The aim of this study was to analyse the heterogeneity of bone mineral density (BMD) reduction across measurement sites in female systemic lupus erythematosus (SLE) patients on glucocorticoid (CS) treatment. The study population consisted of two subgroups: 32 women at a mean (SD) age of 43.2 (12.0) years, SLE duration of 13.4 (6.2) years, treated with a mean cumulative prednisone dose of 34.4 g; and 16 women at a mean age of 36.1 (9.0) years, SLE duration of 3.2 (2.0) years, never treated with glucocorticoids (control group). The participants underwent a standardised interview, medical record review, blood sampling and BMD examination of the lumbar spine, femoral neck and distal forearm by dual-energy X-ray absorptiometry. CS-treated participants were supplemented with daily calcium (1200 mg) and vitamin D (500 UI). During the study mean daily glucocorticoid dose was 10 mg prednisone equivalent. The controls did not receive either corticosteroids or calcium and vitamin D. BMD and laboratory parameters were re-examined at the end of the second year. At baseline 22 (68.7%) of the CS-treated participants had osteoporosis at least at one major site, compared to 18.8% of the controls. The BMD reduction was proportional to the trabecular bone content at the specific measurement site. At baseline mean T scores in the CS-treated group were the highest at the forearm (-1.03 +/- 1.13), followed by the hip (-1.32 +/- 1.26), AP spine (-1.87 +/- 1.46) and lateral spine (-2.90 +/- 1.50). At follow-up lateral spine bone loss was 5.54% per year, the total hip and the forearm lost 3.59% and 0.33%, respectively, compared to annual losses of 1.02% (AP spine), 1.30% (lateral spine), 0.83% (total hip) and 0.11% (forearm) in the control group. The heterogeneity of BMD reduction in our SLE population emphasises the need for the targeted use of bone densitometry in steroid-treated patients. Attention should be paid to trabecular-rich sites, and fracture risk should be specifically determined.