Abstract

Objective. To compare baseline characteristics, responses and drug survival in patients with early RA starting SSZ or MTX in a real-life setting.

Methods. The analyses included DMARD-naïve patients with RA (disease duration ≤ 1 year) starting SSZ or MTX. Three- and 6-month effectiveness was compared by unadjusted analysis and with adjustment for propensity score quintile. In addition, effectiveness in SSZ- and MTX-treated patients matched for RF status and baseline DAS-28 was compared.

Results. SSZ-treated patients (n = 175) had lower baseline disease activity than patients treated with MTX (n = 927) [mean 28-joint DAS (DAS-28) 4.4 vs 5.0, P < 0.001], and were less often RF positive (50 vs 61%, P = 0.006). Six-month mean ΔDAS-28 was smaller with SSZ than MTX (−1.0 vs −1.5, P = 0.003); the difference was not significant after adjustment for propensity score quintile (P = 0.36). For SSZ/MTX, 3-month ACR50 and European League Against Rheumatism (EULAR) good responses were 9/23% (P < 0.001) and 24/31% (P = 0.14), respectively. Three-year drug survival was superior for MTX (P < 0.001) and estimated 1-year survival rates were 42/75% for SSZ/MTX. In patients matched for baseline DAS-28 and RF, mean ΔDAS-28 (MTX −1.2, P = 0.55 vs SSZ) and EULAR good responses (39 vs 37%, P = 0.74) were similar at 6 months; drug survival was superior for MTX (P < 0.001).

Conclusion. Patients treated with SSZ as first DMARD were more often RF negative and had lower baseline disease activity. Drug survival was superior for MTX, and effectiveness was greater with MTX than with SSZ although the difference was reduced when adjusting for differences in baseline characteristics.

Introduction

RA is a chronic inflammatory disease with a potentially disabling outcome, and it is widely accepted that early diagnosis and early institution of treatment with DMARDs is paramount to prevent joint damage and achieve the lowest possible disease activity state [1, 2]. Both SSZ and MTX have documented efficacy on signs and symptoms in patients with RA [3–5]. Two randomized controlled trials (RCTs) on DMARD-naïve patients with early RA have shown similar clinical responses with SSZ and MTX, but the MTX doses used were relatively low (up to 15 mg/week) [6, 7]. A 5-year observational study showed similar clinical responses of SSZ and MTX, but less erosive progression in the MTX group [8]. Several studies have shown that the use of MTX has increased since the 1980s, and that, more recently, the relative use of SSZ vs MTX has declined [8–11].

The most recent treatment recommendations from the European League Against Rheumatism (EULAR) state that MTX should be (part of) the first treatment strategy for patients with active RA; SSZ is mentioned as an alternative drug for patients in whom MTX is contraindicated [1]. MTX and SSZ are also listed as possible first-line therapies in the most recent recommendations from the American College of Rheumatology (ACR) [12]. While the ACR recommends MTX as first-line therapy in all scenarios, irrespective of disease activity level and prognostic factors, SSZ is generally only recommended for patients without unfavourable prognostic markers and with a low to moderate disease activity level [12].

There is a lack of recent studies on the comparative effectiveness of SSZ and MTX in an era where treatment practice has changed with both the introduction of biologic DMARDs and a stronger focus on early and aggressive treatment. The objective of this study was to compare demographics, baseline characteristics, clinical responses and retention to therapy in DMARD-naïve patients with early RA starting SSZ or MTX, who have been included in an ongoing longitudinal observational study in Norway over the past decade.

Methods

Setting and patients

Data for this study were provided by the NOR-DMARD register, which includes adult patients with inflammatory arthropathies starting treatment with synthetic and/or biological DMARDs in five Norwegian rheumatology departments. The register covers approximately1.4 million people (one-third of the Norwegian population) and is set up as a longitudinal observational study with assessments at baseline, after 3, 6 and 12 months and yearly thereafter. The study was conducted with approval by the Regional Ethics Committee of South-Eastern Norway and the Data Inspectorate. Patients gave written informed consent before participation.

For the current study, we included DMARD-naïve patients with RA with disease duration of ≤1 year, enrolled in the NOR-DMARD register between December 2000 and May 2009 and starting treatment with SSZ or MTX as monotherapies. The diagnosis of RA was made by the treating rheumatologist based on clinical judgement. At the time of analysis, follow-up data until February 2010 were available. For the effectiveness analyses we included 3- and 6-month follow-up data, and analysis of drug survival was based on 3-year data with censoring of patients with inadequate follow-up time and patients who were lost to follow-up.

Assessments and outcomes

Assessments included 28 swollen joint count (28-SJC) and 28 tender joint count (28-TJC) performed by rheumatologists or trained research nurses, ESR, CRP, 0–100 mm visual analogue scales (VASs) for physician's assessment of global disease activity and patient's assessment of joint pain, fatigue and global disease activity (100 mm = worst score), and the modified HAQ (MHAQ) [13]. Further, patients completed the Medical Outcomes Study 36-item Short-Form Health Survey (SF-36) questionnaire [14, 15] which was used to calculate physical component summary (SF-36 PCS) and mental component summary scores (SF-36 MCS) and SF-6D [16]. The 28-joint DAS (DAS-28) with ESR [17], the simplified disease activity index (SDAI) [18] and the proportions achieving EULAR good response [19] and DAS-28 remission (DAS-28 < 2.6) were calculated. We also applied the ACR response criteria (ACR20, ACR50 and ACR70 responses) [20] as well as the recently published ACR/EULAR remission criteria, both the Boolean- and the index-based definition (SDAI ≤ 3.3) [21, 22]. To account for differences in retention to therapy, we calculated LUNDEX values for each response and remission category at 3 and 6 months by the following formula: LUNDEX value = (fraction of starters still in study at time T) × (fraction responding at time T) [23]. Information about dates of and reasons for treatment discontinuation was systematically recorded.

Statistical analysis

Continuous variables are presented as means (s.d.) if normally distributed and medians [interquartile ranges (IQRs)] if non-normally distributed. Independent-samples t-test and Mann–Whitney U-test were applied as appropriate. Dichotomous variables are presented as frequencies and percentages and were compared by chi-square test. For continuous variables, we compared the 3- and 6-month completer data, and in addition we used the last observation carried forward (LOCF) approach to replace missing 6-month data with data from the 3-month assessment if available.

