FormalPara Key Points

  • Depressive symptoms are not synonymous with a depressive disorder.

  • There is a wide range of explanations for depressive symptoms, many of which are not related to psychiatric illness.

  • The presence or absence of melancholic symptoms is a core diagnostic distinction, particularly in the older individual.

  • Effective treatment of a major depressive episode includes pharmacological, social and psychological strategies.

  • The evidence for differential efficacy of antidepressants is modest.

Referrals from geriatricians to psychogeriatric services are common. This is a good thing. Geriatric medicine services are well placed to take a holistic view of a patient’s predicament. Attending to mental health issues, including seeking further advice if required, is an essential part of this task. Clinicians, families or carers of a patient (and sometimes the individual themselves) are often unaware of potentially reversible psychiatric morbidity. Erroneous ideas that link the ageing process itself to inevitable psychological distress or decline may underlie this.

The purpose of this chapter is to provide clinicians in geriatric medicine with a guide for confidently approaching a topic of frequent concern: diagnosis and treatment of depressed mood in the older person. Depression in older people is associated with significant morbidity and poor physical health outcomes [1]. Individuals over 70 years have the highest rates of suicide globally [2]. There is not scope here to cover other important and related issues, such as depressive symptoms in the context of established dementia or assessment of the patient with suicidal thoughts, the latter being an indication for specialist referral.

Harry is a 73-year-old retired train driver who lives in a rural area with his wife. He is being assessed by a geriatrician, as his general practitioner is concerned both about functional decline and increasingly difficult to control diabetes. Harry is an ex-smoker with poorly controlled hypertension for many years. Six years ago he was diagnosed with type 2 diabetes; he has been on oral hypoglycaemic agents but his glycaemic control is poor, and he may soon require insulin therapy. He had coronary artery stenting 2 years ago for ischaemic heart disease. There is no history of previous psychiatric problems, and he has only used alcohol very occasionally. His partner reports a history over the past 2 years (though it may be longer) of increasing lethargy and social withdrawal (he has stopped playing darts and attending an exercise group). He appears sad and flat. Around a year ago, his general practitioner commenced sertraline. The dose is now 100 mg mane. His wife feels he is slowly worsening. She is immensely frustrated that he sits on the couch watching television all day, initiates little conversation, shows minimal interest in family affairs and needs prompting to tackle tasks such as washing the dishes or lawn mowing. Harry himself is unconcerned, and is grateful when a meal is prepared for him. Because he has not responded to the sertraline, the geriatrician requests a psychogeriatric review.

7.1 Diagnosis of Depression

Does Harry have depression? Because of multiple meanings, the word itself can be problematic. In a clinical setting, use of the term depressive symptoms may be more useful. This reflects an awareness that multiple medical and psychiatric disorders present with depressed mood, as do non-pathological reactions to the vicissitudes of later life. Bereavement, later life existential distress, executive dysfunction and physical illness need to be considered. A common diagnostic classification system such as the Diagnostic and Statistical Manual of Mental Disorders 5th edition, known as the DSM-5 [3], includes multiple psychiatric disorders that present with depressed mood. It should come as no surprise that psychiatrists are particularly interested in assessing for major depressive episode (MDE—as part of either a unipolar major depressive disorder or bipolar disorder), as missing this common and largely treatable disorder leads to increased morbidity and mortality. This should not, however, imply that subsyndromal depressive symptoms require no attention—these are common and not only cause significant morbidity but are a risk factor for development of MDE [4]. Further, comorbid psychological problems are also common. For example, a subset of the bereaved develops a mood disorder.

‘Depression’ is commonly understood by the medical community to mean a MDE (a depressive episode), and as this is a major focus of an initial psychogeriatric assessment, diagnostic issues around a MDE will be emphasized here.

Because of the wide differential diagnosis and risks of inappropriate (or non) treatment, it is important to attend carefully to the diagnostic process. This is particularly salient given the increasing awareness of the risks of antidepressant medications in the elderly [5].

