You are here
Evaluating disease activity in rheumatoid arthritis: Which composite index is best? A systematic literature analysis of studies comparing the psychometric properties of the DAS, DAS28, SDAI and CDAI
Joint Bone Spine, Volume 79, Issue 2, March 2012, Pages 149 - 155
To evaluate and compare four composite indices for assessing the activity of rheumatoid arthritis (RA).
We conducted a systematic literature review by searching Medline via PubMed and Embase and Cochrane databases for articles published up to March 2009. We selected studies that directly compared at least two of the four composite indices. The DAS (Disease Activity Score), DAS28, Simplified Disease Activity Index (SDAI) and Clinical Disease Activity Index (CDAI) were evaluated in terms of reproducibility, construct validity, discriminative performance, and sensitivity to change.
We included 61 articles. The only study that directly compared the intraobserver reproducibility of the DAS28, SDAI, and CDAI found comparable intraclass correlation coefficients ranging from 0.85 to 0.89. Concordance among indices was good (kappa values of ∼0.7), except between the DAS28 and the other indices in definition of remission (kappa 0.48–0.63). The indices had good construct validity by their similar fair-to-good correlations with the Health Assessment Questionnaire (HAQ) score and structural damage. Discriminative performance was comparable and satisfactory for treatment changes or remission according to the American College of Rheumatology (ACR). Two studies evaluated the sensitivity to change of the SDAI and CDAI; both indices detected a difference between responders and non-responders according to ACR definitions.
The DAS, DAS28, SDAI, and CDAI are valid tools for evaluating the activity of RA. The DAS28 is less conservative in defining remission than are the other three indices. Longitudinal studies of individual patients are needed to confirm these results.
Keywords: Composite indices, Disease activity, Rheumatoid arthritis, Metrologic properties, Reliability, Validity, Responsiveness.
Rheumatoid arthritis (RA) is an inflammatory joint disease that causes structural joint damage, incapacitation, and alterations in quality of life. These adverse consequences can be prevented, at least in part, by early appropriate therapy, in particular a “tight control” strategy . Such strategy necessitates evaluating disease activity and response to treatment with objective and standardized tools. In contrast to other diseases such as hypertension and diabetes, the activity of RA cannot be evaluated by a single clinical or laboratory measurement. Therefore, in the 1990s, composite indices based on several clinical and/or laboratory variables were developed. These indices reduce measurement error, allow for objective evaluation of patients seen in everyday practice, and benefit the analysis and interpretation of clinical trials of new, potential disease-modifying drugs.
The first composite index developed for assessing disease activity in RA patients was the Disease Activity Score (DAS or DAS44) . Then, another group developed a simplified version based on 28 joints instead of 44, the DAS28, which is currently the most widely used index . However, the DAS44 and DAS28 are based on complex equations, and their determination requires a dedicated software program or calculator. This drawback led to the development of simpler indices, the Simplified Disease Activity Index (SDAI)  and its modified version, the Clinical Disease Activity Index (CDAI) .
Which index to use in clinical practice or in clinical research is difficult to determine. One way to compare the indices is to compare their psychometric properties, as defined by the OMERACT filter . The OMERACT filter checks that a potential outcome measure is truthful (reflects what it is supposed to reflect), reliable (reproducible), and discriminant (sensitive to change, over time, and between different severity stages). The composite indices cited above have met the filter requirements. However, for true evaluation, they should be compared with each other.
Thus, the aim of this study was to perform a systematic literature review of studies comparing the DAS, DAS28, SDAI and CDAI in terms of reproducibility, concordance, construct validity, discriminative performance, and sensitivity to change.
We conducted a systematic literature review by searching Medline via PubMed and the Embase and Cochrane databases for articles published in English or French up to March 2009, without date limits, using the following terms: (DAS [ti] OR SDAI [ti] OR CDAI [ti] OR disease activity score [ti] OR disease activity index [ti]) AND arthritis, rheumatoid [MeSH]. We also searched meeting abstracts of the American College of Rheumatology (ACR), European League Against Rheumatism, and French Society of Rheumatology for the past two years. In addition, we hand-searched reference lists of papers initially obtained to identify additional relevant reports. The articles were initially selected on the basis of their titles and abstract, then on the full texts. The inclusion criteria were all studies that directly compared, for RA patients, at least two of the four indices, DAS, DAS28, SDAI, and CDAI, in terms of reproducibility, concordance, construct validity, discriminative performance, and sensitivity to change.
