Excess mortality in depression: a meta-analysis of community studies

Abstract

Background: Although most studies examining the relationship between depression and mortality indicate that there is excess mortality in depressed subjects, this is not confirmed in all studies. Furthermore, it has been hypothesized that mortality rates in depressed men are higher than in depressed women. Finally, it is not clear if the increased mortality rates exist only in major depression or also in subclinical depression. Methods: A meta-analysis was conducted to examine these questions. A total of 25 studies with 106,628 subjects, of whom 6416 were depressed, were examined. Both univariate and multivariate analyses were conducted. Results: The overall relative risk (RR) of dying in depressed subjects was 1.81 (95% CI: 1.58–2.07) compared to non-depressed subjects. No major differences were found between men and women, although the RR was somewhat larger in men. The RR in subclinical depression was no smaller than the RR in clinical depression. Limitations: Only RRs of mortality were examined, which were not corrected for important confounding variables, such as chronic illnesses, or life-style. In the selected studies important differences existed between study characteristics and populations. The number of comparisons was relatively small. Conclusions: There is an increased risk of mortality in depression. An important finding of this study is that the increased risk not only exists in major depression, but also in subclinical forms of depression. In many cases, depression should be considered as a life-threatening disorder.

Introduction

The relationship between mortality and mental disorders has been examined systematically for more than 150 years (Brown, 1997, Pokorski, 1994). Most studies in this area have found mental disorders to be associated with an increased risk of death (Harris and Barraclough, 1998). The excess mortality in psychiatric disorders has been hypothesized to be caused by increased suicide rates (Pokorski, 1994), by hazardous health behaviors (smoking, less physical activity, unhealthy eating habits), and by higher incidence rates of accidental deaths because of hazardous activities.

In depressive disorders, it has been suggested that depression itself may cause physiological changes that enhance susceptibility to disease and, consequently lead to death (Penninx et al., 1999). Depressive disorders have been found to adversely affect endocrine, neurologic, and immune processes by increasing the sympathetic tone, decreasing vagal tone, and causing immunosuppression (Irwin et al., 1990, Lesserman et al., 1997). On the other hand, depressive disorders may be partly caused by physical disorders which are responsible for the increased mortality rates. In patients with established disorders, depression may also increase mortality rates by interfering with the patient’s motivation toward recovery, and by affecting compliance with treatment (Carney et al., 1995).

Most studies in this area have concentrated on patient samples (Harris and Barraclough, 1998). Because many mentally ill people do not seek help for their problems, and because it could be assumed that such patients are more ill than those not seeking help, these studies are biased, and may not be the best for examining the relationship between mortality and mental disorders.

In the last decades, several studies examining excess mortality in depression have been conducted in random samples of people living in the community. Most studies in this area indicate that there is excess mortality in depressed subjects, but this is not confirmed in all studies (Weissman et al., 1986, Bruce et al., 1994). One study even found a decreased mortality in subjects with major depression (Fredman et al., 1989). Because of the ambiguity of these results, we conducted a meta-analysis of these studies. In a meta-analyses the results of all studies in this field can be integrated statistically, and an overall mortality rate can be estimated.

Another issue that is examined in the current study concerns the differences in mortality rates between depressed men and women. In some studies in this field, higher mortality rates are reported for men than for women (Penninx et al., 1999, Murphy et al., 1987, Zheng et al., 1997), but in other studies no significant differences are found (Pulska et al., 1998, Roberts et al., 1990).

We also examine differences in excess mortality between subjects with major depression and subjects with subclinical forms of depression. Intuitively, it would make sense if major depression were associated with higher mortality rates than subclinical depression. But, some studies report equally increased mortality rates in subjects with major and in subjects with minor or subclinical forms of depression (Fredman et al., 1989, Sharma et al., 1998, Penninx et al., 1999).