Initially, physical causes must be addressed and excluded. This is almost always done when referral from a geriatrician is made. Importantly, however, these problems do not preclude psychiatric comorbidity, so an ‘exclusion’ approach is of limited value. An important exception is delirium. Florid, hyperactive delirium is rarely mistaken for depression; however, the less common presentation of a hypoactive delirium can present similarly to a MDE. If there is a history of very recent onset changes in the context of physical illness (or even if no cause has been identified), caution should be taken in diagnosing a depressive episode—particularly as antidepressant therapy can worsen a delirium from its own adverse effects, such as anticholinergic activity or hyponatraemia.

A full history is necessary, with emphasis on presenting psychiatric symptoms, previous mental health problems as well as the developmental history. The latter informs the assessment in terms of personality traits and identification of risk factors for psychiatric disorders and allows for cautious speculation as to why these particular symptoms may be presenting now. Often the information that contributes to understanding the genesis of symptoms that can mimic a depressive illness, such as chronic suicidal ideation, can be found here.

Screening for MDE symptoms using a classification such as the DSM-5 paying attention to duration, impairment of function and a relatively recent departure from baseline is helpful. Chronic depressive symptoms are less likely to indicate a MDE and suggest that other possibilities, such as dysthymia or personality disorder, need to be explored.

It is worthwhile here to comment on diagnostic classifications, and other assessment tools, as their strengths and limitations must be understood before they can be used properly. Disagreements about the diagnosis of depression between medical practitioners are often based around this issue. Psychiatry, as a rule, treats syndromes rather than diseases with identified pathology. Like many other aspects of medicine, it uses consensus-derived definitions of what is ‘pathological’ to define a disorder, such as the DSM-5. Because we are using symptom descriptors rather than measured values, the diagnostic process can appear at times to be particularly subjective. However, if a clinician is trained in both eliciting psychological symptoms and descriptive phenomenology, it is possible to make a reasonably robust diagnosis. Indeed, the DSM itself is an exercise in improving diagnostic reliability, rather than validity. Whilst such diagnostic criteria offer a reliable starting point, they often cannot account for the existence or absence of comorbid psychological symptoms, and they can make no attempt to appreciate the patient’s real predicament. The reason why the individual is unwell now, what meaning these symptoms may have for the patient and what role the patient’s personality may play, both in the genesis and treatment of the problem, is not captured in a psychiatric diagnosis. A checklist approach to diagnosis, whilst useful for reliability in research and as a clinical heuristic, is simplistic and often of limited value.

As will be discussed below, there is a particularly useful diagnostic distinction to make for older patients, even when a MDE has been diagnosed. This is the concept of classification within MDE, namely, the notion of melancholic depression, which is a MDE specifier within DSM-5 [3]. The core distinction is that, in melancholic depression, vegetative symptoms (such as early morning waking, diurnal mood variation and psychomotor changes—either slowing or agitation) predominate. Anhedonia (total loss of pleasure in usual activities) is also frequently present. Thus, if a patient meets the requirement for MDE, some consideration should be given to whether the clinical picture is weighted in this regard.

Depression screening instruments and rating scales are useful for identifying who is likely to have a psychiatric illness and for tracking response to treatment, but they are no substitute for an assessment by an experienced clinician. These often detect symptoms without clinical weighting or reference to duration and severity, which risks exposing individuals with real psychological distress, but no psychiatric disorder, to treatments that are very unlikely to be effective and have real risks. A high score on the Geriatric Depression Scale denotes psychological distress and depressive symptoms but is not a diagnosis of a clinical mood episode. Rather, it is a screening instrument [6].

Risk for developing a depressive episode should be considered. Risk factors for developing late life depression include physical illness or disability, a personal history of depression, loss of spouse, subsyndromal depression, sleep disturbance and comorbid anxiety [7].

The corroborative history is crucial to determine the time course of change, premorbid personality and for the risk assessment.

The Mental State Examination (MSE) may reveal features supportive of a diagnosis of MDE, such as slow and monotonous speech, a restricted affect, pervasive low mood and suicidal ideation. However, there are multiple possible explanations for each of these, and none is pathognomonic for a particular psychiatric syndrome.

