Key messages
Tumour necrosis factor inhibitors (TNFi) – a group of medicines that help reduce inflammation – are probably more effective than placebo (an inactive or 'dummy' medicine) in treating people with psoriatic arthritis who did not benefit from treatment with first-line immunosuppressants (medicines that prevent the activity of the immune system, such as methotrexate).
High-quality studies comparing TNF inhibitors with second-line immunosuppressants, and studies that evaluate the effects of TNFi in people with psoriatic arthritis who didn’t respond to second-line treatments, are needed.
What is psoriatic arthritis?
Psoriatic arthritis is a lifelong inflammatory disease of the joints, affecting about one-third of people with psoriasis, which is an inflammatory skin disease. The cause of psoriatic arthritis is not fully understood. Inflammation of tissues in and around the joints causes pain, swelling, and stiffness of joints and tendons. If untreated, psoriatic arthritis may lead to disability and reduce quality of life.
How is psoriatic arthritis treated?
Psoriatic arthritis is treated with physiotherapy, anti-inflammatory medicines (NSAIDs), corticosteroids, and other immunosuppressants, which are called disease modifying anti-inflammatory drugs (DMARDs) in a step-up approach – that is, a treatment strategy where doctors start with the simplest or least intensive treatment and gradually move to stronger or more complex treatments if needed. DMARDs are divided into three categories: csDMARDs (e.g. methotrexate), bDMARDs (e.g. TNFi, ixekizumab), and tsDMARDs (e.g. upadacitinib).
TNFi selectively block a body protein called tumour necrosis factor (TNF). TNF plays a key role in long-term inflammation of joints and skin in psoriatic arthritis. Therefore, targeting TNF with TNFi may reduce inflammation and improve related symptoms.
What did we want to find out?
We wanted to find out if TNFi are better than placebo, physiotherapy, NSAIDs, corticosteroids, and other immunosuppressants, in treating adults with psoriatic arthritis. We were particularly interested in the effect of treatment on clinical improvement, control of inflammation (disease activity), physical function, health-related quality of life, radiographic damage (changes or harm to bones or joints that can be seen on an X-ray scan), serious side effects, and withdrawals due to side effects.
What did we do?
We searched for studies that investigated TNFi compared with placebo or alternative treatments. We categorised studies according to target participants: those who had never been treated with immunosuppressants; those who were unsuccessfully treated with csDMARDs; and those who were unsuccessfully treated with bDMARDs or tsDMARDs.
What did we find?
We found 25 studies that involved 7857 participants with psoriatic arthritis. The largest study included 1282 adults and the smallest, 47 adults. Studies were conducted in Europe, North America, Asia, Australia, South Africa, and South America. Most studies involved participants who were unsuccessfully treated with csDMARDs.
Compared to placebo:
359 more people per 1000 reported a clinical improvement at 12 weeks with TNFi
• 436 per 1000 reported a clinical improvement with TNFi.
• 77 per 1000 reported a clinical improvement with placebo.
258 more people per 1000 had better disease control at 24 weeks with TNFi
• 351 per 1000 had better disease control with TNFi.
• 93 per 1000 had better disease control with placebo.
Function measured on a 0 to 3 scale (0 is best function) at 24 weeks was 0.33 points better with TNFi
• People who had TNFi rated their function as -0.47 points.
• People who had placebo rated their function as -0.14 points.
Quality of life measured on a 0 to 100 scale (100 is best score) at 24 weeks was 3.29 points better with TNFi
• People who had TNFi rated their quality of life as 5.69 points.
• People who had placebo rated their function as 2.4 points.
Radiographic damage measured on a 0 to 528 scale (0 is best score) at 24 weeks was 0.37 points better with TNFi
• People who had TNFi rated their radiographic damage as -0.12 points.
• People who had placebo rated their radiographic damage as 0.25 points.
No difference in serious side effects were found with TNFi
• 31 per 1000 people experienced a serious side effect with both TNFi and placebo.
10 more people per 1000 withdrew from treatment due to a side effect with TNFi
• 28 per 1000 withdrew treatment with TNFi.
• 18 per 1000 withdrew treatment with placebo.
In participants who had never received immunosuppressants, TNFi probably led to a large clinical improvement, better disease control, but to no difference in physical function, and to a slight reduction of radiographic damage, compared to methotrexate. No study investigated the effect on health-related quality of life. We are uncertain whether there are any differences in harm between TNFi and methotrexate.
What are the limitations of the evidence?
Some studies did not clearly report how they were conducted, or whether the healthcare professionals or participants were aware of what medicines were delivered, which may have affected the validity of the results. For some outcomes, such as serious side effects, the evidence is based on few events.
How current is this review?
The review is current to 28 March 2024.
In csDMARD-inadequate responders, moderate-certainty evidence showed that TNFi probably result in a large clinical improvement, lower disease activity, small decrease in radiographic progression, and better quality of life compared to placebo. Low-certainty evidence showed that TNFi may lead to a slight improvement in physical function compared to placebo. Low-certainty evidence suggested that TNFi may lead to a slight increase in withdrawals due to adverse events, whereas they may result in little to no difference in serious adverse events compared to placebo. No trials assessed TNFi compared to placebo in DMARD-naïve participants or in b/tsDMARD-IR.
