Arthritis & Rheumatism, Volume 63,
November 2011 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Chicago, Illinois November 4-9, 2011.

A New Definition of Treatment Response in Rheumatoid Arthritis: Identification of the Critical Difference in Disease Activity.

Behrens1,  Frank, Koehm1,  Michaela, Scharbatke2,  Eva C., Kleinert2,  Stefan, Weyer3,  Geerd, Alten4,  Rieke, Tony2,  Hans Peter

CIRI/Rheumatology, J.W. Goethe-University, Frankfurt/Main, Germany
Rheumatology, University of Würzburg, Würzburg, Germany
Biostatistics, ICRC, Berlin, Germany
Rheumatology Schlossparkklinik, Berlin, Germany


Situational effects and measurement errors inducing fluctuations in disease activity measurements complicate the evaluation of a clinically meaningful therapeutic response in rheumatoid arthritis (RA). To itemize this complexity, a statistical approach was used to determine a critical difference (dcrit) defining valid criterion for clinical response as assessed by the Disease Activity Score-28 joints (DAS28).


The population comprised a total number of 728 RA patients with stable response to DMARD, steroid or biological therapy (including TNF-inhibitors and Rituximab) derived from three different clinics in Germany (University Würzburg (n=50), University Frankfurt (n=51), Berlin (n=50)) and from a prospective observational study with Adalimumab. Patients were included with stable therapy and disease course from months 12 to 24 after therapy initiation. To evaluate changes in DAS28 score, DAS28 scores at 12, 18 and 24 months were subjected to an ANOVA model to determine the error of measurement which was used to establish a 95% one sided confidence interval (95% CI) for decrease occurring by chance within the range of normal fluctuations. The limit of the confidence region defined the critical difference (dcrit) in disease activity for a reliable change in a single patient.


The overall dcrit value in the 728 patient populations was detected at 1.75. Values for dcrit were comparable in all evaluated subgroups, regardless of treatment centre, class of therapy (DMARDs or biologics including TNF-inhibitors and Rituximab), or baseline disease activity (Table).

 All patientsDAS28 M12Class of therapyAgeGender
  <=3.2>3.2DMARDsBiologics<=60 years>60 yearsmalefemale


Based on our data, a dcrit value of 1.8 (DAS28 improvement of 1.8 points) signifies an individual therapeutic response that exceeds the threshold of random fluctuation. The dcrit value determined by statistical analysis of expected variation in DAS28 scores higher than the DAS28 change required to achieve a good EULAR response (1.2 points) and is independent of baseline activity, which may make it more convenient for clinical use. Further studies in larger populations will be required to confirm the utility of the dcrit value in determining therapeutic response. However, our data suggest that a dcrit value of 1.8 has the potential to guide treatment decisions in daily clinical practice and to facilitate research on treatment responders.

To cite this abstract, please use the following information:
Behrens, Frank, Koehm, Michaela, Scharbatke, Eva C., Kleinert, Stefan, Weyer, Geerd, Alten, Rieke, et al; A New Definition of Treatment Response in Rheumatoid Arthritis: Identification of the Critical Difference in Disease Activity. [abstract]. Arthritis Rheum 2011;63 Suppl 10 :363

Abstract Supplement

Meeting Menu