How TNF Inhibitors Work, What They Treat, and Risks

TNF inhibitors are a class of biologic drugs that block tumor necrosis factor-alpha, an inflammatory signaling molecule that drives tissue damage in autoimmune conditions like rheumatoid arthritis, Crohn’s disease, psoriasis, and ankylosing spondylitis. Introduced in the late 1990s, they became the first widely used biologic therapies and remain among the most prescribed, with a global market exceeding $20 billion annually.¹ But the five approved TNF inhibitors are not interchangeable, and the differences between them have real consequences for infection risk, pregnancy safety, and which diseases they actually help.

How TNF Inhibitors Work

TNF-alpha is a protein released mainly by immune cells called macrophages. In healthy people, it helps coordinate immune defense and inflammation. In autoimmune diseases, the body produces too much of it, and the excess TNF drives chronic inflammation that damages joints, the gut lining, skin, and other tissues. TNF inhibitors intercept this process by binding to TNF-alpha and preventing it from activating its receptors on cell surfaces.

The early understanding was straightforward: block the signal, reduce the inflammation. But research over the past two decades has shown the mechanism is more layered than simple neutralization. TNF inhibitors also appear to trigger cell-cycle arrest and programmed cell death in certain inflammatory cells, suppress production of other inflammatory molecules downstream, and improve the function of regulatory T cells, the immune cells responsible for keeping inflammation in check.² This broader set of effects helps explain why TNF inhibitors work across so many different diseases, not just the one they were originally designed for.

The Five Approved Agents Are Structurally Different

There are five TNF inhibitors approved for clinical use: infliximab, adalimumab, etanercept, golimumab, and certolizumab pegol. They all block TNF-alpha, but they are built differently, and those structural differences matter.

Infliximab and adalimumab are full monoclonal antibodies. Infliximab is a chimeric antibody, meaning part of it is derived from mouse protein and part from human protein. Adalimumab is fully human. Both carry an Fc region, the tail end of an antibody that interacts with other parts of the immune system. Golimumab is also a fully human monoclonal antibody with an Fc region. Etanercept is different: it is a fusion protein that combines two TNF receptor domains with an antibody Fc region, essentially a decoy receptor that soaks up TNF. Certolizumab pegol stands apart from all the others because it lacks the Fc region entirely. It is a PEGylated antibody fragment, a piece of an antibody attached to a polyethylene glycol chain that extends its half-life in the body.³

The Fc region is the key dividing line. It is the part of an antibody that gets recognized by a receptor called FcRn, which actively transports antibodies across cell barriers, including the placenta. Certolizumab’s missing Fc region explains its unique pregnancy profile, discussed below. The Fc region also influences how the drug interacts with immune cells and whether it can trigger complement activation or cell killing, which may contribute to differences in side-effect profiles among the agents.

What Conditions They Treat

The approved uses of TNF inhibitors span rheumatoid arthritis, psoriatic arthritis, plaque psoriasis, ankylosing spondylitis, Crohn’s disease, ulcerative colitis, and juvenile idiopathic arthritis, though not every agent is approved for every condition. In rheumatoid arthritis, combining a TNF inhibitor with methotrexate or another conventional disease-modifying drug typically works better than using the TNF inhibitor alone. A population-based study found that infliximab plus conventional drugs was significantly more effective than etanercept alone at preventing erosion progression and joint-space narrowing in RA patients.⁴

In ankylosing spondylitis, TNF inhibitors remain a strong first-line biologic option. A network meta-analysis comparing TNF inhibitors with interleukin inhibitors and JAK inhibitors in ankylosing spondylitis found that TNF inhibitors had significantly better clinical response rates than every other drug class studied.⁵ In psoriatic arthritis, adalimumab has been studied in patients who previously failed a different TNF inhibitor. Even in that harder-to-treat group, about 71% of patients with psoriatic arthritis met response criteria after 12 weeks, though the rate was somewhat higher (about 79%) in those who had never tried a TNF inhibitor before.⁶

Why Not Every TNF Inhibitor Works for Every Disease

One of the most striking findings in this drug class is that etanercept, which works well in rheumatoid arthritis and psoriasis, is ineffective in Crohn’s disease. A randomized controlled trial found that patients with moderate-to-severe Crohn’s disease who received etanercept had essentially the same response rate as those on placebo.⁷ By contrast, infliximab and adalimumab are cornerstones of Crohn’s treatment.

