The aim of the study was to determine whether 24 wks of robotic locomotor training or activity-based training was sufficient time to induce bone mineral density and body composition changes in individuals with spinal cord injury. This study reports the secondary analysis of a randomized pilot trial. Design: Participants with chronic motor incomplete tetraplegia (N = 16) were recruited. Interventions involved 60-min sessions, 3Â per week, over 24 wks. Robotic locomotor training involved walking in the Ekso GT suit. Activity-based training involved a combination of resistance, cardiovascular, and weight-bearing exercise. Results: Hip bone mineral density was maintained during robotic locomotor training; however, it was significantly reduced ( P = 0.04, effect size = 0.86) during activity-based training by 0.03 (−0.29 to 0.23) g/cm 2 after intervention. Both interventions improved arm fat-free soft tissue mass, but neither group experienced changes in leg fat-free soft tissue mass. The activity-based training group had a significant decrease in visceral adipose tissue ( P = 0.04, effect size = 0.72) and gynoid fat mass ( P = 0.01, effect size = 0.62). Conclusions: Twenty-four weeks of robotic locomotor training is possibly a sufficient duration to prevent the progressive decline of bone mineral density usually occurring in this population. A longitudinal period of activity-based training serves as an effective rehabilitation strategy to reduce indices of fat mass in individuals with spinal cord injury.