OBJECTIVE Osteoclast-mediated bone loss in the hand predicts future bone erosions in patients with rheumatoid arthritis (RA). Osteoclast activity depends on RANKL, which is inhibited by denosumab, an investigational fully human monoclonal antibody against RANKL. We measured metacarpal shaft cortical bone thickness using a novel computer-based technique, digital x-ray radiogrammetry (DXR), to evaluate the effects of denosumab on cortical bone in RA. METHODS Patients (n = 227) with active, erosive RA were randomized to receive subcutaneous denosumab 60 mg or 180 mg or placebo every 6 months. All patients received stable doses of methotrexate and daily calcium and vitamin D. For this blinded post hoc analysis (n = 218), cortical bone loss was determined by DXR using computer-assisted measurement of cortical thickness and shaft width at 21 midshaft levels of the second through fourth metacarpal bones of both hands. RESULTS At 12 months, patients receiving denosumab had significantly less metacarpal bone loss versus placebo (denosumab 60 mg: -0.0034, denosumab 180 mg: 0.0001 gain, placebo: -0.0108; P < or = 0.01 for both denosumab doses). Twelve-month decreases from baseline greater than the smallest detectable change occurred in 2 patients in the denosumab 180 mg group, 9 patients in the denosumab 60 mg group, and 12 patients in the placebo group. Negative correlation was significant between static cortical thickness ratios and static erosion scores (6 and 12 months), and for placebo, between changes in erosion scores and changes in cortical thickness ratio. CONCLUSIONS Twice-yearly injections of denosumab with ongoing methotrexate treatment significantly reduced cortical bone loss in RA patients for up to 12 months. These results add to the growing evidence supporting the clinical utility of DXR.