Article Text


Extended report
The role of methotrexate co-medication in TNF-inhibitor treatment in patients with psoriatic arthritis: results from 440 patients included in the NOR-DMARD study
  1. Karen Minde Fagerli1,
  2. Elisabeth Lie1,
  3. Désirée van der Heijde1,2,
  4. Marte Schrumpf Heiberg1,
  5. Åse Stavland Lexberg3,
  6. Eric Rødevand4,
  7. Synøve Kalstad5,
  8. Knut Mikkelsen6,
  9. Tore K Kvien1
  1. 1Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway
  2. 2Department of Rheumatology, Leiden University Medical Center, Leiden, Netherlands
  3. 3Department of Rheumatology, Vestre Viken, Drammen Hospital, Norway
  4. 4Department of Rheumatology Trondheim, St Olavs Hospital, Norway
  5. 5Department of Rheumatology, University Hospital of Northern Norway, Tromsø, Norway
  6. 6Lillehammer Hospital for Rheumatic Diseases, Lillehammer, Norway
  1. Correspondence to Dr Karen Minde Fagerli, Department of Rheumatology, Diakonhjemmet Hospital, Pb 23, Vinderen, Oslo 0319, Norway; karen.fagerli{at}


Background The role of co-medication with tumour necrosis factor inhibitors (TNFi) is well established in rheumatoid arthritis and ankylosing spondylitis. In psoriatic arthritis (PsA) there is little evidence available on this issue.

Material and methods The analyses were based on data from the Norwegian longitudinal observational study on disease-modifying antirheumatic drugs (NOR-DMARD). Patients with PsA starting their first TNFi, either as monotherapy or with concomitant methotrexate (MTX), were selected. Baseline characteristics, responses after 3, 6 and 12 months, and drug survival were compared between those with and without MTX co-medication. A secondary analysis was performed on patients who had confirmed swollen joints at baseline. Cox regression was used to identify predictors of discontinuation.

Results We included 440 patients, 170 receiving TNFi as monotherapy and 270 receiving concomitant MTX. The groups had similar baseline characteristics, except for number of swollen joints, which was higher in the concomitant MTX group. Responses were similar in the two groups in both analyses. Drug survival analyses revealed a borderline significant difference in favour of patients receiving co-medication (p=0.07), and this was most prominent for patients receiving infliximab (IFX) (p=0.01). In the Cox regression analysis lack of concomitant MTX and current smoking were independent predictors of discontinuation of TNFi.

Conclusions We found similar responses to TNFi in patients with and without concomitant MTX, but drug survival was superior in patients receiving co-medication. The effect of MTX on drug survival was most prominent in patients receiving IFX. Smoking at baseline and use of TNFi as monotherapy were identified as independent predictors of drug discontinuation.

  • TNF-alpha
  • Methotrexate
  • Psoriatic Arthritis

Statistics from


TNF inhibitors (TNFi) have been proven efficacious in psoriatic arthritis (PsA) in reducing signs and symptoms of disease and also inhibiting structural joint damage.1–5 In rheumatoid arthritis (RA), it is well established that co-medication with methotrexate (MTX) improves efficacy of TNFi.6 ,7 Proposed mechanisms of this are synergistic pharmacological effects and immunogenic effects.8 Less information on this subject is available in ankylosing spondylitis (AS), but it is commonly accepted that additional MTX does not increase efficacy.9 This issue has been investigated in a randomised controlled trial (RCT) on infliximab (IFX) in AS without there being any additional benefit.10 However, a recent observational study showed that patients with spondyloarthritis taking concomitant MTX were less likely to develop antibodies to IFX and that antibody formation was associated with poorer clinical response and discontinuation of treatment.11 Some investigators have proposed that patients receiving TNFi should have additional immunosuppressive therapy irrespective of diagnosis to reduce the incidence of formation of antibodies to TNFi.8

The standard design for the major TNFi RCTs for patients with PsA has been that the patients were allowed to continue MTX if they were already on it and no trials have indicated any difference in response to TNFi in subgroups with and without concomitant MTX.1 ,4 Concomitant MTX has been associated with increased drug survival in register studies.12 ,13 However, the role of MTX co-medication in PsA is still unclear, which was also highlighted by the most recent recommendations, and the background literature review, on treatment of PsA.14 ,15

The aim of this study was to further investigate the effect of concomitant MTX on responses and drug survival in patients with PsA starting their first TNFi.

