Aim To determine whether tumour necrosis factor α (TNFα) blockers are more effective than methotrexate in inhibiting the progression of radiographic joint damage in patients with psoriatic arthritis (PsA).
Methods A cohort analysis of patients followed prospectively in a large PsA clinic was conducted. Patients who received a TNFα blocker were compared to those treated with methotrexate. Patients who had records of at least 12 months of treatment with either medication for active peripheral PsA and had radiographic bone erosions were analysed. Radiographs of the hands and feet were performed at baseline, 1–2 years (time 1) and 3–4 years (time 2). Radiographic joint damage was scored according to the modified Steinbrocker score. The outcome of interest was the occurrence of radiographic progression. Multivariate logistic regression analysis using generalised estimating equations for repeated measures was used to compare progression in radiographic joint damage between the two treatment groups.
Results 65 patients treated with TNFα blockers and 70 patients treated with methotrexate were analysed. The proportion of patients who demonstrated progression of radiographic damage score at time 1 and time 2 was higher in the methotrexate group compared to the TNFα blockers group (at time 1: 80% vs 58.9% p=0.005; at time 2: 88% vs 61% p=0.005). In the multivariate regression analysis methotrexate treatment was associated with an increase in radiographic damage compared to TNFα blockers (p=0.001).
Conclusions In a clinic setting, patients with erosive PsA receiving TNFα blockers had a better radiographic outcome compared to those treated with methotrexate.
- Psoriatic Arthritis
Statistics from Altmetric.com
Psoriatic arthritis (PsA) is a chronic inflammatory arthritis that can lead to severe erosions and destruction of joints resulting in physical disability. Radiographic damage among patients with PsA is comparable to that among patients with rheumatoid arthritis.1 The demonstration of elevated levels of tumour necrosis factor α (TNFα) in synovial tissue has provided a rationale for the application of TNFα blockers in PsA.2 TNFα blockers have created a breakthrough in the treatment of PsA. In addition to controlling several aspects of the disease, including peripheral arthritis, spondylitis, psoriasis, enthesitis and dactylitis, all four TNFα blockers (etanercept, infliximab, adalimumab and golimumab) were found to inhibit radiographic joint damage progression better than placebo following 24 weeks of therapy in controlled clinical trials.3–7
Despite the limited evidence-based data supporting its efficacy, methotrexate is widely used as a monotherapy or in combination with TNFα blockers for patients with PsA who have peripheral arthritis.8 Several small placebo controlled trials found that methotrexate improves patient outcomes; however, the level of the evidence is low.9–11 On the other hand, the Methotrexate in Psoriatic Arthritis trial that compared methotrexate to placebo suggested that methotrexate does not have a disease-modifying effect in patients with peripheral PsA.12 Only a single randomised trial compared the efficacy of a TNFα blocker (infliximab) plus methotrexate versus methotrexate alone in PsA patients who were naive to disease-modifying anti rheumatic drugs (DMARD).13 This study showed that higher proportions of the patients achieved American College of Rheumatology 20% improvement (ACR20) and psoriasis area severity index 75 (PASI75) response in the infliximab plus methotrexate treatment group. However, the study did not assess radiographic damage. To the best of our knowledge, no head-to-head comparison of the progression in radiographic joint damage between TNFα blockers and methotrexate has been conducted. The aim of this study was to determine whether TNFα blockers are more effective than methotrexate in inhibiting the progression of radiographic joint damage in patients with PsA in a clinic setting.
A cohort analysis of patients followed prospectively from 1978 to 2010 at the University of Toronto PsA clinic was performed. Patients were evaluated according to a standard protocol at 6–12-month intervals.14 At each visit, symptoms, physical examination (including complete musculoskeletal examination and assessment of psoriasis severity), current use of medications and laboratory findings were recorded. The data were entered and stored in a computerised database. All subjects’ written consent was obtained according to the Declaration of Helsinki. The study has been approved by the University Health Network Ethics Board.
In the present study patients who received a TNFα blocker were compared to those treated with methotrexate.
The TNFα blockers treatment group
The database was searched for patients who had a record of TNFα blocker use. TNFα blockers were prescribed to those patients with high disease activity. Most patients had failed several DMARD. In Ontario, Canada, to be eligible for coverage for a TNFα blocker for peripheral arthritis by the Ontario drug benefit programme, PsA patients are required to have five swollen and tender joints, evidence of erosive disease and to have failed methotrexate and leflunomide (or sulfasalazine if leflunomide is contraindicated). However, private insurance may provide medication with less than five swollen joints, and if a patient has significant axial disease, a TNFα blocker may be provided if they have a Bath ankylosing spondylitis disease activity index score greater than 4 and evidence of sacroiliitis. All patients who had records of at least 12 months of treatment with TNFα blockers for active peripheral arthritis, either as a monotherapy or with concomitant DMARD, were included in the study. Only patients who have had radiographic peri-articular bony erosions at the time of initiation of the medication were included. Patients who initiated the medication before attending the clinic were excluded.
