Objective. While remission is possible in patients with ankylosing spondylitis (AS), it is often unclear what attitude should be adopted once remission has occurred. We investigated whether dosage adjustment is an effective means of maintaining remission. Methods. This was a retrospective study drawn from clinical situations. Remission was defined using clinical measures [Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) ≤ 20/100 and no peripheral joint disease] and biological measures [C-reactive protein (CRP) levels ≤ normal value]. The tumor necrosis factor-α (TNF-α) inhibitors used were infliximab, adalimumab, and etanercept. Response predictors of remission were evaluated by logistic regression (age, CRP, HLA-B27 positivity, sex, duration of disease, and anti-TNF-α naivety). CRP and BASDAI were evaluated before and after dosage adjustment at about 6, 12, 24, and 36 months. Results. One hundred eighty-nine patients with AS were included in the study, with a mean followup of 43.5 (± 17.9) months after the introduction of the first anti-TNF-α inhibitor. Mean age was 45.6 (± 12.5) years. Remission had occurred in 65 patients (35%). Significant response predictors of remission were male sex (p = 0.003) and anti-TNF-α naivety (p < 0.001). Dosage adjustment was observed 49 times, and progressively reducing treatment frequency was effective to maintain remission in a large number of patients for 36 months. The cumulative probability of continuing anti-TNF-α after dosage adjustment was 79.0% at 12 months, 70.5% at 24 months, and 58.8% at 36 months. Conclusion. Remission had occurred in 35% of the patients with AS under anti-TNF-α inhibitor therapy. Dosage adjustment and progressively reducing treatment frequency was effective in maintaining remission.