Elicit: Clinical Outcomes of TNF Inhibition in RA and Psoriasis

Clinical Outcomes of TNF Inhibition in RA and Psoriasis

Identify clinical evidence linking TNF inhibition to outcomes in rheumatoid arthritis and psoriasis

Abstract

TNF inhibitors demonstrate substantial therapeutic efficacy for psoriatic arthritis and psoriasis, with 37% of csDMARD-inadequate responders achieving clinical improvement (ACR50) versus 8% with placebo (RR 5.63, 95% CI 3.98 to 7.96). For moderate-to-severe plaque psoriasis, PASI 75 response rates reach 80-91% with infliximab compared to 3% with placebo, with therapeutic effects appearing within 4 weeks. However, paradoxical new-onset or worsening psoriasis emerges as a recognized adverse event in patients receiving TNF inhibitors for rheumatoid arthritis, occurring at 1.04 per 1000 person-years compared to 0 per 1000 person-years with traditional DMARDs. Adalimumab shows 4.6-fold higher incidence of paradoxical psoriasis than etanercept and 3.5-fold higher than infliximab, though no statistically significant differences in therapeutic effectiveness exist among anti-TNF agents. Most patients (66%) developing paradoxical psoriasis can continue TNF inhibitor therapy with concurrent psoriasis treatment, balancing control of the primary indication against induced skin disease.

Methods

We analyzed 10 sources from an initial pool of 200, using 8 screening criteria. Each paper was reviewed for 8 key aspects that mattered most to the research question.

Records from Elicit search

n = 200
Papers screened using: Target Population - Disease, Intervention Type, Clinical Outcomes, Study Design, Population Age, Study Type - Human Research, Sample Size Adequacy, Publication Type
n = 200
Papers screened out
n = 190
Papers included for extraction
n = 10

Paper search

We performed a semantic search across over 138 million academic papers from the Elicit search engine, which includes all of Semantic Scholar and OpenAlex.

We ran this query: “Identify clinical evidence linking TNF inhibition to outcomes in rheumatoid arthritis and psoriasis”
The search returned 200 total results from Elicit.
We retrieved 200 papers most relevant to the query for screening.

Screening

We screened in sources based on their abstracts that met these criteria:

Target Population - Disease: Does the study include patients diagnosed with rheumatoid arthritis and/or psoriasis?
Intervention Type: Does the study involve TNF inhibitor therapy (adalimumab, etanercept, infliximab, certolizumab pegol, or golimumab)?
Clinical Outcomes: Does the study report clinical outcomes such as disease activity measures, symptom improvement, quality of life, or safety outcomes?
Study Design: Is the study design a randomized controlled trial, cohort study, case-control study, systematic review, or meta-analysis?
Population Age: Does the study population include adult patients (≥18 years of age)?
Study Type - Human Research: Is this a study conducted in human participants (not in vitro, animal, or preclinical research)?
Sample Size Adequacy: Does the study include 10 or more patients (not a case report or small case series with <10 patients)?
Publication Type: Is this a full-text article (not a conference abstract, letter, editorial, or commentary)?

Data extraction

We asked a large language model to extract each data column below from each paper. We gave the model the extraction instructions shown below for each column.

Study Design: Extract the study design and methodology used to investigate TNF inhibition in rheumatoid arthritis and/or psoriasis, including:
- Type of study (RCT, cohort, case-control, registry study, meta-analysis, etc.)
- Comparison groups (placebo, active comparator, historical controls, etc.)
- Blinding status
- Duration of follow-up
- Key methodological features relevant to assessing TNF inhibitor effects
Target Condition: Extract which condition(s) the study addresses in relation to TNF inhibition:
- Rheumatoid arthritis (as primary indication)
- Psoriasis (as primary indication)
- Psoriatic arthritis
- Psoriasis as adverse event in RA patients
- Other related conditions
- Specify if the study covers multiple conditions and how they are analyzed.
Patient Population: Extract key characteristics of the study population relevant to TNF inhibition outcomes in RA and psoriasis, including:
- Sample size
- Age range/mean age
- Disease severity at baseline
- Disease duration
- Prior treatment history (DMARDs, biologics, etc.)
- Any relevant comorbidities
- Geographic/clinical setting
TNF Inhibitor Details: Extract specific information about TNF inhibitor treatment including:
- Name of TNF inhibitor(s) studied (infliximab, adalimumab, etanercept, golimumab, certolizumab)
- Dosage and administration route
- Treatment duration
- Whether used as monotherapy or combination therapy
- Any dose escalations or modifications
- Comparison between different TNF inhibitors if applicable
Primary Outcomes: Extract the primary clinical outcomes used to assess TNF inhibition effects in RA and psoriasis, including:
- Outcome measure names (PASI, ACR response, DAS28, PGA, etc.)
- How outcomes were defined and measured
- Time points of assessment
- Primary endpoint results with effect sizes, confidence intervals, and statistical significance
- Response rates or mean changes from baseline
Safety Findings: Extract safety and adverse event data related to TNF inhibition in RA and psoriasis patients, including:
- Serious adverse events
- Treatment-related adverse events
- Infections (including opportunistic infections)
- Paradoxical reactions (e.g., psoriasis in RA patients)
- Discontinuation rates due to adverse events
- Any condition-specific safety concerns
- Comparison of safety profiles between different TNF inhibitors
Key Evidence: Extract the main findings about the link between TNF inhibition and clinical outcomes in RA and psoriasis including:
- Direction and magnitude of treatment effects
- Clinical significance of observed effects
- Time to response/onset of effect
- Durability of response
- Factors that modify treatment response
- Any unexpected or paradoxical findings
- Comparative effectiveness between TNF inhibitors
- Authors’ conclusions about TNF inhibition efficacy/safety
Study Limitations: Extract key limitations and potential biases that could affect the interpretation of TNF inhibition evidence including:
- Sample size limitations
- Short follow-up duration
- Selection bias or generalizability concerns
- Missing data or dropout rates
- Confounding factors not controlled for
- Measurement bias or outcome assessment issues
- Funding source and potential conflicts of interest
- Authors’ stated limitations