Material and methods

Data for these analyses are from the Norwegian disease modifying anti-rheumatic drug (NOR-DMARD) study, a Norwegian five-centre, longitudinal observational study. The study was started in 2000, and patients with inflammatory arthropathies were included when starting a new DMARD regimen. Patients were assessed at baseline, after 3, 6 and 12 months and yearly thereafter. The assessments include demographics, disease characteristics, patient-reported outcome measures, physician global assessment of disease activity, joint counts (standard 28-joint count with addition of ankles and feet (as one joint) resulting in a 32-joint count), C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR). For this study patients with a clinical diagnosis of PsA made by their rheumatologist, who were receiving their first TNFi, were included. Only patients receiving the TNFi as monotherapy or in combination with MTX were selected.

Baseline characteristics and disease activity were compared between the TNFi monotherapy and MTX co-medication groups. Due to the heterogeneity in clinical presentation, especially regarding presence and extent of peripheral joint swelling, measures without joint counts were used in the primary analysis. Selected outcome measures were physician global, patient global, the Modified Health Assessment Questionnaire (MHAQ) score16 and Short Form 6 Dimensions (SF-6D), calculated from the Medical Outcomes Study 36-item Short-Form Health Survey (SF-36).17 State scores and changes from baseline after 3, 6 and 12 months were compared.

A secondary analysis was performed on patients who had at least one swollen joint at baseline (n=330). In these analyses additional outcome measures were assessed, including joint counts. These measures were developed for RA but have been commonly used for PsA.18 The selected outcome measures were the Disease Activity Score-28 (DAS28) with corresponding European League Against Rheumatism response criteria and American College of Rheumatology (ACR) response criteria. We also applied a modified version of the Disease Activity Index for Psoriatic Arthritis (DAPSA) score. This was originally developed for use in reactive arthritis and is a sum score of the 66-swollen joint count, 68-tender joint count, patient global assessment (visual analogue scale (VAS)), pain (VAS) and CRP, and has subsequently been validated in PsA.19 We used available 32-joint counts as a substitute for the original 66/68-joint counts.

Statistical analyses

The χ2 test, independent t-test and Mann–Whitney U test were applied as appropriate for group comparisons. Drug survival was assessed by Kaplan–Meier analysis, and patients receiving and not receiving concomitant MTX were compared using the log rank test. These analyses were performed for the full group of patients and separately for each of the three most frequently used TNFi (etanercept (ETN), IFX and adalimumab (ADA)).

Multivariate Cox regression analysis was performed to identify predictors of drug survival, including the use of concomitant MTX as a covariate. Other covariates were age, sex, disease duration, presence/absence of swollen joints, number of previous DMARDs, patient global, physician global, type of TNFi, smoking, CRP and education. All of these were tested in univariate Cox regression analyses, and variables with p values<0.25 were included in multivariate Cox regression analysis. The least significant variable was excluded from the model until only variables with a p value<0.10 remained, while age and sex were kept in the model at all stages. Variables not included in the multivariate analysis were then individually retested in the multivariate model. Proportionality of the hazard of the variables was assessed using cumulative hazard plots for dichotomous variables and Schonenfeldt residual plots for continuous variables.

Statistical tests were two sided with level of significance set at 0.05, without correction for multiple testing. All analyses were performed using IBM SPSS V.19.


The analyses included 440 patients: 170 received TNFi monotherapy and 270 received TNFi with concomitant MTX. Patients receiving TNFi with other co-medication were not included (n=18). Baseline characteristics are shown in table 1. Twenty-two percent of patients had no swollen joints at baseline assessed by the 32-joint count and this proportion was significantly higher in patients who did not receive co-medication (68.5% vs 83.3%, p=0.001). Smaller, but statistically significant group differences were also observed in physician global assessment and SF-6D score. Other important baseline characteristics such as age, sex and number of previously used DMARDs were similar. In the co-medication group 80.4% had used MTX prior to starting the TNFi, while 70.6% in the monotherapy group had previously used MTX. Median potential time on treatment was similar between the monotherapy and co-medication groups, reflecting that the proportion of patients with and without MTX co-medication was quite stable over the years (shown in online supplementary figure S1). This was also true within each of the three most commonly used TNFi (data not shown). Similar proportions of patients in the monotherapy versus MTX group were taking non-steroidal anti-inflammatory drugs (64.3% vs 64.0%) and steroids (36.0% vs 39.2%) at base line. Mean (SD) MTX dose in the co-medication group was 14.7 (5.1).

Table 1

Baseline characteristics of patients

In the secondary analysis, including only patients with swollen joints at baseline, more prominent baseline differences were observed. Patients in the monotherapy group had used significantly more DMARDs previously, and there was a trend towards higher disease severity in the monotherapy group, reaching statistical significance for physician global assessment and SF-6D.