3. Data collection
One investigator (CGV) selected the articles and collected the data using a predetermined form. The following data were collected: type of study, number of patients, and rheumatoid arthritis features.
The outcomes were reproducibility (interobserver, intraobserver), concordance, construct validity, discriminative performance, and sensitivity to change (after an effective therapy).
4. Statistical analysis
Interobserver and intraobserver reliability was estimated by calculating intraclass correlation coefficients (ICCs), unweighted kappa statistics, and overall agreement (defined as the percentage of observed exact agreements).
Concordance was the ability of the different indices to classify patients in the same activity grades or in the same response to treatment grades.
4.3. Construct validity
Construct validity was achieved when measures agreed with other measures that evaluate the same phenomenon. We compared the composite index score with other clinical data by correlation (Spearman's rank correlation test) for this variable.
4.4. Discriminative performance
Discriminative performance was the ability of the different indices to discriminate between patients in different activity grades or with different response to treatment or between patients whose treatment was changed or not. This variable was calculated by analysis of the area under the receiver-operating curve (AUC).
4.5. Sensitivity to change
Responsiveness of a composite index score is its ability to change after an intervention. Sensitivity to change was estimated by the effect size (ES, difference of the mean/SD baseline) or the standardized response mean (SRM, mean change/SD of the change).
5.1. Article selection
Our systematic literature search identified 102 articles, of which 50 were excluded on the basis of their titles and abstracts. Finally, 61 articles were included (Fig. 1). Among them, 17 articles met the inclusion criteria and were analyzed.
Flowchart of articles.
5.2. Description of the composite indices
The DAS44 was developed in 1990 in a prospective study of up to three years’ duration of 113 patients with early RA (duration less than one year) . The disease was considered active if a disease-modifying antirheumatic drug (DMARD) was started or stopped because of inadequate effectiveness. Disease activity was considered low if any of the following occurred: DMARD discontinuation because remission was achieved, continued DMARD therapy with no change for more than one year, or no DMARD therapy for more than one year. The DAS44 includes evaluation of the Ritchie articular index (0–78), swollen joint count among 44 joints (SJC44), erythrocyte sedimentation rate (ESR), and general health (0–100 visual analog scale [VAS]). The DAS44 is computed by the following equation:
where √ indicates square root and ln, naperian logarithm.
The DAS44 can range from 0.23 to 9.87, and the values are normally distributed. High disease activity is defined as a DAS44 greater than 3.7, moderate activity 2.4–3.7, low activity as a DAS44 less or equal to 2.4 and greater or equal to 1.6, and remission as a DAS44 less than 1.6.
The DAS28 was developed with a method similar to that used for the DAS44 . The DAS28 includes evaluation of swollen joint count among 28 joints (SJC28), tender joint count among the same 28 joints (TJC28), ESR, and general health (0–100 VAS). The DAS28 is computed by the following equation:
The DAS28 can range from 0.49 to 9.07, and the values are normally distributed. High disease activity is defined as a DAS28 greater than 5.1, moderate activity as a DAS28 greater than 3.2 and less or equal to 5.1, low activity as a DAS28 less or equal to 3.2 and greater than 2.6, and remission as a DAS28 less than 2.6.
The SDAI was published in 2003 to provide a simpler tool than the DAS . The SDAI is obtained by the algebraic sum of five factors: TJC28 + SJC28 + C-reactive protein (CRP) level + overall disease activity on a 0–10 VAS completed by the patient + overall disease activity on a 0–10 VAS completed by the physician. Thus, no calculator is needed. The values can range from 0.1 to 86 and are not normally distributed .
The disease activity criteria were modified in 2005. High disease activity is defined as a SDAI greater than 26, moderate activity as a SDAI greater than 11 and less or equal to 26, low activity as a SDAI less or equal to 11 and greater than 3.3, and remission as a SDAI less or equal to 3.3 .
The CDAI, derived from the SDAI, was published in 2005 . It uses the same factors as the SDAI, except for CRP level: thus, the CDAI is the sum of TJC28 + SJC28 + overall disease activity on a 0–10 VAS completed by the patient + overall disease activity on a 0–10 VAS completed by the physician. No laboratory tests are needed to determine the CDAI. The values can range from 0 to 76. High disease activity is defined as a CDAI greater than 22, moderate activity as a CDAI greater than 10 and less or equal to 22, low activity as a CDAI less or equal to 10 and greater than 2.8, and remission as a CDAI less or equal to 2.8.
