Many studies have shown that clinical depression or depressive symptoms are associated with a risk for increased mortality from any cause—and in particular death from cardiovascular disease—throughout adulthood. Studies have also indicated that association becomes stronger with age. Now, a new longitudinal study from Australia indicates that while depression is more common in elderly women, it is more dangerous for elderly men.

Further, the study hints that depression may actually be the root of the increased risk—that is, it appears to cause a general decline in health that could signal the end is near.

"Depression may be an early sign of impending physical decline," suggested Kaarin Anstey, Ph.D., senior researcher at the Center for Mental Health Research at Australian National University in Canberra, and the study’s first author.

The study, funded through several Australian governmental research grants, as well as a grant from the U.S. National Institutes of Health, began with 1,947 individuals aged 70 and older in 1992 and followed the group through July 2000. The report appeared in the November/December Psychosomatic Medicine.

Subjects were drawn from those participating in the Australian Longitudinal Study of Aging and were stratified by age and sex. Subjects were assessed during a two-hour home interview that included the Center for Epidemiological Studies Depression Scale (CES-D) and a clinical assessment approximately two weeks later. Subjects were assessed for use of tobacco and alcohol, medical conditions and medications, body mass index, and level of education at baseline and then again at two years. Less substantial follow-up, involving a telephone call, was included at one year and at the third, fourth, fifth, and sixth year for all surviving subjects. For all deaths, mortality status and cause of death were determined through official death-certificate databases.

At the end of the study (July 30, 2000) 1,050 of the original 1,947 subjects were alive and had complete CES-D data on file—55 percent of the original sample. Unadjusted results indicated that survivors were more likely to be younger, female, and nonsmokers; less likely to have reported a medical condition; likely to be taking fewer medications; and likely to have a lower CES-D score at baseline.

After adjusting for age, sex, smoking, alcohol, medical conditions, and medications, the only two demographic variables associated with mortality were age and sex, with men having a statistically significant greater risk than women. Use of alcohol and body mass index were not associated; however, smoking was associated with an increased risk of mortality over time for both men and women, as were a history of cancer, transient ischemic attacks, and higher levels of medication use. Diabetes increased mortality, but only for women.

After controlling for what the authors believed were all the confounding variables, "scores above the clinical cutoff for depression on the CES-D were associated with mortality in the full sample," they reported. "However, this effect was fully explained by depressed men, who had a significantly increased risk of mortality."

By the two-year point, a greater proportion of men had died than women. In addition, the deceased subjects had higher levels of depressive symptoms at baseline than the survivors. Interestingly, those lost to follow-up in the initial two-year period also had a higher level of depressive symptoms than survivors.

Men with "incident depression" (clinically depressed at either baseline or at two years, but not both) had the strongest association with increased risk of mortality. Although the increase was small (an odds ratio of 1.38), the authors characterized the risk as "robust, suggesting that depression may play a role in causing health changes in men." Neither chronic depression (high CES-D scores at both baseline and at two years) nor remittent depression (a high baseline but normal score at two years) had a statistically significant effect on mortality, once comorbid medical conditions were factored into the analysis. In fact, there was a slight—but not statistically significant—decrease in mortality in the remittent-depression group.

"This suggests that treating depression in very old adults may reduce the risk of mortality," Anstey said.

An abstract of "Mortality Risk Varies According to Gender and Change in Depressive Status in Very Old Adults" is posted on the Web at www.psychosomaticmedicine.org/cgi/content/abstract/64/6/880. ▪

Psychosom Med  200264880