Background/Purpose: We have reported (A&R 2011) cross-sectional rates of remission in RA clinic patients of 7.5% using ACR/EULAR (AE) criteria, while 19.1% had Physician global scores ²1. AE clinical criteria require scores ²1 for tender and swollen joints and patient global (PtGlobal). Statistical analysis suggested that PtGlobal was the most important determinant of AE remission, but appeared to be unstable, shifting between observations. To understand the effect of PtGlobal, and what role it plays in AE remission, we studied its meaning and reliability and validity in consecutive patients who met AE joint criteria for remission.
Method: We studied 73 consecutive RA clinic patients who met AE joint criteria for remission (²1 swollen and ²1 tender joint), using the AE 28-joint count. Patients completed a VAS global, VAS fatigue and HAQ immediately before seeing the physician (MD), and completed a 2nd VAS global and questions about global severity immediately after the MD visit. MDs determined an MD global, clinical remission, and a complete tender and swollen joint count, including ankles and feet.
Result: Patient mean age was 60.0 (SD 15.1) years, and 41.7% were men. As expected, the VAS scores for pain, 2.9 (2.1), fatigue 3.7 (2.6) and patient global 2.7 (2.1) were low. Of patients meeting joint criteria for AE remission, 56 (76.7%) had PtGlobal >1, including 21.4% who reported no pain and 52.7% who indicated function or fatigue, not pain, were the main determinants of PtGlobal. Low back pain was also reported by 25% of this group. By contrast, 30.0% reported foot pain and 30.0% reported knee pain (compared with 7% on the MD examination) that may not have been identified in the physician's 28-joint count. There was no agreement beyond chance (kappa = 0.0, p=0.530) between ACR and MD remission, nor between high patient and physician global (global >1), kappa = -0.074, p=0.946. In addition to physician-patient disagreement (Figure 1), patients not meeting AE criteria frequently had high fatigue and high pain scores (Figure 1). Test-retest reliability of the patient global indicated an intraclass correlation coefficient (ICC) of 0.84.
Conclusion: PtGlobal has insufficient reliability for accurate measurement of remission in individual clinic patients. In addition, we identified problems with using PtGlobal as a measure of RA activity, for many patients who would otherwise be in AE remission were excluded because of non-RA activity factors, including low back pain, fatigue, functional limitations and low pain threshold. We also noted that physician joint counts may be biased downward by exclusion of foot and ankle pain that is not included in the 28-joint count or failure to identify knee joint activity. While the AE remission criteria appear to work well in clinical trials and groups of patients, problems with reliability and misclassification related to joint pain and patient global suggest caution in the use of the AE criteria in the individual patients.
Disclosure: K. R. Masri, None; F. Wolfe, None; T. S. Shaver, None; S. H. Shahouri, None; S. Y. Wang, None; J. D. Anderson, None; R. E. Busch, None.