5.3. Comparison of metrologic properties
Surprisingly few data are available on reproducibility (Table 1). In the seminal study of the DAS44 , test-retest results showed good correlation, 0.89. Intraobserver reproducibility of the DAS28 was evaluated in two studies, which showed similar ICC values, 0.82 and 0.89  and .
Reproducibility of the four composite indices.
No of patients
|Van der Heijde 1990 ||Longitudinal study
113 RA patients
|DAS||Test-retest correlation = 0.89|
|Vrijhoef 2003 ||Longitudinal study
56 RA patients
|DAS28||Intraobserver reproducibility: ICC = 0.82 (95% CI = 0.72–0.89)|
|Walsh 2008 ||Longitudinal study
12 RA patients
|DAS28||Intraobserver reproducibility: ICC = 0.89
Interobserver reproducibility: ICC = 0.92
|Uhlig 2008 ||Longitudinal study
28 patients with stable RA
ICC DAS28 = 0.85
ICC SDAI= 0.88
ICC CDAI= 0.89
RA: rheumatoid arthritis; ICC: intraclass correlation coefficient; 95% CI: 95% confidence interval; DAS: Disease Activity Index in 44 joints; DAS28: DAS in 28 joints; SDAI: Simplified Disease Activity Index; CDAI: Clinical Disease Activity Index.
The only study that directly compared reproducibility focused on the intraobserver reproducibility of the DAS28, SDAI, and CDAI . ICC values were excellent (0.85 for the DAS28, 0.88 for the SDAI, and 0.89 for the CDAI) and seemed comparable, although this last point remains doubtful because the confidence intervals were not reported. Oddly enough, interobserver reproducibility was evaluated only for the DAS28 and in a single study . The result was excellent, with an ICC of 0.92.
The correlation between the DAS28 and the SDAI was evaluated in five studies , , , , and  and found to be very good, with correlation coefficients (r) ranging from 0.80 and 0.91. Three studies assessed correlations between the DAS28 and the CDAI. Again, the results were very good, with r values ranging from 0.89 to 0.93 , , and .
The concordance between the DAS44 and the DAS28 was evaluated in three studies (Table 2). The first found moderate to good concordance in terms of response to treatment, with agreement between response to treatment defined using the DAS44 and the DAS28 in 77% of 223 patients (kappa statistics not provided) . In a similar study, the treatment response was concordant for 86.7% of 105 patients, and the DAS28 classified responses more conservatively than did the DAS44. In the same study, both response definitions were significantly associated with Health Assessment Questionnaire (HAQ) score changes and radiological evidence of disease progression . Remission classification with the DAS44 and DAS28 was compared for data from the COBRA randomized controlled trial . Of 620 observations in 155 patients, 168 showed remission according to one or both indices, and the discordance rate was 14.7%. Among patients classified as being in remission by the DAS28, 55% were not in remission according to the DAS44. Involvement of the ankles and/or feet was noted in 51% of discordant observations as compared with only 18% of concordant observations (P < 0.0005).
Concordance among the four composite indices.
|Index studied||Comparator index||Study||No of patients||Results|
|DAS28||DAS||Ranganath ||223 with early RA||Response to treatment: concordance = 77%|
|Van Gestel ||105 with early RA||Response to treatment: concordance = 86.7%|
|Landewé ||620 observations in 155 patients||55% of patients in remission according to the DAS28 were not in remission according to the DAS44|
|SDAIa||Aletaha ||767||Disease activity grades: k = 0.70|
|Ranganath ||223 with early RA||Response to treatment: concordance = 80%|
|Khanna ||200||Remission: k = 0.48
Low activity: k = 0.68
|Mierau ||621||Remission: k = 0.63|
|CDAI||Aletaha ||767||k = 0.79|
|Ranganath ||223 with early RA||Response to treatment: concordance = 74%|
|Khanna ||200||Remission: k = 0.52
Low activity: k= 0.67
|Mierau ||621||Remission: k = 0.58|
|Shaver ||849||Remission DAS 28 = 28.5%
Remission CDAI = 6.5%
|DAS||SDAIa||Ranganath ||223 with early RA||Response to treatment: concordance = 80%|
|CDAI||Ranganath ||223 with early RA||Response to treatment: concordance = 74%|
|SDAI||CDAI||Khanna ||200||Remission: k = 0.97|
a The latest SDAI categories were used .
k: kappa coefficient.
