This study determined the effect of relocating workers from traditional to activity-based offices on objectively measured sitting patterns. Office workers (n=493) from five office-sites within a large Swedish government agency were included in a controlled study of a natural intervention (2015-2017). At four sites, traditional offices were replaced by activity-based offices, while workers at one site with no relocation acted as controls. Sitting, standing and walking were measured objectively for 5-8days in a sub-sample (n=110) using accelerometry (Actigraph). Total sitting time (% of working time) and time spent in short (<5min), moderate (5-30min) and prolonged (>30min) uninterrupted periods in sitting were determined. Intervention effects were determined at 3- and 12-month follow-ups using linear mixed models adjusted for baseline age, gender and office type, and stratified by office-site (referencing controls). The relocation to activity-based offices did not result in an overall effect (across sites) on occupational sitting time (all p>0.05), while walking time had increased significantly by 1.4% of the working time at 12months compared with controls. Heterogeneous results were found across offices after 12months on total sitting time compared with controls (estimated change -18.3% time-1.4% time), prolonged sitting (change -18.3% to -3.8%), walking (change 0.5%-3.5%) and standing (change -1.4%-13.9%). In conclusion, relocation to activity-based offices had a limited overall effect on occupational sitting patterns in the studied organization, but differed considerably between office sites. Site-specific determinants of sitting behavior in activity-based offices need be identified.