A priori we suspected that there would be differences in demographics and baseline disease activity, which might have influenced the treating physicians’ decision to treat with SSZ or MTX (i.e. confounding by indication), and we calculated propensity scores for each patient to be able to at least partially adjust for this source of bias [8, 24, 25]. The propensity score was calculated by multivariate logistic regression with treatment (SSZ vs MTX) as the dependent variable and age, sex, centre, RF status, baseline prednisolone use and baseline DAS-28 as covariates. The propensity scores were divided into quintiles, and the balance of demographics and baseline disease activity across treatment groups within each quintile was checked [25]. Further details are available in the supplementary data, available at Rheumatology Online.

We then applied analysis of covariance (ANCOVA) with the changes in outcome measures as the dependent variables and treatment (SSZ vs MTX) and propensity score quintile as factors. For the changes which were not normally distributed, we used the van der Waerden normal scores of the changes instead of the changes themselves. Finally, 3-year retention to therapy (drug survival) was compared by Kaplan–Meier analysis and Cox regression analysis with adjustment for propensity score quintile.

In addition to the comparison of the full cohorts of patients on SSZ and MTX, we performed 1 : 1 matching of patients based on RF status and baseline disease activity (measured by DAS-28). A complete set of DAS-28 values (as described in the supplementary data, available at Rheumatology Online) was used for matching of patients. For matched patients, unadjusted comparison of effectiveness and Kaplan–Meier analysis of 3-year drug survival were performed.

Statistical tests were two-sided, P < 0.05 was considered statistically significant and no correction for multiple testing was made. Statistical analyses were performed using the Predictive Analytics Software (PASW) program, version 17.0 (SPSS Inc., Chicago, IL, USA).

Results

Baseline characteristics

Overall, 1102 patients, of whom 175 (15.9%) were started on SSZ and 927 (84.1%) on MTX, fulfilled the inclusion criteria for these analyses. Other treatments than SSZ and MTX (HCQ n = 38, synthetic DMARD combinations n = 26, TNF inhibitors n = 24, LEF n = 15, sodium aurothimalate n = 6 and auranofin n = 4) for this group of patients were excluded from the analyses.

SSZ-treated patients were somewhat younger and more often RF negative than patients started on MTX (Table 1). The percentage of SSZ treatment within each of the five participating centres ranged from 5.7 to 28.6%, reflecting local differences in practice. The relative use of SSZ vs MTX declined during the study period, with SSZ constituting 25.3% of treatments in patients included from 2000 to 2002, 18.2% from 2003 to 2005 and only 8.4% from 2006 to 2009.

Table 1

Demographics and baseline characteristics

 SSZ
 
MTX
 
  
Baseline characteristic All patients (n = 175) Patients with 3- and/or 6-month data (n = 150) All patients (n = 927) Patients with 3- and/or 6-month data (n = 880) P-valuea P-valueb 
Age, mean (s.d.), years 49.9 (14.8) 49.7 (15.2) 55.9 (13.6) 56.1 (13.3) <0.001 <0.001 
Female sex, n (%) 108 (61.7) 97 (64.7) 629 (67.9) 600 (68.2) 0.11 0.39 
RF positive, n (%) 88 (50.3) 76 (50.7) 569 (61.4) 541 (61.5) 0.006 0.01 
Prednisolone use, n (%) 56 (32.0) 46 (30.7) 515 (55.6) 485 (55.1) <0.001 <0.001 
DAS-28, mean (s.d.4.38 (1.35) 4.35 (1.40) 5.00 (1.34) 5.01 (1.36) <0.001 <0.001 
SDAI, mean (s.d.22.0 (12.0) 21.7 (12.2) 27.2 (14.3) 27.3 (14.6) <0.001 <0.001 
MHAQ score (0–3), median (IQR) 0.50 (0.125–0.75) 0.50 (0.125–0.8125) 0.625 (0.25–1.00) 0.625 (0.25–1.00) 0.001 0.002 
28-SJC, median (IQR) 4 (2–8) 4 (2–8) 6 (3–10) 6 (3–10) 0.001 <0.001 
28-TJC, median (IQR) 6 (3–9) 5 (2–9) 7 (3–13) 7 (3–13) 0.001 0.001 
ESR, median (IQR), mm/h 17 (8–32.5) 17 (8–29) 22 (12–41) 22 (12–41) <0.001 <0.001 
CRP, median (IQR), mg/l 10 (4–22) 9 (4–22) 12 (5–29.75) 12 (5–30) 0.03 0.02 
PhGA VAS, mean (s.d.33.2 (16.7) 32.5 (16.5) 40.7 (18.9) 40.5 (18.9) <0.001 <0.001 
PGA VAS, mean (s.d.43.4 (23.1) 43.0 (22.7) 46.9 (23.8) 46.6 (23.7) 0.08 0.08 
Pain VAS, mean (s.d.42.2 (22.9) 42.3 (22.3) 46.5 (24.6) 46.3 (24.4) 0.03 0.05 
Fatigue VAS, median (IQR) 37.5 (10.75–64.5) 32.5 (8.75–64) 43.5 (16–65) 43 (16–65) 0.10 0.05 
SF-6D, mean (s.d.0.63 (0.13) 0.63 (0.13) 0.59 (0.13) 0.60 (0.13) 0.001 0.004 
SF-36 PCS, mean (s.d.34.3 (9.6) 34.1 (9.7) 32.0 (9.6) 32.0 (9.5) 0.005 0.02 
SF-36 MCS, mean (s.d.46.3 (12.5) 46.2 (12.3) 45.4 (11.5) 45.7 (11.4) 0.36 0.58 
 SSZ
 
