Accurate assessment of disease activity and joint damage in rheumatoid arthritis (RA) is essential in the clinical management of RA patients and in RA clinical trials. A number of studies have described improved sensitivity for detection of joint effusion, synovitis, and bone erosions in RA joints with the use of ultrasound (US) as compared with conventional clinical and radiologic methods (1–3). Because ultrasonography is a relatively easily accessible and lowcost method, the technique can be used not only by radiologists, but also by trained rheumatologists. Examination of RA joints by US therefore has the potential to become a rheumatologic bedside procedure for assessment of early inflammatory and destructive joint pathology. The dependence on skilled operators, potential problems with reproducibility of results, and poor objective documentation of findings are often mentioned as main disadvantages of musculoskeletal US (4, 5).
Despite increasing evidence of the potential applications of ultrasonography in the evaluation of arthritic joints, data on its accuracy and reproducibility remain very limited. In RA, involvement of the small joints of the hands and feet is frequent and occurs early in the course of the disease (6). Consequently, reliable assessment of these joints is of major importance. Both of these joint groups are accessible for assessment by US.
The main aim of the present study was to assess the agreement between independent investigators with different specialty backgrounds regarding US assessment of inflammatory and destructive changes in finger and toe joints in patients with RA. A system of semiquantitative scoring of the analyzed parameters was introduced. Furthermore, the relationship between clinical and US findings was explored.
- Top of page
- PATIENTS AND METHODS
There is accumulating evidence of the usefulness of US for the diagnosis and monitoring of different rheumatic disorders (1). US units of sufficient quality for examination of the small joints of the extremities have become increasingly available, and the interest in ultrasonography among rheumatologists is growing. Yet, US for assessment of arthritis in the small joints is poorly validated in comparison with its many other applications (10–12).
Because operator dependence is considered to be one of the major disadvantages of musculoskeletal US (4, 5), we decided to investigate the interobserver agreement for 4 important parameters in the US assessment of RA in hand and foot joints. Due to the ordinal character of the data in the study, the ICC was considered the primary outcome measure. The unweighted kappa and overall agreement, which only assess presence or absence of absolute agreement, were also calculated. The present study introduced a semiquantitative scoring system for joint effusion, synovitis, bone erosions, and power Doppler signal in the small joints of the extremities, which is illustrated in Figures 1–4.
Assessment of synovitis and bone erosions showed a good interobserver agreement (ICC 0.78–0.81, unweighted kappa 0.63–0.68). Agreements on the assessment of joint effusion and power Doppler signal were moderate to good (ICC 0.61–0.72, unweighted kappa 0.48–0.55). (Table 1). These interobserver agreement values are not inferior to the corresponding values for other methods of assessment of RA joint disease (13). Agreements varied from joint to joint probably reflecting the anatomic localization and availability for examination. The third MCP joint, which is difficult to access by US, showed relatively low agreement rates compared with the easily accessed second MCP joint. However, the second MTP joint, anatomically equally as difficult to access as the third MCP joint, showed markedly better agreement rates, possibly due to fewer variations in the US examination of the joint. Our data are in accordance with the observations by Wakefield et al (14), who reported interobserver agreements (kappa values) for RA bone erosions of 0.76 versus 0.68 in the present study. No other studies, to our knowledge, have examined interobserver variation in the assessment of the remaining parameters in this study.
Different backgrounds and experience of the US investigators did not markedly influence the results, as shown by Bland-Altman plots (Figure 5), which revealed no systematic difference between the investigators. This suggests that with suitable training, a rheumatologist can satisfactorily perform ultrasonography on finger and toe joints with RA. In comparison with clinical assessment of the small joints, US showed many more joints with signs assumed to reflect inflammation (US 94 joints versus clinical 64 joints). Furthermore, there were very few false negative results by US (12%), but the etiology in those joints remained unknown. If the additional information obtained by US is shown to be of clinical significance in longitudinal studies, US may take an important place in the assessment of the RA joints.
As previously described (9), power Doppler flow signal is not frequently detected in small joints with RA. In the present study, power Doppler signal in the synovial membrane was present in 43 of 94 joints which showed signs of inflammation on B-mode US (joint effusion and synovial thickening). With the high-end US equipment now available, or by using US contrast agents (8, 15), more joints may be visualized with power Doppler signal.
In conclusion, this study suggests that ultrasonography is a reproducible method for assessment of inflammatory and destructive changes in the finger and toe joints of patients with RA. After suitable training, this technique can be used by rheumatologists for improved assessment of RA patients. Furthermore, the study introduces a semiquantitative scoring system for the examined parameters.