Depression and the Risk of Fractures in Later Life: the Health In Men Cohort Study

Abstract Introduction Fractures are common and disabling health events, particularly later in life. The presence of clinically significant depressive symptoms has been associated with increased risk of fractures, and we designed the present study to clarify if this association is likely to be causal or due to the confounding effect of measures associated with both fractures and depression. Method Cohort study of a community-derived sample of 4224 men aged 70 to 88 years at the start of the follow-up period of up to 17 years. Clinically significant symptoms of depression were defined as a recorded diagnosis of depressive episode in the Western Australian Data Linkage System (WADLS) or by a total score of 7 or greater on the 15-item Geriatric Depression Scale. Health contacts associated with fractures were retrieved from WADLS. Other measures included age, past history of fractures, education, smoking, frailty, poor vision, falls, medications, and the concentration of vitamin D, homocysteine, hsCRP and testosterone. Death was considered a competing risk for fractures. Results 911 (21.6%) participants had a bone fracture during follow-up. After adjustment for multiple potential confounders, past and current depression were associated with an increase in the risk of novel fractures; respective odds ratios were 1.41 (95%CI = 1.03, 1.93) and 1.64 (95%CI = 1.20, 2.25). Parsimonious competing risk regression showed that both past and current depression were associated with an increase in the risk of novel fractures: sub-hazard ratio = 1.29 (95%CI = 1.03, 1.63) and 1.27 (95%CI = 1.05, 1.55) respectively. Estimation of confounding due to unmeasured factors showed that a small additional effect could potentially dilute the association between depression and fractures. Discussion History of clinically significant symptoms of depression is associated with an increased risk of future fractures. This association may be due to multiple other associated risk factors, both measured and unmeasured, but nevertheless the presence of depression should alert clinicians to the need to develop a management plan that includes the management not only of depression but also of fracture risk.