Treatment with raloxifene or estradiol hampered arthritis development and joint destruction. Female B10.Q-ncf1*/* mice were sham-operated or ovariectomized, and immunized with collagen II and Freund's incomplete adjuvant supplemented with mycobacteria to induce CIA. Three weeks later treatment was started with raloxifene (60 μg/day; n = 15) (turned triangles), estradiol (1 μg/day; n = 11) (prisms), or vehicle control (Miglyol812) for sham-operated mice (n = 11; squares) and OVX controls (n = 15; triangles). Treatment was continued until day 56 after immunization, and the experiment was terminated on day 73. A: Diagram of the experimental timeline. OVX, ovariectomy; sham, sham operation; immunization with collagen II and Freund's complete adjuvant on day 0. B: Frequency of arthritis. Treatment with raloxifene in OVX mice delayed the onset and the frequency of arthritis compared to vehicle-treated OVX controls (P < 0.05 for the entire experiment). The presence of endogenous hormones (sham-operated mice) or estradiol treatment also delayed the onset of arthritis and decreased the frequency (P < 0.001 and P < 0.001, respectively). There was a significant difference between estradiol treatment and raloxifene treatment (P < 0.001). Kruskall-Wallis test with post hoc comparison was used. C: Histopathologic destruction scores of paw sections. Scores were evaluated in a blinded manner, with the proximal part of each paw graded 0-4, and the distal part graded 0-3, yielding a maximum score of 28 per mouse, as follows: 1 = synovial hypertrophy; 2 = pannus, erosions of cartilage and bone; 3 = severe erosions of cartilage and bone; 4 = complete ankylosis. The scatter plot shows the scores of individual mice, and lines show the median in each group. Kruskall-Wallis test with post hoc comparison was used, **P < 0.01. D: Representative images of paw tissue sections, revealing treatment effects on histologic features in each group.