We studied bone mineral metabolism prospectively in 113 children with chronic rheumatic diseases (juvenile arthritis, systemic lupus erythematosus, and juvenile dermatomyositis) to determine the relationship of serum levels of osteocalcin to rheumatic disease activity and corticosteroid usage, and to determine, in part, the cause of osteopenia in this population. Disease activity was quantitated by historical, clinical, and serologic means and an activity score derived. The 113 children were divided according to the expression of their disease, which was active (group 1: mean score 3.42, mean erythrocyte sedimentation rate 28 mm/hr) or inactive (group 2: score 1.7, erythrocyte sedimentation rate 15 mm/hr) (p less than 0.02 group 1 vs group 2 for each value), or which remitted during the study (group 3). We found that serum levels of osteocalcin, but not those of ionized calcium, 25-hydroxyvitamin D, 1,25-dihydroxyvitamin D, and parathyroid hormone, were reduced in group 1 children even before corticosteroid therapy was employed. Children in both group 2 and group 3 had normal osteocalcin levels despite the use of corticosteroids. The reduced levels of osteocalcin were predictive of a reduction in bone mass measured by photon absorptiometry in 16 of 19 children so studied. We conclude that skeletal abnormalities that result in a reduced bone mass occur in the clinical course of the majority of children with active chronic rheumatic diseases, are associated with reduced osteocalcin levels, and are not related to the use of corticosteroids. Serum osteocalcin levels may be a sensitive marker for reduced osteoblast activity and bone formation in children with chronic rheumatic diseases.