerns (see Supplementary Figures S1, S2). The random-effects model was chosen to account for this clinical and methodological heterogeneity. Consistency between direct and indirect evidence was assessed using inconsistency models, we found that the inconsistency models had slightly less DIC than the consistency mode (Figure 4). This illustrates the possibility of inconsistency in the network.

Figure 3

Figure 3. Leverage plots and DIC for fixed and random effects models for TB infection. Regarding outliers, data points falling outside the purple arc suggested they may lead to poor model fit. A lower DIC value indicated a better model fit. DIC, deviance information criterion; Dres, deviance residual; pD, posterior mean deviance; Wik, adjustment for normal distribution of studies and arms; TB, Tuberculosis.

Figure 4

Figure 4. Leverage plots and DIC for consistency and inconsistency model for TB. The inconsistency model had slightly less DIC than the consistency model. Regarding outliers, data points falling outside the purple arc suggested they may lead to poor model fit. A lower DIC value indicated a better model fit. DIC, deviance information criterion; Dres, deviance residual; pD, posterior mean deviance; Wik, adjustment for normal distribution of studies and arms; TB, Tuberculosis.

3.4 Treatment rankings and league table

SUCRA values and rank probabilities consistently identified IFX as the TNFi agent associated with the highest risk of TB, followed by ADA and then ETN. CZP demonstrated the most favorable (lowest risk) profile among the TNFi agents evaluated (Figure 5). The league table (Figure 6) presented pairwise comparisons between treatments. Each cell showed the outcome for the row intervention relative to the corresponding column intervention. In this plot, green cells signified that the row intervention was associated with a lower risk of TB infection compared with the column intervention, whereas red cells signified a higher associated risk. The symbols (**) denoted statistically significant differences between treatments and comparators at a 95% confidence level.

Figure 5

Figure 5. SUCRA plot and plot of treatment rank probabilities. (A) TB SUCRA plot. Higher rankings associated with smaller outcome values. (B) plot of treatment rank probabilities. Treatments: ETN, Etanercept; ADA, Adalimumab; IFX, Infliximab; GOL, Golimumab; CZP, Certolizumab pegol; TB, Tuberculosis; SUCRA, surface under the cumulative ranking curve.

Figure 6

Figure 6. League heat plot for all treatment in the network for TB. The league plot provided a comprehensive summary of the NMA results, indicating the significance of all interventions compared to both the TNF blocker-naive and other treatments. Each cell showed the outcome for the row intervention relative to the corresponding column intervention. Green cells indicated that the row intervention carried a lower risk of tuberculosis infection than the column intervention, whereas red cells signified a higher associated risk. The symbols (**) denoted statistically significant differences between treatments and comparators at a 95% confidence level. The negative values represented beneficial or protective associations, while positive values represented adverse or harmful associations. ETN, Etanercept; ADA, Adalimumab; IFX, Infliximab; GOL, Golimumab; CZP, Certolizumab pegol; TB, Tuberculosis.

There were with significant differences observed between IFX and CZP (logRR = 29.44, 95% CI: 2.31–66.76) and between ADA and CZP (logRR = 28.78, 95% CI: 1.69–65.98). Compared with TNFi-naive patients, IFX (logRR = 2.32, 95% CI: 1.12–3.32), ADA (logRR = 1.72, 95% CI: 0.42–2.65), and ETN (logRR = 1.39, 95% CI: 0.33–2.42) were associated with significantly increased risks of TB. No significant differences were observed for GOL or CZP compared with TNFi-naive patients. Forest plots illustrating these comparisons are shown in Figure 7.

Figure 7

Figure 7. The forest plot of the logRR in different treatment compared to TNF blocker-naive. The result showed the LogRR of patients with IMIDs receiving different TNFi versus TNF blocker-naive. The negative values represented beneficial or protective associations, while positive values represented adverse or harmful associations. ETN, Etanercept; ADA, Adalimumab; IFX, Infliximab; GOL, Golimumab; CZP, Certolizumab pegol.

