Shorter Oxford Textbook of Psychiatry, 6th Ed.

CHAPTER 10. Mood disorders


Clinical features

Transcultural factors

Classification of mood disorders

Differential diagnosis

The epidemiology of mood disorders

The aetiology of mood disorders

Course and prognosis

The acute treatment of depression

The acute treatment of mania

The longer-term treatment of mood disorders

The assessment of depressive disorders

The management of depressive disorders

The assessment of mania

The management of manic patients


The mood disorders are so called because one of their main features is abnormality of mood. Nowadays the term is usually restricted to disorders in which this mood is depression or elation, but in the past some authors have included states of anxiety as well. In this book, anxiety disorders are described in Chapter 9. Mood disorders have in the past been referred to as ‘affective disorders’, a term that is still used fairly widely.


It is part of normal experience to feel unhappy during times of adversity. The symptom of depressed mood is a component of many psychiatric syndromes, and is also commonly found in certain physical diseases (e.g. in infections such as hepatitis, and some neurological disorders). In this chapter we are concerned neither with normal feelings of unhappiness nor with depressed mood as a symptom of other disorders, but with the syndromes known as depressive disorders.

The central features of these syndromes are:

• depressed mood

• negative thinking

• lack of enjoyment

• reduced energy

• slowness.

Of these, depressed mood is usually, but not invariably, the most prominent symptom.


Similar considerations apply to states of elation. A degree of elated mood is part of normal experience during times of good fortune. Elation can also occur as a symptom in several psychiatric syndromes, although it is less widely encountered than depressed mood. In this chapter we are concerned with a syndrome in which the central features are:

• overactivity

• elevated or irritable mood

• self-important ideas.

This syndrome is called mania. Some diagnostic classifications distinguish a less severe form of mania, known as hypomania (see below).

Clinical features

Depressive syndromes

The clinical presentations of depressive states are varied, and they can be subdivided in a number of different ways. In the following account, disorders are grouped by their severity. The account begins with a description of the clinical features of an episode of severe depression, together with certain important clinical variants that can influence the presentation of depressive disorders. Finally, the special features of the less severe depressive disorders are outlined. What constitutes an ‘episode’ of clinical depression is inevitably a somewhat arbitrary concept. The symptoms listed for the diagnosis of ‘depressive episode’ in the ICD-10 classification and the various levels of severity are shown in Table 10.1Table 10.2 shows the criteria for ‘major depressive episode’ in DSM-IV.

Severe depressive episode

In a severe episode of depression, the central features are low mood, lack of enjoyment, negative thinking, and reduced energy, all of which lead to decreased social and occupational functioning.

Table 10.1 Symptoms needed to meet the criteria for ‘depressive episode’ in ICD-10


• Depressed mood

• Loss of interest and enjoyment

• Reduced energy and decreased activity


• Reduced concentration

• Reduced self-esteem and confidence

• Ideas of guilt and unworthiness

• Pessimistic thoughts

• Ideas of self-harm

• Disturbed sleep

• Diminished appetite

Mild depressive episode: at least two of A and at least two of B

Moderate depressive episode: at least two of A and at least three of B

Severe depressive episode: all three of A and at least four of B

Severity of symptoms and degree of functional impairment also guide classification


The patient’s appearance is characteristic. Dress and grooming may be neglected. The facial features are characterized by a turning downward of the corners of the mouth, and by vertical furrowing of the centre of the brow. The rate of blinking may be reduced. The shoulders are bent and the head is inclined forward so that the direction of gaze is downward. Gestural movements are reduced. It is important to note that some patients maintain a smiling exterior despite deep feelings of depression.

Table 10.2 Criteria for major depressive episode in DSM-IV

Five (or more) of the following symptoms have been present during the same 2-week period and represent a change from previous functioning; at least one of the symptoms is either (1) depressed mood or (2) loss of interest or pleasure

1. Depressed mood most of the day, nearly every day, as indicated by either subjective report (e.g. feels sad or empty) or observation made by others (e.g. appears tearful)

2. Markedly diminished interest or pleasure in all, or almost all, activities most of the day, nearly every day (as indicated by either subjective account or observation made by others)

3. Significant weight loss when not dieting or weight gain (e.g. a change of more than 5% of body weight in a month), or decrease or increase in appetite nearly every day

4. Insomnia or hypersomnia nearly every day

5. Psychomotor agitation or retardation nearly every day (observable by others, nor merely subjective feelings of restlessness or being slowed down)

6. Fatigue or loss of energy nearly every day

7. Feelings of worthlessness or excessive or inappropriate guilt (which may be delusional) nearly every day (not merely self-reproach or guilt about being sick)

8. Diminished ability to think or concentrate, or indecisiveness, nearly every day (either by subjective account or as observed by others)

9. Recurrent thoughts of death (not just fear of dying), recurrent suicidal ideation without a specific plan, or a suicide attempt or a specific plan for committing suicide


The mood of the patient is one of misery. This mood does not improve substantially in circumstances where ordinary feelings of sadness would be alleviated—for example, in pleasant company or after hearing good news. Moreover, the mood is often experienced as different from ordinary sadness. Patients sometimes speak of a black cloud pervading all mental activities. Some patients can conceal this mood change from other people, at least for short periods. Some try to hide their low mood during clinical interviews, which makes it more difficult for the doctor to detect. The mood is often worse first thing in the morning when the patient wakes, improving a little as the day wears on. This is called diurnal variation of mood.

Depressive cognitions

Negative thoughts (‘depressive cognitions’) are important symptoms which can be divided into three groups:

• worthlessness

• pessimism

• guilt.

In feeling worthless, the patient thinks that he is failing in what he does and that other people see him as a failure; he no longer feels confident, and he discounts any success as a chance happening for which he can take no credit. Pessimistic thoughts concern his future prospects. The patient expects the worst. He foresees failure in his work, the ruin of his finances, misfortune for his family, and an inevitable deterioration in his health. These ideas of hopelessness are often accompanied by the thought that life is no longer worth living and that death would come as a welcome release. These gloomy preoccupations may progress to thoughts of, and plans for, suicide. It is important to ask about these ideas in every case (the assessment of suicidal risk is considered further in Chapter 16).

Feelings of guilt often take the form of unreasonable self-blame about minor matters—for example, the patient may feel guilty about past trivial acts of dishonesty or letting someone down. Usually these events have not been in the patient’s thoughts for years, but when he becomes depressed they flood back into his memory, accompanied by intense feelings. Preoccupations of this kind strongly suggest depressive disorder. Some patients have similar feelings of guilt but do not attach them to any particular event. Other memories are focused on unhappy events; the patient remembers occasions when he was sad, when he failed, or when his fortunes were at a low ebb. These gloomy memories become increasingly frequent as the depression deepens. The patient blames himself for his misery and incapacity, and attributes it to personal failing and moral weakness (a view not uncommonly held by the wider public).

Goal-directed behaviour

Lack of interest and enjoyment (also known as anhedonia) is frequent, although it is not always complained of spontaneously. The patient shows no enthusiasm for activities and hobbies that he would normally enjoy. He feels no zest for living and no pleasure in everyday things. He often withdraws from social encounters. Reduced energy is characteristic (although depression is sometimes associated with a degree of physical restlessness that can mislead the observer). The patient feels lethargic, finds everything an effort, and leaves tasks unfinished. For example, a normally house-proud person may leave the beds unmade and dirty plates on the table. Work outside the home becomes increasingly difficult. Understandably, many patients attribute this lack of energy to physical illness.

