Leflunomide in psoriatic arthritis: Results from a large European prospective observational study


  • Frank Behrens,

    Corresponding author
    1. Johann Wolfgang Goethe-University, Frankfurt, Germany
    • CIRI/Division of Rheumatology, Johann Wolfgang Goethe-University, Theodor-Stern-Kai 7, D 60590 Frankfurt Main, Germany
    Search for more papers by this author
    • Dr. Behrens has received speaking fees (less than $10,000) from Sanofi-Aventis.

  • Christoph Finkenwirth,

    1. Johann Wolfgang Goethe-University, Frankfurt, Germany
    Search for more papers by this author
  • Karel Pavelka,

    1. Charles University Prague, Prague, Czech Republic
    Search for more papers by this author
    • Dr. Pavelka has received consultant fees, speaking fees, and/or honoraria (less than $10,000 each) from Pfizer, Abbott, Roche, MSD, and Bristol-Myers Squibb.

  • Jiří S̆tolfa,

    1. Charles University Prague, Prague, Czech Republic
    Search for more papers by this author
  • Alenka S̆ipek-Dolnicar,

    1. Ljubljana University Medical Center, Ljubljana, Slovenia
    Search for more papers by this author
  • Diamant Thaçi,

    1. Johann Wolfgang Goethe-University, Frankfurt, Germany
    Search for more papers by this author
    • Dr. Thaçi has received consultant fees, speaking fees, and/or honoraria (less than $10,000 each) from Novartis, Janssen-Cilag, and Astellas, and (more than $10,000 each) from Abbott, Pfizer, and MSD.

  • Harald Burkhardt

    1. Johann Wolfgang Goethe-University, Frankfurt, Germany
    Search for more papers by this author
    • Dr. Burkhardt has received speaking fees (less than $10,000) from Sanofi-Aventis.



To determine the “real-world” clinical effectiveness and safety of leflunomide in patients with psoriatic arthritis (PsA).


This prospective, multinational 24-week observational study involved adult patients with active PsA who initiated treatment with leflunomide. Patients were evaluated at baseline, 12 weeks, and 24 weeks. The primary outcome was response as assessed by the Psoriatic Arthritis Response Criteria (PsARC) in patients with pre- and posttreatment data. A modified PsARC response analysis included patients with joint counts, but no severity scores. Other effectiveness evaluations included global assessments, fatigue, pain, skin disease, dactylitis, and nail lesions. All patients were evaluated for safety.


A total of 514 patients were enrolled in this study (mean age 50.7 years, mean disease duration 6.1 years). In the primary effectiveness analysis, 380 (86.4%) of 440 patients (95% confidence interval 82.8%–89.4%) achieved a PsARC response at 24 weeks. Significant improvements were observed in tender and swollen joint scores and counts, patient and physician global assessments, fatigue, pain, skin disease, dactylitis, and nail lesions. The discontinuation rate was 12.3%. Ninety-eight adverse drug reactions occurred in 62 (12.1%) patients; 3 drug reactions were serious (2 increased liver enzymes, 1 hypertensive crisis).


Leflunomide is an effective and well-tolerated option for PsA in daily clinical practice, with beneficial effects on peripheral arthritis and on other PsA manifestations, including pain, fatigue, dactylitis, and skin disease.


Psoriatic arthritis (PsA) is a chronic inflammatory joint disease that affects up to 30% of patients with psoriasis and has an overall prevalence of approximately 0.2% in Western Europe (1–4). Many patients with PsA remain undiagnosed (2), so the true prevalence of this condition may be much higher. Magnetic resonance imaging indicates that in early stages of PsA (disease duration ≤1 year), disease activity is as high as it is in early rheumatoid arthritis, as assessed by synovitis, bone erosions, subchondral bone edema, and tenosynovitis (5). In patients with established PsA, joint involvement is generally less severe than in rheumatoid arthritis. However, quality of life scores are reduced to a similar degree, perhaps due to the additional impact of skin disease on pain and emotional well-being (6–8).

