The Pathophysiological Basis of Late-Life Depression

K. Ranga Rama Krishnan, M.D., and David C. Steffens, M.D.

Depression that starts for the first time in late life is relatively uncommon: Less than 1% of elderly patients in the community have depression (Blazer et al., 1987). On the other hand, depression is more prevalent in patients who are medically ill (Koenig et al., 1997). The greater the severity of medical illness, the greater the odds of developing depression. Depression in the elderly is seen to a great extent among hospitalized patients. Unlike depression in early life, genetic factors are less important in depression that starts later in life (Mendlewicz and Baron,1981). Life stress and social factors also appear to be less critical for late-onset depression.

A number of medical diseases appear to increase the risks for depression. They include coronary artery disease, hypertension, diabetes and hypothyroidism. About 18% to 20% of patients with coronary artery disease develop depression (Hance et al., 1996).

Among all the patients presenting with depression for the first time in late life, vascular disease appears to be quite common. In the last century the German psychiatrist Gaupp developed a concept called arteriosclerotic depression. This referred to patients in whom the depression was attributed to hardening of the arteries. A generation ago, Felix Post suggested that subtle vascular disease of the brain was a significant factor in the development of depression in late life (1962). With the advent of computer tomography (CT) and magnetic resonance imaging (MRI), it has become possible to evaluate subtle vascular changes in the brain of elderly patients with depression.

Vascular Changes

MRI is a particularly useful tool for identifying vascular pathology in the brain. Vascular changes are seen as hyperintense signals both in the white matter and gray matter. Large lesions are quite rare. The frequency of the changes increases with aging. In most normal subjects the changes are minimal and are thought to represent potential spaces around perforating vessels.

In 1988, in an early study, we demonstrated that older patients, especially those with late-onset depression (after age 50), had hyperintense signals on the T2-weighted MRI images (Krishnan et al., 1988). Subsequent studies have shown that these changes are present in the white matter and gray matter of patients with depression. The main location of the gray matter changes was in the basal ganglia, in the caudate and putamen nuclei (Husain et al., 1991; Krishnan et al., 1992). The findings were also present in the deep white matter in the frontal part of the brain.

In our early study, changes were present in over 70% of patients with late-onset depression. In a subsequent study of patients referred for electroconvulsive therapy, these changes were identified both on CT and MRI brain scans (Coffey et al., 1989). Many of the patients had risk factors for vascular disease. These findings have since been replicated by numerous studies. Most of the studies have found that the changes occur to a greater extent in patients with a later onset of depression (Figiel et al., 1991a). One autopsy study also supports these hypotheses.

These studies of vascular changes can be quite useful for assessing neuroanatomic substrates of depression. Recently, a group of researchers at Long Island Jewish Hospital showed that the lesions in the left frontal lobe and left basal ganglia are more important (Greenwald et al., 1997). Preliminary data from our group also suggest that these locations are important. Robinson and colleagues (1984), in their study of stroke patients, have also shown that left-sided frontal lesions are more related to poststroke depression than lesions in other locations. Other investigators have suggested that basal ganglia lesions are quite important in the development of depression. Studies of patients with head trauma and brain tumors have also implicated these locations in depression.

These lesion-location studies have helped to frame some of the neuroanatomical substrates of depression. We have elaborated on such a model (Krishnan, 1993), which is outlined in the Figure. Damage to the circuit at any of the locations can engender risk for affective disorder. The aging process may bring further damage to the prefrontal cortex, affecting serotonin and norepinephrine neurons-neurotransmitters which are critical in the regulation of these circuits and in regulating mood.

When these hyperintensities were first described they were called "unidentified bright objects." Since that time a variety of descriptors have been proposed. Table 1 lists some of them. The pathologic basis underlying these changes has been delineated more clearly in the last few years. When these changes are punctate, they represent spaces (called Virchow-Robin spaces) around blood vessels.

When vascular lesions are five millimeters in size and are irregular in shape, then they most likely represent ischemic changes. Upon postmortem pathological examination, they appear as areas of myelin pallor or as infarcts or lacunes.

The large lesions are the ones that appear to be increased in patients with late-onset depression. Thus cerebral vascular ischemic changes appear to be quite important in late-life depression. Although the cross-sectional correlation has been established between these changes and late-life depression, the etiological relationship remains to be proven. A single case report is the only data available demonstrating that the changes precede the onset of depression. Table 2 describes the various pathological changes that have been described.

We recently compared depressed patients with vascular changes in the brain to those without such changes, using data from the National Institute of Mental Health Clinical Research Center for the Study of Depression in Later Life. Eighty-nine patients with unipolar disorder were classified as having vascular change or not, using MRI. Patients with vascular changes were older, tended to present with anhedonia, had a lower family history of alcohol and drug abuse, and greater functional disability (Krishnan et al., 1997). We have termed this "vascular depression," and our unpublished data suggests that older patients with vascular depression are less likely to be improved six months after the initial evaluation.

Vascular Depression

The concept of vascular depression remains to be fully clarified. We currently conceptualize vascular depression as major depression occurring in patients with ischemic vascular changes in the brain. The controversy that surrounds the diagnosis of vascular dementia also applies to the evaluation of this entity. Older patients often develop a number of risk factors for vascular disease. These risk factors include diabetes, high blood pressure, high cholesterol or lipid levels. Patients with these factors would therefore be at high risk for developing ischemic changes in the brain (silent strokes). Patients with vascular changes in the brain develop the depressive disorder in the context of life stress and lack of social support. This model of the development of vascular depression can be tested.

A number of issues regarding vascular depression need further study. First, the importance of lesion location remains to be fully studied. Lesion location studies in this population can provide clues to understanding the brain structures and circuits involved in depression. Second, it may be possible to test whether patients with vascular changes in the brain have an increased risk for developing depression following stress using paradigms to assess response to stress. Third, it may be worthwhile to evaluate whether these lesions are predisposed to blood flow changes in the brain similar to that seen in depression. Fourth, the outcome of patients with vascular depression compared to those without vascular depression needs to be addressed.

It is possible that patients with vascular depression will have higher comorbidity and mortality. They may be also be at increased risk for dementia. Studies from Cornell have suggested that late-onset depressed patients are at increased risk for developing dementia (Alexopoulos et al., 1993). The treatment of vascular depression needs to be examined. Preliminary studies have suggested that patients with vascular depression are more prone to confusion and delirium with medications and convulsive therapy (Figiel et al., 1990). Some have suggested that these patients do not respond well to standard treatment for depression. An important aspect may be the use of concomitant medications to prevent future ischemic episodes and progression of vascular changes. This strategy may include simple measures such as use of aspirin, cholesterol-lowering drugs, and so forth. Additional long-term studies with these types of preventive measures may be quite critical.

In the case of women, it is possible that postmenopausal estrogen therapy may be quite critical. Patients on estrogens may be less likely to develop vascular depression. This is an issue that may have fairly important public health consequences, and needs to be evaluated. Just as in the case of Alzheimer's disease, the role of nonsteroidal anti-inflammatory agents and estrogens needs to be studied. Table 3 describes the potential clinical implications for the diagnosis of vascular depression.

Dr. Krishnan is professor and head of the Division of Biological Psychiatry at Duke University Medical Center.

Dr. Steffens is assistant professor and director of clinical services, Geriatric Affective Disorders Program, at Duke University Medical Center.

This article originally appeared in the February 1998 issue of Medicine & Behavior