MTX
 
  
Baseline characteristic All patients (n = 175) Patients with 3- and/or 6-month data (n = 150) All patients (n = 927) Patients with 3- and/or 6-month data (n = 880) P-valuea P-valueb 
Age, mean (s.d.), years 49.9 (14.8) 49.7 (15.2) 55.9 (13.6) 56.1 (13.3) <0.001 <0.001 
Female sex, n (%) 108 (61.7) 97 (64.7) 629 (67.9) 600 (68.2) 0.11 0.39 
RF positive, n (%) 88 (50.3) 76 (50.7) 569 (61.4) 541 (61.5) 0.006 0.01 
Prednisolone use, n (%) 56 (32.0) 46 (30.7) 515 (55.6) 485 (55.1) <0.001 <0.001 
DAS-28, mean (s.d.4.38 (1.35) 4.35 (1.40) 5.00 (1.34) 5.01 (1.36) <0.001 <0.001 
SDAI, mean (s.d.22.0 (12.0) 21.7 (12.2) 27.2 (14.3) 27.3 (14.6) <0.001 <0.001 
MHAQ score (0–3), median (IQR) 0.50 (0.125–0.75) 0.50 (0.125–0.8125) 0.625 (0.25–1.00) 0.625 (0.25–1.00) 0.001 0.002 
28-SJC, median (IQR) 4 (2–8) 4 (2–8) 6 (3–10) 6 (3–10) 0.001 <0.001 
28-TJC, median (IQR) 6 (3–9) 5 (2–9) 7 (3–13) 7 (3–13) 0.001 0.001 
ESR, median (IQR), mm/h 17 (8–32.5) 17 (8–29) 22 (12–41) 22 (12–41) <0.001 <0.001 
CRP, median (IQR), mg/l 10 (4–22) 9 (4–22) 12 (5–29.75) 12 (5–30) 0.03 0.02 
PhGA VAS, mean (s.d.33.2 (16.7) 32.5 (16.5) 40.7 (18.9) 40.5 (18.9) <0.001 <0.001 
PGA VAS, mean (s.d.43.4 (23.1) 43.0 (22.7) 46.9 (23.8) 46.6 (23.7) 0.08 0.08 
Pain VAS, mean (s.d.42.2 (22.9) 42.3 (22.3) 46.5 (24.6) 46.3 (24.4) 0.03 0.05 
Fatigue VAS, median (IQR) 37.5 (10.75–64.5) 32.5 (8.75–64) 43.5 (16–65) 43 (16–65) 0.10 0.05 
SF-6D, mean (s.d.0.63 (0.13) 0.63 (0.13) 0.59 (0.13) 0.60 (0.13) 0.001 0.004 
SF-36 PCS, mean (s.d.34.3 (9.6) 34.1 (9.7) 32.0 (9.6) 32.0 (9.5) 0.005 0.02 
SF-36 MCS, mean (s.d.46.3 (12.5) 46.2 (12.3) 45.4 (11.5) 45.7 (11.4) 0.36 0.58 

Chi-square test, independent-samples t-test or Mann–Whitney U-test were applied as appropriate. aComparison of all patients SSZ vs MTX. bComparison of patients with 3- and/or 6-month data. PhGA: physician's global assessment; PGA: patient's global assessment.

At baseline, patients started on MTX scored significantly worse for most disease measures (Table 1), with mean DAS-28 of 5.00 vs 4.38 for patients started on SSZ (P < 0.001). Further, MTX-treated patients had a higher number of swollen joints (Table 1), but the distribution of joint swelling (large vs small joints, upper vs lower extremities, symmetry) did not differ (data not shown). Baseline data are shown both for all included patients and for patients with available 3- and/or 6-month data included in the effectiveness analyses, and the results were very similar (Table 1).

Medication

For patients treated with SSZ, mean (s.d.) SSZ dose was 1.8 (0.5) g at baseline, 2.0 (0.5) g at 3 months and 2.1 (0.4) g at 6 months, with a median (IQR) dose of 2.0 (2.0–2.0) g at all time points. Mean (s.d.)/median (IQR) doses of MTX were 11.6 (3.6)/10 (7.5–15) mg at baseline, 14.4 (3.9)/15 (12.5–15) mg at 3 months, and 15.1 (4.2)/15 (12.5–20) mg at 6 months. At baseline, prednisolone was used by 32.0% of SSZ- and 55.6% of MTX-treated patients (P < 0.001), with mean doses of 8.4 and 9.4 mg (P = 0.10), respectively. At 3 months, 21.9/49.0% (P < 0.001) of patients on SSZ/MTZ used prednisolone in doses of 6.0/6.2 mg (P = 0.69), and the respective numbers at 6 months were 16.9/41.5% (P < 0.001) and 5.6/5.7 mg (P = 0.91). The proportions that received IA CS injections were similar for SSZ- and MTX-treated patients at baseline, 3 and 6 months.

Propensity scores

Fit of the propensity score model was satisfactory (Hosmer–Lemeshow goodness-of-fit test P = 0.35). Adding more variables to the propensity score model reduced the model fit substantially. The mean/median propensity scores in each group as well as distribution of patients across propensity score quintiles are shown in Table S1, available as supplementary data at Rheumatology online. All baseline characteristics in Table 1, except for the physician global VAS, were balanced within each quintile (data not shown). We therefore performed a sensitivity analysis where we included the physician global as an independent variable in addition to the propensity score quintile.

Changes in disease measures

Patients treated with MTX showed numerically larger improvement in all disease measures at 6 months using the LOCF approach (Table 2). Mean ΔDAS-28 was −1.04 for SSZ vs −1.52 for MTX (P = 0.003). The statistical significance of many of the differences was not maintained when adjusting for propensity score quintile (Table 2). Similar results were observed for the 3 and 6 months as observed data (Table S2, available as supplementary data at Rheumatology Online). The difference between groups was greater at 3 than 6 months, possibly reflecting that more patients discontinued SSZ than MTX due to lack of efficacy (LOE) between 3 and 6 months (see details below). Adjusting for baseline, physician global in addition to propensity score quintile generally resulted in only minor changes in P-values, with a statistically significant difference in favour of MTX for a minority of the outcomes (Table 2 and Table S2, available supplementary data at Rheumatology Online).

