Article Text


A systematic literature review of drug therapies for the treatment of psoriatic arthritis: current evidence and meta-analysis informing the EULAR recommendations for the management of psoriatic arthritis
  1. Zoe Ash1,2,
  2. Cécile Gaujoux-Viala3,
  3. Laure Gossec4,
  4. Elizabeth MA Hensor1,2,
  5. Oliver FitzGerald5,
  6. Kevin Winthrop6,
  7. Désirée van der Heijde7,
  8. Paul Emery1,2,
  9. Josef S Smolen8,9,
  10. Helena Marzo-Ortega1,2
  1. 1Section of Musculoskeletal Disease, Leeds Institute of Molecular Medicine, University of Leeds, Leeds, UK
  2. 2Leeds NIHR Musculoskeletal Biomedical Research Unit, Leeds Teaching Hospitals NHS Trust, Leeds, UK
  3. 3Pierre et Marie Curie University-Paris VI, APHP, Rheumatology, Pitié-Salpétrière Hospital, Paris, France
  4. 4Paris Descartes University, Medicine Faculty, APHP, Rheumatology B Department, Cochin Hospital, Paris, France
  5. 5Department of Rheumatology, St Vincent's University Hospital and Conway Institute, University College Dublin, Ireland
  6. 6Department of Infectious Diseases, Public Health and Preventive Medicine, Oregon Health and Science University, Portland, Oregon, USA
  7. 7Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands
  8. 8Division of Rheumatology, Department of Internal Medicine III, Medical University of Vienna, Vienna, Austria
  9. 92nd Department of Medicine, Hietzing Hospital, Vienna, Austria
  1. Correspondence to Helena Marzo-Ortega, Section of Musculoskeletal Disease, Chapel Allerton Hospital, Chapeltown Road, Leeds LS7 4SA, UK; medhmo{at}


Objectives To review the available evidence for the efficacy and safety of non-steroidal anti-inflammatory drugs (NSAIDs), synthetic and biological drug therapies for the different clinical manifestations of psoriatic arthritis (PsA) in order to provide data for the development of treatment recommendations by the European League Against Rheumatism (EULAR) taskforce.

Methods A systematic literature review (SLR) of available treatments for PsA was performed using the largest electronic databases (MEDLINE, EMBASE and COCHRANE) by two working groups formed within the EULAR taskforce. This comprised a comprehensive sample of rheumatologists, dermatologists, epidemiologists and patients. The available evidence was reviewed for NSAIDs, synthetic disease modifying antirheumatic drugs (DMARDs), local and systemic corticosteroids and biologic drugs. All articles and abstracts published between 1962 and January 2010 were reviewed and considered and a meta-analysis of data on biological therapies was performed.

Results While little data are available on NSAIDs, glucocorticoids and synthetic DMARDs, the available evidence suggests an acceptable efficacy and safety profile of both NSAIDs and synthetic DMARDs (methotrexate, cyclosporine A, sulfasalazine and leflunomide) in PsA. More evidence is available (level 1B) supporting the efficacy of anti-tumour necrosis factor (anti-TNF) agents (adalimumab, etanercept, golimumab and infliximab) in treating the signs and symptoms of PsA as well as reducing radiographic progression. Registry data show no new safety concerns, although the numbers studied to date are relatively small.

Conclusions This SLR reveals some evidence to support the use of NSAIDs and synthetic DMARDs and good evidence for the efficacy of anti-TNF therapy in PsA.

View Full Text

Statistics from


Psoriatic arthritis (PsA) remains one of the most challenging clinical entities in rheumatology. One reason for this is the heterogeneity of clinical presentation with diverse articular and dermatological features as well as varied disease course and outcomes. Treatment options have traditionally been derived largely based on experience in rheumatoid arthritis with few studies using specific PsA-derived instruments. This is now changing as an increased interest in clinical trials on the treatment of PsA over the last decade, the era of biologics, led to study designs that have highlighted areas of research need for this disease.