The methotrexate treatment group
The comparison group consisted of patients who initiated treatment with methotrexate either orally or by subcutaneous injections. In order to minimise selection bias that could have led to oversampling patients with milder disease to the methotrexate group, we selected only patients who initiated the medication before the year 2000 (before TNFα blockers were available in Canada). Similar criteria with respect of duration and timing of initiation of the medication and the presence of radiographic erosions, as specified above for the TNFα blockers group, were applied.
Data collection and outcome measures
Data were collected at 0, 1, 2, 3 and 4 years following initiation of the medication. For the statistical analysis, the collected information was grouped into the following time points: time 0, time 1 and time 2. Time 0 was the initiation of the study medication; time 1 was defined as within 2 years of initiation of the study medication and time 2 was 2–4 years following the initiation of the study medication. Radiographs of the hands and feet were performed at 2-year intervals. The radiographs were assessed by at least two rheumatologists (one being the senior author) who specialise in PsA. The radiographic damage was scored after the readers reached a consensus. As this study was performed in a clinic setting, the readers were not blinded to the treatment group or to earlier readings of radiographic damage. Radiographic damage in 42 joints in the hands and feet was assessed according to a modification of the Steinbrocker score (mSS). Each joint was scored on a 0–4 scale, in which 0 is normal, 1 reflects peri-articular osteopaenia or soft tissue swelling, 2 is the presence of an erosion, 3 denotes erosions and joint space narrowing and 4 is total joint destruction.15 The total mSS is the sum of all scores that are equal to or greater than 2. A score of 1 is considered as 0, therefore only definite erosive damage is considered. This method is used to follow radiographic damage in our cohort as it has proved to be valid, reliable, sensitive to change and feasible to perform in a clinic setting.16 The primary outcome of the study was the proportion of patients who had progression of radiographic damage. Progression of radiographic damage was defined as an increase of at least 2 mSS units compared to time 0.
Baseline descriptive statistics were computed with continuous variables summarised by their means and SD and categorical variables summarised by proportions.
The probability of satisfying the radiographic progression criterion was estimated at time 1 and time 2. Multivariate logistic regression analysis using generalised estimating equations for repeated measures was used to compare progression in radiographic joint damage between the two treatment groups incorporating available information from all time points. The regression model was fit with radiographic progression as a binary outcome and with treatment group (TNFα blockers vs methotrexate) as a covariate, adjusting for age, sex, duration of PsA and radiographic damage, actively inflamed and swollen joint counts at initiation of the treatment. The effect of each factor was measured with OR that were estimated along with their respective 95% CI and reported with p values. The statistical analysis was performed using SAS V.9.2 statistical software.
Sixty-five patients treated with TNFα blockers and 70 patients treated with methotrexate who fulfilled the study criteria were analysed. The average follow-up duration was 2.95±0.95 years. A total of 122 patients who had a record of TNFα blockers use were excluded from the analysis for the following reasons: 45 patients had no evidence of bone erosion at baseline; 62 patients had incomplete data or initiated the medication before the first visit to the clinic; 12 patients were lost to follow-up; three patients received TNFα blockers for other indications without having peripheral arthritis. The study population characteristics are presented in table 1. There were no significant differences between the two treatment groups with respect to their age, sex, PsA duration and severity of arthritis as reflected by their active, swollen and damaged joint counts, baseline radiographic damage and by inflammatory markers. Both groups included patients with long-standing disease who had already sustained significant joint damage. The great majority of patients in the TNFα blockers group did not receive previous biological treatment; however, 6.2% of the patients in the TNFα blockers group received a course of these medications that was stopped before the baseline visit (time 0). Sixteen out of 65 (24.6%) of the patients in the TNFα blockers treatment group used concomitant methotrexate at baseline; 70.1% of the patients were treated with etanercept, 12.3% with infliximab, 12.3% with adalimumab and 4.6% with golimumab. A decline in the tender and swollen joint count, health assessment questionnaire score, erythrocyte sedimentation rate and PASI were noted in the TNFα blockers group at time 1 and time 2. A less consistent decline in these items was noted in the methotrexate group. However, a significant difference between the treatment groups in favour of TNFα blockers was found only in the tender joint count (at time 2), erythrocyte sedimentation rate and PASI (table 1).
Comparison of radiographic progression between the treatment groups
Cumulative probability plots of radiographic progression are presented in figures 1 and 2. The proportion of patients who had progression of radiographic damage score at time 1 and time 2 compared to time 0 was higher in the methotrexate group compared to the TNFα blockers group. At time 1, 34 out of 50 patients (68%) in the methotrexate group developed a new erosion in at least one joint and 40 out of 50 patients (80%) had radiographic progression compared to 22 out of 39 patients (56.4%) with a new eroded joint (p=0.03) and 23 out of 39 patients (58.9%) with radiographic progression in the TNFα blockers group (p=0.005). At time 2, 42 out of 50 patients (84%) in the methotrexate group developed a new erosion in at least one joint and 44 out of 50 (88%) had radiographic progression versus 27 out of 36 patients (75%) who developed a new eroded joint (p=0.005) and 22 out of 36 patients (61%) who progressed radiographically in the TNFα blockers group (p=0.005). In multivariate regression analysis methotrexate treatment was associated with an increase in radiographic damage compared to TNFα blockers therapy (OR 4.45, 95% CI 1.81 to 10.94, p=0.001). A higher radiographic damage score at time 0 was also independently associated with radiographic progression (OR 1.03, 95% CI 1.01 to 1.05, p=0.006) (table 2).