Table 2 shows state and change scores from baseline and at 3 and 6 months; overall values and separate values are given for those on monotherapy and those on concomitant MTX. After 3 months physician global assessment was significantly higher in the group not receiving concomitant MTX. However, changes from baseline were not different between the groups. No significant group differences were observed at 6 and 12 months (table 2 and online supplementary table S1), and there was no numerical trend.

Table 2

State and change from baseline after 3 months: all patients

The secondary analysis including only patients with swollen joints at baseline assessed by the 32-joint count showed similar results regarding the generic response measures (table 3). The RA-specific measures and modified DAPSA did not show any statistically significant group differences. Similar results were shown for state and change scores of swollen joint count, tender joint count, CRP and ESR (data not shown). Analyses adjusted for number of previously used DMARDs and baseline disease activity (DAS28) showed similar results (data not shown).

Table 3

Responses, state and change from baseline in patients with at least one swollen joint at baseline

Drug survival for the two groups (including patients with no swollen joints at baseline) is shown in figure 1A. We observed a trend towards better drug survival in the group receiving concomitant MTX, reaching statistical significance at 1 and 2 years, and borderline significance at 3 years. The stratified analysis for the three major TNFi (ETN, IFX and ADA) (figure 1B–D) showed a statistically significant difference between the monotherapy and the MTX co-medication groups in patients receiving IFX, favouring co-medication. A similar trend was seen for patients treated with ADA. The group difference in the ETN group was negligible. Similar results were seen when only patients with swollen joints at baseline were included in the analyses (data not shown).

Figure 1

Drug survival by co-medication. (A) All patients. N, patients still observed at the end of the interval. (B–D) Drug survival by co-medication for patients starting etanercept, infliximab and adalimumab. *p Value by log-rank test.

Discontinuations were more frequent in the monotherapy versus co-medication group (54.1% vs 45.9%). Reasons for discontinuation were loss/lack of efficacy (20.0% vs 13.7%), adverse events (AEs) (21.2% vs 14.3%) and other (13.0% vs 17.9%). Investigation of discontinuations stratified for the three major TNFi were done by Kaplan–Meier analyses as shown in online supplementary figure S2. These analyses showed that patients receiving IFX without co-medication had markedly higher discontinuation rates due to AEs (p<0.001) (online supplementary figure S2).

The identified predictors of 3-year treatment termination are shown in table 4. Use of concomitant MTX was significant as it decreased hazard of drug termination with a p value of 0.04. Baseline patient global assessment and CRP were borderline significant. Current smoking at baseline was a significant predictor of treatment termination (p=0.02). Smoking was also tested in the model as a tricotomous variable (never/past/current) and dichotomous in the form of never/ever, but current smoking versus past/never came out as the strongest predictor. Disease activity of smokers was not statistically significantly different from non-smokers at baseline, except the MHAQ score was slightly higher, and the 3-month responses were also similar (data not shown). Type of biological drug was not significant when tested as IFX versus other or ETN versus other TNFi to increase statistical power. The analysis was repeated for 1-year and 2-year drug survival. The results were similar, although smoking was not a significant predictor for 1-year drug termination (data not shown). The results from the univariate analyses and details of model selection are shown in online supplementary table S2.

Table 4

Multivariate analysis: predictors of 3-year drug termination


In this study we investigated the potential effects of MTX co-medication on response and drug survival, focusing on patients with PsA treated with their first TNFi in a normal clinical setting. We found no consistent differences in response, but a trend towards better drug survival in patients receiving concomitant MTX. When investigating drug survival separately for the three major TNFi, we found that the difference between those receiving and not receiving concomitant MTX was more prominent in patients on IFX than for other TNFi, with a similar trend in the group receiving ADA.

Patients with no previous exposure to TNFi were selected for the analysis, and we only included those who received TNFi as monotherapy or with concomitant MTX. Other synthetic DMARDs were seldom used. The baseline characteristics were similar to those reported from other observational cohorts,13 ,20 ,21 but with slightly lower disease activity, especially compared with patients from the British Society for Rheumatology Biologics Register.20 This difference might reflect higher availability of TNFi in Norway compared with the UK. Presence of peripheral joint swelling was surprisingly low in both groups. Axial, entheseal disease, dactylitis and possibly skin disease are also likely indications for starting TNFi in patients with no recorded joint swelling. Unfortunately the NOR-DMARD study does not include information about the presence of enthesitis, dactylitis or skin disease.