The aim of this systematic literature review (SLR), which is in line with the standardised operating procedures for developing EULAR (European League Against Rheumatism) recommendations,1 was to assess the efficacy and safety of all available non-topical pharmacological treatments for PsA, from non-steroidal anti-inflammatory drugs (NSAIDs) to synthetic disease modifying antirheumatic drugs (DMARDs) and biological drugs, and to inform the development of treatment recommendations.2 This SLR intends to provide a comprehensive update of all the current evidence for the treatment of PsA, focusing on the musculoskeletal features of the disease and incorporating the latest literature since the last review performed by members of the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis.3 4


The EULAR taskforce consisted of 30 rheumatologists, several of whom are also epidemiologists, one dermatologist, one infectious disease specialist, one health professional, two patients with psoriasis and PsA and two rheumatology fellows. Two working subgroups were formed. The first subgroup focused on the role of NSAIDs and synthetic DMARDs and the second subgroup reviewed the literature on glucocorticoids and biological drugs. Systematic literature searches covering the period between 1962 and January 2010 were performed using PubMed MEDLINE, EMBASE and COCHRANE databases to identify all articles reporting the efficacy of pharmacological therapies in PsA. The taskforce was responsible for outlining the scope of the literature search, specifying research questions designed to facilitate the formulation of the guidelines and evaluating the results of the systematic literature search. Details of the search, the selection process and assessment of risk of bias on study selection can be found in sections 1 and 2 of the online supplementary material.

Study eligibility criteria

Inclusion criteria for this review were: (1) systematic reviews and meta-analyses; (2) double-blind randomised controlled trials (RCTs); (3) patients with PsA; (4) studies involving one of the following treatment strategies: NSAIDs, synthetic DMARDs, glucocorticoids or biological agents; (5) studies recruiting at least five patients; (6) publications in English, French, German, Italian or Spanish.

In view of the paucity of controlled trials for NSAIDs and DMARDs, the group decided to include other publications such as observational studies, cohort studies and registries. With regard to biological agents, data on safety and efficacy were extracted from all studies, but only double-blind RCTs were included in the meta-analysis.

Data extraction

Data were extracted from the articles independently by two reviewers from each subgroup. Discrepancies were resolved by discussion. Details on the outcome measures collected and the studies included in the efficacy meta-analysis can be found in section 4 of the online supplementary material.

Data analysis

Where possible (in particular for biologics), efficacy data were meta-analysed using Revman V.5.0. software5 and expressed as RR for dichotomous variables or mean differences for continuous variables. Random effects were assumed throughout and heterogeneity was assessed using the χ2 test and I2 test. Where the heterogeneity test was significant at p<0.1, data were not pooled. Toxicity data from RCTs were meta-analysed using generic inverse variance analysis of the natural log of the rate ratio (of events per 100 patient years, random effects assumed). For synthetic DMARDs and NSAIDs, in each RCT, the magnitude of the treatment effect versus the comparator was analysed through calculation of Cohen's effect size. Due to heterogeneity, no pooling of results was performed.



An initial 1099 potentially relevant articles were screened. Finally, 125 publications with data concerning efficacy and/or safety were analysed. The results are presented below and in table 1, according to the questions agreed on by the taskforce at the first meeting.

Table 1

Effect size of controlled trials comparing DMARDs or NSAIDs with placebo or other DMARDs in patients with PsA

Are NSAIDs efficacious in PsA?

Two RCTs compared an NSAID with placebo in patients with PsA. Nimesulide 200 and 400 mg/day significantly reduced pain, morning stiffness, global assessment, tender and swollen joint score at 4 weeks (n=80 patients).6 A 12-week RCT compared the efficacy of celecoxib 400 mg versus 200 mg versus placebo in active PsA showing that celecoxib (both doses) was efficacious at 2 weeks. However, the difference between active treatment and placebo was no longer significant at week 12 due to high placebo response.7 Comparative studies did not support any difference in efficacy between different NSAIDs.8,,10

Are DMARDs efficacious in PsA? If so which?

Generally, few RCTs were found investigating the use of synthetic DMARDs in PsA. Most of these had small sample sizes, which limits the interpretability of the analyses.


Three RCTs (n=93 patients) comparing methotrexate (MTX) monotherapy versus placebo11,,13 and seven open or retrospective studies of MTX in PsA were analysed14,,20 showing efficacy of MTX for the treatment of peripheral arthritis and psoriasis (table 1). Data on radiographic progression were not conclusive as it was only analysed in a small case–control study.14


Seven RCTs (n=666 patients) reported comparisons of sulfasalazine (SSZ) monotherapy versus placebo21,,26 or symptomatic treatment (NSAIDs, analgesics and/or prednisolone ≤5 mg/day)27 (table 1). In addition, one open trial28 and one case–control study29 were reviewed. SSZ was effective for the treatment of peripheral arthritis and axial manifestations and appeared effective for psoriasis. Two studies reported data on dactylitis with no significant difference found between SSZ and placebo.21 23 One study reported on enthesitis, again with no significant benefit of SSZ over placebo.21 Radiographic progression was not prevented in a small case–control study of 20 SSZ-treated patients and 20 matched controls.29