Results

Characteristics of Included Studies

The review included 10 studies published between 2000 and 2025, comprising 5 randomized controlled trials, 3 meta-analyses/systematic reviews, 1 prospective registry study, and 1 retrospective cohort study. These studies addressed TNF inhibition in three distinct contexts: therapeutic use for psoriatic arthritis, therapeutic use for psoriasis, and paradoxical psoriasis development during TNF inhibitor therapy for other conditions.

Study	Full text retrieved?	Study Type	Target Condition	TNF Inhibitor(s)	Sample Size	Primary Outcomes
L. Lemos et al., 2014	No	Systematic review and meta-analysis	Psoriatic arthritis	Adalimumab, etanercept, golimumab, infliximab	Not mentioned	ACR20, ACR50, PsARC, PASI75
G. Cagnotto et al., 2025	No	Systematic review and meta-analysis of RCTs	Psoriatic arthritis	TNFi (specific agents not mentioned)	7857	ACR50, HAQ, SF-36 MCS, Sharp/Van der Heijde-PsA
E. Shmidt et al., 2012	No	Retrospective cohort study	Psoriasis as adverse event	Infliximab, adalimumab, etanercept	56	Complete or partial response
Mark Harrison et al., 2008	Yes	Prospective observational cohort (registry)	Rheumatoid arthritis; psoriasis as adverse event	Etanercept, infliximab, adalimumab	9826 anti-TNF, 2880 DMARD	Incidence of new-onset psoriasis
Angelique N. Collamer & D. Battafarano, 2010	No	Systematic literature review	Psoriasis as adverse event in RA and other conditions	Infliximab, adalimumab, etanercept	207	Not mentioned
P. Mease et al., 2000	No	Randomized controlled trial	Psoriatic arthritis and psoriasis	Etanercept 25 mg twice-weekly subcutaneous	60	PsARC, ACR20, PASI
N. Bansback et al., 2009	No	Systematic review and meta-analysis	Psoriasis	Infliximab, adalimumab, etanercept	Not mentioned	PASI
U. Chaudhari et al., 2001	No	Randomized controlled trial	Psoriasis	Infliximab 5 mg/kg and 10 mg/kg intravenous	33	Physician’s Global Assessment
K. Reich et al., 2005	No	Randomized controlled trial	Psoriasis	Infliximab 5 mg/kg via infusion	378	PASI
K. Thorlund et al., 2012	No	Indirect comparison meta-analysis	Psoriatic arthritis	Adalimumab, etanercept, golimumab, infliximab	Not mentioned	PsARC, HAQ, PASI

Studies varied in methodological rigor, with systematic reviews noting high or unclear risk of performance, detection, and reporting biases across included trials. Only one study in the Cagnotto review achieved low risk of bias across all domains. Sample sizes ranged from 33 participants in early-phase RCTs to 9826 patients in registry studies. Follow-up duration ranged from 10 weeks to a median of 2.81 years.

Effects of TNF Inhibitors

Efficacy in Psoriatic Arthritis

TNF inhibitors demonstrated substantial clinical benefit in psoriatic arthritis across multiple outcome domains. In csDMARD-inadequate responders, TNFi produced large improvements in clinical response at 12 weeks, with 37% achieving ACR50 compared to 8% with placebo (RR 5.63, 95% CI 3.98 to 7.96). An earlier meta-analysis found that more patients achieved ACR20, ACR50, and PsARC responses with anti-TNF agents compared to controls at 24 weeks, with etanercept and infliximab showing higher ACR70 response rates.

