Depression is widespread among elderly persons, affecting at least one of every six patients treated in general medical practice and an even higher percentage of those in hospitals and nursing homes. Elderly persons have the highest suicide rate of any age group, and a range of physical disturbances is related to or intensified by depressed states.

This paper describes current knowledge about the prevalence, clinical burden, and treatment of geriatric depression. It discusses the expected state of affairs related to geriatric depression in 2010 and suggests elements of strategic planning needed to optimize outcomes.

Magnitude of the problem

Prevalence estimates for clinically significant depression in later life range from about 10 percent for elderly persons living independently in the community to about 25 percent for those with chronic illness, especially persons with ischemic heart disease, stroke, cancer, chronic lung disease, arthritis, Alzheimer's disease, and Parkinson's disease (1,2,3,4,5,6,7). These data underscore the inseparability of mental and physical health in aged persons. They also point to the need for clinical trials using agents likely to be safe and well tolerated by elderly depressed patients who are heavily burdened with chronic medical illnesses and depletion of psychosocial resources.

Certainly the most compelling consequence of depression in later life is increased mortality from both suicide and medical illness. Elderly persons have the highest suicide rate of any age group. This high rate is largely accounted for by older white males, with rates rising to 67.6 suicides per 100,000 among white males age 85 and older, more than 5.5 times the overall national rate of 12 per 100,000 (8). Suicide among elderly persons is most likely to be a result of depression. Indeed, among patients 75 years of age and older, 60 to 75 percent of those committing suicide have diagnosable depression (9). Hopelessness is strongly linked to the presence of suicidal ideation and tends to persist, even during periods of clinical remission, among those with a history of suicide attempts (10).

These data underscore the need to develop effective and efficient disease management strategies for the prevention of suicide among elderly patients attending primary care clinics. This need has become a priority of the National Institute of Mental Health in its recent initiative in suicide prevention; its feasibility is supported by the work of Katon and associates (11) and Meyers (12).

In addition, the relationship between geriatric depression and nonsuicidal mortality is well supported for myocardial infarction, where depression elevates mortality risk by a factor of 5 (13,14). In nursing home patients, major depression has been found to increase the likelihood of mortality by 59 percent independent of physical health measures (15). Hence the selection of treatment modalities for the short- and long-term management of geriatric depression may be, literally, a matter of life and death.

Some 800,000 older Americans are widowed each year. Grief following the death of a loved one is an important risk factor for both major and minor depression. At least 10 to 20 percent of widows and widowers develop clinically significant depression during the first year of bereavement (16). Without treatment, such depressions tend to persist, becoming chronic and leading to further disability and impairments in general health.

Bereavement depression often coexists with another dimension of emotional distress, which has been termed traumatic grief (17). The symptoms of traumatic grief appear to be a mixture of both separation and traumatic distress; such symptoms are extremely disabling and are associated with functional and health impairment and with persistent suicidal ideation. Appropriate treatments for traumatic grief need to be developed and tested.

Persistent insomnia, which occurs among 5 to 10 percent of elderly persons, is a known risk factor for the subsequent onset of new cases of major depression among both middle-aged and elderly persons (18). Longitudinal fluctuation in sleep complaints among elderly persons residing in the community covaries with the intensity of depressive symptoms after health status, gender, and age are controlled for (19). Sleep disturbances in older men and women have also been recently linked to poor health, depression, angina, limitations in activities of daily living, and the chronic use of benzodiazepines (20). Persistent or residual sleep disturbance among elderly patients with prior depressive episodes augurs a less successful maintenance response to either pharmacotherapy (21) or to interpersonal psychotherapy (22).

The estimated prevalence of chronic, primary insomnia among elderly persons is 5 to 10 percent (23). Relatively little is known about the etiology or pathophysiology of chronic primary insomnia and about why it constitutes a risk factor for depression among adults and geriatric subjects. An important and promising area for further research is whether effective treatment for chronic insomnia would serve to prevent the subsequent development of clinical depression in midlife and later life.