Two RCTs (n=171 patients) reported comparisons of cyclosporine (CsA) monotherapy versus placebo30 or symptomatic treatment (NSAIDs, analgesics and/or prednisone ≤5 mg/day)27 and four open trials were reviewed (table 1). In a 24-week trial (n=99) assessing CsA monotherapy (3 mg/kg/day) versus symptomatic treatment, CsA appeared effective for the treatment of peripheral arthritis and axial involvement.27 In a 12-month RCT, 72 patients with active PsA and an incomplete response to MTX were randomised to either CsA or placebo in addition to MTX.30 The only significant differences between the groups were in synovitis detected by ultrasound and psoriasis area and severity index (PASI) score in favour of the MTX/CsA group. There was no direct evidence that CsA reduces the progression of radiographic damage.31


One RCT (n=190 patients)32 and two open trials33 34 of leflunomide in PsA have been conducted (table 1). The 24-week RCT comparing leflunomide monotherapy (100 mg/day loading dose for 3 days followed by 20 mg/day orally) versus placebo indicated that leflunomide was clinically effective for the treatment of peripheral arthritis and psoriasis. In an open trial of patients with previous DMARD exposure, 8 of the 12 patients showed at least partial response to leflunomide.33

Gold salts (injectable gold and auranofin)

Three RCTs of gold salts in PsA have been conducted: the first trial compared oral gold (auranofin, 3 mg/day) versus placebo (n=238);35 the second trial compared intramuscular gold (50 mg/week), oral gold (3 mg twice daily) and placebo (n=82)36; the third trial compared sodium thiomalate with oral gold (n=42)37 (table 1). These trials indicated the efficacy of injectable gold on peripheral arthritis. A case–control study assessing radiographic progression in 18 patients with PsA treated with intramuscular gold and 36 controls revealed no statistical difference in disease progression at 24 months.38 Gold salts did not appear to be effective for the treatment of psoriasis, but there was no evidence for skin exacerbation in these studies.

Azathioprine, chloroquine, D penicillamine, fumaric acid and colchicine

Limited data are available on these agents with no conclusive evidence for efficacy. A summary of the studies analysed is available in sections 3 and 4 of the online supplementary material.

What is the relative efficacy of DMARDs?

Four RCTs (see sections 3 and 4 of the online supplementary material) and two retrospective studies comparing different DMARDs in PsA were found. One RCT compared CsA (3–5 mg/kg/day) with MTX (7.5–15 mg weekly) in 35 patients. Both were effective at 6 and 12 months for both PsA and skin psoriasis.39 In a 24-week trial (n=99), CsA monotherapy (3 mg/kg/day) appeared more effective for the treatment of peripheral arthritis and skin27 than symptomatic treatment and also more effective for the treatment of skin and potentially more effective for peripheral arthritis than SSZ. Two 24-week RCTs compared injectable gold and auranofin, both finding greater efficacy with the injectable preparation.36 37

When to start a DMARD?

One small randomised trial in early PsA was found comparing step-up versus combination therapy from baseline (NSAIDs plus MTX) showing no differences in outcome between the groups at 6 months.12

Which DMARD should be started first? Is DMARD switching efficacious? If so is there a preferred order?

There were no data in the literature to answer these questions.

Are there specific safety issues for DMARDs and NSAIDs in PsA?

Safety data on synthetic DMARDs are presented in section 3.2 of the online supplementary material, but there were no safety issues particularly related to the skin disease.


Of 5557 published articles reviewed, 38 met the inclusion criteria and an additional 3 were found by hand searches. No RCTs were available.

Despite concerns over the safety of systemic glucocorticoids in patients with psoriasis, they appear to be widely used. The safety and efficacy data are shown in section 3 of the online supplementary material.

Biological treatments

Of 1974 articles screened, 132 articles and 33 abstracts were included in the final review. These included 11 RCTs used for the efficacy meta-analysis (see section 4 of the online supplementary material) on adalimumab, alefacept, efalizumab, etanercept, golimumab, infliximab and ustekinumab.

In patients with PsA, is treatment with biological agents efficacious compared with placebo or active comparator?

All studied tumour necrosis factor (TNF) inhibitors (adalimumab, etanercept, golimumab and infliximab) showed efficacy at 12–16 weeks for PsA Response Criteria (PsARC) response, American College of Rheumatology (ACR) 20, 50 and 70 response criteria and PASI (see figures 15 and section 4 of the online supplementary material).40,,47 Efalizumab was not superior to placebo for PsARC or ACR 20, 50 and 70 responses and this drug has now been withdrawn due to concerns of an increased risk of progressive multifocal leukoencephalopathy.48 Ustekinumab was superior to placebo in achieving ACR 20 and 50 responses.49

Figure 1

Risk ratios for PsA Response Criteria (PsARC) response at 12–14 weeks comparing the use of a biological disease modifying antirheumatic drug versus placebo in patients with psoriatic arthritis.