Table 2

Six-month changes in disease measures (LOCF for missing data)

Variable SSZ (n = 150) MTX (n = 880) P-valuea P-valueb P-valuec 
DAS-28 −1.04 (1.64) −1.52 (1.60) 0.003 0.36 0.71 
SDAI −6.9 (13.6) −13.2 (15.3) <0.001 0.03 0.21 
MHAQ score (0–3) −0.13 (0.45) −0.26 (0.48) 0.002 0.05 0.13 
28-SJCd −2 (−5, 0) −3 (−8, 0) 0.001 0.07 0.005 
28-TJCd −1 (−6, 1) −3 (−8, 0) 0.001 0.05 0.009 
ESRd, mm/h −3 (−14, 3) −7 (−20.5, 0) 0.003 0.06 0.04 
CRPd, mg/l 0 (−10, 2) −5 (−19, 0) <0.001 0.23 0.13 
PhGA VAS −11.8 (20.1) −20.4 (22.6) <0.001 0.007 0.20 
PGA VAS −8.6 (27.1) −12.5 (26.9) 0.10 0.62 0.85 
Pain VAS −9.2 (23.6) −14.7 (26.9) 0.02 0.24 0.41 
Fatigue VAS −0.4 (28.2) −4.4 (29.6) 0.13 0.21 0.24 
SF-6D 0.05 (0.13) 0.09 (0.13) 0.005 0.02 0.05 
SF-36 PCS 4.0 (8.5) 5.4 (9.8) 0.11 0.26 0.42 
SF-36 MCS 2.4 (11.4) 2.8 (11.2) 0.67 0.78 0.74 
Variable SSZ (n = 150) MTX (n = 880) P-valuea P-valueb P-valuec 
DAS-28 −1.04 (1.64) −1.52 (1.60) 0.003 0.36 0.71 
SDAI −6.9 (13.6) −13.2 (15.3) <0.001 0.03 0.21 
MHAQ score (0–3) −0.13 (0.45) −0.26 (0.48) 0.002 0.05 0.13 
28-SJCd −2 (−5, 0) −3 (−8, 0) 0.001 0.07 0.005 
28-TJCd −1 (−6, 1) −3 (−8, 0) 0.001 0.05 0.009 
ESRd, mm/h −3 (−14, 3) −7 (−20.5, 0) 0.003 0.06 0.04 
CRPd, mg/l 0 (−10, 2) −5 (−19, 0) <0.001 0.23 0.13 
PhGA VAS −11.8 (20.1) −20.4 (22.6) <0.001 0.007 0.20 
PGA VAS −8.6 (27.1) −12.5 (26.9) 0.10 0.62 0.85 
Pain VAS −9.2 (23.6) −14.7 (26.9) 0.02 0.24 0.41 
Fatigue VAS −0.4 (28.2) −4.4 (29.6) 0.13 0.21 0.24 
SF-6D 0.05 (0.13) 0.09 (0.13) 0.005 0.02 0.05 
SF-36 PCS 4.0 (8.5) 5.4 (9.8) 0.11 0.26 0.42 
SF-36 MCS 2.4 (11.4) 2.8 (11.2) 0.67 0.78 0.74 

Values are mean (s.d.) changes unless otherwise indicated. aIndependent-samples t-test or Mann–Whitney U-test, applied as appropriate. bANCOVA with adjustment for propensity score quintile. cANCOVA with adjustment for propensity score quintile and physician global VAS. dMedian (IQR). PhGA: physician's global assessment; PGA: patient's global assessment.

Response and remission rates

Three- and 6-month response and remission rates are presented in Table 3. At 3 months, MTX-treated patients showed higher response and remission rates by all criteria, and the differences in ACR response and SDAI remission rates were statistically significant (Table 3). ACR50 response was achieved by 8.7 vs 22.6% of patients treated with SSZ and MTZ, respectively, while 6.3 vs 12.9% achieved SDAI remission. Applying the LUNDEX formula the corresponding numbers were 5.9 vs 21.3% for ACR50 and 4.3 vs 12.1% for SDAI remission due to fewer patients on SSZ (69%) than on MTX (94%) still on therapy at 3 months. As observed for the change scores, differences in crude response rates were less pronounced at 6 months, and statistically significant superior response for MTX over SSZ was only seen for the ACR20 response (51.8 vs 37.2%, P = 0.01) (Table 3). Of the patients assessed at 6 months, the DAS-28 remission rate was actually significantly higher for SSZ (44.7%) than for MTX (32.5%) (P = 0.03), while all LUNDEX-adjusted response and remission rates were superior for MTX since far more patients on MTX (86%) than on SSZ (56%) remained on therapy for ≥6 months (Table 3).

Table 3

Three- and 6-month response and remission rates—observed percentages and LUNDEX values

 3 months
 
6 months
 
Response/remission criterion SSZ (n = 137) MTX (n = 819) P-value SSZ (n = 89) MTX (n = 743) P-value 
ACR20 26.8 50.4 <0.001 37.2 51.8 0.01 
ACR20 LUNDEX 18.3 47.4 NA 20.8 44.5 NA 
ACR50 8.7 22.6 <0.001 16.3 25.1 0.07 
ACR50 LUNDEX 5.9 21.3 NA 9.1 21.6 NA 
ACR70 4.7 14.9 0.002 9.3 16.6 0.08 
ACR70 LUNDEX 3.2 14.0 NA 5.2 14.3 NA 
EULAR good response 24.0 31.2 0.14 39.1 36.0 0.61 
EULAR good response LUNDEX 16.5 29.3 NA 21.9 31.0 NA 
DAS-28 <2.6 21.3 27.2 0.19 44.7 32.5 0.03 
DAS-28 <2.6 LUNDEX 14.6 25.6 NA 25.1 27.9 NA 
SDAI ≤3.3 6.3 12.9 0.04 10.7 14.2 0.41 
SDAI ≤3.3 LUNDEX 4.3 12.1 NA 6.0 12.2 NA 
ACR/EULAR remission 6.2 9.4 0.27 6.7 10.2 0.33 
ACR/EULAR remission LUNDEX 4.2 8.8 NA 3.7 8.8 NA 
 3 months
 
6 months
 
Response/remission criterion SSZ (n = 137) MTX (n = 819) P-value SSZ (n = 89) MTX (n = 743) P-value 
ACR20 26.8 50.4 <0.001 37.2 51.8 0.01 
ACR20 LUNDEX 18.3 47.4 NA 20.8 44.5 NA 
ACR50 8.7 22.6 <0.001 16.3 25.1 0.07 
ACR50 LUNDEX 5.9 21.3 NA 9.1 21.6 NA 
ACR70 4.7 14.9 0.002 9.3 16.6 0.08 
ACR70 LUNDEX 3.2 14.0 NA 5.2 14.3 NA 
EULAR good response 24.0 31.2 0.14 39.1 36.0 0.61 
EULAR good response LUNDEX 16.5 29.3 NA 21.9 31.0 NA 
DAS-28 <2.6 21.3 27.2 0.19 44.7 32.5 0.03 
DAS-28 <2.6 LUNDEX 14.6 25.6 NA 25.1 27.9 NA 
SDAI ≤3.3 6.3 12.9 0.04 10.7 14.2 0.41 
SDAI ≤3.3 LUNDEX 4.3 12.1 NA 6.0 12.2 NA 
ACR/EULAR remission 6.2 9.4 0.27 6.7 10.2 0.33 
ACR/EULAR remission LUNDEX 4.2 8.8 NA 3.7 8.8 NA 

Values are expressed as percentage. Chi-square test was applied. LUNDEX value = (fraction of starters still in study at time T) × (fraction responding at time T).

