We assessed the effect of a four weeks exercise training intervention on bone turnover markers in premature infants. Twenty-four very low birth weight premature infants were matched for gestational age, birth weight, gender, as well as for corrected age and weight at initiation of the study. Then the subjects were randomly divided into an exercise (n = 12) and a control group (n = 12). Exercise consisted of passive range of motion exercise with gentle compression of both the upper and lower extremities lasting 5 - 10 minutes each day, 5 days per week for 4 weeks. This protocol has been shown to increase bone mineral density in premature infants. Bone formation was assessed by measurements of circulating bone specific alkaline phosphatase (BSAP) and the C-terminal procollagen peptide (PICP). Bone resorption was determined by serum measurements of C- terminal cross-links telopeptide of type-I collagen (ICTP). Training led to a significant (P < 0.05) increase in weight gain (767 +/- 49 versus 586 +/- 24 gr in trained and control premature infants, respectively); and to a significant increase in BSAP (37.2 +/- 14.6 versus 4.1 +/- 8.4 % in trained and control premature infants, respectively). PICP increased also following exercise (34.6 +/- 18.9 versus 5.4 +/- 9.1 % in trained and control subjects, respectively), however, this increase was not statistically significant. Exercise led to a significant decrease in ICTP (-24.7 +/- 3.1 versus -5.5 +/- 5.4 % in trained and control subjects, respectively). A relatively brief exercise intervention was associated with a biochemical evidence of bone formation in very low birth weight premature infants.