Figure 2

Risk ratios for ACR 20 response at 12–16 weeks comparing the use of a biological disease modifying antirheumatic drug versus placebo in patients with psoriatic arthritis. ACR, American College of Rheumatology.

Figure 3

Risk ratios for ACR 70 response at 12–16 weeks comparing the use of a biological disease modifying antirheumatic drug versus placebo in patients with psoriatic arthritis. ACR, American College of Rheumatology.

Figure 4

Risk ratios for PASI 50 response at 12–16 weeks comparing the use of a biological disease modifying antirheumatic drug versus placebo in patients with psoriatic arthritis. PASI, psoriasis area and severity index.

Figure 5

Risk ratios for PASI 50 response at 24 weeks comparing the use of a biological disease modifying antirheumatic drug versus placebo in patients with psoriatic arthritis. PASI, psoriasis area and severity index.

Improvements in Health Assessment Questionnaire were greater with adalimumab,40 41 infliximab,45 46 ustekinumab50 and alefacept51 than placebo at 12 weeks, and with adalimumab,41 etanercept,43 golimumab44 and infliximab45 at 24 weeks, but not with alefacept51 at 24 weeks. Radiographic progression as measured by the modified total Sharp or PsA modified van der Heijde Sharp scoring method was lower for patients treated with all studied TNF inhibitors at 12 and/or 6 months compared with placebo.43 52,,55

Is there a different efficacy for biological agents in PsA subtypes of articular involvement (monoarthritis, oligoarthritis, polyarthritis, axial disease, dactylitis, enthesitis)?

No RCTs reported results separately for the different subtypes of PsA.

With regard to spinal disease, only one observational study56 reported on spinal disease associated with PsA, although final results were pooled for patients with and without spinal disease.

For dactylitis and enthesitis, six RCTs reported data (between 12 and 16 weeks) as secondary end points,40 44,,46 49 57 showing significant benefits with golimumab, infliximab and ustekinumab. Studies used different outcome measures with only a proportion of patients having documented baseline dactylitis or enthesitis. Further details are given in section 5 of the online supplementary material.

Is there efficacy of biological agents on the extra-articular manifestations of PsA (ie, skin and nail involvement)?

Skin disease

Nine RCTs reported data for PASI responses. Treatment with all TNF inhibitors was better than placebo at reducing the PASI score at both 12 and 24 weeks (figures 4 and 5)41,,43 45 46 58 Efficacy was also demonstrated for alefacept, efalizumab and ustekinumab.48 49 51

Nail disease

Only one RCT reported on nail disease. Patients treated with golimumab showed a significantly greater percentage change in Nail Psoriasis Severity Index score at both 14 and 24 weeks.44 Two large observational studies with adalimumab also showed clear improvements in nail disease.59 60

In the treatment of PsA, are there differences in efficacy or safety between the different biological agents (spine, skin, nail etc)?

There is evidence for a lower RR for etanercept for a PASI 75 response at 12 weeks (11.00, 95% CI 0.65 to 186.02) (figures 4 and 5) when compared with the other anti-TNF agents. Efficacy of etanercept is demonstrated at 24 weeks, but with a lower RR than adalimumab, golimumab and infliximab (see section 4 of the online supplementary material).

Other efficacy considerations

Further data are available in section 5 of the online supplementary material regarding response to a biological agent in TNF-blocker inadequate responders, combination therapy with DMARDs, predictors of response to biological agents and changes of dosage/dose interval.

Safety issues

Analyses on safety aspects of biological drug therapies, including infection and malignancy and a meta-analysis of data available from the RCTs, are shown in section 6 of the online supplementary material. Overall, there does not appear to be an increased risk of toxicity specific to patients with PsA, although data may still be insufficient both in terms of numbers of patients and duration of follow-up.


The current analysis provided data for the EULAR taskforce to formulate a new set of recommendations for the treatment of PsA.2

Traditional therapeutic options in PsA have included local or systemic glucocorticoids, NSAIDs and synthetic DMARDs. Our SLR, as others previously,4 61,,63 has demonstrated that the number of studies on these agents is limited, although the available data suggest an efficacy of all of them and they continue to be widely used. A study presented in abstract form subsequent to this literature review being performed did not demonstrate clinical efficacy of MTX in PsA, but the findings were limited by the low dosage used and the high placebo response.64 These results also contradict previous studies showing evidence of efficacy as described above. Nevertheless, there is clearly a need for further research in this respect.