Cognitive screening is of limited use for making or refuting a diagnosis of a mood episode. It is well known that a depressive episode can compromise cognitive abilities, but this is not always the case. It could be argued that if clinical suspicion of a mood disorder is high, it is best to delay testing altogether, as an impaired performance can be misinterpreted by future readers of the medical record as an argument for a longstanding cognitive deficit, which may well not be the case.

Routine depression screening should also include checking for anaemia, infection, renal or hepatic failure, malnourishment and thyroid disease. A ‘subclinical’ thyroid disorder is often detected, as mood disorders themselves can alter thyroid function [8].

Brain imaging remains useful for excluding (rare) reversible causes such as a space-occupying lesion. MRI is preferable as the FLAIR sequence provides valuable information about subcortical pathology (particularly white matter disease) which is implicated in executive dysfunction—a common explanation for apparent depressive symptoms.

7.2 Clinical Reasoning Around Depressive Symptoms

Having now progressed through history taking, examination and investigations, this information, as in general medicine, must be interpreted as a whole. Because estimating the likelihood of a reversible psychiatric syndrome is a fundamental task of the initial consultation, the following questions should now be considered.

  1. 1.

    Are the symptoms a departure from the individual’s usual (‘baseline’) state? Many patients are referred with chronic low mood, hopelessness or intermittent suicidal ideation—often stretching back decades. In this situation, the diagnosis is unlikely to be a mood episode, as, by definition, this involves episodic symptoms—lasting weeks, months or longer, but interrupted by asymptomatic periods. If the patient has a personal history of a mood disorder, then an episodic history is usually obtained. This is a crucial point as, whilst longer-term problems such as a personality disorder can be often effectively treated, this is often done without medication. The risks of antidepressant use in the elderly are high enough for the clinician to need to be confident that their use would likely be of significant benefit.

  2. 2.

    Is there functional impairment? Similarly to diagnosing dementia, a mood episode cannot be diagnosed on symptom collection alone. All psychiatric disorders require that the individual’s symptoms result in impairment in function, such as social, occupational, interpersonal or personal care abilities.

  3. 3.

    What is the patient’s mood? A MDE is characterized by a diminishment of pleasure in usual activities. Further, mood is pervasively low, and the patient will often report feeling sad, flat or low. Descriptions of the patient’s usual activities, the enjoyment derived from them and mood itself need to be sought. Again, a departure from usual with an identifiable point in time is very helpful. Often, patients who have attracted a diagnosis of depression are describing subjective states other than low mood. A common example is a late life existential distress or crisis in meaning. Often such a patient describes feeling that their life has been lived and, as there is little more to achieve, life has no purpose. However, when asked about the ability to achieve pleasure from usual pursuits or a favourite meal, this is often not diminished.

  4. 4.

    If there is an episodic pattern to the depressive symptoms and a MDE seems likely, is there evidence of vegetative symptoms (early morning waking, weight loss, psychomotor change, etc.)? This will assist in making a distinction between melancholic and non-melancholic depressive episodes.

  5. 5.

    Is there a recent bereavement? As this is common amongst older people, it is always worth considering. Comorbid depression is possible, but often low mood in an uncomplicated bereavement occurs in ‘waves’ often precipitated by a reminder of the lost loved one, such as a song or photograph. The ability to enjoy some activities is generally preserved. Psychotic and suicidal ideation is rare in bereavement alone; though hallucinations of the lost individual are common and do not, of themselves, imply psychiatric illness.

  6. 6.

    Is there a pre-existing neurodegenerative process? The particular issue here is that the insidious onset of executive impairment can also mimic a depressive episode, particularly in a cross-sectional encounter. The patient will likely have a restricted affect, prominent apathy and impaired insight. Individuals with an apathetic syndrome secondary to executive dysfunction are very unlikely to initiate activity but will respond to prompting. An individual with a depressive episode is less likely to be able to respond to external activity planning. A second issue is the difficulty in making a diagnosis of MDE when there is already advanced dementia (from any cause). An empirical approach (making the diagnosis by assessing response to treatment) is difficult to justify in light of safety findings regarding the most commonly used antidepressants [5]. An instrument for assessing depressive symptoms in dementia, such as the Cornell Scale [9], may be useful in this situation.

