Objective. To investigate, in a cohort of patients with ankylosing spondylitis (AS) adequately treated with infliximab, changes over time in presenteeism and the role of presenteeism relative to that of the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and the Bath Ankylosing Spondylitis Functional Index (BASFI) in predicting sick leave. Methods. Data were analyzed from 71 patients with paid work and taking a stable dose of infliximab participating in a 96-week study with 5 assessment points. Covariates included presenteeism, sick leave, time, sex, age, BASDAI, BASFI, Bath Ankylosing Spondylitis Metrology Index, and part-or full-time work. Presenteeism represented the AS impact on productivity (by visual analog scale, range 0-10, where 10 5 completely unproductive). Sick leave represented the number of days absent from work due to AS in the last 6 months. A linear mixed-effects model for presenteeism, and hurdle and zero-inflated count models for sick leave were explored. Results. Mean 6 SD presenteeism ranged from 2.2 6 2.2 to 3.8 6 7.8, and sick leave occurred in 8-17% of the patients during the 6-month period. Presenteeism positively correlated with BASDAI and BASFI, but was not significantly influenced by time. The chance of incurring sick leave was affected by presenteeism but not by BASDAI and BASFI. Conditional on being absent from work, the effect of presenteeism on the length of sick leave (in days) was much stronger than BASDAI and BASFI. For presenteeism ‡5, an increase of 1 unit in presenteeism yielded an increase by 36-40% (or 2-12 days) in the length of sick leave during the following 6 months. Conclusion. Presenteeism, even measured by a simple visual analog scale, was an important factor to explain future sick leave.