3.5 Subgroup analyses and sensitivity analysis

Subgroup analyses based on follow-up duration (<2 years) and prophylactic anti-TB therapy yielded results consistent with the main analysis (see Supplementary Figures S3 and S4). The sensitivity analysis in the meta-analysis indicated that the exclusion of individual studies had little impact on the results. This suggests that the findings of the meta-analysis are robust and not significantly influenced by any single study (see Supplementary Figures S5).

4 Discussion

This systematic review and NMA provided a comprehensive comparative safety assessment of five TNFi agents regarding the risk of TB infection in patients with IMIDs based on real-world, long-term cohort data.

The principal finding that IFX carries the highest risk of TB, followed by ADA and then ETN, with CZP appearing to have the lowest risk, is consistent with the prevailing hypothesis that monoclonal antibodies confer a greater risk than soluble receptor constructs. This risk hierarchy was robust, remaining consistent across subgroup analyses of follow-up duration and prophylactic anti-TB therapy.

4.1 Mechanistic insights into differential TB risk

The observed differential risk profile can be plausibly explained by distinct mechanisms of action among TNFi agents. TNF-α is a critical cytokine for maintaining the structural integrity of granulomas, which are essential for containing Mycobacterium tuberculosis infection (38–40). Beyond sTNF-α, monoclonal antibodies such as IFX, ADA, GOL also bind tmTNF-α with high affinity. This binding can induce complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC), leading to the lysis of immune cells such as, monocytes and T-cells which express tmTNF-α and are crucial for granuloma stability (41–43). This lytic effect potentially disrupts existing granulomas, facilitating bacterial dissemination and reactivation of latent TB.

In contrast, ETN, a soluble receptor fusion protein, has a lower binding avidity for tmTNF-α and lacks an Fc domain capable of effectively recruiting complement or effector cells, resulting in markedly reduced CDC/ADCC activity (43, 44). CZP, a PEGylated Fab’ fragment, completely lacks an Fc region, which explains its absence of CDC activity and may account for its seemingly favorable risk profile in our analysis (45). Furthermore, evidence suggests that monoclonal antibodies, by compromising the key function of signaling through tmTNF-α, impair the innate immune control of TB. This function is preserved by receptor agonists. This mechanism provides a distinct explanation for the increased TB risk associated with monoclonal antibodies (41, 46). In recent years, researchers have also explored the association between HLA-B subtypes and tuberculosis development induced by anti-TNF therapy from the perspective of genetic susceptibility (47).

4.2 Addressing heterogeneity and limitations

A primary strength of our study is its inclusion of diverse IMIDs and global populations, which enhances the generalizability of our findings. However, this diversity inevitably introduces clinical heterogeneity. Variations in background TB incidence rates ranging from low to high burden across the included countries, differences in standard-of-care practices such as screening protocols and prophylactic therapy, and the spectrum of concomitant immunosuppressants including corticosteroids and csDMARDs could all contribute to the observed statistical heterogeneity (I² > 50%). We addressed this by employing a Bayesian random-effects model, which is explicitly designed to account for such variability, providing more conservative and generalizable effect estimates. Nevertheless, this heterogeneity necessitates cautious interpretation of the point estimates.

Several other limitations warrant consideration. First, the evidence base for CZP and GOL is notably scarce. Particularly for CZP, the estimate relied on only one small study comprising 49 patients. The point estimate for CZP suggests a potentially lower risk, but the exceedingly wide credible intervals indicate very low certainty in this estimate. These findings are therefore primarily hypothesis-generating, and more robust data from large-scale prospective studies are needed to draw definitive conclusions regarding the TB risk associated with CZP and GOL. Second, while we focused on comparative safety, treatment efficacy was not evaluated. A comprehensive clinical decision must balance the TB risk identified here against the known differential efficacy of these drugs for specific IMIDs. Future studies integrating both efficacy and safety outcomes are valuable. Finally, as with any meta-analysis, our results are constrained by the quality and reporting of the original studies.