Retention to therapy

As visualized in Fig. 1, 3-year drug survival was superior for MTX compared with SSZ (P < 0.001). Estimated retention rates for SSZ/MTX were 42/75% at 1 year, 32/66% at 2 years and 28/58% at 3 years. Estimated median (95% CI) drug survival time was 225 (123, 327) days for SSZ, and 1546 (1310, 1782) days for MTX. Adjusted for propensity score quintile, the hazard ratio (95% CI) for drug discontinuation >3 years was 2.81 (2.25, 3.51) for SSZ vs MTX (P < 0.001).

Fig. 1

Kaplan–Meier plots over 3-year retention to therapy. The table below the graph shows the numbers of patients at risk at different time points during follow-up. Log-rank test: P < 0.001.

Fig. 1

Kaplan–Meier plots over 3-year retention to therapy. The table below the graph shows the numbers of patients at risk at different time points during follow-up. Log-rank test: P < 0.001.

As of February 2010, 138 (78.9%) of 175 patients treated with SSZ had discontinued therapy [34.8% LOE, 45.7% adverse events (AEs) and 19.7% other reasons], 7 (4.0%) patients were lost to follow-up and 30 (17.1%) patients were still on therapy. Of the 927 patients starting treatment with MTX, 446 (48.1%) patients had discontinued therapy (45.1% LOE, 32.1% AEs and 22.9% other reasons), 54 (5.8%) patients were lost to follow-up, while the remaining 427 (46.1%) patients were still on therapy. Information about registered AEs is included in supplementary data, available at Rheumatology Online.

Matched patients

Based on RF status and baseline DAS-28, we matched the 175 patients on SSZ with 175 patients from the MTX group. Baseline characteristics for the matched MTX patients are shown in Table 4. The only statistically significant difference vs SSZ patients was age. Six-month changes in disease measures (LOCF for missing data) were generally numerically superior for MTX, but most differences did not reach statistical significance (Table 4). Response rates and LUNDEX values for matched patients are shown in Table S3 (available as supplementary data at Rheumatology Online), and drug survival is compared in Fig. S1 (available as supplementary data at Rheumatology Online). Further details are given in the supplementary data, available at Rheumatology Online.

Table 4

Baseline and 6-month changes in disease measures for matched MTX patients

 Baseline
 
Δ6 months (LOCF)
 
Variable Matched MTX (n = 175) P-valuea Matched MTX (n = 165) P-valuea 
Age, years 55.4 (13.1) <0.001 NA NA 
Sex: female, n (%) 122 (69.7) 0.11 NA NA 
RF positive, n (%) 88 (50.3) NA NA NA 
DAS-28 4.47 (1.31) 0.56 −1.16 (1.56) 0.55 
SDAI 22.7 (11.7) 0.64 −9.1 (12.4) 0.18 
MHAQ score (0–3) 0.50 (0.25, 0.875) 0.26 −0.23 (0.48) 0.04 
28-SJCb 5 (2, 8) 0.87 −2 (−6, 0) 0.23 
28-TJCb 5 (2, 9.25) 0.65 −2 (−6, 0.75) 0.56 
ESRb, mm/h 16 (10, 33.25) 0.75 −4 (−15, 1) 0.24 
CRPb, mg/l 10 (5, 25) 0.52 −3 (−16, 0) 0.005 
PhGA VAS 35.4 (16.8) 0.22 −16.9 (18.7) 0.02 
PGA VAS 44.3 (23.9) 0.73 −9.8 (26.6) 0.69 
Pain VAS 42.3 (24.5) 0.97 −11.2 (26.4) 0.49 
Fatigue VAS 40 (14.5, 64) 0.40 −1.9 (29.0) 0.65 
SF-6D 0.61 (0.13) 0.22 0.07 (0.14) 0.33 
SF-36 PCS 33.7 (10.0) 0.57 4.3 (10.1) 0.75 
SF-36 MCS 45.6 (11.3) 0.60 3.0 (12.6) 0.66 
 Baseline
 
Δ6 months (LOCF)
 
Variable Matched MTX (n = 175) P-valuea Matched MTX (n = 165) P-valuea 
Age, years 55.4 (13.1) <0.001 NA NA 
Sex: female, n (%) 122 (69.7) 0.11 NA NA 
RF positive, n (%) 88 (50.3) NA NA NA 
DAS-28 4.47 (1.31) 0.56 −1.16 (1.56) 0.55 
SDAI 22.7 (11.7) 0.64 −9.1 (12.4) 0.18 
MHAQ score (0–3) 0.50 (0.25, 0.875) 0.26 −0.23 (0.48) 0.04 
28-SJCb 5 (2, 8) 0.87 −2 (−6, 0) 0.23 
28-TJCb 5 (2, 9.25) 0.65 −2 (−6, 0.75) 0.56 
ESRb, mm/h 16 (10, 33.25) 0.75 −4 (−15, 1) 0.24 
CRPb, mg/l 10 (5, 25) 0.52 −3 (−16, 0) 0.005 
PhGA VAS 35.4 (16.8) 0.22 −16.9 (18.7) 0.02 
PGA VAS 44.3 (23.9) 0.73 −9.8 (26.6) 0.69 
Pain VAS 42.3 (24.5) 0.97 −11.2 (26.4) 0.49 
Fatigue VAS 40 (14.5, 64) 0.40 −1.9 (29.0) 0.65 
SF-6D 0.61 (0.13) 0.22 0.07 (0.14) 0.33 
SF-36 PCS 33.7 (10.0) 0.57 4.3 (10.1) 0.75 
SF-36 MCS 45.6 (11.3) 0.60 3.0 (12.6) 0.66 

Patients (SSZ n = 175 and MTX n = 175) were matched on RF status and baseline DAS-28. Values are mean (s.d.) changes unless otherwise indicated. aComparison vs SSZ patients (independent samples t-test or Mann–Whitney U test or chi-square test, applied as appropriate). bMedian (IQR). PhGA: physician's global assessment; PGA: patient's global assessment.