As our data are observational, confounding by indication is a concern. Clinically we considered the degree of peripheral versus axial disease as a potential major confounder. As we did not have any suitable measure of axial disease recorded for these patients, we were unable to investigate this issue further. Presence of peripheral joint swelling was significant in our group comparison of baseline characteristics. Because we were aware that we lacked essential information to contribute to a propensity score or adjusted analysis, we chose to perform secondary analysis on patients with known peripheral joint swelling at baseline. The baseline characteristics in this secondary analysis showed some more differences between the groups, like higher number of previously used DMARDS in the monotherapy group and a trend towards higher disease activity. This difference could reflect a higher degree of drug intolerance and treatment resistance in the monotherapy group. It could also reflect a partial effect of MTX in the co-medication group.

Due to the likely heterogeneity of the type of disease activity in patients, we only used measures that did not include joint counts to assess response in the primary analysis. However, in the secondary analysis we included RA-specific measures commonly used for PsA in patients with peripheral joint swelling. We included a modified version of the DAPSA score as an outcome measure for PsA because we did not have full joint counts to calculate the original DAPSA and we did not have sufficient information to use other PsA-specific measures. Selection of outcome measures in PsA is largely an unresolved issue, which is reflected in the great variety of outcome measures selected in published papers on PsA. This variety complicates comparison of our results with those of other studies. However, the overall ACR responses in our subgroups of patients with swollen joints at baseline are comparable to those observed in RCTs,1 ,3 ,4 with a slightly higher ACR 70 response in our data. Our finding of no difference in responses between the groups receiving monotherapy versus concomitant MTX is consistent with the findings of Kristensen et al.13 The consistency in our results across the number of selected measures and time points also strengthens the finding that there are no major differences in responses between the two groups.

For drug survival, there seems to be a certain difference between the two groups, and this is supported by the multivariate Cox regression analysis. This finding is also in agreement with the findings of Kristensen et al13 and previous findings from our own study.12 The lack of difference in responses does not support a theory of the synergistic effect of MTX and TNFi as is assumed in RA, but survival analyses do indicate a more favourable overall outcome. A possible explanation for this could be discontinuation of medication prior to the first follow-up visit. However, since only 23 patients discontinued medication during this period, this explanation is less likely. Further, the difference in drug survival could be explained by mechanisms related to the potential of MTX to prevent formation of antidrug antibodies. Our results indicated that there is a higher frequency of AEs leading to discontinuation in patients receiving IFX without co-medication. Kristensen et al13 also found AEs to be the main cause of this difference between drug survival in patients receiving and not receiving co-medication. As investigations in this study did not include blood samples other than acute-phase reactants, we could not measure drug concentrations or antidrug antibody levels to explore this issue further. Interestingly, current smoking was found to increase the frequency of discontinuation. In recent years, smoking has been related to disease development in PsA,22 but to our knowledge this is the first time it has been identified to influence drug survival.

In conclusion, we found similar levels of responses to TNFi treatment between patients receiving and not receiving MTX co-medication, but superior drug survival in the group receiving MTX. The superior drug survival in those receiving MTX co-medication was most prominent in patients receiving IFX, and the same trend was seen for ADA. In addition to MTX co-medication, current smoking was identified as an independent predictor of drug discontinuation. The results indicate that there is a role for MTX co-medication with TNFi in PsA, and that this might be more important in patients receiving monoclonal antibodies.


The authors thank the patients for participating in this study and the local rheumatology staff for data collection.


View Abstract

Review history and Supplementary material

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Handling editor Gerd Burmester

  • Contributors Study design: TKK, EL. Data acquisition: SK, KM, ER, ÅL, MSH. Data analysis: KMF, EL, DvdH. Manuscript preparation: KMF, EL, DvdH, MSH, TKK. All authors have approved the final manuscript.

  • Funding This work was supported by the Norwegian South-Eastern Health Board through a grant for Karen M. Fagerli's salary as a PhD-student. The NOR-DMARD study have received funding in form of unrestricted grants from Abbott, Amgen, Aventis, Bristol Myers Squibb, MSD, Roche, Schering-Plough/Centocor and Wyeth.

  • Competing interests KMF has received speaker's honoraria from Abbott and Pfizer. EL has received speaker's/consulting honoraria from Roche, Pfizer, BMS and Abbott. DvdH has received consulting fees and/or research grants from Abbott, Amgen, AstraZeneca, BMS, Centocor, Chugai, Eli-Lilly, GSK, Merck, Novartis, Otsuka, Pfizer, Roche, Sanofi-Aventis, Schering-Plough, UCB, and Wyeth. TKK has received honoraria as speaker and/or consultant from Abbott, BMS, MSD/Schering-Plough, Pfizer/Wyeth, Roche and UCB.

  • Ethics approval Regional Ethics Committee of South-Eastern Norway.

  • 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.