The reason likely comes down to how these agents interact with TNF at the cellular level. Etanercept, as a soluble receptor, binds and neutralizes free-floating TNF but does not effectively target the membrane-bound form of TNF that sits on the surface of inflammatory cells in the gut. The monoclonal antibodies bind both soluble and membrane-bound TNF, and their ability to lock onto TNF-expressing cells may trigger additional effects like programmed cell death of those cells. In a disease like Crohn’s, where the inflammatory process involves dense clusters of immune cells in the intestinal wall, that ability to directly engage and kill inflammatory cells appears to be essential.

Infection Risks, Especially Tuberculosis

Because TNF-alpha plays a central role in the body’s defense against infections, particularly in forming and maintaining granulomas (the organized clusters of immune cells that wall off certain pathogens), blocking it increases susceptibility to specific infections. The most important of these is tuberculosis. TNF is essential for maintaining the granulomas that keep latent TB contained, and starting a TNF inhibitor can destabilize those structures, allowing dormant bacteria to reactivate.⁸ This is why screening for latent TB before starting any TNF inhibitor is standard practice worldwide.

The infection risk is not identical across the five agents. In a large safety study of over 30,000 TNF inhibitor users, about half a percent developed targeted fungal or mycobacterial infections. TB accounted for the majority of mycobacterial cases, followed by histoplasmosis and coccidioidomycosis.⁹ The monoclonal antibodies, particularly infliximab, carry a higher risk of granulomatous infections than etanercept. In the SABER study of over 33,000 new TNF inhibitor users, infliximab was associated with roughly 2.6 times the risk of non-viral opportunistic infections compared to conventional drugs, while etanercept carried a significantly lower risk than infliximab.¹⁰ This difference aligns with the structural distinction: etanercept’s mechanism of neutralizing soluble TNF without fully disabling the membrane-bound signaling that maintains granuloma integrity may preserve some protective immunity.¹¹

Baseline corticosteroid use roughly doubled the risk of opportunistic infections on top of the TNF inhibitor risk, reinforcing the importance of keeping steroid doses as low as possible when using biologics.¹²

Paradoxical Psoriasis

One of the stranger side effects of TNF inhibitors is that they can cause the very skin disease they are used to treat. Patients taking a TNF inhibitor for inflammatory bowel disease or arthritis sometimes develop new-onset psoriasis or see existing psoriasis flare, even as their primary condition improves. This is called paradoxical psoriasis, and it occurs across all five agents.¹³

The leading explanation involves a type of immune cell called a plasmacytoid dendritic cell. Normally, TNF promotes the maturation of these cells, which shuts down their production of type I interferon. When TNF is blocked, the cells stay immature longer and keep pumping out interferon-alpha, a molecule that drives psoriatic skin inflammation. Research has shown that adding a TNF-blocking antibody to stimulated plasmacytoid dendritic cells in the lab significantly increased and prolonged their interferon production, while adding TNF back suppressed it.¹⁴ In IBD patients specifically, the onset of paradoxical psoriasis has been associated with female sex, the presence of other health conditions, adalimumab use, and certain genetic variants in the TNF-alpha gene region.¹⁵

Cancer and Demyelinating Disease Signals

The relationship between TNF inhibitors and cancer has been debated since these drugs launched. TNF-alpha has complex roles in tumor biology: it can promote tumor growth through inflammation but also contributes to immune surveillance. A study of older Americans with rheumatoid arthritis found that TNF inhibitor use was associated with a modestly elevated risk of non-melanoma skin cancer and non-Hodgkin lymphoma, with a particularly strong signal for follicular lymphoma.¹⁶ However, disentangling the drug’s contribution from the underlying disease risk is difficult. Rheumatoid arthritis itself raises lymphoma risk, and patients on TNF inhibitors tend to have more severe disease.