Discussion

In this register-based real-life study including DMARD-naïve patients with early RA, MTX was much more commonly used than SSZ, which is in line with the recent EULAR recommendations for treatment of RA [1]. Patients started on SSZ were more often RF negative and had, on average, lower disease activity when starting therapy. The unadjusted 3- and 6-month effectiveness and 3-year retention to therapy were superior for MTX.

Several placebo-controlled randomized trials published in the 1980s and 1990s showed that MTX reduced signs and symptoms of RA, but all these trials included patients with long-standing disease who had failed on other DMARDs [4]. During the same period, the efficacy of SSZ was established through several placebo-controlled randomized trials, including patients with and without previous use of DMARDs and both early and long-standing disease [3]. MTX has gradually become the most-used first-line DMARD in RA [10], and has also been found to be associated with reduced mortality in RA [26]. RCTs on biological DMARDs in RA have shown that the efficacy of biologics is increased when combined with MTX, and this observation has also been important to establish MTX as the cornerstone in the treatment of RA [27–30].

The main difference between SSZ and MTX in this study was retention to therapy, which was far superior for MTX, both in the short (3–6 months) and long (3 years) term. Further, the retention rate at 5 years was only 20% for SSZ and 46% for MTX. The superior retention rate with MTX implies greater effectiveness and less serious side effect problems and is in line with previous studies comparing retention rates of different DMARDs in various RA populations [8, 31–33] However, a number of patients continued to use either SSZ or MTX even if they were not in remission or a low disease activity state, which would have been the current treatment goal according to the Treat-to-Target recommendations [34].

The frequency of AEs on SSZ vs MTX treatment in this study was relatively similar. However, the percentage of patients discontinuing therapy due to AEs was higher for SSZ than for MTX. This would suggest that the side effects that patients experienced on SSZ were more severe or problematic for them, and/or that the tolerance for side effects was inherently less with SSZ than with MTX. The latter could be due to less experienced (and expected) benefit from treatment as well as the availability of other treatment options. Further, we analysed the influence of LOE and AEs on drug survival separately (censoring patients discontinuing the drug due to other reasons), and we found that the benefit of MTX vs SSZ was highly significant for discontinuations due to both LOE and AEs, but relatively greater for AEs (log-rank P < 0.001 for both, chi-square statistic 19.5 vs 80.3). Drug survival will always be influenced by the expectations both of patients and treating physicians as well as the treatment context, including the availability of biologic DMARDs.

In this study, patients on SSZ had generally less severe disease (lower DAS-28 and less use of prednisolone) and more favourable prognostic factors (RF negative). This is different from the study by Hider et al. [8] and more in line with the findings in a recent study from the UK by Rachapalli et al. [35] and an observational study by Aletaha et al. [31] (the latter included not only DMARD-naïve patients and not only early RA). The group receiving SSZ in our study may thus be relatively representative for RA patients who might be considered candidates for initial therapy with SSZ in light of the most recent treatment recommendations [1, 12].

The diagnosis of RA in this study was made by clinical judgement and based on the 1987 ACR classification criteria for RA [36], but fulfilment of the criteria was not systematically checked since the case record form did not include questions about each of the seven items of these criteria. Further, it would have been of interest to be able to classify the patients retrospectively according to the recent ACR/EULAR RA classification criteria [37], but this was not possible as we did not have information about complete 66/68-joint counts and the levels of RF and anti-CCP. There is a possibility that the RA diagnoses were less certain in the patients prescribed with SSZ and that these patients might have features related to the spondyloarthritis entity. If so, this could have influenced the treating physicians’ choice of SSZ over MTX, and further explain some of the observed differences in demographics and baseline characteristics as well as have influenced the effectiveness results. As mentioned, we found no difference in joint distribution, but a lower frequency of swollen joints in all joint areas for patients prescribed with SSZ. Also, comorbidities and contraindications might have influenced the choice of first DMARD.

We used propensity score adjustment as well as a matching for disease activity and RF status to account for indication bias in this study. Neither method will be fully able to adjust for the differences between patient groups due to confounding by indication. There will likely be sources of bias not captured by the registered variables, and the statistical adjustment will generally be imperfect. Propensity score adjustment has previously been used in a similar manner in observational studies in RA [8, 24]. The adjustment for propensity score quintile reduced the statistical difference between treatment groups for all outcomes, and the effect was particularly strong for DAS-28, which is a consequence of its inclusion in the propensity score. There is the risk of over-adjustment as well as under-adjustment with this approach, and results should be interpreted with some caution.

The percentage that used systemic CSs differed between groups, which is a potential source of bias. We partially accounted for this by including baseline prednisolone use in the propensity score model. Prednisolone was tapered in both groups between baseline and 6 months, but relatively more patients treated with SSZ stopped taking prednisolone, and this may have influenced the effectiveness results. The mean/median doses of MTX in this study were lower than the target dose by modern standards [1, 38], which may have resulted in reduced effectiveness of MTX. However, the current doses were higher than those used in the trials by Haagsma et al. [6] and Dougados et al. [7] as well as in the observational study by Aletaha et al. [31].

More patients on SSZ than on MTX discontinued therapy within the first 6 months and this difference influenced the effectiveness results since patients with a good response are more likely to remain on therapy. The LOCF approach is commonly used to replace missing data, but has obvious limitations and certain prerequisites which are seldom met in clinical studies [39]. Still, we applied this method in this study to get a better impression of 6-month effectiveness in a situation with rather large differences in retention to therapy between the groups. The LUNDEX is a suggested tool combining response rates and retentions rates [23], and we believe that the LUNDEX values presented in Table 3 gives a more true picture of the effectiveness of SSZ than the crude response and remission rates, especially at 6 months.

Systematic radiographs of hands and feet were not part of the data collection in this study, and thus we cannot provide data on the relative benefit of MTX vs SSZ regarding radiographic progression. Both SSZ and MTX have been shown to reduce radiographic progression, but the evidence is generally considered to be stronger for MTX [40–43].

Despite the differences in baseline characteristics for patients treated with SSZ and MTX in this study, the observed superior effectiveness and retention rates for MTX further support the ideas that MTX is the first-line DMARD in RA, and that SSZ should mainly be considered for patients in whom MTX is contraindicated.

graphic

Supplementary data

Supplementary data are available at Rheumatology Online.

Acknowledgements

The authors thank the patients for participating in this study, the local rheumatology staff for data collection and Inge C, Olsen for helpful discussions and input regarding the statistical analyses.

