Reviews/TNF Alpha Inhibitors/TNF-alpha inhibitors increase the risk of serious infections including tuberculosis reactivation
· TN-SAF-01 · v2
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
The evidence in this corpus splits into two parts. On tuberculosis, several independent studies across different countries and designs point the same way: people taking TNF-alpha inhibitors are more likely to develop or reactivate tuberculosis and related mycobacterial infections [16][8][15][11], and laboratory and modelling work offers a biological explanation for why [9][5]. On general serious infections the picture is less uniform: risk is higher when TNFi are compared with some milder treatments [14][3] but no higher when compared with others in psoriasis [17] and rheumatoid arthritis [19]. Most of the human evidence comes from observational cohorts and registries, where other factors such as age, lung disease and steroid use also drive infection risk [1]. The panel concluded the tuberculosis half of the claim is well supported while the serious-infection half depends on what TNFi are compared against.
Independent cohorts, a nested case-control study and a registry consistently associate TNFi with tuberculosis and nontuberculous mycobacterial infection: anti-TNF TB OR 5.04 and NTM OR 2.19 [16], mycobacterial HR 1.98 and serious-infection HR 1.71 versus thiopurine monotherapy in a 190,694-patient IBD cohort [14], active TB above general-population rates in inflammatory arthritis [8], and incident TB despite screening [15][7]. Mechanistic support comes from Markov modelling of FDA reports showing infliximab reactivates latent infection far more than etanercept [9] and an in-vitro model in which anti-TNF reactivated dormant M. tuberculosis [5]. For non-TB serious infection, the signal is comparator-dependent: raised versus thiopurine [14] and versus placebo arms [3], but an adjusted HR of 0.98 versus nonbiologics in psoriasis [17] and aIRRs of 0.84 to 1.2 with all intervals crossing 1 versus cDMARDs in rheumatoid arthritis [19]. A registry shows TB rates falling from 783 to 38 per 100,000 person-years as screening took hold [11]. The contested point: The division is over the unqualified word 'increase' in the serious-infection half of the claim: it holds against milder comparators [14][3] but is null against nonbiologics [17] and cDMARDs [19]. The tuberculosis direction is not contested within the corpus, but its magnitude and present-day absolute size are, given small case counts [16][15] and the sharp fall in rates after screening [11].
All 19 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from TNF Alpha Inhibitors, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The serious-infection half of the claim is directly contradicted within the corpus: [17] reports an adjusted HR of 0.98 (0.80-1.19) in psoriasis and [19] reports TNFi aIRRs of 0.84-1.2 versus cDMARDs with all CIs crossing 1. The unqualified word 'increase' is comparator- and indication-dependent, not universal.
Nearly all human evidence is observational (cohort, registry, case-control) and subject to confounding by indication and co-medication; [1] identifies age, chronic lung disease and oral corticosteroid use as the dominant infection predictors, so disease-severity and steroid confounding are not excluded.
Where it remains uncertain. The serious-infection increase is contradicted within the corpus, being null versus nonbiologics in psoriasis (HR 0.98) [17] and versus cDMARDs in rheumatoid arthritis [19].
Aleatory
The headline mycobacterial estimate rests on only 37 TB cases [16] and 7 incident TB cases [15], producing wide intervals such as OR 5.04 (1.27 to 20.0) [16].
Epistemic
The corpus holds no RCT of TNFi versus a non-biologic baseline for serious infection; the quantitative estimates come from cohorts and registries [14][17][19].
Inferential
Observational populations have intrinsic infection risk; age, chronic lung disease and time-dependent oral corticosteroid use are the top predictors [1], so steroid and severity confounding is not excluded.
The load-bearing papers from a corpus of 19, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
In-vitro mechanistic support that anti-TNF reactivates dormant M. tuberculosis.
“In vitro , anti-TNFα treatment showed increased staining for Auramine-O, decreased Nile red staining and decreased rifampicin resistance, indicative of mycobacterial reactivation.”
Panel auditInternal validity: low · scope of inference: Mechanistic signal that anti-TNF reactivates dormant M. tuberculosis in a laboratory granuloma model.Key limitations: in vitro not clinical · no statistics given · designed as a secukinumab comparator study
Active-TB incidence under TNFi far exceeds background population rates.
“The number of active TB cases/time of exposure to TNFi (person-years, PY) was 8/593.5, 9/1044.0 and 3/555.3, respectively, accounting for an incidence of 1348.0, 862.1 and 540.2 cases per 10^5 PY.”
Panel auditInternal validity: low · scope of inference: Single-centre RA, PsA and AS patients on TNFi, active-TB incidence by baseline screening method.Key limitations: no concurrent comparator · comparator rate not in abstract · few TB events
Mechanistic/modeling support that TNF blockade drives latent-TB reactivation.
“Modeling revealed an apparent median monthly rate of reactivation of LTBI by infliximab treatment of 20.8%, which was 12.1 times that with etanercept treatment (P<0.001).”
Panel auditInternal validity: low · scope of inference: Modelled pathogenesis of TNFi-associated TB, distinguishing LTBI reactivation from new-infection progression.Key limitations: modelling not empirical cohort · based on spontaneous FDA reports · no non-TNFi comparator
Counterweight: null serious-infection increase versus nonbiologics in psoriasis.
“For serious infections, age, gender, and Charlson morbidity index adjusted hazard ratio of exposure to anti-TNFs compared with nonbiologics was 0.98 (95% confidence interval 0.80-1.19), for bacterial cutaneous infections it was 1.00 (95% confidence interval 0.62-1.61), and for granulomatous infections it was 1.23 (95% confidence interval 0.82-1.84).”
Panel auditInternal validity: high · scope of inference: Psoriasis patients in clinical practice, serious, cutaneous and granulomatous infection risk of anti-TNF versus nonbiologics.Key limitations: did not isolate TB · heterogeneous any-infection outcome · psoriasis population only
4 of 19 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
The corpus consistently carries the tuberculosis-reactivation component but shows the serious-infection increase is conditional on comparator and indication, warranting a conservative Likely rather than Established.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
TNFi versus true non-biologic baseline for serious infection. Whether the unqualified 'increase' holds turns on the comparator, and the corpus is split between positive [14][3] and null [17][19] contrasts. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: An active-comparator new-user study or pragmatic trial contrasting TNFi initiation against conventional non-biologic therapy with adjustment for steroid use and disease activity.
Open question: Does TNFi therapy raise serious non-TB infection risk relative to a genuine non-biologic baseline, and does this vary by indication? Why it is still open: Estimates diverge from HR 1.71 versus thiopurine [14] to HR 0.98 versus nonbiologics in psoriasis [17] and null aIRRs versus cDMARDs in RA [19].
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 26 Oct 2026 · protocol v0.19