A more concerning signal involves the central nervous system. A systematic review and meta-analysis found that TNF inhibitor exposure was associated with a 36% increased risk of inflammatory CNS diseases compared to conventional therapies, driven mainly by demyelinating conditions like multiple sclerosis.¹⁷ A Canadian population study also found an increased risk of MS among TNF inhibitor users, particularly in those with rheumatic diseases.¹⁸ This is why existing demyelinating disease is considered a contraindication to TNF inhibitor therapy, and new neurological symptoms in a patient on these drugs warrant prompt evaluation.

When TNF Inhibitors Stop Working

Roughly 30 to 40% of patients starting a TNF inhibitor will eventually discontinue it due to primary failure (it never worked well enough), secondary loss of response (it worked initially but stopped), or intolerance.¹⁹ Secondary loss of response is often driven by immunogenicity: the patient’s immune system produces antibodies against the TNF inhibitor itself. These anti-drug antibodies can neutralize the medication and accelerate its clearance from the body, resulting in declining drug levels and worsening disease.²⁰

The risk of developing anti-drug antibodies is higher when a TNF inhibitor is used without an immunomodulator alongside it. Methotrexate, when taken together with a TNF inhibitor, reduces the likelihood of immunogenicity.²¹ In inflammatory bowel disease, adding an immunomodulator to reverse established anti-drug antibodies has been studied in systematic reviews, though the evidence remains mixed.²²

Blood tests can now measure both drug levels and anti-drug antibody concentrations, a practice called therapeutic drug monitoring. This helps clinicians decide whether to increase the dose, add an immunomodulator, or switch to a different drug entirely. There is growing evidence that this approach improves outcomes and saves money compared to making dose changes based on symptoms alone, though variability in available assays has limited its universal adoption.²³

When a patient fails one TNF inhibitor, there are two strategic options: cycling to a different TNF inhibitor, or swapping to a biologic with a completely different mechanism of action. Cycling can work when the failure was due to immunogenicity against a specific drug, since the new TNF inhibitor is a different protein. But if the patient’s disease simply does not respond well to TNF blockade, switching drug classes makes more sense.

Pregnancy Considerations

Pregnancy raises a particular concern with TNF inhibitors because most of them are based on full IgG antibodies, and IgG antibodies are actively transported across the placenta starting in the second trimester. Adalimumab, infliximab, and golimumab all cross into the fetal circulation and can be detected in the infant for up to a year after birth.²⁴ This matters because live vaccines given to the infant in the first months of life could theoretically be dangerous if biologic drug is still circulating in the baby’s system.

Certolizumab pegol is the exception. Because it lacks the Fc region that FcRn receptors use to move antibodies across the placenta, it shows no to minimal placental transfer. The CRIB study confirmed that certolizumab levels in infants born to treated mothers were negligible, suggesting the fetus is not meaningfully exposed during the third trimester.²⁵ Pharmacovigilance data have reinforced this structural distinction: agents with high placental transfer showed increasing adverse event reporting rates over time, while certolizumab showed a constant, lower pattern.²⁶ For women who need continuous biologic therapy through pregnancy, certolizumab has become the preferred TNF inhibitor in many clinical settings.

How TNF Inhibitors Affect Vaccine Responses

Given that TNF inhibitors dampen immune function, a natural question is whether they blunt vaccine effectiveness. The answer depends on the vaccine type and how much time has passed since vaccination.

For traditional vaccines like influenza and pneumococcal, a meta-analysis of rheumatoid arthritis patients found that TNF inhibitors did not significantly reduce the antibody response to either vaccine. The odds of mounting an adequate response were similar between patients on TNF inhibitors and controls for all influenza strains and pneumococcal serotypes tested.²⁷ This was reassuring and supported annual flu vaccination for patients on these drugs.