A final pathway to geriatric depression, suggested by computed tomography and magnetic resonance imaging studies, may involve structural neuroanatomic factors. Enlarged lateral ventricles, cortical atrophy, increased white matter hyperintensities, decreased caudate size, and vascular lesions in the caudate nucleus appear to be especially prominent in late-life depression associated with vascular risk factors (23). These findings have generated the vascular hypothesis of geriatric depression, which states that even in the absence of a clear stroke, disorders that cause vascular damage, such as hypertension, coronary artery disease, and diabetes mellitus, may induce cerebral pathology that constitutes a vulnerability factor for depression (24).

Clinical issues

A recent review by the Geriatric Psychiatry Alliance concluded that depression is widely linked to increased utilization of health care services (25). For example, Luber and associates (26) reported that depressed patients required almost twice the number of medical appointments required by nondepressed patients, with almost half of the depressed patients making five or more office visits yearly. Luber and colleagues (26) also observed far more frequent use of multiple medications by depressed patients. In addition, they found the frequency of lengths of stay per hospital admission that were over diagnosis-related-group norms to be twice as great among depressed patients as among control subjects. These findings could be due to a greater burden of medical disease among depressed patients, to depression itself, or to the interaction of depression and other medical diseases.

Depression amplifies the disability of medical disease. After controlling for the effects of medical comorbidity, Unutzer and colleagues (27) found that outpatient visits by depressed primary care patients increased in proportion to the severity of depression, with total median annual health care costs twice as high for those with severe depression compared with nondepressed control subjects. In the nursing home environment, Fries and colleagues (28) found that depression significantly increased demands on nursing home time.

Presenters at the 1991 National Institutes of Health (NIH) consensus development conference on the diagnosis and treatment of late-life depression confirmed that one of six elderly patients seen in primary care is affected by depression, but only one in six of those is diagnosed and treated appropriately (29). Because older persons obtain most of their medical and mental health care from the general medical sector, data such as these underscore the need for more effective and efficient treatment of depression in primary care and for studies showing that antidepressant treatment for elderly patients works in geriatric primary care settings. The fundamental problem in the primary care treatment of geriatric depression is undertreatment (26,27). The most common errors are underdosing and discontinuation of medication too quickly. Recent demonstration research in this area includes the Hartford-MacArthur minor depression study and the Hartford depression management study.

Our recent reviews of 30 years of treatment studies of chronic insomnia among nonelderly subjects (30) and elderly subjects (31) have shown only short-term clinical trials evaluating the efficacy of benzodiazepines and other sedative hypnotics. Clinicians' concerns about benzodiazepine dependence have led to a 30 percent decrease in benzodiazepine prescription and a 100 percent increase in the use of antidepressants as hypnotics over the past decade, despite the absence of efficacy data from controlled clinical trials (32). It is clearly a priority to inform geriatric practice clinicians about such data and to determine if appropriate long-term management of insomnia can reduce the incidence of geriatric depression.

Treatment issues

The hallmark of depression among older people is its comorbidity with medical illness (25). There are many pathways to depression in later life, including medical illness and its attendant disability, psychosocial stressors such as bereavement, chronic insomnia, and genetic liability. At the same time, geriatric depression has many costs and consequences, including suicide, anxiolytic dependence and alcoholism, cognitive impairment, disability, medical symptoms, increased health care utilization, and excess mortality.

Researchers consistently find that elderly patients, even the very old, respond to antidepressant medication (33). Patients 75 years old and older typically have a high prevalence of medical comorbidity, and physicians are often reluctant to add another medication to an already complex regimen for a frail individual. However, because very old patients are also at high risk for adverse medical outcomes of depression, risk-benefit considerations favor treatment. Selective serotonin reuptake inhibitors (SSRIs) have become the first-line treatment for depression in later life because of their favorable side-effect profile and safety in case of overdose. In standard controlled studies, SSRIs appear to be as effective as tricyclic antidepressants in the treatment of elderly depressed patients.

