Risedronate reduces the risk of vertebral fractures by up to 70% within the first year of treatment. Increases in bone mineral density or decreases in bone turnover markers explain only a portion of the anti-fracture effect, suggesting that other factors, such as changes in trabecular bone architecture, also play a role. Our objective was to determine the effects of risedronate on bone architecture by analyzing iliac crest bone biopsy specimens using three-dimensional microcomputed tomography (3-D micro CT). Biopsy specimens were obtained at baseline and after 1 year of treatment from women enrolled in a double-blind, placebo-controlled study of risedronate 5 mg daily for the prevention of early postmenopausal bone loss. Trabecular architecture deteriorated in the placebo group (n = 12), as indicated by a 20.3% decrease in bone volume (25.1% vs. 20.0%, P = 0.034), a 13.5% decrease in trabecular number (1.649 vs. 1.426 mm(-1), P = 0.052), a 13.1% increase in trabecular separation (605 vs. 684 microm, P = 0.056), and an 86.2% increase in marrow star volume (3.251 vs. 6.053 mm(3), P = 0.040) compared with baseline values. These changes in architectural parameters occurred in the presence of a concomitant decrease from baseline in lumbar spine bone mineral density (-3.3%, P = 0.002), as measured by dual energy x-ray absorptiometry. There was no statistically significant ( P < 0.05) deterioration in the risedronate-treated group (n = 14) over the 1-year treatment period. Comparing the actual changes between the two groups, the placebo group experienced decreases in bone volume (placebo, -5.1%; risedronate, +3.5%; P = 0.011), trabecular thickness (placebo, -20 microm; risedronate, +23 microm; P = 0.032), and trabecular number (placebo, -0.223 mm(-1); risedronate, +0.099 mm(-1); P = 0.010), and increases in percent plate (placebo, +2.79%; risedronate, -3.23%; P = 0.018), trabecular separation (placebo, +79 microm; risedronate, -46 microm; P = 0.010) and marrow star volume (placebo, +2.80 mm(3); risedronate, -2.08mm(3); P = 0.036), compared with the risedronate group. These data demonstrate that trabecular architecture deteriorated significantly in this cohort of early postmenopausal women, and that this deterioration was prevented by risedronate. Although there is no direct link in this study between fracture and preservation of architecture, it is reasonable to infer that the preservation of bone architecture may play a role in risedronate's anti-fracture efficacy.