Options for the treatment of PsA include conventional disease-modifying antirheumatic drugs (DMARDs) and tumor necrosis factor α (TNFα) inhibitors. Leflunomide, a de novo pyrimidine biosynthesis inhibitor with immunomodulatory properties, is an established DMARD that is approved by the European Medicines Agency for the treatment of rheumatoid arthritis and PsA (9). The efficacy and safety of leflunomide in patients with PsA was demonstrated by Kaltwasser et al in a randomized, placebo-controlled clinical trial (10), and its ability to improve psoriasis symptoms was further documented by Nash et al (11). However, drug effectiveness in daily practice does not always mirror the efficacy established in randomized clinical trials. Prospective observational studies supplement clinical trial data by providing information on drug effectiveness and tolerability in a diverse patient population under routine care conditions in daily practice.

Here we present data from the Observational Study of Psoriatic Arthritis Treated with Leflunomide (OSPAL), a postauthorization study that assessed the effectiveness and safety of leflunomide in adult patients with active PsA. Data were collected in Germany, the Czech Republic, and in Slovenia.

Significance & Innovations

  • In a prospective observational study, 86% of leflunomide-treated psoriatic arthritis (PsA) patients achieved a Psoriatic Arthritis Response Criteria response.

  • Leflunomide had beneficial effects on peripheral arthritis symptoms, pain, fatigue, dactylitis, and skin disease.

  • Leflunomide is an effective and well-tolerated option for PsA in daily clinical practice.


Study design and population.

We conducted a prospective 24-week observational study (OSPAL) involving adult patients at 161 centers in 3 countries (105 centers in Germany, 50 centers in the Czech Republic, and 6 centers in Slovenia). As this was a noninterventional, observational study with pseudonymized data collection, ethics approval and patient consent were not required by German and Czech Republic drug law at the time of study initiation. In Slovenia, a positive ethics committee vote was obtained. Patients who started leflunomide therapy at the clinician's decision and according to daily clinical practice were observed between 2004 and 2006. Patient evaluations were performed prior to initiation of leflunomide therapy, at 12 weeks after the start of therapy, and at the 24-week final visit. In each center, patients were examined by the same physician throughout the study. A standardization of examiners was not performed prior to the study.

Patients ages ≥18 years with a diagnosis of PsA, active disease, and no previous leflunomide treatment were eligible for this study. Diagnosis was made by the treating internist/rheumatologist according to the Moll and Wright criteria for PsA. Patients without active joint involvement, with rheumatoid nodules, a positive test result for rheumatoid factor, or who were pregnant or likely to become pregnant were excluded. Concomitant medications were allowed during the study and changes were documented during the study period.

Study end points.

The primary end point was response to treatment at 24 weeks as assessed by the Psoriatic Arthritis Response Criteria (PsARC) (12) in patients for whom appropriate data were available at baseline (before initiation of leflunomide treatment) and at the 24-week final visit. The PsARC is a composite measure that includes joint pain/tenderness scores, joint swelling scores, and global assessments of PsA activity by the patient and the physician. A PsARC response requires improvement in 2 of the 4 assessments, at least 1 of which must be a joint score, and worsening in none. Seventy-six joints were evaluated for joint pain/tenderness, and 74 joints were evaluated for swelling. Joint scores were assessed on a 4-point scale ranging from 0 (not present) to 3 (severe), with improvement/worsening defined as a change of ≥30% in the sum of the scores. Global status was assessed on a 5-point Likert scale ranging from 0 (very good) to 4 (very severe), with improvement/worsening defined as a change of ≥1 unit. A more complete description of this instrument and the precise joints evaluated can be found in the study by Kaltwasser et al (10). In some patients, joint counts were performed but the severity of involvement was not assessed. These patients were included in a modified analysis of PsARC response, with each marked joint given a score of 1, indicating at least mild involvement; the definition of improvement/worsening remained unchanged (≥30%). Other evaluations of effectiveness included fatigue, pain, skin status, and nail lesions on 5-point Likert scales, and dactylitis on a 4-point Likert scale.

All patients were evaluated for safety. Laboratory values, vital signs, and reports of adverse events were obtained at each visit. An adverse drug reaction (ADR) was defined as an adverse event considered by the investigator to be possibly related to leflunomide treatment.

