For many Americans, aging can be successful, with maintenance of happy, healthy, active lifestyles into the eighth or ninth decade of life. However, this happy scenario is not true for many elderly individuals for whom aging may be associated with medical morbidity, disability, social isolation, and the development of neuropsychiatric disturbances, including depression or a decline in memory and cognition. These disturbances are not minor: both depression and cognitive decline increase disability, mortality rates, and risk of institutionalization while lowering one’s quality of life. When they occur, these neuropsychiatric disorders have devastating outcomes affecting both the individual and the individual’s family. Thus, it is important that we better understand the pathophysiology of how and why these disorders develop, which can potentially guide research toward preventive interventions, wherein we treat the individual before the depression or cognitive decline becomes crippling.

Research examining risk factors for developing depression in later life has provided crucial information about its development and pathogenesis. Unfortunately, much of this literature has limited clinical applicability, as it does not provide information on what factors increase the prospective risk of developing incident depression in a currently nondepressed or mildly depressed older individual. In other words, an astute clinician can identify numerous factors that increase the risk of developing depression, including disability, social isolation, and chronic medical illnesses, but not all individuals with these risk factors will go on to become depressed. Which of these risk factors are the strongest predictors of a new onset of a depressive episode? The answer to this question can inform clinicians and guide intervention studies aimed at developing preventive strategies to reduce rates of incident depression.

In this issue of the Journal, Lyness and colleagues (1) report results from a study examining this very question. The study design involved recruiting over 600 elderly individuals without current major depressive disorder from primary care practices. This population was exceptionally well-characterized across a wide range of clinical and psychosocial variables, and over 400 of them remained in the study for at least 1 year. Using this approach, the authors found that many previously identified risk factors for the development of depression did not prospectively influence incident new-onset depression. However, they found that a specific combination of factors, including subsyndromal depressive symptoms, history of previous depressive episodes, and functional impairment, identified a group at high risk for incident depression.

It is not necessarily surprising that this combination of risk factors substantially increases the risk for incident depression, but what is important is the number-needed-to-treat statistic. The authors report that five individuals with disability, past major depressive disorder, and current low levels of depressive symptoms would need to be treated effectively to prevent one future major depressive episode. Given the complications of depression in an elderly population, a preventive approach for this at-risk population may be quite important to not only prevent psychological suffering but to also avoid the deleterious effects of depression on comorbid medical illness (2). Thus, a preventive approach may be cost-effective based on the number-needed-to-treat statistic, may improve other medical outcomes, and may be acceptable to many patients due to the relative tolerability of selective serotonin reuptake inhibitors and low cost of their generic versions. Alternatively, in those situations where antidepressant medications are not an acceptable option, psychotherapy or interventions targeting functional disability may have preventive efficacy.

There are several subtly important findings in this article that are overshadowed by the primary findings. First, depression is often characterized as a stress response to environmental adversity; however, the authors did not find that stressful life events predicted incident depression. Instead, they found a relationship between the risk of incident depression and perceived family criticism, an interesting finding given how genetic differences may influence the perception of social relationships (3). However, although stressful events did not predict incident depression, stress may be related to memory problems, which are common in late-life depression. The converse of this relationship is also true: depressive symptoms are common in older subjects with mild cognitive impairment. This epidemiological relationship may be particularly salient given how Lyness and colleagues found that almost 75% of their sample exhibited minor or subsyndromal depressive symptoms.

Also in this issue, Peavy and colleagues (4) examine the influence of stress on memory decline. Although others have investigated the relationship between psychosocial stress and memory performance, this study reports on longitudinal assessments of stress, cortisol, and memory in distinct cohorts of older subjects who were either cognitively intact or who showed deficits consistent with mild cognitive impairment. Interestingly, the authors did not find a relationship between measures of salivary cortisol and stressful life events but did find independent effects of each of these measures on cognition. The authors report that stressful life events over 3 years were associated with accelerated cognitive decline, but only in those subjects with mild cognitive impairment. However, the authors also found that increased cortisol levels were associated with a decreased rate of cognitive decline in mildly impaired subjects over the study period. Peavy and colleagues discuss this unexpected finding and propose that the pathophysiological processes underlying mild cognitive impairment may alter cortisol’s effect on hippocampal and memory function, although higher cortisol may ultimately be associated with memory impairment should this population progress from mild impairment to a diagnosis of Alzheimer’s disease (5).

As the authors observe, the association between stress and cortisol is complex. This relationship may depend on numerous environmental and biological factors, including differences in the type or duration of stressors, support systems, individual coping mechanisms and reactivity to environmental stimuli, and genetic differences (6). This relationship becomes even more complex when examined in an older population, who are at risk of exhibiting memory impairment and cognitive changes but also exhibit age-related changes in hypothalamic-pituitary-adrenal axis activity (7). Importantly, when examining these relationships in any adult population, elderly or not, it is becoming evident that it needs to be considered in the context of childhood life events.

Childhood adversity is associated with altered adult hypothalamic-pituitary-adrenal axis reactivity (8), persistent changes in cognitive function through adolescence and into adulthood (9), and increased risk for psychiatric disturbances in adulthood, including depression (10). Interestingly, both the relationship between early-life stress and hypothalamic-pituitary-adrenal axis reactivity and between early-life stress and depression may be moderated by genetic differences (6). Although much of the work examining the relationship between childhood adversity, psychopathology, and genetic differences has been conducted in younger adult populations, there may be comparable relationships in elderly cohorts, where childhood adversity is associated with poorer psychosocial adjustment (11) and risk of depression (12).

The studies by Lyness and colleagues and Peavy and colleagues utilized longitudinal approaches to better understand the development of depression and cognitive dysfunction in older populations. These methodological approaches are particularly crucial in studies of older individuals, who are characterized by ongoing aging effects on the brain and increased risk for dementia and cognitive impairment with advanced age. As demonstrated in this issue, such approaches have strong utility to inform about potential avenues for preventive interventions and also to inform about differences in biological mechanisms contributing to neuropsychiatric disease. Both types of methodological approaches will be crucial as we develop new interventions to improve mental health in later life.

Address correspondence and reprint requests to Dr. Taylor, Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, DUMC Box 3903, Durham, NC 27710; Taylo066@mc.duke.edu (e-mail). Editorial accepted for publication September 2009 (doi: 10.1176/appi.ajp.2009.09091314).

The author reports no financial relationships with commercial interests.