Questioning the Accuracy of a Recent Review of Osteoporosis Medications

TO THE EDITOR: Because the choice of osteoporosis drugs often depends on their nonskeletal effects, it is important that clinicians have access to accurate information about their benefits and risks. In this regard, the article by MacLean and colleagues (1) fails in several respects. MacLean and colleagues report that ibandronate protects against upper gastrointestinal perforations, ulcers, and bleeding (odds ratio, 0.33 [95% CI, 0.14 to 0.74]). This finding is based on 2 studies (2, 3), neither of which reports a difference between ibandronate and placebo. In addition, the authors did not include important benefits and risks of some drugs. Raloxifene is approved by the U.S. Food and Drug Administration for breast cancer prevention, but the article does not mention that raloxifene has been shown to reduce breast cancer risk. The technical report leading to this conclusion (4) cites only 2 studies (5, 6), 1 of which does not specifically report on breast cancer (5) and omits data from the pivotal MORE (Multiple Outcomes of Raloxifene Evaluation) (7), RUTH (Raloxifene Use for the Heart) (8), and STAR (Study of Tamoxifen and Raloxifene) (9) trials. The article also omits any mention of the risk for endometrial cancer associated with tamoxifen. Here too, the technical report erroneously concludes that endometrial cancer is not an adverse effect of tamoxifen, citing the article by Fisher and colleagues (10). However, Fisher and colleagues reported an increased risk for endometrial cancer associated with tamoxifen (relative risk, 2.53 [CI, 1.35 to 4.97]). Finally, neither the paper nor the technical report mentions the increased risk for hot flashes that has been conclusively demonstrated for tamoxifen (10) and raloxifene (7). Although not mentioned in MacLean and colleagues' article, the technical report concludes that risedronate protects against musculoskeletal events, citing 10 studies. However, only 4 are listed in the appendix of the report as having a protective effect. In the largest of these studies (11), the direction of the effect reported in the appendix is opposite that reported in the original article. Also, the technical report concludes that teriparatide reduces cancer risk. Two of the 3 studies cited found identical incidence rates for treatment and placebo, whereas the third (12) shows a barely significant difference (P= 0.05). This makes it improbable that a pooled estimate could be statistically significant. This article is based on a technical report that is in the public domain (4). We urge the authors of the article and technical report, as well as other interested scientists, to carefully review these materials and their conclusions. Making appropriate corrections should be a priority.