Objectives: The 28-Joint Disease Activity Score (DAS28) using C-reactive protein (CRP) and DAS28 using erythrocyte sedimentation rate (DAS28-ESR) may not be interchangeable. We sought to compare and estimate optimal thresholds for the DA28-CRP for use in early rheumatoid arthritis (ERA).
Methods: Patients from the Canadian Early Arthritis Cohort with baseline and 12 months' data for both DAS28-ESR and DAS28-CRP were examined for correlations and differences between DAS28-CRP and DAS28-ESR across their range of values. Receiver operating characteristic analysis identified thresholds for DAS28-CRP that best corresponded to established thresholds for the DAS28-ESR using the total sample, then stratified by age and sex. Agreement between DAS28-CRP and DAS28-ESR thresholds was assessed with the kappa statistic.
Results: The sample included 995 patients with mean (SD) age of 53.7 (14.5) years, 5.8 (2.9) months of symptom duration and 74% were female. DAS28-CRP and DAS28-ESR scores were highly correlated (r= 0.92, p<0.0001), however DAS28-CRP values were consistently lower than DAS28-ESR values. Calculated thresholds for DAS28-CRP were lower with 2.5 for remission, 2.9 for low disease activity, and 4.6 for high disease activity but showed moderate agreement with the DAS28-ESR thresholds (kappa=0.70).
Conclusions: In this large sample of ERA patients, newly estimated thresholds for DAS28-CRP were consistently lower than DAS28-ESR thresholds across the spectrum of disease activity. This may have important clinical implications if inflammatory markers are used interchangeably. Additional external validation of our findings is needed.