Background/Purpose:

Simplified measures to quantify rheumatoid arthritis (RA) disease activity are increasing in use in clinical practice. However, the absolute minimally important difference (MID) in some of these measures, such as the clinical disease activity index (CDAI), has not been well defined in real-world settings, especially for early RA patients (ERA) with low or moderate disease activity. To determine the MID for improvement of the CDAI in a “real world” setting using an observational cohort of patients with ERA.

Methods:

Data from the Canadian CATCH cohort of patients with ERA were used to identify pairs of rheumatology visits spaced 3 months apart occurring within 12 months of cohort entry. Patients with concomitant fibromyalgia (8% of cohort) were excluded. Absolute change in CDAI was examined between visit pairs and correlated with relevant changes corresponding to MIDs in patient self-reported improvement (better vs. same/worse), patient pain (>1 vs. <= 1, and >2 vs. <= 2) and HAQ (>0.22 vs. <= 22, and >0.30 vs. <= 0.30). The 10^th and 90^th percentile of the CDAI distributions for EULAR good response vs. non-response in DAS28 ESR (>1.2 vs. < 0.6 units improvement) were used to determine proposed CDAI cutpoints to define MID, overall and stratified by initial CDAI disease activity categories (low < 10; moderate 10<=-<=22; high > 22). Discrimination of these cutpoints was examined using area under receiver operator curves (AUROC). These CDAI cutpoints were used to describe the Sensitivity (Se), Specificity (Sp), positive predictive value (PPV) and negative predictive value (NPV) of the CDAI cutpoints for patient pain and HAQ.

Results:

A total of 1191 unique CATCH patients (mean+- SD age 53.5+- 15.1 years; 72% women, median RA disease duration 5.8 +- 3.0 months and contributed 3262 visit pairs within 12 months of enrollment.  . Among those with available serologic results (91% of cohort for RF, 67% of cohort for anti-CCP), 70%  were RF positive  and 60% anti-CCP antibody positive.

Overall, there was excellent discrimination between DAS28 responders using a CDAI cutpoint of 5 units; the AUROC was 0.87. CDAI cutpoints for patients who started in low, moderate, or high disease were 2, 6, and 10, respectively. The Se, Sp, PPV and NPVs shown in the Table showed acceptable performance characteristics using these proposed CDAI cutpoints.

Conclusion:

These minimally important absolute differences in CDAI can be used to evaluate improvement and increase the usefulness of this clinical simplified disease activity measure in real-world settings.

 

Improvement in	CDAI Cutpoint and Starting Disease Activity	Sensitivity	Specificity	PPV	NPV
Patient self-reported improvement (better vs. same/worse)	5 (all data) 2 (low) 6 (moderate) 11 (high)	76	53	56	74
		84	58	42	90
		76	82	57	72
		82	73	88	63
Pain (>1 vs. <=1)	5 (all data) 2 (low) 6 (moderate) 11 (high)	45 33 46 52	85 92 75 77	55 54 52 71	80 83 70 61
Pain (>2 vs. <=2)	5 (all data) 2 (low disease) 6 (moderate) 11 (high)	29 15 24 40	93 96 88 85	62 54 55 73	77 80 66 58
HAQ (> 0.22 vs. <= 0.22)	5 (all data) 2 (low disease) 6 (moderate) 11 (high)	57 28 46 70	80 87 73 64	69 40 58 77	71 79 64 55
HAQ (> 0.30 vs. <= 0.30)	5 (all data) 2 (low disease) 6 (moderate) 11 (high)	48 17 36 63	87 93 81 72	74 44 59 80	69 78 62 53

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/determining-the-absolute-change-in-the-clinical-disease-activity-index-cdai-to-define-a-minimally-important-difference/