Physical function improved with TNFi treatment at 24 weeks, with a mean change of -0.14 points on the Health Assessment Questionnaire with placebo versus 0.33 points lower with TNFi (scale 0-3, MCID 0.35). For health-related quality of life, TNFi produced a clinically important improvement on the SF-36 Mental Component Summary at 24 weeks: 2.4 points with placebo versus 3.29 points higher with TNFi (95% CI 2.18 to 4.40; MCID 1.7).

At 24 weeks, 32% of TNFi-treated patients achieved minimal disease activity compared to 9% with placebo (RR 3.76, 95% CI 2.39 to 5.92). TNFi slightly reduced radiographic progression at 24 weeks on the Sharp/Van der Heijde-PsA score (0-528 scale): mean change 0.25 points with placebo versus 0.37 points lower with TNFi (95% CI 0.48 lower to 0.25 lower). Radiographic endpoints appeared time-dependent, with longer drug use associated with better results.

In the Mease trial of etanercept for psoriatic arthritis, 87% of etanercept-treated patients met PsARC criteria at 12 weeks compared to 23% with placebo, while 73% achieved ACR20 compared to 13% with placebo. Median time to response was 4 weeks for both etanercept and infliximab.

An indirect comparison meta-analysis found no statistically significant differences in effectiveness between available anti-TNFs for psoriatic arthritis, though golimumab yielded the highest relative risk for PsARC response, with etanercept second highest. For HAQ improvement among PsARC responders, etanercept and infliximab yielded the largest mean differences. Effect estimates proved sensitive to analytic approach.

Efficacy in Psoriasis

TNF inhibitors demonstrated high efficacy for moderate-to-severe plaque psoriasis, with differences in potency among agents. A mixed-treatment comparison meta-analysis found TNF inhibitors had a mean relative risk of 15.57 (95% CI 12.46-19.25) for achieving PASI 75. Infliximab showed the highest probability of PASI 75 achievement at 81%, followed by adalimumab at 71% and etanercept at 50%. Dosage emerged as an important determinant of outcome.

In the Reich trial of infliximab 5 mg/kg, 80% of patients achieved PASI 75 at week 10 compared to 3% with placebo (p<0.0001), while 57% achieved PASI 90 versus 1% with placebo. These improvements were maintained at week 24 (82% PASI 75, 58% PASI 90) and week 50 (61% PASI 75, 45% PASI 90), demonstrating durability of response over one year.

The Chaudhari trial of infliximab found 82% of patients receiving 5 mg/kg achieved good, excellent, or clear ratings on the Physician’s Global Assessment at week 10 versus 18% with placebo (difference 64%, 95% CI 20-89, p=0.0089). The 10 mg/kg dose achieved 91% response versus 18% placebo (difference 73%, 95% CI 30-94, p=0.0019). Median time to response was 4 weeks for both doses.

In the Mease etanercept trial for patients with both psoriatic arthritis and psoriasis, 26% of etanercept-treated patients achieved PASI 75 versus none with placebo (p=0.015), with median PASI improvement of 46% versus 9%.

For PASI improvement across agents, infliximab had the largest mean difference, followed by golimumab, with etanercept showing the smallest improvement.

Safety Profile

The safety profile of TNF inhibitors in psoriatic arthritis appeared favorable. At trial end, 4% of TNFi-treated patients experienced serious adverse events versus 3% with placebo (RR 1.00, 95% CI 0.70 to 1.42). However, TNFi slightly increased withdrawals due to adverse events: 3% with TNFi versus 2% with placebo (RR 1.53, 95% CI 1.01 to 2.33). Etanercept and infliximab showed worse results for application site reactions compared to other agents. No serious adverse events occurred in the Chaudhari infliximab trial, with the drug described as well tolerated.

Paradoxical Psoriasis During TNF Inhibitor Therapy

Despite TNF inhibitors’ efficacy in treating psoriasis, paradoxical new-onset or worsening psoriasis emerged as a recognized adverse event in patients receiving these agents for other indications, particularly rheumatoid arthritis and inflammatory bowel disease.

In the British Society for Rheumatology Biologics Register study of 9826 anti-TNF-treated and 2880 DMARD-treated RA patients, the crude incidence rate of psoriasis was 1.04 per 1000 person-years (95% CI 0.67 to 1.54) in the anti-TNF cohort versus 0 per 1000 person-years (upper 97.5% CI 0.71) in the DMARD cohort. The incidence rate varied significantly by agent: adalimumab showed a 4.6-fold higher rate than etanercept (95% CI 1.7 to 12.1) and 3.5-fold higher than infliximab (95% CI 1.3 to 9.3). Median time to onset was 6 months.

