Nearly one in six individuals aged 65 or older suffers from depression. They are treated primarily with antidepressants, which have been associated with an increased risk of hip fracture in studies using Medicare claims data. Although these claims data tend to overestimate the relationship between antidepressant use and hip fractures, there remains a significant association, according to a recent study that was supported in part by the Agency for Healthcare Research and Quality (HS10881).
Sebastian Schneeweiss, M.D., Sc.D., and Philip S. Wang, M.D., Dr.P.H., of Brigham and Women's Hospital and Harvard Medical School, used the Medicare Current Beneficiary Survey to determine the association between use of selective serotonin reuptake inhibitor (SSRI) antidepressants and five factors also known to affect hip fracture risk. These five confounding factors, which were not measured in the Medicare claims data, include body mass index (BMI, heavier weight tends to strengthen bones leading to lower risk of fracture), smoking (which increases fracture risk), activity of daily living (ADL) score (limited activity can weaken bones and instability can lead to falls), cognitive impairment (which can lead to falls), and physical impairment (impairment can lead to instability and falls).
Comparing SSRI users with nonusers, the researchers found considerable overestimation of an association between SSRI use and hip fractures. Unmeasured activities of daily living scores (+21.5 percent bias) and impairment scales (+10.6 percent bias) accounted for a moderate proportion (30 percent) of the apparent risk of hip fracture associated with SSRIs. After correction for this bias, SSRIs were associated with a relative risk of hip fracture of 1.8 (nearly a doubling of risk). This risk, which is still significant, was similar to the association found between SSRIs and risk of fracture in a recent clinical study of nursing home residents.
See "Association between SSRI use and hip fractures and the effect of residual confounding bias in claims database studies," by Drs. Schneeweiss and Wang, in the December 2004 Journal of Clinical Psychopharmacology 24(6), pp. 632-638.