Predictive Risk Factors of Cervical Spine Instabilities in Rheumatoid Arthritis

Study Design. A prospective multicenter cohort study for more than 10 years of outpatients with rheumatoid arthritis (RA). Objective. To identify predictive risk factors of cervical spine instabilities, which may induce compression myelopathy in patients with RA. Summary of Background Data. Many reports described the natural course of cervical spine involvement in RA. Only a few studies, however, conducted comprehensive evaluation of its prognostic factors. Methods. Cervical spine instability was radiographically defined as atlantoaxial subluxation with the atlantodental interval greater than 3 mm, vertical subluxation (VS) with the Ranawat value less than 13 mm, and subaxial subluxation with irreducible translation of 2 mm or higher. The "severe" category of instability was defined as atlantoaxial subluxation with the atlantodental interval of 10 mm or lower, vertical subluxation with the Ranawat value of 10 mm or higher, and subaxial subluxation with translation of 4 mm or higher or at multiple levels. Of 503 "definite" or "classical" patients with RA without baseline "severe" instability, 143 were prospectively followed throughout for more than 10 years. The Cox proportional hazards regression analysis was performed to determine predictors for the development of "severe" instabilities. To exclude biases from the low follow-up rate, similar assessments were performed in 223 patients followed for more than 5 years from baseline. Results. The incidence of cervical spine instabilities and "severe" instabilities significantly increased during more than 10 years in both 143 and 223 cohorts (all P < 0.01). Multivariable Cox proportional hazards models found that baseline mutilating changes (hazard ratio [HR]=19.15, 95% confidence interval [95% CI] = 3.96-92.58, P < 0.01), corticosteroid administration (HR = 4.00, 95% CI = 1.76-9.11, P < 0.01), and previous joint surgery (HR = 1.99, 95% CI = 1.01-3.93, P = 0.048) correlated with the progression to "severe" instability in 143 cases and also in 223 cases (HR = 8.12, 95% CI = 2.22-29.64, P < 0.01; HR = 3.31, 95% CI = 1.68-6.53, P < 0.01; and HR = 2.07, 95% CI = 1.16-3.69, P = 0.014, respectively). Conclusion. Established mutilating changes, concomitant corticosteroid treatment, and previous joint surgery are relatively robust indicators for a poor prognosis of the cervical spine in patients with RA, based on the consistency in more than 10-year analysis of two different settings.