ARE OBLIQUE FORE-FOOT RADIOGRAPHS USEFUL TO DETECT EROSIONS IN RECENT-ONSET ARTHRITIS? RESULTS OF THE ESPOIR COHORT

Background: The diagnosis of recent-onset arthritis usually includes postero-anterior hand and fore-foot radiographs. They allow the detection of erosions, a criterion classically associated with the diagnosis of rheumatoid arthritis (RA). Their prognostic and diagnostic values in recent-onset arthritis are now well established. By contrast, the usefulness of oblique fore-foot radiographs in the diagnostic of erosions have never been demonstrated in recent-onset arthritis.Objectives: Our objective was to determine the usefulness of oblique fore- foot radiographs to detect erosions in combination with postero-anterior hand and foot radiographs in recent-onset arthritisMethods: 814 patients were included in 14 centres during the prospective national French cohort, the so- called "ESPOIR cohort".Inclusions criteria were as follow: diagnosis of RA possible, presence of at least two swollen joints (confirmed by the investigator at inclusion), evolution of less than 6 months and more than 6 weeks, patients without corticosteroid or DMARDS. All the patients underwent clinical biological and radiographs examination every 6 month. In each patient, standardized radiographs of hand wrist and fore-foot in postero-anterior view and oblique fore- foot were performed at inclusion and all the films were sent to the imaging reference center. All the radiographs were examined by two readers (IS, TJ; who had no information about the patients) for typical erosions, and Sharp score modified by van der Heijde. The readers examined the three different radiographs (called "a set") simultaneously for each patient. Inter and intraobserver variabilities were assessed using the inter and intraclass correlation coefficients for quantitative variables and the kappa coefficient for categorical variables.Results: The radiographs of 797 patients were referred to the centre. 715 sets were available. The inter and intra-observer reliabilities were considered as correct. Intra-observer correlation coefficients were 0.94 for the Sharp's global score, 0.97 for the Sharp's Joint space narrowing score and 0.98 for the Sharp's erosion score for one reader and 0.95, 0.83 et 0.71 for the second reader. Inter-observer correlations were 0.96, 0.95 and 0.89.The quality of the radiographs was considered as excellent by both readers. At inclusion, the global Sharp score modified by van der Heijde was 3.6±6.6, 2.5±6.3, et 1.8±5 for the hand and wrist, foot and oblique fore-foot. 22% (160/715) of the patients had, at least, one erosion on the set of radiographs (hand and wrist: 96, foot: 91, oblique fore-foot 84). In the group of patients with erosions, 28% (44/160) had erosions only detected on hands and wrists, 46% (74/160) had erosion only on the feet (postero-anterior view: 22; oblique: 16; both: 36). The other 26% (42/160) had hand wrist and foot erosions.Conclusion: 22% of the patients of our cohort of recent-onset arthritis are erosive. The 3 different incidences of hand and wrist, foot in postero-anterior view andoblique fore- foot are usefulness to detect erosions. Omitting to do hand and wrist, in postero-anterior view, foot (postero-anterior and oblique view), foot in postero-anterior view or oblique fore-foot exclude respectively 28%, 36%, 14% and 10% of erosive patients.Citation: Ann Rheum Dis, volume 66, supplement II, year 2007, page 324Session: RA – epidemiology, prognosis and predictors