The concordance between the DAS28 and SDAI was assessed in four studies (Table 2). Overall, the concordance was satisfactory, with 80% of patients being classified in the same groups according to EULAR response  and, in one study, a kappa coefficient of 0.7 for classification of the patients in three levels of disease activity based on experts’ opinion . Concordance for classifying patients in a low disease activity state was good, with a kappa coefficient of 0.68 . The concordance for classifying patients in remission was moderate in two studies (kappa values of 0.48 and 0.63)  and , the percentages of patients classified as being in remission were significantly greater with the DAS28 (15% and 42.7%) than the SDAI (11% and 33.5%) (P < 0.01)  and .
Five studies evaluated the concordance between the DAS28 and CDAI (Table 2). Among them, two found satisfying results, with 74% of patients being classified in the same groups according to EULAR response  and a kappa value of 0.79 for classification of the patients in the tenth percentile of disease activity . Concordance for low disease activity was also satisfactory (kappa value of 0.67) . In contrast, the kappa values for remission were only 0.52 and 0.58  and . In another study, the proportion of patients classified as being in remission was 28.5% with the DAS28 and 6.5% with the CDAI .
The concordance between the DAS and SDAI was assessed in one study and was satisfactory, with 80% of patients being classified in the same groups of response to treatment . The same study evaluated the concordance between the DAS and CDAI with satisfying results; 74% of patients were classified in the same groups of response to treatment .
As expected, concordance for remission at six months was excellent (kappa 0.97) between the SDAI and CDAI in a longitudinal study of 200 patients .
5.3.3. Construct validity
The correlations between the HAQ score and the DAS28, SDAI, and CDAI were evaluated in a cohort of patients with established RA  (Table 3). All three-disease activity scores showed similar fair correlations with the HAQ score (r = 0.45–0.47). In the same study, the correlations were weaker in patients with early RA but again similar (r = 0.26–0.31) . A study of the pooled results from three randomized controlled trials showed fair-to-good correlations between the SDAI and HAQ score, at baseline and after six months . The correlations between the 0- to 6-month changes in the HAQ score and changes in the SDAI and CDAI were similar .
Construct validity of the four indices.
No of patients
|Score||Correlation with the DAS28||Correlation with the HAQ score||Correlation with structural damage|
|Aletaha 2005 ||Cross-sectional: 767||SDAI||r = 0.91|
|Leeb 2005 ||Cross-sectional: 399||SDAI||Cross-sectional: r = 0.90|
115 (1 year)
r at 3–6 M = 0.90
r at 6–9 M = 0.85
r at 9–12 M = 0.84
|Vander Cruyssen 2005 ||RCT: 474||SDAI
|r = 0.91
r = 0.89
|Smolen 2003 ||3 RCTs: 399||SDAI||D0: r1 = 0.91
r2 = 0.91
r3 = 0.80
M6: r1 = 0.93
r2 = 0.91
r3 = 0.91
|D0: r1 = 0.46
r2 = 0.36
r3 = 0.44a
M6: r1 = 0.63
r2 = 0.53
r3 = 0.66a
r1 = 0.56
r2 = 0.57
r3 = 0.48a
r1 = 0.56
r2 = 0.47
r3 = 0.56a
|Aletaha 2005 ||Cross-sectional: 767; 106 early RA (inception cohort)||SDAI||r = 0.91 (all) and 0.90 (early RA)*||r = 0.46 (all) and 0.31 (early RA)||r = 0.59|
|CDAI||r = 0.89 (all) and 0.89 (early RA)||r = 0.45 (all) and 0.30 early RA)||r = 0.54|
|DAS28||r = 0.47 (all) and 0.26 (early RA)||r = 0.58
(Δ Larsen over 3 years)
a modified Health Assessment Questionnaire.
HAQ: Health Assessment Questionnaire; D: day; M: month; RCT: randomized controlled trial.
All r values are statistically significant: P < 0.05.
Finally, the mean DAS28, SDAI, and CDAI values over three years of follow-up in 56 patients showed comparable fair correlations with changes in radiographic score over the same period .