Response to acute and continuation treatment with combined nortriptyline and interpersonal psychotherapy is similar between so-called "young old" patients (primarily in their 60s and early 70s) and patients in their 30s and 40s, with 80 percent success rates in samples of patients who completed treatment in both age groups (34). However, the time to remission is about two weeks longer among elderly patients, and elderly patients have twice the rate of relapse during continuation treatment than do younger patients, about 15 percent versus 7 percent.

We have estimated the rate of treatment resistance to combined treatment with nortriptyline and interpersonal psychotherapy to be about 18 percent among elderly patients with recurrent nonpsychotic unipolar depression (35). Finally, we have shown that nortriptyline and interpersonal psychotherapy both singly and in combination are effective maintenance treatments for late-life depression over a three-year period and that the combined treatment is more effective than either alone in preventing recurrence. The value of combined antidepressant medication and psychotherapy is particularly apparent among patients aged 70 and above (36).

A related issue is that although 80 percent of elderly patients with moderate to severe unipolar depression can be expected to respond well to antidepressant treatment—especially combined treatment with medication and psychotherapy—the clinical response to antidepressant treatment in later life follows a variable temporal response trajectory, with a median time to remission of 12 weeks. In addition, reliable prediction of whether elderly patients have responded to antidepressant medication is generally not possible before four to five weeks of treatment.

The delayed onset of clinical antidepressant activity among elderly patients is particularly problematic, prolonging the duration of suffering and disability, potentially reducing compliance (and thus diminishing ultimate recovery rates), not infrequently shifting therapy from pharmacotherapy to electroconvulsive therapy, and increasing the risk for attempted or completed suicide. Thus the development of strategies to accelerate treatment response and to improve the early recognition of patients who are not responding to antidepressant treatment would be an important advance.

The determinants of the temporal profile of the response trajectory in late-life depression are the subject of much current research interest. We have found that the determinants of slower and more brittle response include older age, the presence of acute and chronic stressors, higher levels of pretreatment anxiety, lower levels of perceived social support, and greater biologic dysregulation as evidence by higher levels of rapid eye movement sleep (37). The temporal profile of the initial response trajectory may also provide important clues about which patients are likely to remain well on maintenance treatment and which are likely to have a brittle treatment response and a stormy course of long-term illness.

A recent update of the NIH consensus development conference on the diagnosis and treatment of late-life depression underscored the still unmet need for more data to guide long-term treatment planning, especially in the treatment of patients with major depression who are older than age 70 (38). The question is whether there is any difference in speed and rate of remission, relapse, recovery, and recurrence between patients age 60 to 69 and those age 70 and above.

To address this question, we recently analyzed a data set on 180 patients available through the Pittsburgh study of maintenance therapies in late-life depression, funded by the National Institute of Mental Health (NIMH). The two groups of patients —those age 60 to 69 (N=113) and those age 70 and older (N=67)—showed comparable times to remission and recovery, as well as similar absolute rates of remission during acute therapy, relapse during continuation therapy, and recovery. However, patients age 70 and older experienced a significantly higher rate of recurrence during the first year of maintenance therapy—60.5 percent (23 of 38 patients), compared with 30.4 percent (21 of 69 patients) for patients age 60 to 69. Thus, although response to acute and continuation treatment with combined nortriptyline and psychotherapy was strikingly similar among patients in the two age groups, those age 70 and older showed a much more brittle long-term treatment response, as evidenced by a higher rate of recurrence during the first year of maintenance therapy (39). This observation poses the major therapeutic and public health challenge of how best to protect those age 70 and older from recurrence of major depression.

Data from naturalistic studies have identified several predictors of relapse and recurrence among patients with geriatric depression, including a history of frequent episodes, first episode after age 60, supervening medical illness, a history of myocardial infarction or vascular disease, high pretreatment severity of depression and anxiety, and cognitive impairment, especially frontal-lobe dysfunction. These factors appear to interact with a low level of treatment intensity in determining more severe courses of illness.

Despite the evidence that a high level of treatment intensity is effective in preventing relapse and recurrence (36), naturalistic studies have shown that the intensity of treatment prescribed by psychiatrists begins to decline within 16 weeks of patients' entry into treatment and about ten weeks before recovery (40). In this context, also, residual symptoms of excessive anxiety and worrying among elderly depressed patients predict early recurrence after continuation treatment is tapered (12).

