Londonderry, Northern Ireland - The Disease Activity Scale (DAS) 28, a shortened form of the DAS that excludes the ankles and feet, is widely used as an indicator of RA disease activity and response to treatment. Criticism of the DAS28 surfaced in a number of presentations at the 2004 American College of Rheumatology annual meeting [ 1 , 2 , 3 ]. Dr Phillip Gardiner (Altnagelvin Hospital, Londonderry, Northern Ireland) reports in the March 2005 issue of the Annals of the Rheumatic Diseases that the DAS28 may change significantly even when there is little or no clinical improvement, that this is largely due to the contribution of the erythrocyte sedimentation rate (ESR) to the DAS28 score, and that this disjuncture is most evident in patients with low initial ESR [ 4 ].
The formula for the DAS28 score explains much of the problem: ESR scores are log-transformed, so changes at the top end of the scale have little influence on DAS28, whereas a small change in ESR within the normal range can have a large effect on the score.
"Although this was the exception rather than the rule, we did find that in some patients the DAS28 score changed significantly when we didn't find a significant clinical improvement and even when the patients themselves didn't feel any better. The formula for the DAS28 score explains much of the problem: ESR scores are log-transformed, so changes at the top end of the scale have little influence on DAS28, whereas a small change in ESR within the normal range (ie, under 20 mm/hour) can have a large effect on the score. If someone has an initial ESR of 20 mm/hour, a decrease of 5 or 10 mm/hour will have a much greater effect on the DAS28 score than an ESR change from 70 to 60 mm/hour," Gardiner tells rheumawire .
Conversely, Gardiner says, "RA patients with fibromyalgia may achieve high DAS scores without clinical signs of very active disease because the DAS28 score puts more emphasis on the tender joint score than the swollen joint score (whereas intuitively it should be the other way round)."
These problems have immediate clinical importance for physicians and patients in the UK, where guidelines advise discontinuing TNF-inhibitor therapy if a reduction in DAS28 of >1.2 or a reduction in DAS28 to <3.2 are not achieved at 3 months.
Sometimes DAS28 and clinical findings don't add up
"In the course of reviewing the decision on whether or not to continue anti-TNF therapy, we identified several patients in whom the change in DAS28 score did not correlate with clinical findings," Gardiner says. One example was a patient treated with infliximab who had a baseline DAS28 of 5.25 (including 10 swollen and 5 tender joints and ESR 13 mm/hour). After 3 months of treatment his DAS28 had fallen to 3.8 ("An impressive improvement of 1.45," Gardiner notes), but he still had 8 swollen and 2 tender joints. His ESR was then 4 mm/hour, and the effect of a relatively small ESR change at these low levels is what drove the change in DAS28.
"Thus, even though there had been little or no discernible improvement, the DAS28 response exceeded the cutoff point of 1.2 stipulated in the . . . guidelines," Gardiner writes.
"From a practical viewpoint, in the UK we operate with limited funds for anti-TNF therapy, and some objective tool is necessary to help decide who needs the treatment most. The DAS28 score still seems to be the best tool currently available to assess disease activity as well as change in disease activity, and we all still use it," Gardiner says. However, he advises against the use of the DAS28 as the main response criterion in an individual patient when the score conflicts with the patient's and physician's clinical assessments. He also suggests that the time has come to modify the guidelines, "taking into account situations where the change in DAS28 score is not representative of the clinical situation."
"One way of doing that would be to have a spreadsheet or database present a color-coded bar, with each segment representing the component of the scale (tender joints score, swollen joint count, ESR, and visual analog scale). This can be generated automatically and gives a quick visual check: if nearly all the score is made up from tender joints with normal ESR and swollen joint count, most of us would be cautious about interpretation. Second, if the ESR is within the normal range, why not set a limit, eg, ESR 20 mm/hour. If the ESR falls below this value, the new ESR is just entered as 20, because an ESR change within the normal range is generally regarded as insignificant," Gardiner says.
Gardiner also tells rheumawire that Dr Piet van Riel (University Medical Centre Nijmegen, the Netherlands) and colleagues have developed a DAS28 using C-reactive protein (CRP) rather than ESR. This approach requires a sensitive CRP assay and also uses a log transformation but may be more reliable, since the CRP scale is more continuous than the ESR.
Real-world use of DAS28 already changing
"One of the problems is that those who develop the scores usually use clinical trial data sets rather than real-life clinic patients," Gardiner says.
We also noted that some patients with atypical or seronegative aggressive arthritis may not fulfill the DAS28 criteria but would be worthy of [TNF-inhibitor] treatment.
"As clinicians in Northern Ireland, we met to discuss a series of 'paper cases' based on real patients we had considered for anti-TNF. We had some basic clinical information as well as the DAS28 scores including the components. Each clinician decided if they would have treated the patient and whether or not they would have stopped the treatment at 3 months. The outcomes of these deliberations will be published in the near future, but I think we all were surprised that we were already taking account of the influence of tender points on the score (regarding a score as less significant if not accompanied by high ESR or swollen joints). We also noted that some patients with atypical or seronegative aggressive arthritis may not fulfill the DAS28 criteria but would be worthy of [TNF-inhibitor] treatment," Gardiner says.
Landew¿ RB, van der Heijde D, Voskuyl A, et al. Condensed 28-joint counts jeopardise the construct validity of DAS28 remission: a comparison with the original DAS. American College of Rheumatology meeting; San Antonio, TX; Oct 16-21, 2004; Abstract 1155.
Wolfe F, Michaud K, Pincus T, et al. The DAS score is not suitable as the criterion for TNF initiation in the clinic: evidence of Disagreement between physician and patient assessments and DAS scores. American College of Rheumatology meeting; San Antonio, TX; Oct 16-21, 2004; Abstract 1159.
Keenan R, Yazici Y. Erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels are poorly correlated with each other and with clinical measures of disease outcome in rheumatoid arthritis (RA) patients. American College of Rheumatology meeting; San Antonio, TX; Oct 16-21, 2004; Abstract 289.
Gardiner PV, Bell AL, Taggart AJ, et al. A potential pitfall in the use of the Disease Activity Score (DAS28) as the main response criterion in treatment guidelines for patients with rheumatoid arthritis. Ann Rheum Dis 2005; 64:506-507.
Medscape Medical News © 2005
Cite this: Janis Kelly. DAS28 not always a reliable indicator of treatment effect in RA - Medscape - Feb 22, 2005.