Psychomotor changes

Psychomotor retardation is frequent (although, as described later, some patients are agitated rather than slowed down). The retarded patient walks and acts slowly. Slowing of thought is reflected in their speech; there is a significant delay before questions are answered, and pauses in conversation may be unusually prolonged. Agitation is a state of restlessness that is experienced by the patient as inability to relax, and is seen by an observer as restless activity. When it is mild, the patient is seen to be plucking at his fingers and making restless movements of his legs; when it is severe, he cannot sit for long, and instead paces up and down.

Anxiety is frequent, although not invariably present, in severe depression. (As described later, it is common in less severe depressive disorders.) Another common symptom is irritability, which is the tendency to respond with undue annoyance to minor demands and frustrations.

Biological symptoms

There is an important group of symptoms that are often described as ‘biological’ (also referred to as ‘melancholic’, ‘somatic’ or ‘vegetative’). These symptoms include sleep disturbance, diurnal variation in mood, loss of appetite, loss of weight, constipation, loss of libido, and, among women, amen-orrhoea. They are very common in depressive disorders of severe degree (but less usual in mild depressive disorders). Some of these symptoms require further comment.

Sleep disturbance in depressive disorders is of several kinds. The most characteristic type is early-morning waking, but delay in falling asleep and waking during the night also occur. Early-morning waking occurs 2 or 3 hours before the patient’s usual time of waking. He does not fall asleep again, but lies awake feeling unrefreshed and often restless and agitated. He thinks about the coming day with pessimism, broods about past failures, and ponders gloomily about the future. It is this combination of early waking and depressive thinking that is important in diagnosis. It should be noted that some depressed patients sleep excessively rather than waking early, but they still report waking unrefreshed.

Weight loss in depressive disorders often seems to be greater than can be accounted for merely by the patient’s reported lack of appetite. In some patients the disturbances of eating and weight are towards excess—that is, they eat more than usual and gain weight. Usually eating brings temporary relief of their distressing feelings.

Complaints about physical symptoms are common in depressive disorders. They take many forms, but complaints of constipation, fatigue, and aching discomfort anywhere in the body are particularly common. Complaints about any pre-existing physical disorder usually increase, and hypochondriacal preoccupations are common.

Other features

Several other psychiatric symptoms may occur as part of a depressive disorder, and occasionally one of them dominates the clinical picture. They include depersonalization, obsessional symptoms, panic attacks, and dissociative symptoms such as fugue or loss of function of a limb. Complaints of poor memory are also common; depressed patients commonly show deficits on a wide range of neuropsychological tasks, but impairments in the retrieval and recognition of recently learned material are particularly prominent. Sometimes the impairment of memory in a depressed patient is so severe that the clinical presentation resembles that of dementia. This presentation, which is particularly common in the elderly, is sometimes called depressive pseudodementia (see p. 503).

Psychotic depression

As depressive disorders become increasingly severe, all of the features described above occur with greater intensity. There is complete loss of function in social and occupational spheres. Inattention to basic hygiene and nutrition may give rise to concern about the patient’s well-being. Psychomotor retardation may make interviewing difficult or impossible. In addition, there may be delusions and hallucinations, in which case the disorder is referred to as psychotic depression.

The delusions of severe depressive disorders are concerned with the same themes as the non-delusional thinking of moderate depressive disorders. Therefore they are termed mood congruent. These themes are worthlessness, guilt, ill health, and, more rarely, poverty. Such delusions have been described in Chapter 1, but a few examples may be helpful at this point. The patient with a delusion of guilt may believe that some dishonest act, such as a minor concealment when filling in a tax return, will be discovered and that he will be punished severely and humiliated. He is likely to believe that such punishment is deserved. A patient with hypochondriacal delusions may be convinced that he has cancer or venereal disease. A patient with a delusion of impoverishment may wrongly believe that he has lost all of his money in a business venture.

Persecutory delusions also occur. The patient may believe that other people are discussing him in a derogatory way or are about to take revenge on him. When persecutory delusions are part of a depressive syndrome, typically the patient accepts the supposed persecution as something that he has brought upon himself. In his view, he is ultimately to blame. This can be a useful diagnostic feature for distinguishing severe depression from non-affective psychosis (see Chapter 11). Some depressed patients experience delusions and hallucinations that are not clearly related to themes of depression (i.e. they are ‘mood-incongruent’). Their presence appears to worsen the prognosis of the illness.

Particularly severe depressive delusions are found in Cotard’s syndrome, which was described by the French psychiatrist, Jules Cotard, in 1882. The characteristic feature is an extreme kind of nihilistic delusion. For example, some patients may complain that their bowels have been destroyed, so they will never be able to pass faeces again. Others may assert that they are penniless and have no prospect of ever having any money again. Still others may be convinced that their whole family has ceased to exist and that they themselves are dead. Although the extreme nature of these symptoms is striking, such cases do not appear to differ in important ways from other severe depressive disorders.

Other clinical variants of depressive disorders

Agitated depression

This term is applied to depressive disorders in which agitation is prominent. As already noted, agitation occurs in many severe depressive disorders, but in agitated depression it is particularly severe. Agitated depression is seen more commonly among middle-aged and elderly patients than among younger individuals.

Retarded depression

This name is sometimes applied to depressive disorders in which psychomotor retardation is especially prominent. There is no evidence that they represent a separate syndrome, although the presence of prominent retardation is said to predict a good response to electroconvulsive therapy (ECT). If the term is used, it should be in a purely descriptive sense. In its most severe form, retarded depression merges with depressive stupor.

Depressive stupor

In severe depressive disorder, slowing of movement and poverty of speech may become so extreme that the patient is motionless and mute. Such depressive stupor is rarely seen now that active treatment is available. Therefore the description by Kraepelin (1921, p. 80) is of particular interest:

The patients lie mute in bed, give no answer of any sort, at most withdraw themselves timidly from approaches, but often do not defend themselves from pinpricks.… They sit helpless before their food; perhaps, however, they let themselves be spoon-fed without making any difficulty.

Kraepelin commented that recall of the events that took place during stupor was sometimes impaired when the patient recovered. Nowadays, the general view is that on recovery patients are able to recall nearly all of the events that took place during the period of stupor. It is possible that in some of Kraepelin’s cases there was clouding of consciousness (possibly related to inadequate fluid intake, which is common in these patients). Patients in states of depressive stupor may exhibit catatonic motor disturbances (see p. 24).

Atypical depression

The term atypical depression is generally applied to disorders of moderate clinical severity. The precise meaning of the term has varied over the years, but currently it is applied to disorders characterized by:

• variably depressed mood with mood reactivity to positive events

• overeating and oversleeping

• extreme fatigue and heaviness in the limbs (leaden paralysis)

• pronounced anxiety.

Many patients with these clinical symptoms have a lifelong tendency to react in an exaggerated way to perceived or real rejection (rejection sensitivity), although this character trait can be exacerbated by the presence of a depressive disorder. Patients with atypical depression also have an earlier onset of illness and a more chronic course. The importance of recognizing atypical depression is that, because of their interpersonal sensitivity, patients with this disorder can be hard to manage and may be regarded as having ‘difficult’ personalities rather than depressive disorder. Atypical depression also seems to be associated with a poor response to tricyclic antidepressant treatment but a better response to monoamine oxidase inhibitors (MAOIs) and perhaps selective serotonin reuptake inhibitors (SSRIs) (see Stewart et al., 2009).

Mild depressive states

It might be expected that mild depressive disorders would present with symptoms similar to those of the depressive disorders described above, but with less intensity. To some extent this is the case, but in mild depressive disorders there are frequently additional symptoms that are less prominent in severe disorders. These symptoms have been characterized in the past as ‘neurotic’, and they include anxiety, phobias, obsessional symptoms, and, less often, dissociative symptoms. In terms of classification, both DSM-IV and ICD-10 have categories of mild depression where criteria for a depressive episode are met but the depressive symptoms are fewer and less severe (see Table 10.1).