A recent landmark study by the World Health Organization concluded that unipolar major depression and suicide accounted for 5.1 percent of the total global burden of disease in 1990, measured in disability-adjusted life years, making depression the fourth most important cause of global burden (41). The study also showed that the significance of illness burden attributable to depression increases with age weighting and is projected to grow further by the year 2020, based on demographic shifts toward a greater proportion of elderly in the population, especially of the "old old."

These demographic data underscore the central point of this report: developing clinical preventive strategies to reduce the incidence of geriatric depression and treatment strategies to accelerate treatment response, to improve early recognition of treatment resistance, and to maintain treatment response, especially among patients age 70 and older, is of first-rank public-health importance. Preventive and treatment strategies are needed to enhance the effectiveness and the efficiency of care along the continuum of settings in which elderly depressed patients are to be found, from primary care outpatient services and acute hospital settings to long-term-care settings. A recent report from the NIMH National Advisory Mental Health Council on ways of bridging science and services underscores these priorities (42).

Several elements of strategic planning necessary for optimizing outcomes for management of geriatric depression in the year 2010 can be outlined.

Current strengths

A major strength in the current state of affairs is the availability of treatments that have demonstrated efficacy. Effective pharmacotherapeutic options include the secondary tricyclic antidepressants and the SSRIs (36). Effective psychotherapeutic options include interpersonal psychotherapy (43), cognitive-behavioral therapy (44), and problem-solving therapy (45,46).

Another strength is an improved understanding of risk factors for geriatric depression and of factors that complicate or slow treatment response. Risk factors such as vascular disease and chronically poor sleep are amenable to intervention and may therefore represent opportunities for preventing late-life depression. The interplay of medical and psychosocial factors in geriatric depression can be effectively dealt with through the use of treatment approaches that combine medication and psychotherapy. Combined treatment approaches are probably the most potent for maintaining wellness in those age 70 and older, but this hypothesis remains to be tested definitely (36).

Current weaknesses

An important weakness is the existence of barriers to diagnosis. Some of these barriers reflect the nature of the disorder: geriatric depression occurs in a complex medical and psychosocial context. Clinicians commonly misattribute significant symptoms such as anorexia, weight loss, sleep impairment, and fatigue to physical illness rather than to depression. Psychosocial antecedents such as loss of a loved one, combined with decrements in physical health and sensory impairment, can also divert attention from clinical depression.

Other barriers to diagnosis are patient related. Depression can and frequently does amplify physical symptoms, and many elderly patients may deny psychological symptoms of depression or reject the diagnosis because of the stigma attached to it. The effects of stigma on acceptance of the diagnosis seem particularly significant among men, who also have the highest rates of completed suicide in later life.

For several reasons, providers may be reluctant to inform an elderly patient of a diagnosis of depression. They may be uncertain about the diagnosis, reluctant to stigmatize, or uncertain about optimal treatment. They may lack access to psychiatric care and may have continuing concern about the cost-benefit ratio of treatment intervention.

Finally, because the health care system itself is increasingly restricting the time spent in patient care, attention to mental health concerns must compete with attention to comorbid conditions. The growth of Medicare health maintenance organizations, with restrictions in drug formularies and limitations in psychotherapy services, also impedes the provision of geriatric mental health services. Given the inseparability of mental and physical health in later life, this state of affairs is particularly egregious.

Opportunities

Major opportunities and priorities for research in geriatric depression derive clearly from the current state of affairs, from what is known and practiced, and from projections of the likely state of affairs in 15 to 20 years. These priorities must include:

• Improvement of the recognition and treatment of depression among elderly patients seen in the general medical sector, where most elderly patients are treated, including the development of information on potential ethnic or racial differences in treatment response

• Prevention of suicide among elderly patients seen in primary care

• Development of methods to accelerate the onset of antidepressant treatment response

• Improvement of the early recognition of treatment resistance in late-life depression and development of treatment strategies for improving response among patients whose depression does not respond to first-line treatments

• Development of maintenance treatments with long-term efficacy in late-life depression, especially for patients age 70 and older, who are more likely to have a brittle response to treatment

• Development of preventive strategies to reduce the incidence of late-life depression.

