A depressive disorder may be diagnosed by DSM-IV (or DSM-IV TR) clinical criteria only if the depressive symptoms adversely affect work, family or social functioning, or if the subject seeks or has received treatment for depression. The clinical entities in the cur-rent nomenclature identified in young adults that also occur in older adults are major depression (unipolar or bipolar, single episode or recurrent, with or without melancholia, with or without psychotic features); dysthymic disorder; and atypical depression. Disorders that are common in the elderly include mood syndromes secondary to a general medical condition, and subsyndromal depression including minor depression. Depression with cognitive impairment is a com-mon presentation in elderly patients.

Clinical observations and systematic studies suggest that patients with late-life major depression may well be different in important dimen-sions from younger adult patients with major depression. Such differ-ences may reflect age-associated physiological changes, the changes associated with depressive illness (the ravages of a depressed life) and/or the impact of co-morbid medical conditions^5–7. For example, the differences in young versus old people may manifest in how functional impairment as a symptom of depression is expressed dif-ferently through the life cycle. Depression-associated dysfunction in childhood can present as poor performance in school or antisocial behaviour, in young adults as dissatisfaction in work or love, and for older patients as hypochondrias and increased utilization of medical resources⁸.

The cumulative impact of depressive illness itself in patients with onset of illness early in life may be a critical factor. By the time a patient with early onset depression reaches age 60, they have already suffered through recurrent episodes. There is increasing evidence that the number of depressive episodes as well as the total amount of time spent ‘depressed’ in a lifetime is associated with less robust response to treatment and the development of chronic symptoms⁶. Furthermore, some data suggest that the cumulative time spent in a depressed state is correlated with changes in brain structure that may subsequently affect the phenomenology and treatment response in future episodes⁵.

Moreover, changes in the brain that occur with age may also affect how the individual interacts with the environment, and can result in increased vulnerability to depression. Regional brain dysfunction, especially in the pre-frontal areas, is associated with instability of temperament, vulnerability to and the intensification of the experi-ence of distress, and a degrading of coping mechanisms. Such alter-ations in an individual’s psychic capacities may result in increased social isolation, loss of independence, and related behaviours that can precipitate, intensify, or prolong a depressive episode. Thus, by the time a patient with recurrent depression reaches late life, the con-sequences of the disorder itself, and maybe the impact of treatments received, will influence the presentation, psychobiology, treatment response and course of illness during late life.

Alternatively, the differences that exist in younger versus older depressed patients may indicate that late-life depression is a dis-tinct disease entity and, despite some commonality, early onset and late-life depressive illness are significantly different in fundamen-tal aetiological ways. An example of such aetiological heterogeneity is the distinction between the late-life patient with recurrent major depressive illness with onset of the first episode early in life, and the late-life patient with first onset of major depressive illness after the age of 60. The hypothesis that early-onset (before age 50 or 60) and late-onset depression may be, despite similarities, distinct entities is suggested by data from epidemiological, brain imaging, treat-ment, and long-term outcome studies^9–11 . Patients with late-onset depression have a lower rate of family history of depression, and a greater frequency of MRI abnormalities compatible with ischaemic cerebrovascular disease¹⁰. These lesions have been associated with symptoms of cognitive impairment, a decreased and slower response to antidepressant medication and a deteriorating course of illness^12,13.

In summary, the label late-life depression may be attached to a group of patients who share some commonalties but who, nonethe-less, suffer from a heterogeneous group of disorders that differ in aetiology, pathophysiology and prognosis. Indeed, some of these dis-orders may be relatively unique to older adults.

Diagnostic Features of Major Depression in Late Life

Whatever the reason, and whether in a medical or psychiatric setting, the age-associated changes in the presentation of depression undoubt-edly contribute to the problem of under-diagnosis and inadequate treatment of depression in late-life patients. In general, older patients report fewer emotional symptoms, for example depressed mood, and present with more cognitive/somatic symptoms, for instance sleep disturbance, fatigue, pain and memory or concentration difficulties. In fact, the older-age depressed patient may never complain of, or indeed may deny, being ‘depressed’. This presentation of late-life depression has been labelled ‘depression without sadness’¹⁴. Thus, if when making a diagnosis a physician adheres strictly to DSM-IV criteria, than depressive disorder may well be omitted from the dif-ferential diagnosis for an older patient presenting with non-specific physical complaints.

Diagnostic Criteria for Major Depression in the Elderly

Despite the compelling evidence that late-life depression and older people are different, the diagnostic criteria for unipolar depression in DSM-IV are not adjusted for age¹⁵. One of the five required symp-toms must be either (i) depressed mood most of the day, nearly every day; or (ii) markedly diminished interest or pleasure in all, or almost all activities. The other symptoms that are counted to the total of five, but not required, are: (iii) significant weight loss, or decrease or increase in appetite nearly every day; (iv) insomnia or hypersomnia; (v) psychomotor agitation or retardation; (vi) fatigue or loss of energy; (vii) feelings of worthlessness or excessive or inap-propriate guilt; (viii) diminished ability to think or concentrate, or indecisiveness (older patients with depression often have mild cogni-tive impairment, including disturbances in attention, speed of mental processing, and executive function); (ix) recurrent thoughts of death, suicidal ideation with or without a plan.

