Article Text

Download PDFPDF

Are poor prognostic factors a realistic basis for treatment decisions in patients with rheumatoid arthritis? Lessons from the IMPROVED study
  1. Joy Ardjuna van der Pol,
  2. Cornelia F Allaart,
  3. Tom W J Huizinga and
  4. Sytske Anne Bergstra
  1. Rheumatology, Leiden University Medical Center, Leiden, The Netherlands
  1. Correspondence to Joy Ardjuna van der Pol; j.a.van_der_pol{at}

Statistics from

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.

The European Alliance of Associations for Rheumatology (EULAR) recommendations for the treatment of rheumatoid arthritis (RA) recommend the use of poor prognostic factors (PPFs) in treatment decisions.1 PPFs are: (1) persistent moderate/high disease activity after conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), (2) high C reactive protein (CRP)/erythrocyte sedimentation rate (ESR) levels at baseline, (3) high swollen joint counts (SJC) at baseline, (4) presence of Rheumatoid Factor (RF) and/or Anti-Citrullinated Protein Antibodies (ACPA), especially at high levels, (5) presence of early erosions or (6) failure of two or more csDMARDs.2 PPFs are associated with an increased risk of radiographic damage over time. It is recommended, if a treatment target is not achieved with the first csDMARD strategy, to treat patients without PPFs with a second csDMARD, but to switch to treatment with a biological DMARD (bDMARD) if PPFs are present. These considerations are based on expert opinion, and an investigation into the association between stratification and treatment outcomes is recommended.2 We aim to explore the presence of PPFs in early RA and to investigate their relationship with treatment response after failure of the first csDMARD strategy.

In the IMPROVED study, 479 patients with RA (2010 criteria) started treatment with MTX and prednisone bridging. Patients not in early remission (Disease Activity Score (DAS) ≥1.6) after 4 months were randomised to Methotrexate (MTX)+Hydroxychloroquine (HCQ)+ Sulfasalazine (SSZ) + continued low-dose prednisone (arm 1) or to MTX+adalimumab (arm 2).3 In these patients, we performed a post hoc analysis, aiming to compare the clinical response on either csDMARD combination or bDMARD therapy in relation to the presence or absence of PPFs.

We found that 98–99% of patients, depending on the definition of ‘high swollen joint count’ (>4 or >10), had PPFs. The presence of each PPF was equally distributed among randomisation arms (figure 1A,B). The presence of antibodies and high CRP/ESR were the most frequent PPFs (present in 366 (76%) and 390 (81%) patients, respectively). Only 10 patients (2%) had no PPFs at all in the main analysis (with ‘high SJC’ defined as >10) and only 5 (1%) in the sensitivity analysis (with SJC >4). Of these, 2/10 and 2/5 patients did not achieve early remission and were randomised to the two treatment arms. Thus, a comparison of treatment response in patients without PPFs was unrealistic. In the PPF-positive patients in both randomisation arms, we found similar outcomes for mean DAS, mean HAQ and damage progression over time.

Figure 1

(A and B) Erosions: presence of baseline erosions, DAS>2.4: moderate/high DAS after 4 months of treatment with MTX+prednisone bridging, CRP/ESR: high C reactive protein or erythrocyte sedimentation rate at baseline, SJC-4: presence of >4 swollen joints at baseline, SJC-10: presence of >10 swollen joints at baseline, antibodies: presence of ACPA or RF at baseline. Patients can be present in multiple categories; numbers do not add up to 479 patients with rheumatoid arthritis in (A). ACPA, anti-citrullinated protein antibodies; CRP, C reactive protein; DAS, disease activity score; ESR, erythrocyte sedimentation rate; MTX, methotrexate; RF, rheumatoid factor; SJC, swollen joint count.

In retrospect, the high percentage of PPF in our population is not surprising, due to a large overlap between the 2010 classification criteria for RA and PPFs. Both the classification criteria and PPFs aimed to identify patients deemed to be ‘at risk for damage progression’. Thus, patients who meet the 2010 classification criteria will have at least one PPF, as demonstrated in our cohort. To validate these numbers, we assessed the presence of PPFs in the BeSt study. In this study, 508 patients with RA (1987 criteria positive) were treated to target low disease activity and followed for 10 years.4 Of these 508 patients, 492 also fulfilled the 2010 classification criteria for RA, and all of them were PPF positive (489 even had >1 PPF). This would suggest that almost all classifiable patients with RA should proceed to bDMARDs and not more csDMARDs as the next treatment step (after insufficient response to the first csDMARD treatment). Yet, the IMPROVED data suggest that first adding sulfasalazine and hydroxychloroquine to methotrexate and continuing this combination for 4 months with prednisone 7.5 mg/daily was as effective as starting a bDMARD as a second treatment in patients with PPFs. More fundamentally, previous results of the IMPROVED study showed that after 5 years of continued DAS-remission-steered treatment adjustments, radiological damage was prevented in virtually all patients.5 In univariable logistic regression models, only baseline erosions were associated with damage progression at year 5. In multivariable models, this was no longer significant. In patients with PPFs, there was a non-significant increase in the risk of progression as the number of PPFs increased (using 1 PPF as the reference category, OR 1.31 (95% CI 0.68 to 2.53) for 2 PPFs, 1.51 (0.73 to 3.14) for 3 PPFs and 1.89 (0.59 to 6.06) for 4 PPFs. Only one patient had all 5 PPFs but experienced no progression at year 5. The p values for these associations ranged between 0.26 and 0.43 depending on the amount of PPFs). However, it should be noted that patients from a randomised trial do not always represent the general patient population. In addition, the DAS-remission-steered treatment may not be implemented as strictly in daily practice as in trials and the initial dose of prednisone (60 mg) is higher than is usually implemented presently, which may reduce the generalisability of our PPF rates and detected outcomes. Nevertheless, other outcomes than ‘radiographic damage’ as 'poor’ outcome for prognostic factors and treatment choices may be reconsidered, such as quality of life, absence of remission, failure of multiple DMARDs, pain, physical functioning, participation (ie, work, social), long-term symptoms and emotional well-being.6 Thereafter, new PPFs should be identified and tested in order to allow new, tailored, treatment recommendations.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by CCMO: NL15233.058.06 (countrywide)METC LDD: P06.210 (Leiden region). Participants gave informed consent to participate in the study before taking part.


We thank all patients for their participation in our trial, contributing to improvements in clinical care.



  • Contributors JAvdP did the analyses and wrote the paper. CFA was the principal investigator of the IMPROVED trial. TWJH contributed patients to the IMPROVED trial. SAB co-authored the paper and is the guarantor of the paper. All authors critically assessed the paper and read and approved the final manuscript.

  • Funding The IMPROVED study was financially supported by AbbVie in the first year.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.