4.3 Clinical and research implications

Despite these limitations, our findings have tangible clinical implications. In patients with significant risk factors for TB reactivation (e.g., from high-burden regions, prior latent TB infection, or on concomitant steroids) (48), opting for a soluble receptor inhibitor or CZP may be a prudent choice when clinically appropriate, potentially mitigating the risk of this serious infection. This decision must be made within the context of individual patient factors, disease severity, and drug availability. Our study also underscores the non-interchangeable nature of TNFi agents from a safety perspective.

From a research perspective, our work highlights the critical need for large, prospective pharmacovigilance studies that directly compare newer agents like GOL and CZP against established ones. Furthermore, translational research exploring the precise immunologic mechanisms, especially those involving the role of Fc-mediated functions and tmTNF-α signaling in granuloma biology, will be crucial for understanding the differential risks observed in epidemiological studies and for guiding the development of safer biologic therapies.

5 Conclusion

In conclusion, this NMA demonstrates a gradient of TB risk among TNFi agents used for IMIDs. The risk is highest with the monoclonal antibody IFX, intermediate with ADA, and lower with the soluble receptor fusion protein ETN. This spectrum aligns with understood differences in their mechanisms of action, particularly their capacity to induce cytolytic effects on tmTNF-α-expressing cells. While heterogeneity exists and evidence for some agents remains limited, these findings provide valuable guidance for clinicians in stratifying TB risk and making individualized treatment decisions, especially in TB-endemic areas. Ultimately, vigilant screening for latent TB and maintaining a high index of suspicion for active infection remain paramount, regardless of the chosen TNFi agent.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

Author contributions

XL (1st author): Writing – original draft, Writing – review & editing, Data Curation, Formal Analysis, Funding Acquisition. YuL: Writing – review & editing, Conceptualization, Supervision, Methodology. YL (3rd author): Writing – original draft, Data Curation. QZ: Writing – original draft, Software. SC: Writing – review & editing, Data Curation. XL (6th author): Writing – original draft, Software. GS: Writing – review & editing, Conceptualization, Project Administration. YL (8th author): Writing – review & editing, Conceptualization, Supervision, Funding Acquisition.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the grants from the Fujian Provincial Health Commission (Grant No. 2021GGB025) to YL (8th author), and the Fujian Province Science and Technology Plan Project (Project No.2022J011368) to YL (8th author), Xiamen Health Commission (Project NO. 2024GZL-QN030) to XL.

Acknowledgments

We are grateful to the Department of Rheumatology and Immunology of the First Affiliated Hospital of Xiamen University for their work on this research.

Conflict of interest

The authors declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2026.1726299/full#supplementary-material

References

1. Schett G, McInnes IB, and Neurath MF. Reframing immune-mediated inflammatory diseases through signature cytokine hubs. N Engl J Med. (2021) 385:628–39. doi: 10.1056/NEJMra1909094

PubMed Abstract | Crossref Full Text | Google Scholar

2. Smolen JS, Aletaha D, and McInnes IB. Rheumatoid arthritis. Lancet. (2016) 388:2023–38. doi: 10.1016/S0140-6736(16)30173-8

PubMed Abstract | Crossref Full Text | Google Scholar

3. Bradley JR. TNF-mediated inflammatory disease. J Pathol. (2008) 214:149–60. doi: 10.1002/path.2287

PubMed Abstract | Crossref Full Text | Google Scholar

4. Chaabo K and Kirkham B. Rheumatoid arthritis - anti-TNF. Int immunopharmacol. (2015) 27:180–4. doi: 10.1016/j.intimp.2015.04.051

PubMed Abstract | Crossref Full Text | Google Scholar

5. Mitoma H, Horiuchi T, Tsukamoto H, and Ueda N. Molecular mechanisms of action of anti-TNF-alpha agents - Comparison among therapeutic TNF-alpha antagonists. Cytokine. (2018) 101:56–63. doi: 10.1016/j.cyto.2016.08.014

