When is a DAS28 NOT a disease activity score?

I’m a big fan of the Disease Activity Score (DAS) and if you’re interested at all you should check out this excellent article in ‘The Rheumatologist’ about how Piet van Riel and Desiree van der Heijde developed the score. As much as I like this tool, I think people need to be aware of its limitations in clinical settings. There is no such thing as a perfect disease activity score, and similar problems arise with other scores. I use the DAS28 (ESR) regularly in clinic, and for the majority of patients it is a very useful tool that matches what the patient and physician think is happening to disease activity. I am written this article mainly for clinical health professionals in Rheumatology.

…when the patient has Fibromyalgia (as well as RA)

This one may be a bit obvious, but it can catch out the unwary. I have made this mistake more than once. As many of you know, the incidence of FMS is increased in chronic diseases like RA and SLE and it can seriously skew the DAS28. In a busy clinic we can be so busy doing the DAS28 and sorting out the patient’s problems that we don’t check for muscle tenderness in a patient with known RA. So how can we avoid this mistake? If you see a patient where the tender joint count and VAS scores are very high but the ESR and swollen joint counts are low/normal – look out! This is the typical pattern with FMS symptoms. A quick check for muscular tenderness will usually confirm your suspicions. Of course, the symptoms of pain and fatigue are genuine and need to be addressed, but just not with anti-TNF drugs or more Methotrexate! You are welcome to check out my little shared spreadsheet tool on Google Drive that shows the COMPONENTS of the DAS28 as slices in a pie chart. This tool uses the official formula for DAS28 (ESR) but splits up the components and allows you to experiment with different values and see the results. It makes it a lot easier to see when the Tender joint score and VAS score are dominating the DAS28.

…When the patient has more tender than swollen joints

I’m not going to make a big issue of this, but it is good to be aware that the DAS28 score puts twice as much emphasis on tender compared to swollen joints. This is good in that it makes the DAS28 more of a ‘patient reported outcome measure’ but not so good when you see a patient with quite a few swollen joints and an inappropriately low DAS28 score. And do remember that the DAS28 was not developed to be used in a system where there would be a clear incentive from having a higher score (i.e. access to biologic drugs) – it is open to some bias on the part of both patient and physician.

…When the patient has a low esr

The DAS scores were developed using data from clinical trials. With handy calculators around, most people don’t pause to check the formula. The ESR score is log-transformed to take account of the way an ESR rises. This is all very reasonable, but it has to be remembered that patients were selected for these clinical trials because they had very active disease with a lot of tender and swollen joints. In most of the trials the starting ESR was very high. The ESR during the trials fell significantly but often not to ‘normal’ levels. So the DAS28 was not really properly validated in the population of RA patients we see day to day in clinic: those with fewer active joints and low/normal ESRs.  Which do you think would cause a greater change in DAS28 score: a change in ESR from 40 to 20 or from 15 to 2mm/h? The change in DAS28 score is actually three time larger in the latter example – a reduction of 1.51 – greater than the NICE/BSR ‘response’ criterion for biologics. There is no cut-off for ESR values within the normal range, and even insignificant changes.

In the UK, NICE guidelines dictate which of our RA patients can get access to biologics. The ‘cut-off’ for access is a DAS28 score of over 5.1. This may be no problem for most patients, but for some men with particularly low ESRs this is an issue. I have a couple of patients who have definite ultrasound proven synovitis in multiple small joints but a low ESR of 2mm/h. Now I challenge you to try and get a DAS28 of 5.1 with that! Have a go on the DAS component spreadsheet. Try entering 15 tender, 15 swollen joints and a VAS score of 90. Now this doesn’t affect a lot of patients, but there’s a handful in our department who need treatment but can’t get it because of this.

…When the patient is in ‘Remission’ (DAS28 less than 2.6)

I would guess that most of you are already aware that the DAS28 doesn’t work too well when it comes to defining remission. This is partly because it completely misses out joints such as the feet and ankles which can be important to people with partly controlled arthritis. It is also partly because of the under-valuing of swollen joints (mentioned above). One of my patients had 5 swollen joints, no tender joints, ESR of 8, and VAS of 10. His DAS28 is 2.23 (i.e. in remission). I think it is better to use the new ACR/EULAR remission criteria.

…When it is done by someone inexperienced

Once people have learned how to examine the patient properly to complete the score and practiced using it regularly for a few weeks they will be able to get quite repeatable results. There is an instructional video on the official website http://www.das-score.nl/ and this site also contains a list of the ‘official’ reference articles on the various DAS scores. However, there are grey areas and there can be quite a bit of difference between observers. So if we’re checking DAS28 scores regularly in clinic, it is a good idea to have it checked by an experienced observer rather than a trainee. If we’re relying on a crucial result (e.g. to determine if a patient is to get biologic treatment or not) we need to make sure that the person doing it is experienced. To measure ‘response’ it is best for the observer to be the same person before and after.

So that’s it. All my moans about the DAS28 all in one page. Am I the only one who has these issues?

4 thoughts on “When is a DAS28 NOT a disease activity score?

  1. I have found that a significant proportion of patients with many tender joints but without any swollen actually have synovitis evident on ultrasound. We have good access to US and find it really helpful in this situation. David O'Reilly @Rheuma_Suffolk

  2. Philip. Excellent post. I agree with you 100%. DAS score can be ¨perverse¨ if you are nor aware of this issues. In our patients I can remember other problems in clinical practice. We receive patients labs with ESR from many laboratories, most with wintrobe assessment an not westergreen and not infrequently with z sedimentation and the score its intended to be use with westergreen. Also the visual analogue scale from patients assessment can be difficult to take because of literacy problems. I find better understanding from likert scales than visual analogue scales in our patients. I use it to follow my patient with DAS 28 but an abnormal score means a change adjusting therapy but not always prescribing a biologic taking in account all this issues.
    Ref. http://es.scribd.com/Carvicab/d/73058667-Clinimet

    1. David,
      You are absolutely right. I scanned a patient last week with a low DAS28 and she had effusions and doppler +ve synovitis in 8 of her MTP joints. I didn't mean to imply that tender joints without swelling means that there is no synovitis. The weighting of tender joints is there for a good reason, but there are a few exceptions which can catch us out!
      Thanks for your comment.

    2. Carlo,
      Thanks for your comment. I'm not very familiar with the differences between the Westergreen and Wintrobe methods for ESR. Do you think this will affect the score? I agree with you about difficulties people have scoring the VAS. It is quite common for people to say they feel 8/10 or 80/100 when they mean that their arthritis is not too bad – of course the VAS score is the other way round. I think the Likert approach is probably better than using a continuous scale and I don't know too many people who actually measure a mark on the line (apart from researchers).