Studies have previously demonstrated that brief (4 weeks) passive range-of-motion exercise is beneficial for bone development in very low birth weight (VLBW) preterm infants. However, the optimal duration of exercise for bone development in preterm infants is yet unknown. The aim of the present study was to examine the effect of 8 weeks of assisted exercise on bone strength and metabolism in VLBW premature infants. Sixteen infants (mean +/- standard error of the mean birth weight 1,009 +/- 55 g and gestational age 27.3 +/- 0.3 weeks) were randomly assigned into exercise (n = 8) and control (n = 8) groups. The intervention started at the first week of life and involved 8 weeks of daily passive extension and flexion range-of-motion exercise of the upper and lower extremities. Biochemical markers of bone turnover were measured at enrollment and after 8 weeks. Bone strength was measured weekly by quantitative ultrasound measurement of tibial bone speed of sound (SOS). Bone SOS decreased significantly in the control group (-108.1 +/- 33.7 m/second, P < 0.0001) during the study period, while remaining stable in the exercise group (11.3 +/- 22.8 m/second). The main beneficial effect of exercise occurred in the first 4 weeks of the intervention. There were no significant differences in the bone turnover marker changes between the groups. There is a significant postnatal decrease in bone SOS in VLBW preterm infants. Eight weeks of assisted range-of-motion exercise attenuates the decrease in bone strength and may decrease the risk of osteopenia in premature infants.