Prevalence and predictors of atlanto-axial subluxation in rheumatoid arthritis after 12-years' follow-up (ESPOIR Cohort).

OBJECTIVES: Anterior atlanto-axial subluxation (AAS), defined as an anterior atlanto-dental interval ≥3 mm, can occur in RA and carries a risk of severe neurological impairments. Our objective was to determine the prevalence and predictors of radiographic aAAS after 12 years' follow-up of patients with early polyarthritis. METHODS: We studied patients enrolled in the early polyarthritis cohort ESPOIR (Study and Monitoring of Early Undifferentiated Arthritis) between 2002 and 2005 (at least two swollen joints for >6 weeks and <6 months, no other diagnosis than RA, and no previous exposure to glucocorticoids or DMARDs). All patients still in the cohort after 12 years had dynamic cervical-spine radiographs taken then read by two blinded observers. To evaluate how well combinations of tests performed at baseline and 10 years predicted aAAS after 12 years, univariate analysis and multiple logistic regression procedure were applied. RESULTS: Of 323 patients followed for 12 years, 15 (4.6%; 95% CI 2.8, 6.4) had aAAS. Among baseline variables, only IgA RFs were associated (P < 0.05) with aAAS (sensitivity 60%, specificity 75%). Among data collected after 10 years, oral CS therapy during the 10-year interval, treatment by DMARDs, CRP (mg/dl) and positive tests for RFs were associated with aAAS after 12 years, but only CRP and RFs remained in a model of logistic regression (combination predicted aAAS with a sensitivity of 60% for a specificity of 90%). CONCLUSION: In conclusion, the prevalence of aAAS after 12 years was 4.6% in the ESPOIR cohort, with no patients having severe aAAS. Although some factors were found to be statistically associated to AAS, the event is too rare to allow a clinical relevance.