PubMed Abstract | Crossref Full Text | Google Scholar

6. Su K, Li X, Jiang Z, and Mei Y. Screening, prophylaxis, and challenges: Tumor necrosis factor inhibitors and latent tuberculosis infection nexus in rheumatology. Int J Rheum Dis. (2024) 27:e14996. doi: 10.1111/1756-185X.14996

PubMed Abstract | Crossref Full Text | Google Scholar

7. Harris J, Hope JC, and Keane J. Tumor necrosis factor blockers influence macrophage responses to Mycobacterium tuberculosis. J Infect Dis. (2008) 198:1842–50. doi: 10.1086/593174

PubMed Abstract | Crossref Full Text | Google Scholar

8. Ai JW, Zhang S, Ruan QL, Yu YQ, Zhang BY, Liu QH, et al. The risk of tuberculosis in patients with rheumatoid arthritis treated with tumor necrosis factor-alpha antagonist: A metaanalysis of both randomized controlled trials and registry/cohort studies. J Rheumatol. (2015) 42:2229–37. doi: 10.3899/jrheum.150057

PubMed Abstract | Crossref Full Text | Google Scholar

9. Liao H, Zhong Z, Liu Z, and Zou X. Comparison of the risk of infections in different anti-TNF agents: a meta-analysis. Int J Rheum Dis. (2017) 20:161–8. doi: 10.1111/1756-185X.12970

PubMed Abstract | Crossref Full Text | Google Scholar

10. Park HJ, Choi BY, Sohn M, Han NY, Kim I-W, and Oh JM. Effects of tumor necrosis factor-alpha inhibitors on the incidence of tuberculosis. Korean J. Clin. Pharm. (2018) 28:333–41. doi: 10.24304/kjcp.2018.28.4.333

Crossref Full Text | Google Scholar

11. Zhang Z, Fan W, Yang G, Xu Z, Wang J, Cheng Q, et al. Risk of tuberculosis in patients treated with TNF-α antagonists: a systematic review and meta-analysis of randomised controlled trials. BMJ Open (2017) 7:. doi: 10.1136/bmjopen-2016-012567

PubMed Abstract | Crossref Full Text | Google Scholar

12. Guan X, Zhao Z, Xin M, Xia G, Yang Q, and Fu M. Long-term efficacy, safety, and cumulative retention rate of antitumor necrosis factor-alpha treatment for patients with Behcet’s uveitis: A systematic review and meta-analysis. Int J Rheum Dis. (2024) 27:e15096. doi: 10.1111/1756-185X.15096

PubMed Abstract | Crossref Full Text | Google Scholar

13. Dixon WG, Watson K, Lunt M, Hyrich KL, Silman AJ, Symmons DP, et al. Rates of serious infection, including site-specific and bacterial intracellular infection, in rheumatoid arthritis patients receiving anti-tumor necrosis factor therapy: results from the British Society for Rheumatology Biologics Register. Arthritis Rheum. (2006) 54:2368–76. doi: 10.1002/art.21978

PubMed Abstract | Crossref Full Text | Google Scholar

14. Seong Ss CCBWJH. Incidence of tuberculosis in korean patients with rheumatoid arthritis (RA): effects of RA itself and of tumor necrosis factor blockers. J. Rheumatol. (2007) 54:2368–76.

