Session Information
Session Type: ACR Poster Session C
Session Time: 9:00AM-11:00AM
Background/Purpose:
The original DAS28 is
based on the erythrocyte sedimentation rate (ESR). An alternative
formulation uses the C-reactive protein (CRP). Previous research1,2,3
suggests that in comparison the DAS28-CRP underestimates disease
activity, and therefore new thresholds4,5 for levels of
disease activity should be used. In this study, we compare the
DAS28-ESR and DAS28-CRP data of patients from our outpatient center
and derive a formula to convert between the two formulations.
A total of 996 patients
with RA (n=399, 307 female), PsA (n=356, 235 female) and
undifferentiated arthritis (n=242, 176 female) were included in this
retrospective study. A sub-cohort with early-arthritis (disease
duration ≤ 2 years) includes 248 subjects (206 female) with a
definite diagnosis of RA (n=200) or PsA (n=48). For each patient and
each visit, the DAS28-ESR and the DAS28-CRP were computed. The mean,
standard deviation and skewness were used to obtain a normalization
formula to convert the DAS28-CRP scores via
DAS28-conv = 13.1 *
(((DAS28-CRP – 2.53) * 1.15 + 2.91)0.2 – 1)
For 5254 individual
DAS28 records, we computed the number of patients with remission,
low, medium and high disease activity for the DAS28-BSG and the
converted DAS28-CRP (“DAS28-conv”) using the standard DAS28-ESR
thresholds, and the number of patients in remission and with low
disease activity using the cut-off-points for DAS28-CRP (<2.4, ≤ 2.9) suggested by Fleischmann et al4.
The results are summarized in the following table
(r/l/m/h = remission/low/medium/high disease activity)
|
cohort / n. entries |
DAS28-ESR |
DAS28-CRP |
correlation |
DAS28-conv |
DAS28-ESR r/l/m/h |
Fleischmann et al. cut-offs r/l |
DAS28-conv r/l/m/h |
|
all patients n=5254 |
2.91 ± 1.22 |
2.53 ± 1.06 |
0.89 p < 0.001 |
2.91 ± 1.31 |
2331/995/1652/276 |
2013/871 |
2292/813/1848/301 |
|
female n=4154 |
2.99 ± 1.89
|
2.57 ± 1.04 |
0.88 p < 0.001 |
2.97 ± 1.29 |
1697/819/1419/219 |
1482/691 |
1702/661/1559/232 |
|
male n=1100 |
2.58 ± 1.27 |
2.39 ± 1.10 |
0.89 p < 0.001 |
2.70 ±1.35 |
634/176/233/57 |
531/180 |
590/152/289/69 |
|
RA n=2285 |
3.04 ± 1.26
|
2.62 ± 1.12 |
0.90 p < 0.001 |
3.00 ± 1.36 |
911/453/760/161 |
830/340 |
941/328/848/168 |
|
PsA n=1467 |
2.93 ± 1.22 |
2.61 ± 1.06 |
0.89 p < 0.001 |
3.01 ± 1.29 |
659/270/459/79 |
531/232 |
605/223/550/89 |
|
undiff. arthritis n=1502 |
2.69 ± 1.12 |
2.33 ± 0.92 |
0.85 p < 0.001 |
2.67 ± 1.21 |
761/272/433/36 |
652/299 |
746/262/450/44 |
|
early arthritis n=3241 |
2.77 ± 1.14 |
2.41 ± 0.96 |
0.87 p < 0.001 |
2.77 ± 1.14 |
1560/632/941/108 |
1356/586 |
1540/538/1040/123 |
Our data confirms
earlier findings that the DAS28-CRP underestimates the disease
activity. On the other hand, the new DAS28-CRP cut-off points of
Fleischmann et al. seem to be too strict with significantly fewer
patients in remission and low disease activity.
The data suggests that
a normalization formula can be used to convert between the two
formulations, resulting in higher classification agreement within our
cohort. More research is needed to validate the formula against
larger datasets.
References:
-
Matsu et al, Ann Rheum Dis. 2007;66:1221-1226
-
Wells et al, Ann Rheum Dis. 2009;68:954-960
-
Fleischmann et al, Ann Rheum Dis. 2015 Jun;74(6):1132-7
-
Favalli et al, Ann Rheum Dis 2015;74:e5
To cite this abstract in AMA style:
Langer A, Wiemann O, Werner SG, Spiecker F, Mettler S, Lind-Albrecht G, Langer HE. DAS28-ESR and DAS28-CRP Are Not Exchangeable in Daily Clinical Practice: Data from a Large Outpatient Cohort [abstract]. Arthritis Rheumatol. 2015; 67 (suppl 10). https://acrabstracts.org/abstract/das28-esr-and-das28-crp-are-not-exchangeable-in-daily-clinical-practice-data-from-a-large-outpatient-cohort/. Accessed .« Back to 2015 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/das28-esr-and-das28-crp-are-not-exchangeable-in-daily-clinical-practice-data-from-a-large-outpatient-cohort/