Study Design. Retrospective cohort analysis.
Objective. To clarify the effect of biological agents (BAs) on the development and progression of cervical lesions in patients with rheumatoid arthritis (RA) and to identify biomarkers that accurately predict disease progression.
Summary of Background Data. The introduction of BAs changed the paradigm of RA treatment. However, their effects on cervical lesions in patients with RA have not been studied.
Methods. Ninety-one subjects who had received BAs for 2 years or more were enrolled. Mean radiographical interval was 3.9 years. Disease activity was evaluated by disease activity score–C-reactive protein levels, and matrix metalloproteinase-3 levels. Cervical lesions were defined as an atlantodental interval more than 3 mm for atlantoaxial subluxation (AAS), Ranawat value less than 13 mm for vertical subluxation (VS), and anterior or posterior listhesis more than 2 mm for subaxial subluxation. Disease progression was defined radiographically as an increase in the atlantodental interval more than 2 mm for AAS, a decrease in both Ranawat and Redlund-Johnell values more than 2 mm for VS, and an increase in listhesis more than 2 mm for subaxial subluxation. We used multivariate regression techniques to assess predictors of disease progression.
Results. Baseline radiographical evaluation showed no pre-existing cervical lesion in 44 patients, AAS in 29, and VS in 18. Radiological progression occurred in 7% patients without baseline lesions, 79% in the AAS group, and 72% in the VS group. The incidence of progression was significantly lower in patients without lesions at baseline. Multivariate regression analysis demonstrated pre-existing cervical lesions, disease activity score–C-reactive protein levels at baseline and metalloproteinase-3 levels at final visit as good predictors of RA progression.
Conclusion. BAs prevented de novo cervical lesions in patients with RA but failed to control progression in patients with pre-existing cervical lesions. Disease activity score–C-reactive protein levels at baseline were related to pre-existing joint destruction, and metalloproteinase-3 levels accurately predicted ongoing bone destruction during BA treatment.
Level of Evidence: 3