5.3.4. Discriminative performance
Two studies investigated whether the SDAI, CDAI, and DAS28 discriminated between patients with or without treatment change  and . Both studies used the AUC and demonstrated good discriminant capability of the three indices. In the first study, the performance of the indices in discriminating patients with and without an increase in infliximab dosage were similar: DAS28 (AUC = 0.840 [0.791–0.889]), SDAI (AUC = 0.824 [0.776–0.873]), and CDAI (AUC = 0.821 [0.772–0.870]) . In the second study, the SDAI tended to discriminate patients with and without change in DMARDs better than did the DAS28 (AUC = 0.923 [0.895–0.965] vs. AUC = 0.872 [0.831–0.925]) . The decision to change DMARD treatment was made when the rheumatologist considered that control of disease activity was not sufficient (DAS28 and SDAI were not calculated during the visit).
One study of 378 RA patients directly compared the DAS44 and the DAS28 in terms of their ability to classify patients as being in remission according to American Rheumatism Association (ARA) criteria. The AUC values were similar and excellent (DAS44, 0.96 [0.95–0.97]; and DAS28, 0.93 [0.92–0.94]) . However, this study used the same cohort used for developing the DAS28.
5.3.5. Sensitivity to change
Sensitivity to change was compared in two studies  and , both of which used the ACR response as the reference standard. In one study, one year ESs was estimated for the DAS28, SDAI, and CDAI according to the type of ACR response. The sensitivity to change of the DAS28 and the SDAI were comparable, but the CDAI tended to be more responsive (DAS28: the ES increased 2.4-fold for patients with ACR20/50 response [20% and 50% reduction, respectively, in tender and swollen joints], and 4.4-fold for those with ACR70 response [70% reduction], as compared with non-responders; SDAI: the ES increased 2.7- and 4.1-fold; CDAI: the ES increased 3.3- and 6.5-fold). However, the difference was difficult to interpret because 95% confidence intervals were not provided. In addition, only the CDAI showed a statistically significant difference between ACR20 responders and ACR20-69 responders . The other study  found that both the SDAI and the CDAI detected significant differences between non-responders and ACR20, ACR50, and ACR70 responders and between ACR50 and ACR70 responders. In contrast, neither index significantly differentiated ACR20 and ACR50 responders. The SRMs for the DAS44, DAS28, SDAI, and CDAI were computed in the various ACR responder categories over the two-year follow-up, with slight differences between the indices; the signification is unclear because of the absence of confidence intervals (Table 4) .
Standardized response means of the various indices according to the American College of Rheumatology (ACR)-defined response.
ACR20, ACR50, ACR70: 20% 50%, 70% reduction, respectively, in number of swollen and tender joints.
The management of RA has changed radically over the last 10 years, with the introduction of new drugs and treatment strategies and with the emergence of new concepts of disease severity, treatment targets, and means of evaluating treatment effects.
In particular, the necessity to evaluate disease activity using an objective and accurate instrument has been demonstrated. Because numerous indices have been developed, we conducted an exhaustive review of comparative data on the metrological properties of the most widely used indices (DAS44, DAS28, SDAI, CDAI): reproducibility, construct validity, discriminative performance, and sensitivity to change.
Surprisingly few studies have compared the reproducibility of these indices. A single study directly compared the reproducibility of the DAS28, SDAI, and CDAI and did not demonstrate any difference. The DAS44 is often considered less reproducible than is the DAS28 because the former requires the evaluation of a larger number of joints, including the metatarsophalangeal (MTP) joints, which may be particularly difficult to assess. An unexpected finding from our study is that this supposed difference has never been demonstrated or evaluated.
Concordance among the indices was acceptable in randomized controlled trials and in cohort studies. However, acceptable concordance may be insufficient, because treatment decisions are often based on composite index values. In particular, remission has become a reasonable treatment objective and has been used as a primary outcome measure in some RCTs or treatment strategy studies. However, observing relevant discrepancies between the DAS28 and the DAS44, SDAI and CDAI is difficult because the DAS28 is less conservative in defining remission. One study found that the DAS28 performed poorly for assessing disease activity in individual patients because it does not require an evaluation of the ankles and feet, which are often involved in early RA . However, a recent study suggests that, despite the importance of foot involvement in the clinical evaluation of RA patients, the 28-joint assessment may provide a valid evaluation of disease activity . During 2,754 visits to 767 patients, the authors compared a 28- to a 32-joint assessment, obtained by adding the ankles and MTP joints (considered collectively). Absence of joint swelling by the 28-joint assessment had 98.1% specificity and by the 32-joint assessment, 94.1% positive predictive value. Moreover, the remission rate was the same with and without counting the ankles and MTP joints: patients in whom the only tender and swollen joints were the ankles and/or MTP joints were never classified as being in remission by the DAS28, in particular because they had high general health scores. The authors concluded that the absence of lower extremity assessment in the DAS28 is problematic only in a few patients. A committee consisting of members of the ACR, the European league against rheumatism, and the outcome measures in rheumatology Initiative confirmed that the widely used definition of remission based on a DAS28 of less than 2.6 better represented minimal disease activity than remission. Furthermore reaching remission according to the DAS28, both at the traditional cut point (< 2.6) and at a more stringent cut point (< 2.0), was associated only with the likelihood of HAQ stability, and not radiographic stability unlike remission according to the SDAI or CDAI. So they have recently proposed two new definitions of remission, both of which can be uniformly applied and widely used in RA clinical trials:
- when scores on the tender joint count, swollen joint count, CRP (in mg/dL), and patient global assessment (0–10 scale) are all less or equal to one;
- when the score on the SDAI is less or equal to 3.3 .