Statistical analyses.

Categorical data were analyzed by frequency tables and adjusted relative frequencies were calculated. Continuous data were described by mean ± SD values. Wilcoxon's signed rank tests were used to analyze change between start of therapy and the final visit. The primary end point analysis was the PsARC response in patients who had pre- and posttreatment data in which joints had been scored for severity of pain/tenderness and swelling. The response rate was calculated with exact 95% confidence limits for binomial proportion using the F distribution method (13).


Study population.

A total of 514 patients received leflunomide during the course of this observational study (330 patients in Germany, 159 patients in the Czech Republic, and 25 patients in Slovenia) and 511 were evaluated for effectiveness; the remaining 3 patients did not meet entry criteria (1 did not have tender or swollen joints [spondylitis only] and 2 tested seropositive for rheumatoid factor, with 1 also having rheumatoid nodules). Patients were generally treated with a leflunomide loading dosage of 100 mg/day for 3 days followed by 20 mg/day orally. Most patients (89.2%) were between ages 18 and 65 years and had received previous DMARD treatment (83.5%). Patient characteristics are presented in Table 1.

Table 1. Baseline demographic and disease variables in the study population (n = 514)*
  • *

    Categorical values are expressed as adjusted relative frequencies following adjustment for missing values. Some patients had taken multiple previous disease-modifying antirheumatic drugs (DMARDs). PsA = psoriatic arthritis.

Female, %52.2
Age, mean ± SD years50.7 ± 11.4
Duration of PsA, mean ± SD years6.1 ± 6.8
Duration of psoriasis, mean ± SD years13.9 ± 10.8
Axial disease, %25.9
HLA–B27 positivity, %25.5
Dactylitis, %61.7
Nail lesions, %65.5
C-reactive protein level, mean ± SD mg/dl25.4 ± 33.6
Previous DMARDs, mean ± SD no.1.4 ± 1.0
Previous DMARD treatment, no. of patients (%)429 (83.5)
 Methotrexate390 (75.9)
 Sulfasalazine207 (40.3)
 Cyclosporine48 (9.3)
 Hydroxychloroquine/chloroquine35 (6.8)
 Azathioprine12 (2.3)
 Biologic agents6 (1.2)

At the start of leflunomide therapy, the majority (62.1%) of patients received concomitant medication for PsA, primarily antiinflammatory/antirheumatic drugs (29.4%), systemic corticosteroids (22.9%), and DMARDs (22.4%), predominately methotrexate. Six (1.2%) patients were parallel treated with biologic drugs. Concomitant medications for PsA were unchanged during the study in 75.9% of the patients. At the final visit, 28.1% of the patients received antiinflammatory/antirheumatic drugs and 21.2% received corticosteroids as concomitant medication.

During the 24-week study, 63 (12.3%) of 514 patients discontinued treatment for the following reasons (some patients gave multiple reasons): adverse event(s), 32 patients; lack of efficacy, 21 patients; patient's request, 13 patients; lost to followup, 5 patients; noncompliance, 3 patients; and missing information, 1 patient.

PsARC response to leflunomide therapy.

Data on pre- and posttreatment severity of joint involvement were available for 440 of the 511 patients evaluable for effectiveness. Response to treatment, as assessed by PsARC, was observed in 380 (86.4%) of these patients (95% confidence interval [95% CI] 82.8%–89.4%) at 24 weeks (Figure 1). Response rates in the individual countries were 84.8% (223 responders) for Germany, 89.7% (139 responders) for the Czech Republic, and 81.8% (18 responders) for Slovenia. The likelihood of achieving a response was not influenced by sex, age, spondylitis, or previous DMARD therapy. In 340 (80.0%) of 425 patients, a PsARC response was achieved by 12 weeks. The modified PsARC response analysis, which included patients whose joints had been marked but not scored for severity, also demonstrated a high rate of response to leflunomide treatment. Of 492 patients in this analysis, 382 (77.6%; 95% CI 73.7%–81.3%) achieved a modified PsARC response at 24 weeks (Figure 1).