  7. 7.

    Is there a personal history of a depressive disorder? This raises clinical suspicion of a MDE diagnosis and could be persuasive in a decision to treat when otherwise the picture is not entirely convincing. A family history of depression is not as helpful as the term is commonly used in different ways, and even if a MDE is clearly present in a first-degree relative, the increase in relative risk is modest compared to other psychiatric disorders [10].

  8. 8.

    Is there a history of mania? This is suggestive of bipolar disorder and has treatment implications for the depressive episode. Referral to a psychiatrist in this situation is advised.

  9. 9.

    Is the situation urgent? The presence of suicidal ideation or plans and severely impaired oral intake are two indications for referral to psychiatric services. This may also have treatment implications, as faster acting treatments such as electroconvulsive therapy (ECT) may be preferred if the diagnosis of MDE is likely.

  10. 10.

    Is the Mental State Examination consistent with a diagnosis of a major depressive episode? Whilst there are alternate explanations for each finding, it would be usual to expect a degree of psychomotor symptoms, poor eye contact, slow and/or monotonous speech, pervasive low mood and nihilistic themes. Psychotic or suicidal ideation may also be present. The individual with a MDE often has preserved insight, though this is unusual in executive impairment alone.

We can now return to the case of Harry. From the vignette, clinical suspicion is already raised for a MDE by the fact that Harry is socially withdrawn, his wife is concerned, this appears to be a departure from usual and he appears sad and flat. There is functional impairment. Against this diagnostic argument, however, is the fact that Harry himself is unconcerned (this is less usual for a depressive episode), he has not responded to a trial of treatment, his appetite is preserved and he responds to prompting (often not the case in MDE). Further he has plausible aetiology for significant executive impairment—in this case a high risk of subcortical ischaemic pathology given his medical history.Applying the above principles, the following is established on further history. There is a departure from baseline, though it is insidious and this is unusual for a depressive episode. Harry’s mood is explored—he does not actually feel sad or low, he just ‘can’t be bothered’. He enjoys the TV he is watching, loves seeing his grandchildren and enjoys his wife’s famous chocolate cake. Regarding vegetative symptoms, Harry says he sleeps poorly (but always has, as he was a shift worker) and he has a good appetite. If a meal is put before him, he enjoys it. There is no recent bereavement. There is no personal or family history of depression or mania, and the situation is not urgent.The MSE findings include restricted affect and mild psychomotor slowing, but no change in speech, no evidence of low mood and no hopeless or nihilistic ideation. Restricted affect can occur in subcortical vascular disease, ‘vascular parkinsonism’ [11] as well as MDE, so this finding is not discriminatory.Brain MRI demonstrated significant deep white matter (and extensive periventricular) pathology.In summary, the preservation of mood reactivity, sleep and appetite patterns make a diagnosis of MDE unlikely. The insidious onset, multiple risk factors for cerebrovascular disease, lack of insight and prominent apathy mean that executive impairment is the more likely explanation for Harry’s symptoms.

7.3 Treatment of a Depressive Episode

For depressive symptoms outside of a major depressive episode (such as dysthymia or personality disorders), there are effective treatment options available, but they are outside the scope of this chapter. For a presentation such as Harry’s in the case above, neuroprotective options (smoking cessation, aspirin, etc.) along with activity planning and education for the patient and carer are often valuable in minimizing further decline and optimizing coping. Monitoring cognition for the development of a possible vascular (or other) dementia is important.

If a diagnosis of MDE is made, treatment options include psychological interventions (such as supportive psychotherapy or cognitive behavioural therapy—CBT), biological interventions (such as pharmacotherapy and ECT) and social interventions—including activity scheduling, exercise and optimizing social supports.