PubMed Abstract | Google Scholar

15. Egger M, Davey Smith G, Schneider M, and Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. (1997) 315:629–34. doi: 10.1136/bmj.315.7109.629

PubMed Abstract | Crossref Full Text | Google Scholar

16. Beliveau A, Boyne DJ, Slater J, Brenner D, and Arora P. BUGSnet: an R package to facilitate the conduct and reporting of Bayesian network Meta-analyses. BMC Med Res Methodol. (2019) 19:196. doi: 10.1186/s12874-019-0829-2

PubMed Abstract | Crossref Full Text | Google Scholar

17. van Ravenzwaaij D, Cassey P, and Brown SD. A simple introduction to Markov Chain Monte-Carlo sampling. Psychon Bull Rev. (2018) 25:143–54. doi: 10.3758/s13423-016-1015-8

PubMed Abstract | Crossref Full Text | Google Scholar

18. Higgins JP, Thompson SG, Deeks JJ, and Altman DG. Measuring inconsistency in meta-analyses. BMJ. (2003) 327:557–60. doi: 10.1136/bmj.327.7414.557

PubMed Abstract | Crossref Full Text | Google Scholar

19. Wallis RS, Broder MS, Wong JY, Hanson ME, and Beenhouwer DO. Granulomatous infectious diseases associated with tumor necrosis factor antagonists. Clin Infect Dis. (2004) 38:1261–5. doi: 10.1086/383317

PubMed Abstract | Crossref Full Text | Google Scholar

20. Listing J, Strangfeld A, Kary S, Rau R, von Hinueber U, Stoyanova Scholz M, et al. Infections in patients with rheumatoid arthritis treated with biologic agents. Arthritis Rheumatol. (2005) 52:3403–12. doi: 10.1002/art.21386

PubMed Abstract | Crossref Full Text | Google Scholar

21. Sichletidis L, Settas L, Spyratos D, Chloros D, and Patakas D. Tuberculosis in patients receiving anti-TNF agents despite chemoprophylaxis. Int J Tuberc Lung Dis. (2006) 10:1127–32.

PubMed Abstract | Google Scholar

22. Fernandez-Nebro A, Irigoyen MV, Urena I, Belmonte-Lopez MA, Coret V, Jimenez-Nunez FG, et al. Effectiveness, predictive response factors, and safety of anti-tumor necrosis factor (TNF) therapies in anti-TNF-naive rheumatoid arthritis. J Rheumatol. (2007) 34:2334–42.

PubMed Abstract | Google Scholar

23. Favalli EG, Desiati F, Atzeni F, Sarzi-Puttini P, Caporali R, Pallavicini FB, et al. Serious infections during anti-TNFalpha treatment in rheumatoid arthritis patients. Autoimmun Rev. (2009) 8:266–73. doi: 10.1016/j.autrev.2008.11.002

PubMed Abstract | Crossref Full Text | Google Scholar

24. Fidder H, Schnitzler F, Ferrante M, Noman M, Katsanos K, Segaert S, et al. Long-term safety of infliximab for the treatment of inflammatory bowel disease: a single-centre cohort study. Gut. (2009) 58:501–8. doi: 10.1136/gut.2008.163642

PubMed Abstract | Crossref Full Text | Google Scholar

25. Dewedar AM, Shalaby MA, Al-Homaid S, Mahfouz AM, Shams OA, and Fathy A. Lack of adverse effect of anti-tumor necrosis factor-alpha biologics in treatment of rheumatoid arthritis: 5 years follow-up. Int J Rheum Dis. (2012) 15:330–5. doi: 10.1111/j.1756-185X.2012.01715.x

PubMed Abstract | Crossref Full Text | Google Scholar

26. Lee SK, Kim SY, Kim EY, Jung JY, Park MS, Kim YS, et al. Mycobacterial infections in patients treated with tumor necrosis factor antagonists in South Korea. Lung. (2013) 191:565–71. doi: 10.1007/s00408-013-9481-5

PubMed Abstract | Crossref Full Text | Google Scholar

27. Yoo IK, Choung RS, Hyun JJ, Kim SY, Jung SW, Koo JS, et al. Incidences of serious infections and tuberculosis among patients receiving anti-tumor necrosis factor-alpha therapy. Yonsei Med J. (2014) 55:442–8. doi: 10.3349/ymj.2014.55.2.442

