Antidepressant Drugs and the Risk of Hip Fracture in the Elderly: Is There More to It Than Confounding by Indication?

Abstract

The risk of bone fracture is increased in the elderly, in persons with depression, and in those who receive antidepressant drugs. Plausible mechanisms explain these risks. It is not clear whether the risk in depression is related to the disorder alone or to its treatment with antidepressant drugs as well; this is because observational studies of the fracture risk associated with antidepressant exposure may adjust analyses for confounding variables but cannot eliminate residual confounding by inadequately measured, unmeasured, and unknown confounds. A recent observational study examined hip fracture in the context of depression and antidepressant treatment using a resourceful method to address confounding by indication. The authors studied the risk of fracture not only in the year after antidepressant initiation but also in the year before antidepressant initiation. They found a significantly increased risk of fracture in all of 10 time windows during the 2 years; the highest risks, in fact, lay in the weeks to months before antidepressant initiation. These findings suggest that unremitted depression and persistence of depression-related mechanisms in remitted depression may together drive the risk of hip fracture in observational studies of antidepressant-treated patients; however, there is evidence from other sources to suggest that a contribution from antidepressant-related mechanisms cannot be ruled out. An additional message is that investigators need to use resourceful methods to address confounding by indication, as is now being done in several fields of research. Finally, readers who evaluate the findings of studies that relate exposures to outcomes must consider how the investigators addressed confounding and must draw conclusions based not merely on the findings but also on issues related to research design.