Apart from the ‘neurotic’ symptoms that are found in some cases, mild depressive disorders are characterized by the expected symptoms of low mood, lack of energy and interest, and irritability. There is sleep disturbance, but not the early-morning waking that is so characteristic of more severe depressive disorders. Instead, there is more often difficulty in falling asleep, and periods of waking during the night, usually followed by a period of sleep at the end of the night. ‘Biological’ features (poor appetite, weight loss, and low libido) are not usually found. Although mood may vary during the day, it is usually worse in the evening than in the morning. The patient may not appear obviously dejected, or slowed in their movement. Delusions and hallucinations are not present.

In their mildest forms, these cases merge into the minor mood disorders considered below. As described later, they pose considerable problems of classification. Many of these mild depressive disorders are brief, starting at a time of personal misfortune and subsiding when fortunes have changed or a new level of adjustment has been achieved. However, some cases persist for months or years, causing considerable suffering, even though the symptoms do not increase. These chronic depressive states are termed dysthymia. The term cyclothymic disorder refers to a persistent instability of mood in which there are numerous periods of mild elation or mild depression. It is seen as a milder variant of bipolar disorder. It is not unusual for episodes of more severe mood disorder to supervene in patients who experience dysthymia or cyclothymia, and the converse can also occur (Judd et al., 1998; Judd et al., 2002). This suggests that, in some individuals, milder symptoms are a limited expression of an underlying major mood disturbance.

Minor mood disorders

We have already seen that anxiety and depressive symptoms often occur together. Indeed, earlier writers considered that anxiety and depressive disorders could not be separated clearly even in patients who had been admitted to hospital with severe disorders. Although most psychiatrists now accept that the distinction can usually be made among the more severe forms that present in psychiatric practice, the distinction is less easy to make in the milder forms that present in primary care.


As psychiatrists have worked increasingly with general practitioners, the importance of minor anxiety–depressive disorders has been recognized, but without any agreement about classification.

ICD-10 includes a category of ‘mixed anxiety and depressive disorder’ which can be applied when neither anxiety symptoms nor depressive symptoms are severe enough to meet the criteria for an anxiety disorder or a depressive disorder, and when the symptoms do not have the close association with stressful events or significant life changes that is required for a diagnosis of acute stress reaction or adjustment disorder.

According to ICD-10, patients with this presentation are seen frequently in primary care, and there are many others in the general population who are not seen by doctors. In ICD-10 this diagnosis appears among the anxiety disorders, although some psychiatrists consider that the condition is more closely related to the mood disorders, a view that is reflected in the alternative term, minor affective disorder.

In DSM-IV, no comparable diagnosis appears in the classification. The appendix to the classification contains two provisional diagnoses that might be used for these cases, namely mixed anxiety and depressive disorder and minor depressive disorder. It is stated that there is insufficient factual information to justify the inclusion of either in the classification. Although little is known about these conditions or about their relationship to other disorders, patients commonly present to primary care doctors with this group of symptoms. A suitable category is needed even if it is not possible to write strict criteria for diagnosis.

Clinical picture

One of the best descriptions of minor affective disorder is that given by Goldberg et al. (1976), who studied 88 patients from a general practice in Philadelphia. The most frequent symptoms were:

• fatigue

• anxiety

• depression

• irritability

• poor concentration

• insomnia

• somatic symptoms and bodily preoccupation.

A very similar range of symptoms was found was found in the National Psychiatric Morbidity Household Survey (Jenkins et al., 1997), which surveyed the frequency of ‘neurotic’ symptomatology in a community sample.

Patients with minor affective disorders commonly present to medical practitioners with prominent somatic symptoms. The reason for this is uncertain; some symptoms are autonomic features of anxiety, and it is possible that patients expect somatic complaints to be viewed more sympathetically than emotional problems (see Chapter 15). Another point of clinical relevance is that minor affective disorders can be prolonged and in some cases may cause quite disabling difficulties in personal and occupational function. Thus the term ‘minor’ may not capture the serious consequences of the disorder for an individual. As noted above, in some patients minor affective disorders may represent a residual form of a major mood disturbance (Judd et al., 1998, 2002).


As already mentioned, the central features of the syndrome of mania are elevation of mood, increased activity, and self-important ideas.


When the mood is elevated, the patient appears cheerful and optimistic, and they may have a quality described by earlier writers as ‘infectious gaiety.’ However, other patients are irritable rather than euphoric, and this irritability can easily turn to anger. The mood often varies during the day, although not with the regular rhythm that is characteristic of many severe depressive disorders. In patients who are elated, it is not uncommon for high spirits to be interrupted by brief episodes of depression.

Appearance and behaviour

The patient’s appearance often reflects their prevailing mood. Their clothing may be brightly coloured and ill assorted. When the condition is more severe, the patient’s appearance is often untidy and dishevelled. Manic patients are overactive. Sometimes the persistent overactivity leads to physical exhaustion. Manic patients start many activities but leave them unfinished as new ones attract their attention. Appetite is increased, and food may be eaten greedily with little attention to conventional manners. Sexual desires are increased, and sexual behaviour may be uninhibited and quite out of character. Women may neglect precautions against pregnancy, a point that calls for particular attention if the patient is of childbearing age. Sleep is often reduced. Patients wake early, feeling lively and energetic, and often get up and busy themselves noisily, to the surprise (and sometimes annoyance) of other people.

Speech and thought

The speech of manic patients is often rapid and copious as thoughts crowd into their minds in quick succession. When the disorder is more severe, there is flight of ideas (see p. 23), with such rapid changes that it is difficult to follow the train of thought. However, the links are usually understandable if the speech can be recorded and reviewed. This is in contrast to thought disorder in schizophrenia, where changes in the flow of thought may not be comprehensible even on reflection.

Expansive ideas are common. Patients believe that their ideas are original, their opinions important, and their work of outstanding quality. Many patients become extravagant, spending more than they can afford (e.g. on expensive cars or jewelry). Others make reckless decisions to give up good jobs, or embark on plans for ill-considered and risky business ventures.

Sometimes these expansive themes are accompanied by grandiose delusions. Some patients may believe that they are religious prophets or destined to advise statesmen about major issues. At times there are delusions of persecution, when the patient believes that people are conspiring against them because of their special importance. Delusions of reference and passivity feelings also occur. Schneiderian first-rank symptoms (see Chapter 11Table 11.3) have been reported in around 10–20% of manic patients. Neither the delusions nor the first-rank symptoms last for long, most of them disappearing or changing in content within a period of days.

Perceptual disturbances

Hallucinations occur. These are usually consistent with the mood, taking the form of voices speaking to the patient about his special powers or, occasionally, of visions with a religious content.

Other features

Insight is invariably impaired in more severe manic states. Patients see no reason why their grandiose plans should be restrained or their extravagant expenditure curtailed. They seldom think of themselves as ill or in need of treatment.

Most patients can exert some control over their symptoms for a short time, and many do so when the question of treatment is being assessed. For this reason it is important to obtain a history from an informant whenever possible. Henry Maudsley (1879, p. 398) expressed the problem well:

Just as it is with a person who is not too far gone in intoxication, so it is with a person who is not too far gone in acute mania; he may on occasion pull his scattered ideas together by an effort of will, stop his irrational doings and for a short time talk with an appearance of calmness and reasonableness that may well raise false hopes in inexperienced people.

The symptoms that are required to fulfil the criteria for ‘manic episode’ in DSM-IV are listed in Table 10.3. The criteria for manic episode in ICD-10 are very similar, although the number of manic symptoms required for diagnosis is not specified so precisely.