PubMed Abstract | Crossref Full Text | Google Scholar

28. Kim M, Won J-Y, Choi SY, Ju JH, and Park Y-H. Anti-TNFα Treatment for HLA-B27-positive ankylosing spondylitis–related uveitis. Am. J. Ophthalmol. (2016) 170:32–40. doi: 10.1016/j.ajo.2016.07.016

PubMed Abstract | Crossref Full Text | Google Scholar

29. Lim CH, Lin CH, Chen DY, Chen YM, Chao WC, Liao TL, et al. One-year tuberculosis risk in rheumatoid arthritis patients starting their first tumor necrosis factor inhibitor therapy from 2008 to 2012 in Taiwan: A nationwide population-based cohort study. PloS One. (2016) 11:e0166339. doi: 10.1371/journal.pone.0166339

PubMed Abstract | Crossref Full Text | Google Scholar

30. Cagatay T, Bingol Z, Kıyan E, Yegin Z, Okumus G, Arseven O, et al. Follow-up of 1887 patients receiving tumor necrosis-alpha antagonists: Tuberculin skin test conversion and tuberculosis risk. Clin. Respir. J. (2017) 12:1668–75. doi: 10.1111/crj.12726

PubMed Abstract | Crossref Full Text | Google Scholar

31. Shobha V, Chandrashekara S, Rao V, Desai A, Jois R, Dharmanand BG, et al. Biologics and risk of tuberculosis in autoimmune rheumatic diseases: A real-world clinical experience from India. Int. J. Rheum. Dis. (2018) 22:280–7. doi: 10.1111/1756-185X.13376

PubMed Abstract | Crossref Full Text | Google Scholar

32. Sousa M, Ladeira I, Ponte A, Fernandes C, Rodrigues A, Silva AP, et al. Screening for latent tuberculosis in patients with inflammatory bowel disease under antitumor necrosis factor: data from a Portuguese center. Eur. J. Gastroenterol. Hepatol. (2019) 31:1099–102. doi: 10.1097/MEG.0000000000001469

PubMed Abstract | Crossref Full Text | Google Scholar

33. Argüder E, Yanık Üstüner G, Ekici R, Kılıç H, Erten Ş, and Karalezli A. Tuberculosis risk in patients with rheumatologic disease treated with biologic drugs. Tuberk. Toraks (2020) 68:236–44. doi: 10.5578/tt.69967

PubMed Abstract | Crossref Full Text | Google Scholar

34. Lee JY, Oh K, Hong HS, Kim K, Hong SW, Park JH, et al. Risk and characteristics of tuberculosis after anti-tumor necrosis factor therapy for inflammatory bowel disease: a hospital-based cohort study from Korea. BMC Gastroenterol. (2021) 21:390. doi: 10.1186/s12876-021-01973-5

PubMed Abstract | Crossref Full Text | Google Scholar

35. Koo BS, Lim YC, Lee MY, Jeon JY, Yoo HJ, Oh IS, et al. The risk factors and incidence of major infectious diseases in patients with ankylosing spondylitis receiving tumor necrosis factor inhibitors. Mod Rheumatol. (2021) 31:1192–201. doi: 10.1080/14397595.2021.1878985

PubMed Abstract | Crossref Full Text | Google Scholar

36. Slouma M, Athimni S, Dhahri R, Gharsallah I, Metoui L, and Louzir B. Tuberculosis infection under anti-TNF alpha treatment. Curr. Drug Saf. (2022) 17:235–40. doi: 10.2174/1574886316666211109092354

PubMed Abstract | Crossref Full Text | Google Scholar

37. Boqaeid A, Layqah L, Alonazy A, Althobaiti M, Almahlawi AZ, Al-Roqy A, et al. The risk of tuberculosis infection in Saudi patients receiving adalimumab, etanercept, and tocilizumab therapy. J Infect Public Health. (2024) 17:1134–41. doi: 10.1016/j.jiph.2024.04.016

PubMed Abstract | Crossref Full Text | Google Scholar

38. Godfrey MS and Friedman LN. Tuberculosis and biologic therapies: anti-tumor necrosis factor-alpha and beyond. Clin Chest Med. (2019) 40:721–39. doi: 10.1016/j.ccm.2019.07.003

PubMed Abstract | Crossref Full Text | Google Scholar

39. Benucci M, Saviola G, Manfredi M, Sarzi-Puttini P, and Atzeni F. Tumor necrosis factors blocking agents: analogies and differences. Acta Biomed. (2012) 83:72–80.