All four indices showed similar convergent validity, with fair-to-good correlations with the HAQ score and structural damage, and satisfactory discriminative performance, both for treatment decisions and for ACR-defined remission.
The DAS44, DAS28, SDAI, and CDAI are all valid tools for assessing disease activity in RA patients. Each has advantages and drawbacks. The results of our systematic review do not allow for ranking these indices in terms of their metrological properties because, except for defining remission, the available studies suggest that the four indices may have similar metrological properties. As well, to our surprise, we found only a few studies that compared these indices. Further studies are clearly needed.
Until additional data are available, several points may help rheumatologists choose among the four indices. Although the DAS44 has been extensively validated in therapeutic trials, the evaluation of 44 joints is time-consuming. The DAS28 can be determined more rapidly and has been extensively studied. The main limitation of the DAS28 is a broader definition of remission than with the other indices. This limitation may not be problematic in therapeutic trials, which include large numbers of patients, but may be of concern for individual patients seen in everyday practice. To take this concern into account, the recommendations issued by the French High Health Authority  indicate that the DAS28 may be used to define absence or inadequacy of the treatment response but that treatment decisions can be modulated on the basis of a number of factors, including involvement of the feet and ankles.
Another property of an outcome is feasibility. The DAS44 and DAS28 require a special calculator or dedicated software. The SDAI, in contrast, is a simple sum, yet exhibits useful metrological characteristics. The need for a laboratory test (for ESR or CRP level) to obtain the DAS44 and DAS28 and SDAI may limit the use of these indices in everyday practice. The CDAI, in contrast, can be obtained at any time and in any setting and shows validity similar to that of the other indices. The SDAI and CDAI may also be easier to understand by patients and may therefore enhance adherence to the treatment regimen.
Disclosure of interest
The authors declare that they have no conflicts of interest concerning this article.
This work was supported by an unrestricted educational grant from Abbott France.
-  Y.P. Goekoop-Ruiterman, J.K. de Vries-Bouwstra, P.J. Kerstens, et al. DAS-driven therapy versus routine care in patients with recent-onset active rheumatoid arthritis. Ann Rheum Dis. 2010;69:65-69 Crossref
-  D.M. van der Heijde, M.A. van’t Hof, P.L. van Riel, et al. Judging disease activity in clinical practice in rheumatoid arthritis: first step in the development of a disease activity score. Ann Rheum Dis. 1990;49:916-920 Crossref
-  M.L. Prevoo, M.A. van’t Hof, H.H. Kuper, 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;38:44-48 Crossref
-  J.S. Smolen, F.C. Breedveld, M.H. Schiff, et al. A simplified disease activity index for rheumatoid arthritis for use in clinical practice. Rheumatology. 2003;42:244-257 Crossref
-  D. Aletaha, V.P. Nell, T. Stamm, et al. Acute phase reactants add little to composite disease activity indices for rheumatoid arthritis: validation of a clinical activity score. Arthritis Res Ther. 2005;7:R796-806 Crossref
-  M. Boers, P. Brooks, C.V. Strand, P. Tugwell. The OMERACT filter for outcome measures in rheumatology. J Rheumatol. 1998;25:198-199
-  B.F. Leeb, I. Andel, J. Sautner, et al. Disease activity measurement of rheumatoid arthritis: comparison of the simplified disease activity index (SDAI) and the disease activity score including 28 joints (DAS28) in daily routine. Arthritis Rheum. 2005;53:56-60 Crossref