Table 10.3 Criteria for manic episode in DSM-IV

A. A distinct period of abnormally and persistently elevated, expansive, or irritable mood, lasting at least 1 week (or any duration if hospitalization is necessary)

B. During the period of mood disturbance, three (or more) of the following symptoms have persisted (four if the mood is only irritable) and have been present to a certain degree:

1. Inflated self-esteem or grandiosity

2. Decreased need for sleep (e.g. feels rested after only 3 hours of sleep)

3. More talkative than usual, or pressure to keep talking

4. Flight of ideas or subjective experience that thoughts are racing

5. Distractibility (i.e. attention too easily drawn to unimportant or irrelevant external stimuli)

6. Increase in goal-directed activity (either socially, at work or school, or sexually), psychomotor agitation

7. Excessive involvement in pleasurable activities that have a high potential for painful consequences (e.g. engaging in unrestrained buying sprees, sexual indiscretions, or foolish business investments)

C. The symptoms do not meet criteria for a mixed episode

D. The mood disturbance is sufficiently severe to cause marked impairment in occupational functioning or in usual social activities or relationships with others, or to necessitate hospitalization to prevent harm to self or others, or there are psychotic features

E. The symptoms are not due to the direct physiological effects of a substance (e.g. a drug of abuse, a medication, or other treatment) or a general medical condition (e.g. hyperthyroidism)

Hypomania refers to a state of elevated mood that is of lesser extent than mania. In DSM-IV, the criteria are similar to those for mania, with the following distinctions:

• The duration of elevated, expansive, or irritable mood need be only 4 days.

• The mood disturbance, although associated with an unequivocal change in function, is not sufficiently severe to cause marked impairment in social or occupational activities, or to necessitate hospital admission.

• Psychotic features are absent.

Mixed mood (affective) states

Depressive and manic symptoms sometimes occur at the same time. Patients who are overactive and over-talkative may be having profoundly depressive thoughts. In other patients, mania and depression follow each other in a sequence of rapid changes—for example, a manic patient may become intensely depressed for a few hours and then return quickly to his manic state. These changes were mentioned in early descriptions of mania by Griesinger (1867), and have been re-emphasized in recent years because they appear to predict a better response to certain mood stabilizers, such as valproate.

Rapid cycling disorders

Some bipolar disorders recur regularly with intervals of only days or weeks between episodes. In the nineteenth century these regularly recurring disorders were designated folie circulaire (circular insanity) by the French psychiatrist Jean-Pierre Falret (Falret, 1854). At present, the frequent recurrence of mood disturbance in bipolar patients is usually termed rapid cycling disorder. These recurrent episodes may be depressive, manic, or mixed. The main features are that recurrence is frequent (by convention at least four distinct episodes a year), and that episodes are separated by a period of remission or a switch to an episode of opposite polarity. A number of clinical features of rapid cycling disorder are important in management and prevention.

• They occur more frequently in women.

• Concomitant hypothyroidism is common.

• They can be triggered by antidepressant drug treatment.

• Lithium treatment as sole therapy is relatively ineffective.

The lifetime risk of rapid cycling in bipolar populations varies between studies, but is probably in the range 15–30%. Rapid cycling is often a rather temporary phenomenon, and in most patients it remits within about 2 years (Goodwin, 2009).

Manic stupor

In this unusual disorder, patients are mute and immobile. Their facial expression suggests elation, and on recovery they describe having experienced a rapid succession of thoughts typical of mania. The condition is rarely seen now that active treatment is available for mania. Therefore an earlier description by Kraepelin (1921, p. 106) is of interest:

The patients are usually quite inaccessible, do not trouble themselves about their surroundings, give no answer, or at most speak in a low voice … smile without recognizable cause, lie perfectly quiet in bed or tidy about at their clothes and bedclothes, decorate themselves in an extraordinary way, all this without any sign of outward excitement.

On recovery, patients can remember the events that occurred during their period of stupor. The condition may begin from a state of manic excitement, but at times it is a stage in the transition between depressive stupor and mania.

Transcultural factors

Mania appears to be present in all cultures. However, caution in diagnosis is needed, because culture-specific means of expressing distress can lead to behaviours that may be misdiagnosed as mania (Kirmayer and Groleau, 2001). There are cultural variations in the clinical presentation of depressive states, but in most countries depression appears to be under-diagnosed, particularly in primary care. Although somatic features are undoubtedly found in all societies, they are more frequent and prominent in non-Western cultures. However, it is necessary to distinguish between somatization (see Chapter 15) and somatic metaphors for an emotional state. For examples, Punjabi women living in London have been found to use expressions such as ‘weight on my heart’ and ‘feelings of heat’ to express emotional suffering, the presence of which they were well aware. It should also be noted that sadness, joylessness, anxiety, and lack of energy are common symptoms of depression in all cultures (for a review, see Bhugra and Mastrogianni, 2004).

Classification of mood disorders

The fundamental clinical distinction in the classification of mood disorders is between depression and mania. Depressive episodes have been subdivided in various ways, which are outlined below, but the results of much of this work have not been particularly useful. However, studies of the longitudinal course of mood disorders have indicated that there are useful clinical distinctions to be made between people who never experience mania or hypomania (recurrent depressive disorder) and those who do (bipolar disorder).

Classification of depression

There is no general agreement about the best method of classifying depressive disorders. A number of approaches have been tried, based on the following:

• presumed aetiology

• symptomatic picture

• course.

Classification by presumed aetiology

Historically, depressive disorders were sometimes classified into two kinds—those in which the symptoms were caused by factors within the individual, and were independent of outside factors (endogenous depression), and those in which the symptoms were a response to external stressors (reactive depression). However, it has been recognized for many years that this distinction is unsatisfactory. For example, Lewis (1934) wrote:

every illness is a product of two factors—of the environment working on the organism—whether the constitutional factor is the determining influence or the environmental one … is never a question to be dealt with as either/or.

As noted in Chapter 5, when considering the aetiology of individual cases of depression, the relative contributions of a variety of aetiological factors must be considered. Neither ICD-10 nor DSM-IV contains categories of reactive or endogenous depression.

Classification by symptomatic picture

Melancholic depression. It is well recognized that episodes of depression vary in symptomatic profile both within and between subjects. In the section on clinical description (see p. 207) it was noted that some depressive conditions are characterized by ‘biological’ symptoms, such as loss of appetite, psychomotor changes, weight loss, constipation, reduced libido, amenorrhoea, and early-morning waking. These symptoms have sometime been termed melancholic, and they have been used to delineate a specific subgroup of depressive disorders, namely major depression with melancholia in DSM-IV, or depressive episode with somatic symptoms in ICD-10 (see Table 10.4). The difficulty with this classification is that most patients have melancholic symptoms of some kind, though a careful search may be required to reveal them. Therefore the number of symptoms that are needed to fulfil the criterion for depression with melancholia is somewhat arbitrary. Despite this caveat, it is generally agreed that clear-cut melancholic depression is associated with the following clinical correlates (see Coryell, 2007):

• more severe symptomatology

• family history of depression

• poor response to placebo medication

• more evidence of neurobiological abnormalities (e.g. decreased latency to rapid eye movement sleep, cortisol hypersecretion).

Table 10.4 Clinical features of melancholic and somatic depression

Melancholic features (DSM-IV)

Loss of interest or pleasure in usual activities*

Lack of reactivity to pleasurable stimuli,* plus at least three of the following:

• Distinct quality of mood (unlike normal sadness)

• Morning worsening of mood*

• Early-morning waking*

• Psychomotor agitation or retardation*

• Significant anorexia or weight loss*

• Excessive guilt

• Marked loss of libido* (ICD-10 only)

* Somatic symptoms of depression in ICD-10 (at least four are required for diagnosis).

