Objective: To evaluate the loss of trabecular and cortical bone mineral density in radius, ulna and tibia of spinal cord injured persons with di erent levels of neurologic lesion after 6, 12 and 24 months of spinal cord injury (SCI). Design: Prospective study in a Paraplegic Centre of the University Hospital Balgrist, Zurich. Subjects and methods: Twenty-nine patients (27 males, two females) were examined by the highly precise peripheral quantitative computed tomography (pQCT) soon after injury and subsequently at 6, 12 and in some cases 24 months after SCI. Using analysis of the bone mineral density (BMD), various degrees of trabecular and cortical bone loss were recognised. A rehabilitation program was started as soon as possible (1 ± 4 weeks) after SCI. The in¯uence of the level of neurological lesion was determined by analysis of variance (ANOVA). Spasticity was assessed by the Ashworth Scale. Results: The trabecular bone mineral density of radius and ulna was signi®cantly reduced in subjects with tetraplegia 6 months (radius 19% less, P50.01; ulna 6% less, P40.05) and 12 months after SCI (radius 28% less, P50.01; ulna 15% less, P50.05). The cortical bone density was signi®cantly reduced 12 months after SCI (radius 3% less, P50.05; ulna 4% less, P50.05). No changes in BMD of trabecular or cortical bone of radius and ulna were detected in subjects with paraplegia. The trabecular BMD of tibia was signi®cantly reduced 6 months (5% less, P50.05) and 12 months after SCI (15% less, P50.05) in all subjects with SCI. The cortical bone density of the tibia only was decreased after a year following SCI (7% less, P50.05). No signi®cant di erence between both groups, subjects with paraplegia and subjects with tetraplegia was found for tibia cortical or trabecular BMD. There was no signi®cant in¯uence for the physical activity level or the degree of spasticity on bone mineral density in all subjects with SCI. Conclusions: Twelve months after SCI a signi®cant decrease of BMD was found in trabecular bone in radius and in tibia of subjects with tetraplegia. In subjects paraplegia, a decrease only in tibia BMD occurred. Intensity of physical activity did not signi®cantly in¯uence the loss of BMD in all subjects with para-and tetraplegia. However, in some subjects regular intensive loading exercise activity in early rehabilitation (tilt table, standing) can possibly attenuate the decrease of BMD of tibia. No in¯uence was found for the degree of spasticity on the bone loss in all subjects with SCI. Spinal Cord (2000) 38, 26 ± 32