Though there are no separate diagnostic subtypes unique to late life, there are some considerations that are more relevant to late-life depressed patients than younger patients with the same disorder. The DSM-IV precludes the diagnosis of major depression if it can be established that the symptoms counted towards the diagnosis of depression are a direct physiological effect of a different medical disorder. However, this rule is not supported by evidence, and to the contrary, studies have suggested that the inclusive approach (i.e. if a symptom is present it is counted towards the diagnosis of depression) is justified¹⁶.

There is also the belief that unipolar major depression with psy-chotic features may be a more frequently occurring subtype in older than younger patients. Psychotic depression occurs in 20 to 45% of elderly depressives hospitalized for psychiatric care, and in 3.6% of elderly depressives living in the community¹⁷.

Hickie etal.¹² hypothesized that there may exist a vascular depres-sion subtype that is unique to late life; he wrote that ‘cerebral vascular insufficiency in elderly people leads to major changes in their subcor-tical and basal ganglia structure. The resultant late-onset depressive disorders are characterized by deficits in functions that are depen-dent on intact cortical striatal connections (e.g. psychomotor speed), as well as subcortical hyperintense lesions and reduced basal gan-glia volumes.’ The hypothesis emanated from studies that reported that patients with late-onset major depressive disorder (MDD) had higher rates of hyperintensities on MRI compared to patients with early onset MDD^10,12,18. The patients with late-onset depression and hyperintensities also showed deficits on neuropsychological testing including, but not limited to, deficits in executive function^18,19. Thus, structural brain damage secondary to ischaemia creates a vulnerabil-ity to MDD that may be precipitated by negative life events or other stresses. Given that vascular disease of sufficient severity is the cause of the ischaemia, this form of depression will be significantly more prevalent in older adults²⁰.

Much about the vascular depression hypothesis is compelling. First, the evidence that vascular disease is the cause of the hyperintensities prominent in late-onset depression supports the hypothesis. Second, vascular risk factors, such as smoking, hypertension and high lipid levels, are prevalent in patients with late-onset depression. Actually, early onset depression is itself a very significant risk factor for both ischaemic heart disease and stroke later in life. Increased cortisol, insulin resistance and platelet reactivity, all of which can contribute to vascular damage and ischaemia, may mediate the increased risk of myocardial infarction and stroke associated with depression⁷.

Diagnostic Criteria for Vascular Major Depression

Criteria for vascular depression include vascular disease, cognitive impairment, late age of onset, and depression. However, the various research groups who study this syndrome have proposed the variables in many different combinations. For example, one group considers clinical and/or laboratory evidence of vascular disease and depression onset after age 65 as cardinal features and executive dysfunction as secondary²¹, whereas another group believes neuroimaging evidence of cerebrovascular disease to be the cardinal feature and age-of-onset over 50 as an additional secondary feature. Krishnan etal.²² have further refined the notion of vascular depression, requiring only MRI evidence of cerebrovascular pathology to define it, and have referred to this syndrome as subcortical ischaemic depression. Alex-opoulos has proposed a depression executive dysfunction disorder of late life that only requires executive dysfunction to meet diag-nostic criteria^21,23. Although the proposed criteria sets are important and have moved the field toward recognizing a potentially impor-tant subtype of late-life major depression, this recognition might be premature because there is no agreement on how the construct is defined. Furthermore, most studies only include patients who meet criteria for MDD; however, many patients with vascular disease have dysthymia or subsyndromal symptoms.

Studies consistently show that elderly males, particularly white males, are the demographic group with the highest risk of com-mitting suicide. Also, the majority of elderly patients who commit suicide have visited their primary care physician a few days to weeks before the act²⁴. These findings appear to be present across countries in several continents, although there are cultural differences in the preferred method of committing suicide²⁵.

Little attention has been paid to the atypical subtype in samples of late-life depressed patients. The term ‘atypical’ was originally coined to describe a group of depressed patients who characteristi-cally present with ‘reverse’ vegetative symptoms such as hyperphagia and hypersomnia²⁶. The diagnostic criteria and acceptance of this subtype of major depression have evolved over the past 25 years so that they are included in DSM-IV with the atypical features speci-fier requiring that a patient has (i) mood reactivity and (ii) two or more of the following symptoms: significant weight gain or increase in appetite, hypersomnia, leaden paralysis, or a long-standing pat-tern of interpersonal rejection sensitivity¹⁵. Studies have documented that patients with atypical depression differ from other patients with major depression in a number of parameters besides phenomenol-ogy, including biological variables, treatment response, demographic characteristics and age of onset^27,28.

There are two studies of the atypical subtype in late-life patients. The first study has many methodological limitations but is notable because it brought attention to this subtype of depression in late life. In this study, a single practitioner in a private practice reported on 358 consecutive DSM-IV unipolar (major depressive and dysthymic disorder N=179) and bipolar II (N=179) outpatients presenting for treatment of a major depressive episode over the last two years²⁹. Depressed patients over the age of 60 had a significantly lower rate of the atypical subtype compared to younger depressed patients, 28 vs. 55% respectively (p0.0001), and also a significantly lower rate of bipolar II atypical subtype in the late-life patients compared to the younger sample, 35 vs. 67% respectively (p=0.01). Furthermore, in comparison to atypical patients under age 60, atypical patients over age 60 had an older age of onset of illness (44 years vs. 23 years) and a lower rate of rejection sensitivity (55 vs. 83%).

The second study reported on an open trial of venlafaxine in late-life depressed patients with the atypical subtype³⁰.The sample was 17 patients, mean age 65.6 ±7.3 years, 77% female, mean Cumu-lative Illness Rating Scale-Geriatric score³¹