It is still not clear whether melancholic depression is a distinct subtype, or whether it represents a point on a continuum of severity, towards the more severe end. Kendler (1997) attempted to answer this question using a population sample of twins. He found evidence that melancholic depression did represent a valid subtype in that it identified a group of individuals with a particularly high familial risk of depression. However, the diagnosis of melancholia indicated the presence of a quantitatively more severe form of depression, rather than a distinct aetiological subtype.

Psychotic depression. As noted above, severe depression can also be manifested with psychotic features (although in depressive psychosis the features of melancholia are almost invariably present as well). The presence of psychotic features indicates that treatment with antidepressant medication alone is unlikely to be successful, and that combination with antipsychotic drugs is usually needed (Anderson et al., 2008).

Non-melancholic depression. In this classification by symptom profile, the remaining forms of major depression (‘non-melancholic’ depression) include several different kinds of clinical disorder—for example, mild depressive episodes, atypical depression, and dysthymia. These depressions are more likely to have a relative prominence of features such as anxiety, hostility, phobias, and obsessional symptoms. In the past, because of these symptoms, non-melancholic depressions were sometimes called ‘neurotic depression’, but this term does not appear in current diagnostic classifications. As noted above (see p. 209), atypical depression has particular clinical characteristics and response to treatment; it therefore appears to merit a specific category within the non-melancholic depressions (Stewart et al., 2009).

Classification by course

Unipolar and bipolar disorders. Mood disorders are characteristically recurrent, and Kraepelin was guided by the course of illness when he brought mania and depression together as a single entity. He found that the course was essentially the same whether the original disorder was manic or depressive, and so he put the two together in a single category of manic–depressive psychosis.

This view was widely accepted until 1962, when Leonhard and colleagues suggested a division into three groups:

• patients who had had a depressive disorder only (unipolar depression)

• those who had had mania only (unipolar mania)

• those who had had both depressive disorder and mania (bipolar).

Nowadays, it is the usual practice not to use the term ‘unipolar mania’, but to include all cases of mania in the bipolar group on the grounds that nearly all patients who have mania eventually experience a depressive disorder.

In support of the distinction between unipolar and bipolar disorders, Leonhard et al. (1962) described differences in heredity between the groups, which have been confirmed by later studies (see p. 220). The question of whether or not unipolar and bipolar depression show important differences in their clinical presentation is controversial. Epidemiological studies suggest that depression with a bipolar course is associated more with psychotic features, diurnal mood variation, and hyper-somnia (Forty et al., 2008). However, the differences are not great. There must be some overlap between the two groups, because a patient who is classified as having unipolar depression at one time may have a manic disorder subsequently. In other words, the unipolar group inevitably contains some bipolar cases that have not yet declared themselves. Despite this limitation, the division into unipolar and bipolar cases is a useful classification because it has implications for treatment (see p. 252).

The ability to predict which patients who present with depression will eventually develop bipolar illness is currently limited. The presence of any hypomanic or mixed symptomatology at initial presentation has some predictive value, but the majority of depressed patients who convert to bipolar illness do not have hypomanic symptoms during the initial episodes of depression. Other clinical features associated with subsequent development of bipolar illness include early age of onset and clinical severity, particularly the presence of psychosis (see Fiedorowicz et al., 2011).

Seasonal affective disorder. Some patients repeatedly develop a depressive disorder at the same time of year, usually the autumn or winter. In some cases the timing reflects extra demands placed on the person at a particular season of the year, either in work or in other aspects of their life. In other cases there is no such cause, and it has been suggested that seasonal affective disorder is related in some way to the changes in the seasons (e.g. to the number of hours of daylight). Although these seasonal affective disorders are characterized mainly by the time at which they occur, some symptoms are said to occur more often than in other mood disorders. These symptoms include:

• hypersomnia

• increased appetite, with craving for carbohydrate

• an afternoon slump in energy levels.

The most common pattern is onset in autumn or winter, and recovery in spring or summer. This condition is also called ‘winter depression.’ Some patients show evidence of hypomania or mania in the summer, which suggests that they have a seasonal bipolar illness. This pattern has led to the suggestion that shortening of daylight hours is important in the pathophysiology of winter depression, and treatment methods include exposure to bright artificial light during hours of darkness. For a review, see Eagles (2004). The use of bright light treatment is reviewed on p. 564.

Brief recurrent depression. Some individuals experience recurrent depressive episodes of short duration, typically 2 to 7 days, that are not of sufficient duration to meet the criteria for major depression or depressive episode. These episodes recur with some frequency, about once a month on average. There is no apparent link with the menstrual cycle in female sufferers. Although the depressive episodes are short, they are as severe as the more enduring depressive disorders, and can be associated with suicidal behaviour. Thus recurrent brief depression is associated with much personal distress and social and occupational impairment. There appears to be little comorbidity with manic illness or dysthymia. Individuals with recurrent brief depression often receive treatment with antidepressant medication, but its value is questionable (Pezawas et al., 2005).

Classification in DSM and ICD

The main categories in the sections on mood disorders in DSM-IV and ICD-10 are shown in Tables 10.4 and 10.5. Broad similarities are evident, together with some differences. The first similarity is that both systems contain categories for single episodes of mood disorder as well as categories for recurrent episodes. The second is that both recognize mild but persistent mood disturbances in which there is either a repeated alternation of high and low mood (cyclothymia) or a sustained depression of mood (dysthymia). In neither case are the mood disturbances sufficiently severe to meet the criteria for a hypo-manic or depressive episode. In DSM-IV, mood disorders that are judged to be secondary to a medical condition are included as a subcategory of mood disorders, whereas in ICD-10 these conditions are classified as mood disorders under ‘Organic mental disorders.’

Bipolar disorder

Both classifications delineate hypomania from mania on the basis of duration of symptoms, absence of psychotic features, and lesser degree of social and occupational impairment. A manic episode can be further subdivided according to severity and whether or not psychotic symptoms are present. In DSM-IV, the presence of a single episode of hypomania or mania is sufficient to meet the criteria for bipolar affective disorder. In ICD-10, however, at least two episodes of mood disturbance are needed for this diagnosis. DSM-IV also categorizes bipolar disorder as follows:

• bipolar I (in which mania has occurred on least one occasion)

• bipolar II (in which hypomania has occurred, but mania has not).

The diagnosis of bipolar II disorder is intended to indicate the importance of detecting mild hypomanic episodes in patients who might otherwise be diagnosed as having recurrent major depression. The presence of such episodes may have implications for treatment response. In DSM-IV, if a hypomanic or manic episode appears to have been precipitated by a somatic treatment (e.g. antidepressant drug therapy), it is not counted towards the diagnosis of bipolar I or bipolar II disorder, although there is a growing body of opinion that antidepressant-induced mania or hypomania should be considered to indicate an underlying bipolar disorder (Goodwin, 2009).

Table 10.5 Classification of bipolar disorder


Bipolar spectrum disorders. There is debate about where the line should be drawn between unipolar depression and bipolar disorder, because some patients with definite depressive episodes appear to have features of bipolarity but do not meet the DSM-IV criteria for mania or hypo-mania. Such patients may, for example, have elevations of mood that last for less than the required 4 days, or that have little discernible effect on functioning. In DSM-IV, the diagnosis of bipolar disorder ‘not otherwise specified’ can be applied in this situation. The term bipolar spectrum is also used. The treatment of bipolar depression differs to some extent from that given to patients with recurrent unipolar depression (see below), so making the diagnostic distinction between unipolar and bipolar depression can have important implications for management. However, it is not clear whether depressed patients in the bipolar spectrum do better when treated as having bipolar depression rather than unipolar depression (for a discussion, see Goodwin, 2009).