Figure 1.

Response to treatment as assessed by the Psoriatic Arthritis Response Criteria (PsARC) at week 24.

Additional assessments of effectiveness.

Tender and swollen joint scores, as assessed on a 4-point scale (0 = not present to 3 = severe), decreased significantly in response to leflunomide treatment (P < 0.0001 by Wilcoxon's signed rank test) (Figure 2). Mean tender and swollen joint counts also decreased significantly (18.5 to 8.9 and 12.9 to 5.4, respectively; P < 0.0001) during 24 weeks of leflunomide treatment.

Figure 2.

Tender and swollen joint scores over time in psoriatic arthritis patients treated with leflunomide. Seventy-six joints were assessed for tenderness and 74 joints for swelling. All assessments utilized a 4-point scale ranging from 0 (not present) to 3 (severe). * = (SD).

Physicians and patients were asked to rate the patient's PsA-related general condition on a 5-point scale (very poor, poor, fair, good, and very good). In both assessments, the global status of patients significantly improved in response to leflunomide treatment (P < 0.001) (Table 2). The proportion of patients who were rated as good or very good by physicians increased from 11.4% at baseline to 77.9% at 24 weeks. Similarly, the proportion of patients who considered their condition to be good or very good increased from 6.5% at baseline to 70.7% at the final visit.

Table 2. Proportion of patients experiencing an improvement of ≥1 category from baseline to final visit*
Disease parameterNo.Improvement, %
  • *

    P < 0.001 for all pre- to posttreatment comparisons. PsA = psoriatic arthritis.

  • Disease parameters were assessed on a 5-point Likert scale except for dactylitis, which was assessed on a 4-point scale. All patients with values for a given parameter at the start of therapy and at the final visit were included in the analysis.

Physician assessments  
 General PsA-related condition47380.8
 Nail lesions46632.0
Patient assessments  
 General PsA-related condition47284.5

Significant improvements were observed in additional manifestations of PsA, including fatigue, pain, skin disease, dactylitis, and nail lesions (Figure 3 and Table 2). The lower proportion of patients with improvements in dactylitis and nail lesions is likely due to the fact that approximately one-third of the patients did not have these conditions at baseline (Table 1), and therefore fewer patients had the potential for improvement. The proportion of patients with “no change” was 46.7% and 61.6% for dactylitis and nail lesions, respectively.

Figure 3.

Change in percentage of patients categorized from none to very severe dactylitis, psoriasis, pain, and fatigue from baseline to week 24.

Leflunomide treatment led to a decrease of the mean ± SD C-reactive protein levels from 25.38 ± 33.62 mg/dl at baseline to 11.48 ± 17.98 mg/dl after 24 weeks of treatment, while the mean daily steroid dose was reduced from 7.2 mg to 6.4 mg after 24 weeks of treatment (data from an analysis of the German subgroup).


A total of 107 adverse events were documented in 65 (12.6%) patients. Ninety-eight of the adverse events in 62 (12.1%) patients were judged to be an ADR (possibly related to leflunomide treatment). ADRs were slightly more common in patients between ages 18 and 65 years than in those ages ≥66 years (12.4% versus 9.1%), in patients with no previous DMARD therapy compared to those who had received DMARD treatment (15.3% versus 11.4%), and in women compared to men (13.5% versus 10.7%). There is evidence that adding leflunomide to concomitant DMARD therapy did not lead to an increase in adverse events (Supplementary Table 1, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21848/abstract).

The most frequent ADRs were diarrhea (16.3% of all ADRs), alopecia (9.2%), hypertension (8.2%), and pruritus (5.1%). Three ADRs were serious, affecting 2 patients (increased alanine aminotransferase levels in one and hypertensive crisis plus increased transaminase levels in the other). There were 8 unexpected ADRs: hyperhidrosis, tremor, nocturia, weight increase, abdominal distension, arrhythmia, dyspepsia, and apathy. None of the unexpected ADRs was considered as serious by the investigator. There were no deaths in the study.

