Statins and Fracture Risk

To the Editor: Dr van Staa and colleagues¹ concluded that "use of statins . . . was not associated with a reduction in risk of fracture." This conclusion is contrary to the results of our study² as well as others.^3- 4 Our results, like those of van Staa et al, were derived from the General Practice Research Database (GPRD). However, the study by van Staa et al encompassed 683 practices originally considered to have data of satisfactory quality; by contrast, the Boston Collaborative Drug Surveillance Program conducts ongoing internal (unpublished) validation studies to identify practices that provide unsatisfactory data. As a result, several practices were excluded and our recent study encompassed only 340 practices. Van Staa et al also failed to exclude patients at high risk for fracture, for instance patients with osteoporosis, osteomalacia, cancer, alcoholism, and those receiving bisphosphonates. We excluded such patients since it is well recognized that including persons who are already at high risk for an outcome will dilute the effect of a drug on the outcome.⁵ These methodological differences resulted in van Staa et al identifying 81 880 fracture cases, while we identified only 3940.