Funding: The work was supported by Eastern Norway Regional Health Authority. The Norwegian DMARD study has received unrestricted grant support from Abbott, Amgen, Wyeth, Aventis, MSD, Schering-Plough/Centocor, BristolMyers Squibb, UCB, Roche and the Norwegian Directorate for Health and Social Affairs. The NOR-DMARD register is financially supported by pharmaceutical companies but the sponsors are not involved in the analyses and presentation of data.

Disclosure statement: The authors have declared no conflicts of interest.

References

1
Smolen
JS
Landewe
R
Breedveld
FC
, et al.  . 
EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs
Ann Rheum Dis
 , 
2010
, vol. 
69
 (pg. 
964
-
75
)
2
Quinn
MA
Conaghan
PG
Emery
P
The therapeutic approach of early intervention for rheumatoid arthritis: what is the evidence?
Rheumatology
 , 
2001
, vol. 
40
 (pg. 
1211
-
20
)
3
Suarez-Almazor
ME
Belseck
E
Shea
B
, et al.  . 
Sulfasalazine for rheumatoid arthritis
Cochrane Database Syst Rev
 , 
2000
, vol. 
2
 pg. 
CD000958
 
4
Suarez-Almazor
ME
Belseck
E
Shea
B
, et al.  . 
Methotrexate for rheumatoid arthritis
Cochrane Database Syst Rev
 , 
2000
, vol. 
2
 pg. 
CD000957
 
