Background/Purpose: Exercise is an important component of Ankylosing Spondylitis (AS) management.  The purpose of this study was to determine if the amount patients exercise associates with better functional outcomes and to assess if medication use contributes to this relationship.

Methods: This is a prospective cohort of 623 AS patients, meeting modified New York criteria followed up to 4 years.  We collected demographic, clinical and self-reported outcomes every 6 months. Participants were queried about exercise habits including number of days per week and minutes per time exercised. A moderate exercise dose was defined as ³120 minutes per week.  Using a mixed effects Poisson regression model, we assessed multivariable associations between independent variables and the Bath Ankylosing Spondylitis Functional Index (BASFI).  This accounted for correlation of repeated measures over time.  Potential confounding, including baseline function, and effect modifications were examined and addressed while developing a final longitudinal multivariable model.

Results: The mean age of patients was 43 ± 14.0 years. The cohort was 72% male and 78% of patients were white.  Mean disease duration was 19 ± 13.4 years. Findings from our final multivariable model indicated BASFI scores for the moderate exercise group were significantly lower than those who exercised less than 120 minutes per week at 1, 2, 3, and 4 years of follow-up (Table 1).  Additionally, there was a significant interaction between exercise and Tumor Necrosis Factor (TNF) inhibitor use on function (p=0.046). When we looked at the effect based on level of exercise, there was a significant association between TNF inhibitor use and better function regardless of exercise status, but a greater benefit in function in those who were both on a TNF inhibitor and exercising above the defined threshold (Figure 1).

Conclusion: AS patients who exercise for at least 120 minutes per week have significantly better long-term function than those who do not.  Additionally, there is a greater likelihood of functional improvement in those on both a TNF inhibitor and moderate exercise program compared to those using a TNF inhibitor alone.

Table 1. Adjusted Associations of Exercise and Other Covariates on Function (BASFI)

^a Rate Ratio of a higher BASFI score based on the independent variables in column 1 adjusted for BASFI at baseline, and all other covariates above.

^b Report of exercise for at least 120 minutes per week.

Figure 1. A: In TNF inhibitor users: estimated BASFI Scores by Exercise Group Over Time

               B: In TNF inhibitor non-users: estimated BASFI Scores by Exercise Group Over Time