COVID-19 vaccination told a more nuanced story. Shortly after vaccination, patients on TNF inhibitors mounted an antibody response comparable to controls. But the durability of that response was reduced. One study found that the proportion of patients on TNF inhibitors with adequate antibodies dropped from about 88% at four weeks to roughly 45% at six months, a steeper decline than in patients on other treatments.²⁸ Another study found that only about 62% of patients on TNF inhibitors reached maximum immune response after COVID-19 vaccination, compared with about 77% of controls.²⁹ The practical takeaway is that patients on TNF inhibitors should stay current with recommended boosters and should not assume one or two doses confer long-lasting protection.

Live vaccines remain contraindicated during TNF inhibitor therapy because the suppressed immune system may not contain the attenuated pathogen effectively. Inactivated vaccines are safe to administer, and most guidelines recommend giving them on schedule.

Biosimilars

As patents on the original TNF inhibitors expired, biosimilar versions entered the market. Biosimilars are not generic drugs in the traditional sense; because biologics are produced by living cells, a biosimilar must demonstrate that it is highly similar to the originator in structure, function, and clinical performance. A large French observational study including nearly 87,000 patients compared biosimilar and originator versions of infliximab, etanercept, and adalimumab. Biosimilar users showed no increased risk of discontinuation, infections, or hospitalization compared to originator users across all three drugs and all disease categories studied.³⁰ The cost savings from biosimilar availability have been substantial, broadening access to biologic therapy in healthcare systems worldwide.

Stopping Treatment and the Risk of Relapse

For patients who achieve stable remission on a TNF inhibitor, the question of whether and when to stop arises regularly, especially given the cost, injection burden, and ongoing infection risk. The evidence suggests that stopping entirely is possible but carries a meaningful relapse risk. A systematic review of IBD patients found that about 44% of Crohn’s disease patients and 38% of ulcerative colitis patients relapsed after stopping their TNF inhibitor.³¹ Encouragingly, about 80% of those who relapsed regained remission when the same drug was restarted.

In Crohn’s disease, the basis for stopping mattered. When the decision was based on clinical remission alone, the one-year relapse rate was about 42%. When endoscopic remission (confirmed healing of the intestinal lining) was also required before stopping, the rate dropped to about 26%, suggesting that deeper remission offers better odds of staying well off therapy.³²

In rheumatoid arthritis, a gradual tapering strategy fares better than abrupt cessation. One study randomized RA patients in stable remission into groups that either continued full treatment, tapered the biologic alone, tapered both drugs, stopped the biologic while tapering conventional drugs, or stopped everything. Relapse rates climbed steeply with each escalation: about 42% relapsed when the biologic was halved, compared to nearly 77% when all medications were stopped. Only about 7% of those who continued unchanged therapy relapsed. Patients whose disease activity score was very low at baseline (below 2.0) were more likely to maintain remission after tapering.³³ Ultrasound monitoring of joints and certain blood markers of autoimmunity helped predict who would flare, supporting a personalized approach rather than a one-size-fits-all stopping rule.³⁴

Where TNF Inhibitors Stand Among Newer Biologics

The biologic landscape has expanded considerably since TNF inhibitors debuted. Drugs targeting interleukin-17, interleukin-23, interleukin-12/23, and the JAK signaling pathway now compete for similar indications. In some conditions the newer agents have edged ahead: a network meta-analysis of nail psoriasis found that the IL-17 inhibitor ixekizumab was more effective than infliximab, adalimumab, and several IL-12/23 and IL-23 inhibitors.³⁵ In ankylosing spondylitis, by contrast, TNF inhibitors still outperform the alternatives as noted above.

TNF inhibitors retain several practical advantages: decades of long-term safety data, well-established biosimilar options that reduce cost, and familiarity among clinicians across specialties. For many patients, particularly those with rheumatoid arthritis or axial spondyloarthritis, they remain the first biologic tried. The decision to start with a TNF inhibitor versus a newer agent increasingly depends on the specific disease, the patient’s comorbidities (avoiding TNF inhibitors in someone with a history of demyelinating disease, for instance), and whether the cost differential from biosimilar availability tips the balance.