In this cohort study we compared the effect of treatment with TNFα blockers versus methotrexate on radiographic damage progression in patients with PsA. We have found that in a clinic setting, patients with erosive PsA receiving TNFα blockers had a better radiographic outcome compared to those treated with methotrexate.
It is widely accepted that TNFα blockers can effectively inhibit radiographic joint damage progression.8 However, it is unknown whether methotrexate can also slow radiographic progression in PsA, because in the clinical trials (except for RESPOND) patients were included if they failed their current DMARD including methotrexate. Furthermore, the comparative effectiveness of these medications in reducing radiographic damage progression has not been evaluated in PsA patients. It was recently suggested that methotrexate does not have a disease-modifying effect, based on a lack of reduction in inflammatory markers and only a limited control of symptoms compared to placebo in patients with PsA.12 No radiographic outcome was assessed in that study. On the other hand, Chandran et al17 observed a reduction in radiographic progression in patients with PsA who were using higher doses of methotrexate compared to historical controls using lower-dose or later-initiated methotrexate, supporting the notion that methotrexate may modulate inflammation and prevent damage. Our study results support the superiority of TNFα blockers over methotrexate treatment even though patients in the TNFα blockers group probably had more resistant arthritis as the great majority of them had to fail two DMARD in order to qualify for TNFα blockers under the local guidelines. However, it should be noted that data on patients on methotrexate were obtained at a time when TNFα blockers were not available.
We provide information about the long-term radiographic outcome of patients following 4 years of treatment with methotrexate and TNFα blockers. In clinical trials only 9–21.2% of the patients who were treated with TNFα blockers showed radiographic progression at week 24.4 ,5 ,7 ,18 In the open-label extension period of the Adalimumab Effectiveness in Psoriatic Arthritis (ADEPT) trial, only 20% of the patients in the adalimumab group were found to have radiographic damage at 144 weeks.4 In our study, the probability of radiographic progression at 1–4 years time was significantly higher. Several reasons may account for this difference: first, the baseline radiographs were performed within a year before the initiation of the medication rather than at the time of initiation. Therefore, it is possible that some of the damage had accrued before the initiation of the study medication. This is a limitation of our study that stems from its observational nature. However, it does not explain the difference between the two treatment groups. Second, only patients who had had radiographic damage at baseline were selected for this study resulting in an overrepresentation of patients with more severe and destructive arthritis, which accounts for the higher frequency of radiographic progressors compared to clinical trials. We and others have shown that the presence of erosions is predictive of poor outcome in PsA.19 ,20 Other patient characteristics did not differ significantly from those of the patients who were included in clinical trials and represent moderate to severe, long-standing PsA. However, while the average duration of PsA in our cohort was 13 years, the average duration in the ADEPT study was 10 years. Finally, in general it is likely that the adherence to the study medication is better in clinical trials compared to that in a real-life setting such as in the present study. This may also explain the high frequency of progressors in our study, although we cannot prove this claim.
Limitations of the study include the lack of blinding to clinical data, the timing of radiographs by the assessors and the limited generalisability due to the inclusion of only patients with erosive disease, although this would not have an effect on the analysis. Another potential limitation is the use of historical controls. This may have increased the difference between the two groups because lower doses of methotrexate were used earlier in our cohort.17 We have found in our cohort that among methotrexate users over several decades there was a trend for methotrexate to be prescribed earlier and at higher doses with greater clinical improvement and less radiographic damage.17 In this study the mean maximum doses of methotrexate were moderate to high averaging 19.2 mg per week, and despite this there was a better beneficial effect of anti-TNF inhibitors with regard to radiographic progression. The use of historical controls in our study provided the opportunity to evaluate the long-term effect of methotrexate. We specifically avoided using patients treated with methotrexate alone during the same period as the anti-TNF agents because those may not have been as severe to start with.
In summary, the results of the present study reveal that in a clinic setting, patients with erosive PsA receiving TNFα blockers had a better radiographic outcome compared to those treated with methotrexate alone, supporting an early use of the former group. Randomised head-to-head trials comparing methotrexate to TNFα blockers in DMARD-naive PsA patients are needed to provide better quality information about the comparative short and long-term efficacy of the two classes of medication.
Handling editor Tore K Kvien
Contributors 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 published. All authors were likewise involved in the study conception and design, acquisition of data as well as analysis and interpretation of data. DDG 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.
Funding The Psoriatic Arthritis Program is funded in part by the Arthritis Society, Canadian Institutes of Health Research and the Krembil Foundation.
Competing interests None.
Ethics approval The study was approved by the University Health Network Ethics Board.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
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.