The majority (60 of 98, 61.2%) of ADRs had resolved by the end of the study. One (1.0%) resolved with sequelae, 24 (24.5%) were ongoing, and the outcome was unknown for 13 (13.3%). Two of the 3 serious ADRs were resolved at the end of the study; the status of the third ADR was unknown.


Studies of PsA are frequently hampered by small numbers and mixed populations that include patients with other types of rheumatic diseases. Our study utilized data from 514 patients, one of the largest cohorts of PsA patients treated with conventional DMARDs reported to date. The clinical features of these patients, such as the high proportion with dactylitis or nail involvement, were typical of a true PsA cohort. Patients who were seropositive for rheumatoid factor were excluded from the study. We therefore have a high level of confidence that this cohort is representative of the PsA population as a whole.

The efficacy of leflunomide in patients with PsA had been previously demonstrated in a randomized, placebo-controlled clinical trial (10). However, such trials do not always provide insights into the challenges faced when drugs are used in everyday clinical practice. Compared to controlled trials, clinical practice typically involves a more diverse patient population and a wide variety of comorbid conditions and concomitant medications. Observational studies provide a more accurate view of the effectiveness and safety profile of a drug under “real world” conditions (14), and are therefore important adjuncts to randomized controlled trials.

In this prospective observational study involving 161 clinical centers in Germany (105 centers), the Czech Republic (50 centers), and Slovenia (6 centers), we evaluated the effectiveness and safety of leflunomide in adult patients with PsA. Leflunomide treatment resulted in a PsARC response in 80% of patients at 12 weeks and 86% at 24 weeks in the primary analysis (n = 440 patients) (Figure 1). It is of particular note that leflunomide was effective in a patient population consisting primarily of those failing DMARD therapy with longstanding disease, and that the response rate was similar regardless of whether patients had received previous DMARD therapy. Most PsA patients receive first-line therapy with methotrexate or sulfasalazine (15), and it is thus critically important that DMARDs not typically used as first-line agents demonstrate effectiveness in patients who have failed DMARD therapy. A recent double-blind randomized controlled trial (Methotrexate in Psoriatic Arthritis) did not meet the primary outcome for effectiveness of a methotrexate treatment compared to placebo (16), so that robust evidence for efficacy of methotrexate in active PsA is still missing. In this context, the present large study, performed under routine care conditions, confirmed the effectiveness of leflunomide, as earlier demonstrated in a randomized controlled trial (Treatment of Psoriatic Arthritis Study), and might thus impact future treatment recommendations for PsA.

Some of the clinicians in this study marked involved joints but did not score the severity of involvement. To allow the inclusion of these patients in the evaluation of response, joints that were marked but not scored were assigned a score of 1. These patients thus needed a ≥30% reduction in the number of involved joints to meet the criteria for improvement in a joint score. As might be expected, the PsARC response rate determined by this more comprehensive analysis (n = 492) was slightly lower (78%), but still quite favorable.

The PsARC response rate observed in our study is similar to the 85% PsARC response rate reported for etanercept in a 26-week observational study of 20 DMARD-refractory patients with PsA (17). Our data also correspond well to those from a recent observational study reported by Heiberg et al of PsA patients treated with methotrexate or TNFα inhibitors (18). Although PsARC response rates were not reported for this study, mean swollen joint counts improved by 43% with either methotrexate or TNFα inhibitors (18). In comparison, we observed an improvement of 58% in swollen joint counts. An interesting finding from the study by Heiberg et al was that methotrexate and TNFα inhibitors had comparable effects on swollen and tender joint counts over 6 months in patients with PsA, despite the fact that TNFα inhibitors were associated with significantly greater improvements in other measures, including fatigue, pain, and Disease Activity Score in 28 joints assessments. This finding may reflect the significantly greater disease activity at baseline in patients treated with TNFα inhibitors, or it may suggest that DMARDs and biologic therapies share an upper limit in their beneficial effects on joints in patients with PsA (18).