Psoriatic arthritis (PsA) is a chronic arthritis affecting people with psoriasis. If untreated, it may lead to disability. Recommended drugs are non-steroidal anti-inflammatory drugs (NSAIDs), corticosteroids, conventional synthetic disease-modifying anti-rheumatic drugs (csDMARDs), biologic DMARDs (bDMARDs), and targeted synthetic DMARDs (tsDMARDs). Tumour necrosis factor inhibitors (TNFi) are the first choice bDMARDs.
To assess the benefits and harms of TNFi in adults with psoriatic arthritis.
We searched CENTRAL, MEDLINE, Embase, ClinicalTrials.gov and the World Health Organization trials portal up to 28 March 2024.
We included randomized controlled trials (RCTs) in adults with PsA, comparing TNFi to placebo, physiotherapy, NSAIDs, corticosteroids, and cs/b/tsDMARDs. Major outcomes included clinical improvement, minimal disease activity, physical function, health-related quality of life, radiographic progression, serious adverse events, and withdrawals due to adverse events.
We used standard Cochrane methods. The primary comparison was TNFi versus placebo. The primary time point was 12 weeks for clinical improvement; 24 weeks for minimal disease activity, function, quality of life, and radiographic progression; and the end of the trial period for serious adverse events and withdrawals due to adverse events.
We included 25 RCTs randomizing 7857 participants. Four studies compared TNFi to methotrexate and one to ustekinumab in DMARD-naïve participants. In csDMARD-inadequate responders, 11 studies compared TNFi to placebo; four studies compared TNFi to placebo and ixekizumab, bimekizumab, tofacitinib, or upadacitinib; and three studies compared TNFi to ixekizumab, secukinumab, and ustekinumab. Two studies compared different TNFi. We found no studies with b/tsDMARD-inadequate responders (b/tsDMARD-IR). No studies compared TNFi to NSAIDs, corticosteroids, or physiotherapy. Performance (32%), detection (56%) and reporting (80%) biases were at high or unclear risk across studies. Only one study had a low risk of bias in all domains.
We limit reporting to the primary comparison, TNFi versus placebo.
DMARD-naïve
We found no studies comparing TNFi with placebo in DMARD-naïve participants.
csDMARD-inadequate responders
TNFi probably result in a large clinical improvement compared to placebo. At 12 weeks, 149/1926 (8%) participants in the placebo group showed a clinical improvement (ACR50) compared to 784/2141 (37%) participants in the TNFi group (risk ratio (RR) 5.63, 95% confidence interval (CI) 3.98 to 7.96; I2 = 65%; 14 studies, 4067 participants; moderate-certainty evidence).
TNFi probably result in a higher proportion of participants in minimal disease activity. At 24 weeks, 95/1017 (9%) participants in the placebo group were in minimal disease activity compared to 428/1336 (32%) participants in the TNFi group (RR 3.76, 95% CI 2.39 to 5.92; I2 = 72%; 5 studies, 2353 participants; moderate-certainty evidence).
At 24 weeks, TNFi may improve function compared to placebo. The mean change in function from baseline (assessed with the Health Assessment Questionnaire; score from 0 to 3, 0 = no disability; minimal clinically important difference (MCID) = 0.35) was -0.14 points with placebo and 0.33 points lower (0.41 lower to 0.25 lower) with TNFi (I2 = 72%; 8 studies, 2949 participants; low-certainty evidence).
TNFi probably result in a clinically important improvement in health-related quality of life. The mean change in quality of life from baseline (assessed with the Short Form 36-item Mental Component Summary questionnaire; score from 0 to 100, 100 = best score; MCID = 1.7) was 2.4 points with placebo and 3.29 points higher (2.18 points higher to 4.40 points higher) with TNFi (I2 = 52%; 8 studies, 2928 participants; moderate-certainty evidence).
TNFi probably slightly reduce radiographic progression. The mean change in radiographic progression (assessed with the Sharp/Van der Heijde-PsA score; scale from 0 to 528, 0 = no damage) was 0.25 points with placebo and 0.37 points lower with TNFi (0.48 lower to 0.25 lower) (I2 = 32%; 7 studies, 2478 participants; moderate-certainty evidence) at 24 weeks.
We downgraded the evidence to moderate certainty for clinical improvement, minimal disease activity, quality of life, and radiographic progression due to risk of bias. For function, we downgraded the evidence to low certainty for risk of bias and imprecision.
TNFi may result in little to no difference in serious adverse events, but may slightly increase withdrawals due to adverse events, compared to placebo. At the end of follow-up: 56/1826 participants (3%) given placebo and 69/1900 (4%) participants given TNFi experienced serious adverse events (RR 1.00, 95% CI 0.70 to 1.42; I2 = 0%; 13 studies, 3866 participants; low-certainty evidence); and 35/1926 (2%) participants given placebo and 65/2140 (3%) given TNFi withdrew due to adverse events (RR 1.53, 95% CI 1.01 to 2.33; I2 = 0%; 14 studies, 4066 participants; low-certainty evidence). We downgraded the evidence to low certainty for risk of bias and imprecision.