In the Mayo Clinic retrospective series of 56 patients who developed psoriasis during anti-TNF therapy, the mean age at onset was 48.1 years, with 73% female. The underlying conditions were Crohn’s disease (39%) and rheumatoid arthritis (25%). Patients were treated with infliximab (54%), adalimumab (34%), or etanercept (12%). New-onset or worsening psoriasis occurred after a mean treatment duration of 17.1 months. Clinical presentations included plaque psoriasis (48%), palmoplantar pustulosis (45%), scalp psoriasis (21%), generalized pustular psoriasis (13%), erythrodermic psoriasis (4%), and inverse psoriasis (4%).

Among 39 patients with complete treatment response data, 85% achieved complete or partial response. Response rates were higher among those who discontinued anti-TNF therapy (94%) compared to those who continued with concomitant psoriasis treatment (77%).

A comprehensive literature analysis identified 207 cases of TNF antagonist-induced psoriasis between 1996 and 2009. The population included 43% with rheumatoid arthritis, 26% with seronegative spondyloarthropathy, and 20% with inflammatory bowel disease. Mean age was 45 years and 65% were female. Lesion morphology comprised pustular psoriasis (56%), plaque psoriasis (50%), and guttate lesions (12%), with 15% experiencing multiple lesion types. No statistically significant predisposing factors for new-onset psoriasis were identified. Importantly, 66% of patients continued TNF antagonist therapy successfully with concurrent psoriasis treatments.

Synthesis

The evidence reveals an apparent paradox: TNF inhibitors demonstrate substantial efficacy in treating both psoriatic arthritis and moderate-to-severe psoriasis, yet simultaneously induce or exacerbate psoriasis in patients treated for other inflammatory conditions. This contradiction can be reconciled through several mechanistic and contextual considerations.

The immunopathogenesis differs between therapeutic response and paradoxical induction. The pathogenesis of TNF antagonist-induced psoriasis appears to involve disruption of the cytokine milieu with unopposed interferon-α production by plasmacytoid dendritic cells in genetically predisposed individuals. This mechanism suggests that while TNF-α plays a pivotal role in established psoriasis pathogenesis, its inhibition may trigger compensatory pathways that, in susceptible individuals, precipitate psoriatic lesions through different immunologic mechanisms.
Population differences explain divergent outcomes. Patients developing paradoxical psoriasis during anti-TNF therapy for RA or inflammatory bowel disease represent a distinct population from those with pre-existing psoriasis or psoriatic arthritis. The Mayo Clinic series found no statistically significant predisposing factors, suggesting that genetic polymorphisms may determine susceptibility. Registry data reporting remains necessary to define true incidence and prevalence, as current estimates may underestimate actual rates due to non-reporting of mild cases.
Agent-specific differences emerged in paradoxical psoriasis incidence but not therapeutic efficacy. Adalimumab showed significantly higher rates of incident psoriasis in RA patients compared to etanercept and infliximab, yet meta-analyses found no statistically significant differences in therapeutic effectiveness for psoriatic arthritis. For psoriasis treatment, infliximab and adalimumab demonstrated higher PASI 75 achievement rates than etanercept, but all agents proved superior to control. These discordant safety and efficacy profiles across agents suggest that differential binding characteristics, pharmacokinetics, or immunogenicity may influence paradoxical reactions independently of therapeutic effect.
Temporal patterns provide insight. Paradoxical psoriasis typically developed after 6 months (median) to 17.1 months (mean) of treatment, whereas therapeutic responses occurred within 4 weeks for etanercept and infliximab. The delayed onset of paradoxical reactions compared to rapid therapeutic effects suggests distinct underlying mechanisms—immediate TNF blockade produces benefit in TNF-driven disease, while chronic TNF suppression may eventually trigger compensatory pathways in susceptible individuals.
Management strategies differ based on clinical context. In patients developing paradoxical psoriasis, discontinuation of anti-TNF therapy achieved higher complete response rates (94%) than continuation with concurrent psoriasis treatment (77%), though 66% could successfully continue therapy with appropriate management. This contrasts with the sustained efficacy seen in therapeutic use, where infliximab maintained PASI 75 response in 61% at week 50. The decision to continue or switch therapy depends on balancing control of the primary indication against severity of induced psoriasis.

The evidence supports concluding that TNF inhibitors are highly effective for psoriatic arthritis and psoriasis when used as indicated, with ACR50 response rates of 37% versus 8% placebo and PASI 75 rates reaching 80-91% for infliximab. Simultaneously, paradoxical psoriasis represents a genuine but relatively uncommon adverse event (1.04 per 1000 person-years) that occurs through distinct pathophysiologic mechanisms in genetically susceptible individuals. Both phenomena accurately reflect TNF biology: central involvement in psoriasis pathogenesis explains therapeutic benefit, while complex cytokine network disruption explains paradoxical induction in certain contexts.