Depressive disorders

Both ICD-10 and DSM-IV classify depressive episodes on the basis of severity and whether or not psychotic features are present. It is also possible to specify whether the depressive episode has melancholic (DSM-IV) or somatic (ICD-10) features. In DSM-IV, an episode of major depression with appropriate clinical symptomatology (see above) can be specified as atypical depression. In ICD-10, atypical depression is classified separately under ‘Other depressive episodes.’ Both ICD-10 and DSM-IV allow the diagnosis of recurrent brief depression, but under slightly different headings (see Table 10.6).

Table 10.6 Classification of depressive disorders


Classification and description in everyday practice

Although neither DSM-IV nor ICD-10 is entirely satisfactory, it seems unlikely that further re-arrangement of descriptive categories would be an improvement. A solution will only be achieved when we have a better understanding of aetiology. Meanwhile, either ICD-10 or DSM-IV should be used for statistical returns. For most clinical purposes, it is better to describe disorders systematically than to classify them.

This can be done for every case by referring to the severity, the type of episode, the distinguishing symptomatic features, and the course of the disorder, together with an evaluation of the relative importance of known aetiological factors (see Table 10.7). It has become conventional practice to record all cases with a manic episode as bipolar, even if there has been no depressive disorder, on the grounds that most manic patients develop a depressive disorder eventually, and in several important respects manic patients resemble patients who have had both types of episode. This convention is followed in this textbook.

Table 10.7 A systematic scheme for the clinical description of mood disorders


Differential diagnosis

Depressive disorder

Depressive disorders have to be distinguished from the following:

• normal sadness

• anxiety disorders

• schizophrenia

• organic brain syndromes.

As already explained on p. 205, the distinction from normal sadness is made on the basis of the presence of other symptoms of the syndrome of depressive disorder. Depressive disorders also have rates of comorbidity with a wide range of other disorders—for example, anxiety disorders, eating disorders, substance misuse, and personality disorder. In all of these cases it is important to recognize and treat the depressive disorder.

Anxiety disorders

Mild depressive disorders are sometimes difficult to distinguish from anxiety disorders. Accurate diagnosis depends on assessment of the relative severity of anxiety and depressive symptoms, and on the order in which they appeared. Similar problems arise when there are prominent phobic or obsessional symptoms, or when there are dissociative symptoms with or without histrionic behaviour. In all of these cases, the clinician may fail to identify the depressive symptoms and so prescribe the wrong treatment.


The distinction from schizophrenia can be most difficult. Auditory hallucinations and delusions, including some that are characteristic of schizophrenia, such as delusions of reference, can occur in manic disorders. However, these symptoms usually change quickly in terms of content, and seldom outlast the phase of overactivity. When there is a more or less equal mixture of features of the two syndromes, the term schizoaffective is often used. This term is discussed further in Chapter 11. Further clues to diagnosis can often be elicited by a careful personal and family psychiatric history.

Dementia and other organic conditions

In middle and late life, depressive disorders are sometimes difficult to distinguish from dementia, because some patients with depressive symptoms complain of considerable difficulty in remembering. In fact, patients with severe depression can perform very badly on tests of cognitive function, and distinction between the two conditions purely in terms of the nature of the cognitive impairment may not be possible. Here the presence of depressive symptoms is the key to diagnosis, which should be confirmed with improvement of the memory disorder as normal mood is restored. Numerous other general medical conditions can present with depressive features (see p. 384). The key to diagnosis is a careful history and physical examination, supplemented by special investigations where appropriate.


Manic disorders have to be distinguished from the following:

• schizophrenia

• organic brain disease involving the frontal lobes (including brain tumour and HIV infection)

• states of brief excitement induced by amphetamines and other illegal drugs.


The distinction from schizophrenia depends on a careful search for the characteristic features of this condition (see Chapter 11). Difficult diagnostic problems may arise when the patient has depressive psychosis, but here again the distinction can usually be made on the basis of a careful examination of the mental state, and of the order in which the symptoms appeared. Information about the past psychiatric history may also be useful. Particular difficulties also arise when symptoms characteristic of depressive disorder and of schizophrenia are found in equal measure in the same patient; these so-called schizoaffective disorders are discussed on p. 266.

Organic brain disorder and drug misuse

An organic brain lesion should always be considered, especially in middle-aged or older patients with expansive behaviour and no past history of affective disorder. In the absence of gross mood disorder, extreme social disinhibition (e.g. urinating in public) strongly suggests frontal lobe pathology. In such cases, appropriate neurological investigation is essential. In younger adults, infection with HIV or head injury may lead to the manifestation of mania.

The distinction between mania and excited behaviour due to drug misuse depends on the history, and on an examination of the urine for drugs, which is needed before treatment with psychotropic drugs is started. Drug-induced states usually subside quickly once the patient is in hospital (see Chapter 17). However, it should be remembered that a significant proportion of patients who have bipolar disorder misuse alcohol and other drugs.

The epidemiology of mood disorders

It is difficult to determine the prevalence of depressive disorder, partly because different investigators have used different diagnostic definitions. More modern investigations have used structured diagnostic interviews linked to standardized diagnostic criteria such as the Diagnostic Interview Schedule (DIS) and the Composite International Diagnostic Interview (CIDI) (see p. 65). Of all the mood disorders, bipolar cases are probably identified more reliably because mania is somewhat easier to define and diagnose.

Bipolar disorder

More recent community surveys in industrialized countries have suggested that:

• the lifetime risk for bipolar disorder is in the range 0.3–1.5%

• the 6-month prevalence of bipolar disorder is not much less than the lifetime prevalence, indicating the chronic nature of the disorder

• the prevalence in men and women is the same

• the mean age of onset is about 17 years in community studies

• bipolar disorder is highly comorbid with other disorders, particularly anxiety disorders and substance misuse.


Defining the boundaries of depressive episodes in community surveys presents a number of difficulties. However, if the DSM-IV criteria for major depression are applied, recent surveys in industrialized countries suggest that:

• the 12-month prevalence of major depression in the community is around 2–5%

• the lifetime rates in different studies vary considerably (in the range 4–30%). The true figure probably lies in the range 10–20%

• the mean age of onset is about 27 years

• rates of major depression are about twice as high in women as in men, across different cultures

• there may be increased rates of depression in people born since 1945

• rates of depression are higher in the unemployed and divorced

• major depression has a high comorbidity with other disorders, particularly anxiety disorders and substance misuse.

The reasons for higher rates of major depression among women are uncertain. This increase starts to become apparent at puberty, and could be due in part to a greater readiness of women to admit to having depressive symptoms. However, such selective reporting is unlikely to be the whole explanation. It is possible that some depressed men misuse alcohol and are diagnosed as suffering from alcohol-related disorders rather than depression, with the consequence that the true number of major depressive disorders is underestimated (Branney and White, 2008). Again, misdiagnosis of this kind is unlikely to account for the whole of the difference. The possibility that gender-related differences in neurobiology (e.g. the organization of central 5-HT pathways) might underlie differences in susceptibility to depression requires further investigation. Furthermore, in many societies women are subject to various kinds of social disadvantage and, for example, are more likely than men to experience sexual abuse and domestic violence. Factors of this sort are also likely to be implicated in their increased risk of depression.

Although it used to be thought that the risk of depressive disorders increased with age, recent surveys suggest that major depression is most prevalent in the 18–44 years age group. A number of studies have suggested that people born since 1945 in industrialized countries have both a higher lifetime risk of major depression and an earlier age of onset. These studies have mainly been retrospective, and it is possible that the apparent increased rate of major depression in young people is due to the fact that older people forget (or are less willing to reveal) that they have been depressed. For a review of the epidemiology of mood disorders, see Joyce (2009).

