Objective
Depression may be associated with increased mortality risk, but there are substantial limitations to existing studies assessing this relationship. We sought to overcome limitations of existing studies by conducting a large, national, longitudinal study to assess the impact of depression on all-cause and cause-specific risk of death.
Methods
We used Cox regression models to estimate hazard ratios associated with baseline depression diagnosis (N = 849,474) and three-year mortality among 5,078,082 patients treated in Veterans Health Administration (VHA) settings in fiscal year (FY) 2006. Cause of death was obtained from the National Death Index (NDI).
Results
Baseline depression was associated with 17% greater hazard of all-cause three-year mortality (95% CI hazard ratio [HR]: 1.15, 1.18) after adjusting for baseline patient demographic and clinical characteristics and VHA facility characteristics. Depression was associated with a higher hazard of three-year mortality from heart disease, respiratory illness, cerebrovascular disease, accidents, diabetes, nephritis, influenza, Alzheimer's disease, septicemia, suicide, Parkinson's disease, and hypertension. Depression was associated with a lower hazard of death from malignant neoplasm and liver disease. Depression was not associated with mortality due to assault.
Conclusions
In addition to being associated with suicide and injury-related causes of death, depression is associated with increased risk of death from nearly all major medical causes, independent of multiple major risk factors. Findings highlight the need to better understand and prevent mortality seen with multiple medical disorders associated with depression.
Depression is well-known to be a prevalent [1], costly [2], and burdensome disorder [3] leading to substantial disability in the United States [4] and worldwide [5]. Although prior research has reported a relationship between depression and mortality, there have been substantial limitations associated with most existing studies. Many studies that have examined depression and mortality have been limited to specific subgroups of patients with illnesses comorbid with depression (e.g. depressed patients with concurrent coronary artery disease [6], diabetes [7], COPD [8], inpatients [9]). Other studies have examined whether depression is associated with a single cause of death (e.g. stroke [10], cardiac disease [11]). Studies are also frequently restricted to include only older adult populations [12], [13] and [14]; broader population-based studies have typically been conducted outside the United States [15] and [16]. Finally, meta-analyses have assessed the relationship between depression and mortality, but include heterogeneous measures of depression, confounders, and approaches to accounting for mortality [17] and [18].
The present study is unique in several ways. First, it examines how depression influences mortality in a large, US population-based, longitudinal study, using analyses that control for a variety of potential confounders. Second, it examines a large range of causes of death in a single population, VA users, who have similar levels of access to medical care. The size of this sample potentially allows the detection of associations that were previously too small to detect, and it allows the examination of different causes of death by age group, which has been infrequently reported in the existing literature.
Prior research by our team has shown unadjusted estimates of age at death and early mortality among Veterans with and without depression, in which we found that depression was associated with earlier mortality and more years of potential life lost (YPLLs) with both natural and unnatural causes of death [19]. Capitalizing on the largest integrated healthcare system in the US, the VHA's expansive electronic medical record (EMR) system, this study sought to quantify the association between depression and mortality from all causes and cause-specific risks of death among all adults, stratified by age, and controlling for a comprehensive set of potential confounders. Findings from this work can serve to inform future clinical and health system planning, as strategies may differ depending on whether and to what extent depression is associated with individual causes of mortality (e.g., depression is associated with a few specific causes of death that could be specifically targeted) compared to multiple or all causes of death (e.g. indicating a broader need for strategies to decrease mortality associated with depression across medical conditions.).