Psychological therapy alone may be adequate for mild to moderate depression [12]. This is a very important principle in treating elderly patients as avoiding medications is prudent if possible. It is important to note that for therapies which have a robust evidence base—such as CBT—the therapy should be performed by a trained clinician (usually a psychologist or psychiatrist), and the patient needs to be grossly cognitively intact to gain maximum benefit. This approach is different from what is commonly offered which is supportive therapy. This work can be useful (e.g. in assisting with problem solving, psychoeducation and validating emotional states); however, the evidence base for this alone is less robust.

Geriatricians will often be called on to make decisions regarding pharmacotherapy. Options include antidepressant use (monotherapy and combination therapy) and augmentation (with agents such as lithium or mood stabilizers, atypical antipsychotics or thyroxine). As combination and augmentation approaches usually involve psychiatric input, only an approach to antidepressant monotherapy is covered here.

There are some important principles to consider.

Firstly, if the depressive episode is occurring in the context of bipolar disorder (‘bipolar depression’), specialist psychiatric advice should be sought.

Secondly, the known risks when using antidepressants in the elderly (and their limited evidence base in mild and moderate depression) imply that older patients with a lack of vegetative symptoms should, ideally, try non-medication approaches first. These risks vary with each medication but include falls, hyponatraemia and, for some agents, all-cause mortality. These risks are usually higher in the month after either starting or stopping the antidepressant.

Third, the evidence for differential efficacy between agents is modest [13], and it is reasonable to select on the basis of previous effectiveness, side effect profile and tolerability. An example may be the selection of mirtazapine, the side effects of which (often immediate sedation and appetite improvement) may be advantageous for patients with poor sleep and weight loss. This is distinct from the antidepressant effect itself, which may take some weeks to occur.

Reserving pharmacological therapy for older patients with severe depression will mean that many (if not most) will have a melancholic picture. In specialist practice, it is rare to meet a patient with depressive symptoms who has not already been treated with a selective serotonin reuptake inhibitor (SSRI) so dual-action agents (often with noradrenergic properties—such as venlafaxine) are often trialled next. These agents (and broader-spectrum medications such as the tricyclics) may theoretically be superior to SSRIs for patients with melancholic depression, as it has been elegantly postulated that noradrenergic (rather than serotonergic) neural pathways are implicated in this syndrome [14]. Whilst there is some therapeutic evidence for this amongst elderly patients [15, 16], the larger meta-analyses of antidepressant efficacy in older patients do not specifically address the melancholic subtype of depression.

It is essential that each trial of therapy is long enough to be effective. This is usually at least 4 weeks in duration, though some authors suggest longer [17], at the maximum dose tolerated within the manufacturer’s guidelines. A file review of ‘treatment-resistant’ patients often finds that multiple agents were previously used at either subtherapeutic doses or for a short time only.

Tricyclic antidepressants are no longer commonly prescribed for depressive indications given potential lethality in overdose, propensity to cause confusion and other anticholinergic side effects. However, they do have a role, preferably with specialist supervision.

Regular review after commencing an antidepressant is required, looking for evidence of response or any adverse effects. These can include (varying with the agent) sedation, nausea, diarrhoea, akathisia, hyponatraemia, agitation and emergent suicidal ideation (in a small proportion of those who did not have this symptom as a part of the depressive presentation). Prescribers should monitor all patients for the first few weeks of treatment, including for new-onset suicidal ideation. The risk of suicide related to commencing an antidepressant in this age group is thought to be low and needs to be weighed against the benefits of treatment. The main concerns relating to antidepressant-induced suicide risk are in those aged under 25, with a World Psychiatry Association consensus statement indicating that even risk in this group was likely to be small [18].

ECT is indicated if trials of pharmacotherapy are ineffective and if there are urgent issues such as active suicidal ideation or sustained poor oral intake and also should be considered if the patient requests it.

Referral to specialist psychiatric services should be considered if there are issues with urgency or suicide risk or if trial of a second antidepressant has failed.

Longer-term antidepressant therapy benefits recovery [19], but the indication for ongoing use should be reviewed after 12 months, particularly for those individuals with their first depressive episode.