Dysthymia and recurrent brief depression

The lifetime risk for dysthymia is around 4% (Alonso et al., 2004). Rates of dysthymia are higher in women and in the divorced. There is less epidemiological information about recurrent brief depression, but in the Zurich prospective study the 12-month prevalence for recurrent brief depression was about 2.6%, very similar to the rate found for dysthmia (2.3%) (Pezawas et al., 2003).

Minor mood disorders

Estimates of the frequency of minor mood disorders show wide variations because the different studies have not defined cases in the same way. However, these are probably the most prevalent psychiatric disorders in the community. For example, in the National Psychiatric Morbidity Household Survey (Jenkins et al., 1997), the overall 1-week prevalence of neurotic disorder was 16% (12.3% in males and 19.5% in females). Nearly half of this group met the ICD-10 criteria for mixed anxiety and depression. Similar findings were obtained in the Adult Psychiatric Morbidity Survey in England, |in which McManus et al. (2007) found that mixed anxiety and depression was the most common mental disorder in the community, with a 1-week prevalence of 9%.

The aetiology of mood disorders

There have been many different approaches to the aetiology of mood disorders. There is substantial knowledge about the genetic epidemiology of depression and bipolar disorder, and how certain childhood experiences can lay down a predisposition to mood disorders in adulthood. There is also a good understanding of the role of current life difficulties and stresses in provoking mood disorders in predisposed individuals. There is much less knowledge about the mechanism involved in the translation of these predisposing and provoking factors into clinical symptomatology.

When trying to elucidate these mechanisms (which, of course, have important implications for treatment), investigators have employed two main conceptual approaches, which can be broadly termed psychological and biological. It is likely, of course, that these approaches represent different levels of enquiry that will eventually inform each other. The main kinds of investigations that will be discussed are listed in Table 10.8. In most research areas there are many more data available concerning the aetiology of depression than on that of mania. The aetiology section in the current chapter is structured so as to illustrate to the reader the many ways in which research into the causation of psychiatric disorder can be approached (see Box 10.1). It may therefore be helpful for this section to be read in conjunction with Chapter 5.

Genetic causes

Family and twin studies

Familial aggregation

The risk of mood disorders is increased in first-degree relatives of both bipolar and unipolar probands, with the risk being about twice as great in relatives of bipolar patients (see Table 10.9). Relatives of bipolar probands have increased risks of unipolar depression and schizoaffective disorder as well as bipolar disorder. By contrast, relatives of patients with unipolar depression do not have increased rates of bipolar disorder or schizoaffective disorder.

Table 10.8 The aetiology of mood disorders


Box 10.1 Multifactorial origin of mood disorders

• An important genetic contribution to mood disorder is made by multiple genes of small individual effect. This genetic contribution may be expressed directly through modification of relevant cortical circuitry, or indirectly through effects on personality and psychological coping mechanisms.

• Adverse early life experiences also shape personality and limit subsequent attachment behaviour and ability to access social support. In addition, adverse early experience may affect development of the HPA axis and neurobiological responses to stress in adulthood.

• Mood disorders are often triggered by current life events, particularly in people who lack social support. The impact of life events is modified by early life experience, personality, and genetic inheritance. The interaction of these factors determines the resilience or vulnerability of an individual to a life event, and the subsequent risk of clinical mood disturbance.

• The neurobiology of episodes of mood disorder is associated with changes in the activity of monoamine neurons and the HPA axis, which together modify the activity of the neural circuitry involved in emotional regulation. At a cellular level this may involve a loss of synaptic plasticity and dendrite formation. Some structural and functional brain abnormalities in mood disorder suggest persistent biological vulnerability, probably produced by genetic inheritance or early developmental factors.

Twin studies

Twin studies suggest that the aggregation of mood disorders in families is due to genetic factors, with the concordance rate for both bipolar and unipolar disorder being higher in monozygotic than in dizygotic twins. For example, the concordance rate for mood disorder in the monozygotic co-twin of a proband with bipolar disorder is 60–70%, but for dizygotic twins the rate is only about 20%. In unipolar depression the concordance rate is also higher in monozygotic twins (about 45%) than in dizygotic twins (about 20%). Overall, the genetic influence seems to be greater in bipolar disorder than in unipolar disorder. Thus a recent estimate of the heritability of bipolar disorder (85%) was substantially higher than the corresponding estimate for major depression (37%) (Bienvenu et al, 2011).

Table 10.9 Genetic epidemiology of bipolar disorder and major depression


Genetic evidence on classification of mood disorders

Genetic studies can also throw light on the aetiological relationship between different subtypes of mood disorder. For example, in an epidemiological sample of twins, Karkowski and Kendler (1997) obtained results which suggested that, in terms of genetic susceptibility, unipolar and bipolar depression lie on a continuum of severity, rather than being aetiologically distinct. However, McGuffin et al. (2003) found evidence that the genetic susceptibility to maniaappeared to be inherited more specifically, and did not seem to represent simply a more severe form of mood disorder. There is also evidence that partly distinct and partly overlapping genes contribute to bipolar disorder and schizophrenia (Cole et al., 2008).

Twin studies have suggested that susceptibility to major depression and generalized anxiety disorder involves similar genes but different environmental risk factors (Hettema et al., 2005). In addition, the effect of the environment on susceptibility to depression predominantly involves experiences specific to an individual, rather than shared (family) experiences.

Mode of inheritance

The familial segregation of mood disorders does not fit a simple Mendelian pattern. The female preponderance of depression is well established, and it is possible that genetic factors play a somewhat greater aetiological role in women. However, the data are inconsistent, and meta-analyses reveal rather similar heritability estimates (Lau and Eley, 2010). Overall, it seems likely that the genetic liability to mood disorders results from the combined action of multiple genes of modest, or small effect—so-called polygenic inheritance—in a context of gene–enviroment interaction. Although many important environmental contributions to depression have been identified (see below), the level at which they interact with genetic predisposition is largely unclear.

Molecular genetics

Linkage studies

Molecular linkage studies of mood disorders have so far not been particularly revealing, perhaps because the genes involved are of small effect, or because of genetic heterogeneity. Positive findings have emerged, but have often proved difficult to replicate even in large-scale studies and meta-analyses. For reviews, see Craddock and Sklar (2009) and Lohoff (2010).

Association studies

The monoamine theory of depression suggests that allelic variation in genes coding for monoamine synthesis or metabolism or specific receptors may contribute to the risk of mood disorders. There have been numerous association studies of such candidate genes, although so far the results have not been compelling. The gene coding for the serotonin transporter has a number of allelic variants, one of which, in the promotor region, influences the expression of transporter sites. The evidence implicating this polymorphism in major depression in population studies is inconsistent. However, in an influential study, Caspi et al. (2003) found evidence which suggested that subjects carrying a particular allele for the serotonin transporter were more likely to experience a subsequent episode of major depression when exposed to an adverse life event. The robustness of this particular gene–environment interaction has proved controversial (Risch et al., 2009; Caspi et al., 2010). Similarly, the evidence for other plausible candidates, such as the genes for tryptophan hydroxylase and brain-derived neurotrophic factor, has been inconsistent (Craddock and Sklar, 2009; Lohoff, 2010).

As noted earlier it is now possible to conduct genome-wide association studies (GWAS) which do not rely on hypothesized candidate genes. Such studies face statistical difficulties with regard to the issue of multiple testing, but results from large investigations are becoming available. As with other kinds of molecular genetic investigation, although associations have been identified there is a lack of consistency between different studies, even those that have very large data sets (Craddock and Sklar, 2009; Lohoff, 2010). In addition, GWAS studies have not confirmed the importance of loci implicated by previous candidate-gene studies.