The key to improving recognition of late-life depression is a high index of suspicion. We need to help our general medical colleagues learn to elicit symptoms proactively through the use of screening and direct questions. Scales particularly well suited to general medical practice include the PRIME-MD (47), the Center for Epidemiological Studies Depression Scale (CES-D) (48), and the Geriatric Depression Scale (49). Useful questions in uncovering depression include "Are you sad?" "Are you sleeping poorly?" "Do you worry too much?" and "What have you enjoyed doing lately?"

Treatment strategies involving the SSRIs and short-term psychotherapies, such as problem-solving therapy, cognitive-behavioral therapy, and interpersonal psychotherapy, are particularly appropriate for treatment-transfer studies of efficacy, effectiveness, mitigation of hopelessness, and suicide prevention in the primary care sector.

The application of strategies such as therapeutic sleep deprivation (50) combined with studies of brain metabolism, blood flow, structural integrity, and neurotransmitter binding may further clarify correlates of treatment resistance. In addition, the use of combined medication and psychotherapy for improving long-term treatment response and maintaining the gains of acute therapy seems to be particularly promising in treatment of patients age 70 and older, but this strategy requires further controlled evaluation. Finally, the mitigation of vascular risk factors and treatment of chronic insomnia may represent effective approaches to preventing late-life depression.

Several challenges to implementation need to be faced—for example, how to help general medical colleagues learn to elicit symptoms of depression, how to get rating scales into physicians' offices, and how to help medical colleagues avoid errors in practice without offending them. Collaborative approaches involving nurses who specialize in treatment of depression, with rapid, direct access to psychiatric backup, may offer practical, effective, and affordable solutions to these challenges.

Threats

Current reimbursement schedules discount services rendered for depression and other mental disorder diagnoses. This strategy may be penny wise, but it is definitely pound foolish.

Problems in diagnosis and treatment also constitute threats to our ability to meet the public health challenges posed by geriatric depression. Depression among elderly patients may have a complex, unclear presentation. It may be mixed with dementia, anxiety, or alcohol or benzodiazepine overuse, or it may manifest itself in severe or psychotic forms. Suicide risk factors may also complicate diagnosis and treatment planning. Elderly patients may be unable to tolerate first-line treatment, may not respond to an adequate trial of treatment, may have poor compliance, and may need more than drug therapy to achieve and maintain wellness. All of these factors represent threats to successfully meeting the public health challenges posed by late-life depression.

In another sense, errors common in general medical practice also represent threats to the improvement of care, as well as opportunities for education. The principal medication management errors include underdosing, failure to consider possible drug interaction, discontinuing treatment too soon, and unskilled polypharmacy. Other related barriers to successful treatment are use of benzodiazepines or anxiolytics as the primary or sole drug; use of tertiary tricyclic antidepressants; failure to monitor outcome, side effects, and compliance; and failure to educate patients and their family members. All constitute threats to successfully meeting the public health challenges of geriatric depression.

Geriatric depression is widespread, has serious health consequences, and entails increased health care costs and increased mortality related to suicide and medical illness. Nonetheless, geriatric depression can be effectively diagnosed and treated. Clinicians need to take a long-term view of its management to keep patients well and to prevent complications such as suicide.

This work was partly supported by grants P30-MH-52247, P30-MH-30915, R01-MH-43832, and K05-MH-00295 from the National Institute of Mental Health.

Dr. Reynolds is professor of psychiatry and neuroscience and Dr. Kupfer is Thomas Detre professor and chairman of the department of psychiatry at the University of Pittsburgh School of Medicine. The authors are also affiliated with the Clinical Research Centers for Mid- and Late-Life Mood Disorders in the department of psychiatry at Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, 3811 O'Hara Street, Pittsburgh, Pennsylvania 15213 (e-mail, reynoldscf@msx.upmc.edu). This paper is part of a special section on meeting the mental health needs of the growing population of elderly persons.