This prospective study aimed to clarify the effects of disease activity in rheumatoid arthritis (RA) on cervical and global spine alignment over a mid- to long-term follow-up period. A total of 70 patients with RA (15 males and 55 females) were prospectively followed for more than 4 years. At enrollment, patients presenting with atlanto-axial subluxation, vertical subluxation, or subaxial subluxation were classified as Group C, while those without cervical disorders were classified as Group N. Clinical data, including medication use and disease activity markers (C-reactive protein [CRP] and matrix metalloproteinase-3 [MMP-3]), as well as radiographic parameters of the cervical and global spine, were assessed annually. Longitudinal changes in radiographic parameters were analyzed using linear mixed-effects models with repeated measurements for each patient, adjusting for potential confounders. Cervical disorders were identified in 22 patients at enrollment (Group C) and in 20 patients at the final follow-up. There was no new case with any cervical involvements. Linear mixed-effects modeling demonstrated that higher MMP-3 levels were significantly associated with an increased atlantodental interval (ADI) (estimate = 0.0017, p < 0.001). A significant interaction between time and group was observed (p < 0.05), indicating a more pronounced longitudinal progression of ADI in Group N compared with Group C. In contrast, age, sex, CRP levels, and disease duration were not significantly associated with ADI changes. During the 4-year follow-up period, no new cervical disorders developed in patients with RA; however, elevated MMP-3 levels were significantly associated with worsening ADI. Notably, patients without cervical disorders at baseline exhibited worse progression of ADI over time compared with those with pre-existing cervical involvement. These findings suggest that biochemical disease activity, particularly MMP-3, may play an important role in cervical spine instability in RA, even in the absence of overt radiographic progression.