In addition to mediating significant improvements in measures of peripheral arthritis, leflunomide treatment also resulted in significant improvements in PsA-related global status and other manifestations of PsA, including fatigue, pain, skin disease, dactylitis, and nail lesions (Figure 3 and Table 2). Although not always assessed in studies of PsA, pain and skin disease were recently included in the 6 core domains considered essential to evaluations of PsA by an international consensus committee, and fatigue, dactylitis, and nail disease were listed as other important domains to be considered (19). There is evidence that some biologic agents reduce fatigue in patients with PsA (18, 20), but this is only the second study, to our knowledge, to address the impact of conventional DMARDs on fatigue. In the other study, methotrexate treatment resulted in negligible reductions in fatigue at 24 weeks compared to baseline (18). In contrast, we found that the majority of patients experienced improvements in fatigue of at least 1 category on a 5-point scale. Significant improvements were also observed in dactylitis in response to leflunomide therapy. Dactylitis is a common manifestation of PsA, occurring in approximately half of patients, and is associated with erosive joint damage in the affected joints (21). In a recent review of the impact of PsA therapies on dactylitis, very few of the commonly used therapies had been evaluated with respect to dactylitis, and only infliximab mediated significant improvements (22).

During this study, change in axial disease was not part of the efficacy assessments due to the fact that DMARDs have demonstrated ineffectiveness in axial disease of other spondyloarthritides. But existing axial symptoms of the patient did not influence the primary outcome of the study. The percentage of patients achieving PsARC response was not significantly different from those with or without axial PsA.

In observational studies, the use of concomitant therapy is allowed and could be a potential confounding factor. The stable percentage of patients with concomitant antirheumatic drugs and a stable dose of those therapeutics during the trial period reflects and strongly suggests that the decrease in disease activity was due to the initiation of leflunomide and not to an increased use of concomitant therapy.

The favorable tolerability profile of leflunomide in daily practice was confirmed in this study. The 12.3% discontinuation rate reported here compares favorably with the 17.1% and 18.6% rates recently reported for methotrexate and TNFα inhibitors, respectively, in the study by Heiberg et al (18). Common adverse events and ADRs, including diarrhea, alopecia, hypertension, and pruritus, were expected from the known safety profile of leflunomide therapy alone (9) and in combination with other DMARDs (such as methotrexate) (21), as were the 3 serious ADRs (2 elevations in liver enzymes and 1 hypertensive crisis). Careful monitoring of liver enzymes during therapy may help limit hepatotoxicity. Elevated liver enzymes have also been associated with methotrexate, sulfasalazine, and adalimumab (22). Other common ADRs in patients with PsA include gastrointestinal side effects for methotrexate and sulfasalazine, renal toxicity for cyclosporine, and injection site reactions for infliximab (reviewed by Soriano and McHugh) (23). In our study, no increased numbers of adverse events were seen in the leflunomide add-on to ongoing methotrexate group compared to the leflunomide monotherapy (Supplementary Table 1, available in the online version of this article at http://onlinelibrary.wiley.com/doi/10.1002/acr.21848/abstract). Also, no significant differences in discontinuation rates in patients with add-on to an existing DMARD therapy were seen. Clinically relevant increases in liver enzymes were part of the adverse events reporting only. A systemic collection of exact laboratory values was not part of the study protocol. Therefore, with the available data, we cannot rule out that a combination therapy will lead to higher levels of liver enzymes in comparison to monotherapy.

On the basis of these data, we conclude that leflunomide is effective and has a favorable safety profile in the therapy of PsA in daily clinical practice. Its ability to alleviate joint, skin, and other disease manifestations, and its manageable safety profile and relatively low price make leflunomide a valuable option for the treatment of PsA.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Behrens had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Behrens.

Acquisition of data. Behrens, Pavelka, S̆tolfa, S̆ipek-Dolnicar.

Analysis and interpretation of data. Behrens, Finkenwirth, Thaçi, Burkhardt.


The study design was created by Sanofi-Aventis together with European Investigators. Sanofi-Aventis played no role in the interpretation of data analysis. The content of the manuscript is independent of Sanofi-Aventis. Sanofi-Aventis was allowed to read the manuscript before submission.


The authors would like to thank the investigators, patients, nurses, and support staff who contributed to the OSPAL study, and Sharon L. Cross, PhD, for writing and editorial assistance.