Regarding the efficacy of biological agents, this analysis clearly demonstrates the efficacy of TNF blocking therapies in controlling the signs and symptoms of PsA as well as reducing radiographic progression (Oxford level of evidence 1B) and this was demonstrated in both NSAID and DMARD inadequate responders. Data from RCTs (secondary analyses) also showed efficacy of TNF blocking therapy in reducing dactylitis and enthesitis in patients with predominantly articular disease, although there is variability in the outcomes used and more research needed. This report expands the results of previous SLRs,4 61,,63 adding data regarding the efficacy of golimumab and a meta-analysis. However, the majority of studies have been performed in cases with polyarticular joint involvement, with data remaining sparse on the efficacy of these agents in other disease phenotypes such as those with predominant axial involvement or predominant dactylitis or enthesitis.

With regard to the different TNF blockers, similar efficacy for articular manifestations of PsA has been found. However, there may be a difference in the level of response of skin disease as shown by the lower risk ratios with etanercept at 50 mg weekly, which appears to be a dose effect since the skin response was better at 100 mg weekly.57

Many questions on the efficacy of biologics remain unanswered, including the identification of predictors of response, the possible efficacy of combination therapy with synthetic DMARDs and the implementation of treatment strategies after failure of one biological agent. Data from registries and observational studies suggest that response rates to a second agent may be lower, although the majority of patients will still respond.

Although safety data available from RCTs, registries and long term follow-up studies do not raise any particular new signals, these data are limited in scope and the numbers of patients studied are substantially smaller than those available for rheumatoid arthritis. Interpretation is also made more difficult by the fact that background data regarding infections and malignancy risk are not sufficiently available for PsA. Further PsA disease specific data are needed and registries should continue to gather this. Until then, we must rely on published data from the rheumatoid arthritis and psoriasis literature that suggest an increased risk of some adverse events (including serious infections) in those who use anti-TNF therapies.

Our SLR has some limitations. Ideally, studies should be homogeneous in their design and use appropriate outcome measures to evaluate the efficacy of all current DMARD agents in PsA. However, PsA is a complex disease and this is reflected in the varying entry criteria of the different trials published. Much work has been done recently to improve standardisation of outcome measures to be used in PsA trials,65 66 but many of the studies included in this analysis pre-dated this work. Because of this heterogeneity, we are theoretically limited in our ability to pool risk estimates between studies, so we believe these pooled estimates should be interpreted with caution and comparisons between the agents should not be made.

In summary, this SLR demonstrates the efficacy of NSAIDs, several synthetic DMARDs and TNF blocking therapies in the management of PsA. Given the heterogeneity of PsA, the treatment used should be tailored to the severity of disease in each individual patient. Ongoing monitoring of the safety and effectiveness of biological therapies particularly with the help of registries is vital, as are further trials to answer questions regarding treatment strategies and pharmaco-economic aspects. Importantly, although there are less published data regarding the efficacy of DMARDs for PsA, this is due primarily to the lack of robust clinical trials in this field, and it should not be interpreted as a lack of efficacy. PsA remains a difficult disease to study due its inherent heterogeneity and further work is still needed in many areas.


The authors thank Liz Neilly from the University of Leeds Library for her invaluable advice and help with the literature search.


View Abstract

Supplementary materials


  • ZA and CG-V have contributed equally to the study (joint first authors).

  • See the EULAR recommendations for the management of psoriatic arthritis with pharmacological therapies by Gossec L et al in the January issue (Ann Rheum Dis 2012;71:4–12.)

  • Competing interests ZA has received an unrestricted educational grant from MSD, and speaking fees and conference expenses from Abbott, Chugai, MSD and Pfizer. LG has received speaking fees from Abbott, BMS, Chugai, MSD, Pfizer, Roche, Schering Plough, and UCB. OF receives grant support from Abbott and BMS, is on an international advisory board for UCB and also gives lectures at pharmaceutical sponsored meetings for Abbott, Pfizer, Amgen and UCB. KW has received consulting fees from Genentech and Amgen. DvdH has received consulting fees and/or research grants from Abbott, Amgen, AstraZeneca, BMS, Centocor, Chugai, Eli-Lilly, GSK, Merck, Novartis, Osuka, Pfizer, Roche, Sanofi-Aventis, Schering-Plough, UCB and Wyeth. PE has undertaken clinical trials and provided expert advice for Merck, Pfizer, Abbott, UCB Roche and BMS. JSS has received consulting fees, speaking fees and/or grants from Abbott, BMS, Chugai, MSD, Pfizer, Roche and UCB. HMO has received speaking honoraria and/or consulting fees from Abbott, MSD and Pfizer. CG-V and EMAH have nothing to declare.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.