Background Patient-based disease activity score (PDAS1 [with ESR] & PDAS2 [without ESR]) in rheumatoid arthritis (RA) are validated patient reported outcome measures of disease activity. They have been shown to correlate highly with DAS28 and CDAI, PDAS1 and PDAS2.
Objectives To develop and examine the performance of status and responder criteria, based on PDAS1 and PDAS2: low, moderate and high disease activities and European League Against Rheumatism (EULAR) good and moderate responses to treatment.
Methods Data from 299 RA patients (originally used to develop PDAS) were analysed using receiver operator characteristic (ROC) curves to determine optimal cutpoints for PDAS1 and PDAS2 that correspond to DAS28 and CDAI defined criteria for remission, low, medium and high disease activity. Data from 56 RA patients initiated on Disease Modifying Anti-Rheumatic Drugs (DMARDs) before and 6 months after treatment were used to determine optimal thresholds for PDAS1 and PDAS2 corresponding to EULAR good or moderate responses. Optimal cut-off points were obtained by maximising the average of sensitivity and specificity. Agreement with DAS28 and CDAI response criteria was assessed with kappa (κ) statistics.
Results Table 1 shows criteria for PDAS1- and PDAS2-based remission, low, moderate and high disease activity. Key cutpoints for PDAS1/PDAS2 were, respectively, 3.5, 4.5, 4.8, and 3.8, 4.6, 5.0. Area under curve (AUC) for the ROC curves ranged from 0.89 to 0.95. Sensitivities ranged from 79% to 99%, and specificities from 61% to 89%. Moderate to good agreement with DAS28 categories was observed: respectively, κ =0.44 and 0.31 for PDAS1 and PDAS2. Corresponding agreements with CDAI were κ =0.3 and 0.4. Crucially, these agreements are comparable to those of CDAI and DAS28 in the same group of patients (κ =0.54).
The criteria that correspond to EULAR moderate and good response were 0.4, 0.8 for PDAS1 and 0.3, 1.2 for PDAS2. Area under the ROC curve ranged from 0.88 to 0.93. Sensitivities ranged from 72% to 100% and specificities from 77% to 94%. Agreement of DAS28 response with PDAS1 and PDAS2 were κ =0.46 and 0.38, respectively. Again, these were comparable to the agreement between DAS28 and CDAI in this patient group (κ =0.55).
Conclusions We have established and validated criteria for defining high, medium, and low disease activity as well as remission, good and moderate response for PDAS1 and PDAS2. They have comparable agreement and performance to standard criteria based on DAS28 and CDAI, and should facilitate the use of PDAS1 and PDAS2 in routine practice and research.
Disclosure of Interest None Declared