-  D. Aletaha, M.M. Ward, K.P. Machold, et al. Remission and active disease in rheumatoid arthritis: defining criteria for disease activity states. Arthritis Rheum. 2005;52:2625-2636 Crossref
-  H.J. Vrijhoef, J.P. Diederiks, C. Spreeuwenberg, et al. Applying low disease activity criteria using the DAS28 to assess stability in patients with rheumatoid arthritis. Ann Rheum Dis. 2003;62:419-422 Crossref
-  C.A. Walsh, R.H. Mullan, P.B. Minnock, et al. Consistency in assessing the Disease Activity Score-28 in routine clinical practice. Ann Rheum Dis. 2008;67:135-136 Crossref
-  Uhlig T, Kvien TK, Pincus T. Test-retest reliability of ACR core data set measures and indices derived from these measures (DAS28, CDAI, SDAI, RAPID3). [Abstract ACR 2008]
-  B. Vander Cruyssen, S. Van Looy, B. Wyns, et al. DAS28 best reflects the physician's clinical judgment of response to infliximab therapy in rheumatoid arthritis patients: validation of the DAS28 score in patients under infliximab treatment. Arthritis Res Ther. 2005;7:R1063-R1071 Crossref
-  V.K. Ranganath, J. Yoon, D. Khanna, et al. Comparison of composite measures of disease activity in an early seropositive rheumatoid arthritis cohort. Ann Rheum Dis. 2007;66:1633-1640 Crossref
-  A.M. van Gestel, C.J. Haagsma, P.L. van Riel. Validation of rheumatoid arthritis improvement criteria that include simplified joint counts. Arthritis Rheum. 1998;41:1845-1850 Crossref
-  R. Landewe, D. van der Heijde, S. van der Linden, et al. Twenty-eight-joint counts invalidate the DAS28 remission definition owing to the omission of the lower extremity joints: a comparison with the original DAS remission. Ann Rheum Dis. 2006;65:637-641 Crossref
-  D. Khanna, M. Oh, D.E. Furst, et al. Evaluation of the preliminary definitions of minimal disease activity and remission in an early seropositive rheumatoid arthritis cohort. Arthritis Rheum. 2007;57:440-447 Crossref
-  M. Mierau, M. Schoels, G. Gonda, et al. Assessing remission in clinical practice. Rheumatology (Oxford). 2007;46:975-979 Crossref
-  T.S. Shaver, J.D. Anderson, D.N. Weidensaul, et al. The problem of rheumatoid arthritis disease activity and remission in clinical practice. J Rheumatol. 2008;35:1015-1022
-  M. Soubrier, D. Zerkak, L. Gossec, et al. Which variables best predict change in rheumatoid arthritis therapy in daily clinical practice?. J Rheumatol. 2006;33:1243-1246
-  J. Fransen, M.C. Creemers, P.L. Van Riel. Remission in rheumatoid arthritis: agreement of the disease activity score [DAS28] with the ARA preliminary remission criteria. Rheumatology (Oxford). 2004;43:1252-1255 Crossref
-  T. Kapral, F. Dernoschnig, K.P. Machold, et al. Remission by composite scores in rheumatoid arthritis: are ankles and feet important?. Arthritis Res Ther. 2007;9:R72 Crossref
-  D.T. Felson, J.S. Smolen, G. Wells, et al. American college of rheumatology/european league against rheumatism provisional definition of remission in rheumatoid arthritis for clinical trials. Ann Rheum Dis. 2011;70:404-413 Crossref
-  www.has-sante.fr/.
a Rheumatology department, groupe hospitalier Pitié-Salpêtrière, Paris-6 university–Pierre-et-Marie-Curie, AP–HP, 83, boulevard de l’Hôpital, 75651 Paris cedex 13, France
b Rheumatology department, CHU Lapeyronie, Montpellier, France
c Rheumatology department, CHU hôpital Sud, Echirolles, France
d Rheumatology department, CHU Henri-Mondor, AP–HP, Créteil, France
e Rheumatology department, CHU de Rouen, Rouen, France
f Rheumatology department, Dijon university hospital, 21078 Dijon, France
g Inserm U887, university of Burgondy, 21079 Dijon , France
⁎ Corresponding author. Tel.: +33 1 42 17 78 21; fax: +33 1 42 17 79 55.
© 2011 Société française de rhumatologie, Published by Elsevier B.V.