Longterm administration of steroid drugs, particularly prednisone, is known to induce osteoporosis, as well as bone growth inhibition and delayed fracture union. Recently deflazacort, an oxazoline prednisone derivative, has been developed to reduce such deleterious effects. We carried out a comparative study in premenopausal patients with rheumatoid arthritis (RA). Sixteen cases whose mean age was 36.5 years and mean disease duration 29 months, all fulfilling ARA criteria, were evaluated in a randomized, double blind trial. Visually identical deflazacort or prednisone capsules were given and patients were instructed to maintain an adequate calcium intake. Laboratory tests focussed on bone mineral density in lumbar spine, femoral neck and Ward's triangle and whole body mineral content. Differences between baseline and 12-month values were processed statistically. Persistent synovitis control proved similar for both drugs and features suggestive of Cushing's syndrome were only found in the prednisone group. The difference in whole body bone mineral content between the deflazacort and prednisone groups just failed to reach statistical significance. In the deflazacort group, the difference between the nonsignificant bone mineral density increase at the femoral neck and the significant decrease in the prednisone group proved statistically significant. Ward's triangle was the most sensitive area to bone mineral density changes in patients receiving prednisone, with a highly significant intergroup difference (p < 0.01). We believe this is the first study on corticosteroid induced osteoporosis, as evaluated by whole body mineral content measurements in premenopausal patients with short term RA, showing that deflazacort is a promising alternative in cases severe enough to require steroid therapy.