The 1987 ACR criteria for RA consists of several clinical features, RF and also erosive damage on hand radiographs. The new ACR/EULAR criteria have again clinical features and RF and anti CCP antibodies, but erosive disease is no longer a part of the list of features contributing to the scoring system that defines if a patient is fulfilling the new ACR/EULAR criteria. The reason for this is that developers of the criteria consider erosive disease as a late phenomenon. And as one of the goals was to make it possible to have an early diagnosis to be able to start treatment early, it was felt that it was inappropriate to incorporate structural damage in the table of the criteria set. At the other hand, it was recognized that erosions are a typical feature of RA and that it is indeed likely that a patient is suffering from RA if “typical erosions” are present. This might be the case in patients with longstanding inactive disease. But it might also be that radiographs are available as another physician already orders them. A separate task force was formed to define what is conceived by “typical erosions”. The starting point by the task force was that the definition should be highly specific as patients can be classified as fulfilling the new ACR/EULAR criteria bypassing the scoring system just based on the presence of typical erosions. A second premise was that we wanted to have a data driven definition. Finally, it has been decided to take both radiographs of hands and feet into account, as joints in the feet frequently show erosions very early in the disease course.
Two cohorts of patients with early arthritis were analysed for this purpose: the Leiden Early Arthritis Clinic and the French ESPOIR cohort. Both include patients with arthritis of short symptom duration and were followed according to a fixed protocol with radiographs of hands and feet and clinical outcome data. There were three gold standards used as a surrogate for the diagnosis of RA by the rheumatologist: start of MTX within the first year, start of any DMARD within the first year and persistency of arthritis after 5 years. Patients were grouped according to the number of erosive joints: one, two, three, four or more than four joints. These were analysed for all joints included in hands and feet and grouped for PIPs, MCPs, wrist, or MTPs only and various combinations of these joints groups. The analyses were performed in all patients included in the cohorts as well as only in the patients that did not fulfil the new ACR/EULAR criteria.
The results of both the Leiden EAC and the ESPOIR cohort were very similar. This was also the case for the various golden standards used. There were no joint groups that performed better than others, so it was decided that films of hands and feet should be used. Overall, a high specificity was depicted if the number of erosive joints was between two and five. Just one erosive joint was not sufficiently specific. The task force is now in the process of selecting the best cut-off for the definition of “what is erosive disease in the criteria”.
Disclosure of Interest None Declared