5
Gaujoux-Viala
C
Smolen
JS
Landewe
R
, et al.  . 
Current evidence for the management of rheumatoid arthritis with synthetic disease-modifying antirheumatic drugs: a systematic literature review informing the EULAR recommendations for the management of rheumatoid arthritis
Ann Rheum Dis
 , 
2010
, vol. 
69
 (pg. 
1004
-
9
)
6
Haagsma
CJ
van Riel
PL
de Jong
AJ
, et al.  . 
Combination of sulphasalazine and methotrexate versus the single components in early rheumatoid arthritis: a randomized, controlled, double-blind, 52 week clinical trial
Br J Rheumatol
 , 
1997
, vol. 
36
 (pg. 
1082
-
8
)
7
Dougados
M
Combe
B
Cantagrel
A
, et al.  . 
Combination therapy in early rheumatoid arthritis: a randomised, controlled, double blind 52 week clinical trial of sulphasalazine and methotrexate compared with the single components
Ann Rheum Dis
 , 
1999
, vol. 
58
 (pg. 
220
-
5
)
8
Hider
SL
Silman
A
Bunn
D
, et al.  . 
Comparing the long-term clinical outcome of treatment with methotrexate or sulfasalazine prescribed as the first disease-modifying antirheumatic drug in patients with inflammatory polyarthritis
Ann Rheum Dis
 , 
2006
, vol. 
65
 (pg. 
1449
-
55
)
9
Aletaha
D
Smolen
JS
The rheumatoid arthritis patient in the clinic: comparing more than 1,300 consecutive DMARD courses
Rheumatology
 , 
2002
, vol. 
41
 (pg. 
1367
-
74
)
10
Sokka
T
Increases in use of methotrexate since the 1980s
Clin Exp Rheumatol
 , 
2010
, vol. 
28
 (pg. 
S13
-
20
)
11
Welsing
PM
Fransen
J
van Riel
PL
Is the disease course of rheumatoid arthritis becoming milder? Time trends since 1985 in an inception cohort of early rheumatoid arthritis
Arthritis Rheum
 , 
2005
, vol. 
52
 (pg. 
2616
-
24
)
12
Saag
KG
Teng
GG
Patkar
NM
, et al.  . 
American College of Rheumatology 2008 recommendations for the use of nonbiologic and biologic disease-modifying antirheumatic drugs in rheumatoid arthritis
Arthritis Rheum
 , 
2008
, vol. 
59
 (pg. 
762
-
84
)
13
Pincus
T
Summey
JA
Soraci
SA
Jr
, et al.  . 
Assessment of patient satisfaction in activities of daily living using a modified Stanford Health Assessment Questionnaire
Arthritis Rheum
 , 
1983
, vol. 
26
 (pg. 
1346
-
53
)
14
Ware
JE
Jr
Sherbourne
CD
The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection
Med Care
 , 
1992
, vol. 
30
 (pg. 
473
-
83
)
15
Loge
JH
Kaasa
S
Hjermstad
MJ
, et al.  . 
Translation and performance of the Norwegian SF-36 Health Survey in patients with rheumatoid arthritis. I. Data quality, scaling assumptions, reliability, and construct validity
J Clin Epidemiol
 , 
1998
, vol. 
51
 (pg. 
1069
-
76
)
16
Brazier
J
Roberts
J
Deverill
M
The estimation of a preference-based measure of health from the SF-36
J Health Econ
 , 
2002
, vol. 
21
 (pg. 
271
-
92
)
17
Prevoo
ML
van't Hof
MA
Kuper
HH
, et al.  . 
Modified disease activity scores that include twenty-eight-joint counts. Development and validation in a prospective longitudinal study of patients with rheumatoid arthritis
Arthritis Rheum
 , 
1995
, vol. 
38
 (pg. 
44
-
8
)
18
Smolen
JS
Breedveld
FC
Schiff
MH
, et al.  . 
A simplified disease activity index for rheumatoid arthritis for use in clinical practice
Rheumatology
 , 
2003
, vol. 
42
 (pg. 
244
-
57
)
19
van Gestel
AM
Haagsma
CJ
van Riel
PL
Validation of rheumatoid arthritis improvement criteria that include simplified joint counts
Arthritis Rheum
 , 
1998
, vol. 
41
 (pg. 
1845
-
50
)
20
Felson
DT
Anderson
JJ
Boers
M
, et al.  . 
American College of Rheumatology. Preliminary definition of improvement in rheumatoid arthritis
Arthritis Rheum
 , 
1995
, vol. 
38
 (pg. 
727
-
35
)
21
Felson
DT
Smolen
JS
Wells
G
, et al.  . 
American College of Rheumatology/European League Against Rheumatism provisional definition of remission in rheumatoid arthritis for clinical trials
Arthritis Rheum
 , 
2011
, vol. 
63
 (pg. 
573
-
86
)
22
Felson
DT
Smolen
JS
Wells
G
, et al.  . 
American college of rheumatology/european league against rheumatism provisional definition of remission in rheumatoid arthritis for clinical trials
Ann Rheum Dis
 , 
2011
, vol. 
70
 (pg. 
404
-
13
)
23
Kristensen
LE
Saxne
T
Geborek
P
The LUNDEX, a new index of drug efficacy in clinical practice: results of a five-year observational study of treatment with infliximab and etanercept among rheumatoid arthritis patients in southern Sweden
Arthritis Rheum
 , 
2006
, vol. 
54
 (pg. 
600
-
6
)
24
Wiles
NJ
Lunt
M
Barrett
EM
, et al.  . 
Reduced disability at five years with early treatment of inflammatory polyarthritis: results from a large observational cohort, using propensity models to adjust for disease severity
Arthritis Rheum
 , 
2001
, vol. 
44
 (pg. 
1033
-
42
)
25
D'Agostino
RB
Jr
Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group
Stat Med
 , 
1998
, vol. 
17
 (pg. 
2265
-
81
)
26
Choi
HK
Hernan
MA
Seeger
JD
, et al.  . 
Methotrexate and mortality in patients with rheumatoid arthritis: a prospective study
Lancet
 , 
2002
, vol. 
359
 (pg. 
1173
-
7
)
27
Breedveld
FC
Weisman
MH
Kavanaugh
AF
, et al.  . 
The PREMIER study: A multicenter, randomized, double-blind clinical trial of combination therapy with adalimumab plus methotrexate versus methotrexate alone or adalimumab alone in patients with early, aggressive rheumatoid arthritis who had not had previous methotrexate treatment
Arthritis Rheum
 , 
2005
, vol. 
54
 (pg. 
26
-
37
)
28
Klareskog
L
van der
HD
de Jager
JP
, et al.  . 
Therapeutic effect of the combination of etanercept and methotrexate compared with each treatment alone in patients with rheumatoid arthritis: double-blind randomised controlled trial
Lancet
 , 
2004
, vol. 
363
 (pg. 
675
-
81
)
29
Edwards
JC
Szczepanski
L
Szechinski
J
, et al.  . 
Efficacy of B-cell-targeted therapy with rituximab in patients with rheumatoid arthritis
N Engl J Med
 , 
2004
, vol. 
350
 (pg. 
2572
-
81
)
30
Maini
RN
Taylor
PC
Szechinski
J
, et al.  . 
Double-blind randomized controlled clinical trial of the interleukin-6 receptor antagonist, tocilizumab, in European patients with rheumatoid arthritis who had an incomplete response to methotrexate
Arthritis Rheum
 , 
2006
, vol. 
54
 (pg. 
2817
-
29
)
31
Aletaha
D
Stamm
T
Kapral
T
, et al.  . 
Survival and effectiveness of leflunomide compared with methotrexate and sulfasalazine in rheumatoid arthritis: a matched observational study
Ann Rheum Dis
 , 
2003
, vol. 
62
 (pg. 
944
-
51
)
32
Maetzel
A
Wong
A
Strand
V
, et al.  . 
Meta-analysis of treatment termination rates among rheumatoid arthritis patients receiving disease-modifying anti-rheumatic drugs
Rheumatology
 , 
2000
, vol. 
39
 (pg. 
975
-
81
)
33
Papadopoulos
NG
Alamanos
Y
Papadopoulos
IA
, et al.  . 
Disease modifying antirheumatic drugs in early rheumatoid arthritis: a longterm observational study
J Rheumatol
 , 
2002
, vol. 
29
 (pg. 
261
-
6
)
34
Smolen
JS
Aletaha
D
Bijlsma
JW
, et al.  . 
Treating rheumatoid arthritis to target: recommendations of an international task force
Ann Rheum Dis
 , 
2010
, vol. 
69
 (pg. 
631
-
7
)
35
Rachapalli
SM
Williams
R
Walsh
DA
, et al.  . 
First-line DMARD choice in early rheumatoid arthritis–do prognostic factors play a role?
Rheumatology
 , 
2010
, vol. 
49
 (pg. 
1267
-
71
)
36
Arnett
FC
Edworthy
SM
Bloch
DA
, et al.  . 
The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis
Arthritis Rheum
 , 
1988
, vol. 
31
 (pg. 
315
-
24
)
37
Aletaha
D
Neogi
T
Silman
AJ
, et al.  . 
2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative
Ann Rheum Dis
 , 
2010
, vol. 
69
 (pg. 
1580
-
8
)
38
Visser
K
Katchamart
W
Loza
E
, et al.  . 
Multinational evidence-based recommendations for the use of methotrexate in rheumatic disorders with a focus on rheumatoid arthritis: integrating systematic literature research and expert opinion of a broad international panel of rheumatologists in the 3E Initiative
Ann Rheum Dis
 , 
2009
, vol. 
68
 (pg. 
1086
-
93
)
39
Boers
M
Missing data in trials: do we have to keep carrying the last observation forward?
Arthritis Rheum
 , 
2008
, vol. 
59
 (pg. 
2
-
3
)
40
Smolen
JS
Kalden
JR
Scott
DL
, et al.  . 
Efficacy and safety of leflunomide compared with placebo and sulphasalazine in active rheumatoid arthritis: a double-blind, randomised, multicentre trial. European Leflunomide Study Group
Lancet
 , 
1999
, vol. 
353
 (pg. 
259
-
66
)
41
Choy
EH
Scott
DL
Kingsley
GH
, et al.  . 
Treating rheumatoid arthritis early with disease modifying drugs reduces joint damage: a randomised double blind trial of sulphasalazine vs diclofenac sodium
Clin Exp Rheumatol
 , 
2002
, vol. 
20
 (pg. 
351
-
8
)
42
Strand
V
Cohen
S
Schiff
M
, et al.  . 
Treatment of active rheumatoid arthritis with leflunomide compared with placebo and methotrexate. Leflunomide Rheumatoid Arthritis Investigators Group
Arch Intern Med
 , 
1999
, vol. 
159
 (pg. 
2542
-
50
)
43
Pincus
T
Ferraccioli
G
Sokka
T
, et al.  . 
Evidence from clinical trials and long-term observational studies that disease-modifying anti-rheumatic drugs slow radiographic progression in rheumatoid arthritis: updating a 1983 review
Rheumatology
 , 
2002
, vol. 
41
 (pg. 
1346
-
56
)

Supplementary data

Comments

0 Comments