PubMed Abstract | Google Scholar

40. Wallis RS. Tumour necrosis factor antagonists: structure, function, and tuberculosis risks. Lancet Infect Dis. (2008) 8:601–11. doi: 10.1016/S1473-3099(08)70227-5

PubMed Abstract | Crossref Full Text | Google Scholar

41. Plessner HL, Lin PL, Kohno T, Louie JS, Kirschner D, Chan J, et al. Neutralization of tumor necrosis factor (TNF) by antibody but not TNF receptor fusion molecule exacerbates chronic murine tuberculosis. J Infect Dis. (2007) 195:1643–50. doi: 10.1086/517519

PubMed Abstract | Crossref Full Text | Google Scholar

42. Wang Q, Oryoji D, Mitoma H, Kimoto Y, Koyanagi M, Yokoyama K, et al. Methotrexate enhances apoptosis of transmembrane TNF-expressing cells treated with anti-TNF agents. Front Immunol. (2020) 11:2042. doi: 10.3389/fimmu.2020.02042

PubMed Abstract | Crossref Full Text | Google Scholar

43. Arora T, Padaki R, Liu L, Hamburger AE, Ellison AR, Stevens SR, et al. Differences in binding and effector functions between classes of TNF antagonists. Cytokine. (2009) 45:124–31. doi: 10.1016/j.cyto.2008.11.008

PubMed Abstract | Crossref Full Text | Google Scholar

44. Mitoma H, Horiuchi T, Tsukamoto H, Tamimoto Y, Kimoto Y, Uchino A, et al. Mechanisms for cytotoxic effects of anti-tumor necrosis factor agents on transmembrane tumor necrosis factor alpha-expressing cells: comparison among infliximab, etanercept, and adalimumab. Arthritis Rheum. (2008) 58:1248–57. doi: 10.1002/art.23447

PubMed Abstract | Crossref Full Text | Google Scholar

45. Nesbitt A, Fossati G, Bergin M, Stephens P, Stephens S, Foulkes R, et al. Mechanism of action of certolizumab pegol (CDP870): in vitro comparison with other anti-tumor necrosis factor alpha agents. Inflammation Bowel Dis. (2007) 13:1323–32. doi: 10.1002/ibd.20225

PubMed Abstract | Crossref Full Text | Google Scholar

46. Horiuchi T, Mitoma H, Harashima S, Tsukamoto H, and Shimoda T. Transmembrane TNF-alpha: structure, function and interaction with anti-TNF agents. Rheumatol (Oxford). (2010) 49:1215–28. doi: 10.1093/rheumatology/keq031

PubMed Abstract | Crossref Full Text | Google Scholar

47. Albayrak F, Keser G, Oztuzcu S, Coskun BN, Pehlivan Y, Yildirim TD, et al. HLA B subtype analysis in patients developing tuberculosis during anti-TNF treatment. Sci Rep. (2025) 15:15152. doi: 10.1038/s41598-025-00290-1

PubMed Abstract | Crossref Full Text | Google Scholar

48. Brehm TT, Reimann M, Kohler N, and Lange C. (Re-)introduction of TNF antagonists and JAK inhibitors in patients with previous tuberculosis: a systematic review. Clin Microbiol Infect. (2024) 30:989–98. doi: 10.1016/j.cmi.2024.04.011

PubMed Abstract | Crossref Full Text | Google Scholar