Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement
The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.
Periarticular Bone Mineral Density Does Not Distinguish Between Healthy Controls, Established RA and Early Arthritis Patients
Alves, C., Colin, E. M., van Oort, W. J., Sluimer, J., Hazes, J. M. W., Luime, J. J.
Background:
Early recognition of rheumatoid arthritis (RA) is important, but difficult due to lack of a distinct feature. Loss of periarticular bone mineral density (pBMD) in the hand is thought to be an early feature. The difficulties in pBMD measurement are dilution of BMD loss in the whole hand and less precise measurements in small regions of interest (ROI's). Choosing the size of a ROI is a trade-off between area size and precision error.
Purpose:
To identify periarticular ROI's relevant to RA with a low precision-error and sufficient interrater reliability and to test the validity of these ROI's.
Methods:
Five ROIs were defined around the MCP and/or PIP joints II-V, II-IV and mid-metacarpal to mid-phalangeal. BMD was estimated using a Lunar Prodigy. Precision was estimated by measuring 5 healthy adults 7 times, after repositioning, using the Root-Mean-Square-Coefficient of Variation (RMS-CV). To determine the interrater reliability 20 patients were analyzed on separate occasions by 2 readers using the ICC. Validity of the ROI's was tested using 2 patient groups and healthy controls matched on sex, age and menopausal status. Patients with established RA were recruited from the rheumatology outpatient clinic, patients with early arthritis via the Rotterdam Early Arthritis CoHort (REACH). Simple descriptive analyses, scatter plots, paired t-tests and ROC-curves were done.
Results:
The RMS-CV varied from 0.45% to 1.07%. The ICC was 0.99 for all measurements. Between September 2006 and October 2008 44 patients and 33 healthy controls were recruited. Mean BMDs of the ROI's ranged from 0.321 to 0.372 g/cm2 in established RA, 0.321 to 0.382 g/cm2 in early arthritis and 0.342 to 0.401 g/cm2 in controls. The mean differences ranged from 0.011 to 0.032 g/cm2 for established RA and 0.028 to 0.033 g/cm2 for early arthritis patients compared to matched controls. For ROI 1 and 3–5 in early arthritis and the whole hand in established RA the mean differences were significant. The ROC-curves indicated low discriminative power of all ROI's, with AUC's from 0.60 to 0.66.
Conclusion:
Periarticular BMD seems not to be a useful diagnostic feature due to the wide distribution in BMD values resulting in strong overlap between healthy controls, established RA and early arthritis patients.
Table 1. One sample T-test on matched differences between BMD patients and controls
Figure 1. ROC curve for ROI 1 in early arthritis (AUC=0.62)
To cite this abstract, please use the following information:
Alves, C., Colin, E. M., van Oort, W. J., Sluimer, J., Hazes, J. M. W., Luime, J. J.; Periarticular Bone Mineral Density Does Not Distinguish Between Healthy Controls, Established RA and Early Arthritis Patients [abstract]. Arthritis Rheum 2009;60 Suppl 10 :776
DOI: 10.1002/art.25856
