Cotreatment with methotrexate in routine care patients with rheumatoid arthritis receiving biological treatment yields better outcomes over time

This study adds to the knowledge that cotreatment with MTX in routine care patients with RA treated with biologicals lowers 28-joint Disease Activity Score (DAS28) over time and improves the likelihood to be in DAS28 remission.How might this impact on clinical practice?


Still, there is evidence that many rheumatologists do not adhere to a strategy of MTX added to bDMARD in daily clinical practice. There are several reasons to explain non-adherence, such as poor tolerability and emerging contraindications, but this points to a discrepancy between clinical trials and usual practice. In fact, while RCT data on the efficacy of MTX added to bDMARD therapy are convincingly positive, evidence of such an effect in daily clinical practice is still scarce.Patients and methodsConsecutive adult patients with a clinical diagnosis of RA according to their treating rheumatologist that also fulfilled the ACR 1987 classification criteria13 were treated with conventional synthetic DMARD and/or bDMARDs. These patients were included in this prospective observational study that was executed in a large rheumatology centre in the Netherlands between April 2013 and April 2016. The local medical ethics committee reviewed and approved the study protocol. All patients provided informed consent before inclusion and were treated according to the routine care schedule.Patient’s age, gender, disease duration (in years) and the number of previous bDMARD failures (if any) were assessed at baseline. Data on disease activity measures, laboratory parameters and medication status were collected during routine visits every 3 months by rheumatologists and research nurses. A 28-joint Disease Activity Score (DAS28) based on four variables (tender joint count (TJC), swollen joint count (SJC), patient global assessment (PGA), erythrocyte sedimentation rate (ESR) based (DAS28-ESR)) and Routine Assessment of Patient Index Data 3 (RAPID3) were used in this study to reflect daily clinical practice.14 15 RAPID3 is a patient-driven questionnaire for RA outcome reporting on functioning (0–10), pain (0–10) and global health (0–10). RAPID3 (0–30) scores of >12, 6.1–12, 3.1–6 and ≤3 give high, moderate and low disease activity and remission, respectively.

Prescription of comedication was used as a proxy for the presence of comorbidities and was collected using the Anatomical Therapeutic Chemical (ATC) Classification System for the most frequently reported indications,16 namely cardiovascular diseases (ATC C*), type 2 diabetes mellitus (ATC C10*), chronic obstructive pulmonary disease and asthma (ATC R03*). The effect of combination treatment with MTX and bDMARDs as compared with bDMARD monotherapy on disease activity measures over time was tested in longitudinal linear/binomial (depending on the outcome) generalised estimating equation models with both the treatment and the outcome modelled as time-varying variables. Treatment was added to the models as a ‘dummy’ variable reflecting the four options encountered in clinical practice: (1) bDMARD+MTX; (2) bDMARD alone (reference); (3) MTX alone; and (4) none, but analyses focused on the comparison of interest (ie, combination treatment vs bDMARD monotherapy). The exchangeable ‘working’ correlation structure was used to take into account the repeated measures over time within patients. Disease activity was modelled in three complementary ways: (1) DAS28 and RAPID3 as continuous variables; (2) DAS28 and RAPID3 as binary definitions of remission (ie, <2.6 and ≤3, respectively); and (3) individual DAS28 components: that is, TJC (0–28), SJC (0–28), PGA (0–10) and ESR (mm/hour) as continuous variables. The effects of treatment on each outcome were tested in separate models, first with no adjustment (‘crude model’), and second after adjustment for variables a priori selected on clinical grounds as potential confounders (age (years), gender, use of drugs for comorbidities (yes/no), oral glucocorticoids (yes/no) and use of non-steroidal anti-inflammatory drugs (NSAID; yes/no)). Variables that vary over time (eg, treatment with NSAIDs) were modelled as such. All analyses were done with Stata V.15.Patients and follow-upTable 1Longitudinal effect of combination therapy versus bDMARD monotherapy on disease activity over timeTable 2

Figure 1
Figure 1
Adjusted estimates of DAS28 (95% CI) and RAPID3 (95% CI) scores over time based on multivariate models a priori adjusted for possible confounders: age, gender, drugs for comorbidities (1/0), oral glucocorticosteroids (1/0) and use of non-steroidal anti-inflammatory drugs. *Number of patients per visit. bDMARD, biological disease-modifying antirheumatic drug; DAS28, 28-joint Disease Activity Score; MTX, methotrexate; RAPID3 (0–30), Routine Assessment of Patient Index Data (three domains: functioning, pain, patient global)

This benefit of MTX comedication was also seen with regard to the likelihood of achieving DAS28 remission, which was on average 55% higher with combination treatment. When looking into DAS28 individual components, the added benefit of MTX could only be demonstrated for the SJC but not for the ESR, TJC and PGA. RAPID3, both as a continuous variable and as a binary definition of remission, was not significantly different between the two different treatment strategies (table 2).


Previous observational studies assessing the possible benefits/hazards of maintaining combination therapy in RA have yielded conflicting results. Gabay et al have shown that on average DAS28 was lower with combination treatment compared with bDMARD monotherapy, up to 4 years of follow-up. However, the same study has also shown no meaningful differences between the two strategies on the likelihood of being in DAS28 remission. Listing et al have previously shown a comparable result for combination versus monotherapy in a prospective study from the German biologics registerR. Other studies are less informative due to their cross-sectional design.  Our study includes several paired measurements of DAS28 and RAPID3 collected over time in the same patients, and allows a better estimation of the long-term benefit of bDMARD therapy plus MTX comedication versus bDMARD therapy alone.  Of note, the positive effects of MTX comedication on disease activity measures were still present when taking into account the patients’ comorbidities, age, treatment with NSAIDs and oral glucocorticoids (ie, factors that might confound the association of interest). This finding further argues in favour of combination treatment even in more complex patients with comorbidities and long disease duration.Interestingly, the effect of combination therapy on disease activity was only detected when using DAS28, but not RAPID3. Also, when looking at the various individual components of DAS28, the positive effect was only seen on the number of swollen joints (ie, an objective measure). This finding suggests that objective clinical outcome measures have merits in demonstrating treatment effects in daily clinical practice, and that only measuring subjective, patient-reported aspects of the disease, as is done with RAPID3, does not suffice despite strong messages by advocates.Our study has some limitations worth noticing. The most important one is the expected a priori prognostic dissimilarity that may drive differences between the two strategies (confounding by indication).

Among others, patients who are on MTX since they tolerate it may have better prognosis and thus better response to therapy. Tapering or discontinuation of bDMARD therapy in these patients could be considered. However, this is a common issue of all observational studies, that is, increase external validity (patients from clinical practice in whom, for example, comorbidities may have impaired cotreatment with MTX), at the cost of the internal validity (lack of random and blinded treatment allocation). Second, residual confounding cannot be completely ruled out, nevertheless after the adjustment for, arguably, the most important confounders the effect of MTX was still consistently seen in all models.In conclusion, our results show patients with RA on biological therapy who have maintained their MTX dose have lower disease activity because they have less swollen joints.


We are grateful to the research nurses and medical doctors of the Zuyderland Medical Centre for seeing patients and collecting the data used in this study.References


From the BMJ