Certain kinds of personality development could be associated with predisposition to mood disorder. For example, it is a common clinical observation that patients with depression often seem to have high levels of pre-morbid anxiety. Aspects of personality could be associated with mood disorder in a number of ways.

1. What is recognized as a personality characteristic may in fact represent a mild form of the illness. For example, Kraepelin (1921) suggested that people with cyclothymic personality (i.e. those with repeated and sustained mood swings) were more prone to develop manic–depressive disorder. Nowadays this personality type is classified as a mood disorder (cyclothymic disorder), and is considered to be a mild form of bipolar disorder.

2. Some personality features might influence the way in which people respond to adverse circumstances, and thus make depressive disorders more likely. For example, a cognitive style that is characterized by sociotropy (a strong need for approval) is associated with increased risk of depression after adverse life events (Mazure and Maciejewski, 2003).

3. Certain kinds of personality development and psychiatric disorder may share common genes. For example, neuroticism, as measured by the Eysenck Personality Questionnaire, predisposes to major depression, but twin studies suggest that neuroticism and major depression have genes in common (Fanous and Kendler, 2004).

Overall, current findings suggest that part of the genetic risk of depression takes the form of inheritance of particular character traits and cognitive styles, which predispose to psychiatric illness in the presence of specific kinds of life stresses (Rihmer et al., 2010).

Early environment

Parental deprivation

Psychoanalysts have suggested that childhood deprivation of maternal affection through separation or loss predisposes to depressive disorders in adult life. Overall, however, epidemiological studies do not suggest that loss of a parent by death in childhood increases the risk of depressive disorder in adulthood. By contrast, there is more support for the proposal that depressive disorder in later life is associated with parental separation, particularly divorce. The key factor here appears to be not so much the loss itself as the discord and diminished care that result from it. Indeed family discord and lack of adequate care predispose to depression even in families where separation does not occur (Harris, 2001).

Relationships with parents

It is clear that gross disruption of parent–child relationships, as occurs, for example, in physical and sexual abuse, is a risk factor for several kinds of adult psychiatric disorder, including major depression. It is less certain whether more subtle differences in parental style may also predispose to depression. One problem is the difficulty of determining retrospectively what kind of relationship a patient may have had with their parents in childhood. The patient’s recollection of the relationship may be distorted by many factors, including the depressive disorder itself. However, it appears that both non-caring and overprotective parenting styles are associated with non-melancholic depression in adulthood (Parker and Hadzi-Pavlovik, 1992). Mothers with postnatal depression may manifest a rearing style that is characterized by neglect and emotional indifference. This could lead to longer-term deleterious effects on self-esteem and attachment style in the child, thus increasing the risk of depression in the subsequent generation. Similar effects may occur when a father is depressed (Ramchandani et al., 2005).

Precipitating factors

Recent life events

Methodological considerations

It is an everyday clinical observation that depressive disorders often follow stressful events. However, several other possibilities must be discounted before it can be concluded that stressful events cause the depressive disorders that succeed them. First, the association might be coincidental. Secondly, the association might be non-specific—there might be as many stressful events in the weeks preceding other kinds of illness. Thirdly, the association might be spurious—the patient might have regarded the events as stressful only in retrospect when seeking an explanation for his illness, or he might have experienced them as stressful only because he was already depressed at the time. Finally, the depression itself might have caused the life event.

Research workers have tried to overcome each of these methodological difficulties. The first two problems—whether the events are coincidental, and whether any association is non-specific—require the use of control groups suitably chosen from the general population and from people with other illnesses. The third problem—whether the association is spurious—requires two other approaches. The first approach is to separate events that are undoubtedly independent of illness (e.g. losing a job because a whole factory closes) from events that may have been secondary to the illness (e.g. losing a job when no one else is dismissed). The second approach is to assign a rating to each event according to the consensus view of healthy people about its stressful qualities.


Methodologically reliable research has shown that:

• there is a sixfold excess of adverse life events in the months before the onset of depressive disorder

• an excess of similar events also precedes suicide attempts, and the onset of anxiety disorders and schizophrenia

• in general, ‘loss’ events are associated with depression and ‘threat’ events are associated with anxiety

• life events are important antecedents of all forms of depression, but appear to be relatively less important in established melancholic-type disorders and where there is a strong family history of depression.

Events that lead to feelings of entrapment and humiliation may be particularly relevant to the onset of depression. In contrast, remission from depression is often associated with ‘fresh-start’ life events (e.g. establishing a new relationship or starting an educational course) (Harris, 2001). It is also important to note that genetic factors may be involved in the liability of an individual to experience life events. Thus certain individuals seem to be more prone to select risky environments, and genetic factors also play a role in how life events are perceived by a particular individual, perhaps through the personality mechanisms discussed above (Kendler et al., 1999).

In general, the importance of life events in the onset of a depressive episode decreases as the number of episodes increases. This suggests that once a depressive disorder is clearly established, depressive episodes can occur in the absence of major environmental precipitants. This relationship is much less clear where there is a strong family history of depression, raising the possibility that one of the mechanisms by which a family history increases the risk of depression is by diminishing the need for a major environmental stressor during the first episodes of the illness (Kendler et al., 2001).


It is less certain whether mania is provoked by life events. In the past, mania was thought to arise entirely from endogenous causes. However, clinical experience suggests that a proportion of cases are precipitated, sometimes by events that might have been expected to induce depression (e.g. bereavement). As in the case of recurrent depression, it is possible that the impact of life events may be more important early in the course of a recurrent manic–depressive illness, because once the illness is established environmental precipitants are less important. However, adverse life events increase symptomatology in the course of bipolar illness, although the effect seems to be predominantly on depression. Interestingly, manic symptoms may be triggered when people attain an important goal, which suggests that ‘normal’ feelings of happiness and well-being can develop into manic illness in predisposed individuals (Proudfoot et al., 2011).

Vulnerability factors and life difficulties

It is a common clinical impression that the events immediately preceding a depressive disorder act as the ‘last straw’ for a person who has been subjected to a long period of adverse circumstances, such as an unhappy marriage, problems at work, or unsatisfactory housing. Brown and Harris (1978) divided predisposing events into two categories. The first kind are prolonged stressful circumstances which can themselves cause depression as well as adding to the effects of short-term life events. Brown and Harris gave the name long-term difficulties to these circumstances.

The second kind of predisposing circumstance does not itself cause depression, but rather it acts by increasing the effects of short-term life events. This kind of circumstance is known as a vulnerability factor. In practice, the distinction between the two kinds of circumstance is not clear-cut. Thus a long history of marital problems (a long-term difficulty) is likely to be associated with lack of a confiding relationship, and the latter was identified by Brown as a vulnerability factor.

Overall, the studies suggest that the following social vulnerability factors increase the risk that life events will trigger a depressive episode:

• being responsible for the care of young children

• not working outside the home

• having no one to confide in.

In general, there is good evidence that poor social support, measured as lack of intimacy or social integration, is associated with an increased risk of depression. The mechanism of this association is unclear, and is open to different interpretations. First, it may be that a lack of opportunities to confide makes people more vulnerable. Secondly, it may indicate that depressed people have a distorted perception of the degree of intimacy that they achieved before becoming depressed. Thirdly, some other factor, presumably an abnormality in personality, may result in difficulty in confiding in others, and thus lead to vulnerability to depression (Harris, 2001). For a further discussion of this work, see p. 94.

The effects of physical illness

All medical illnesses and their treatment can act as non-specific stressors which may lead to mood disorders in predisposed individuals. However, sometimes certain medical conditions are believed to play a more direct role in causing the mood disorder (e.g. brain disease, certain infections, including HIV, and endocrine disorders). The resulting mood disorders are known as organic mood disorders (for a further discussion of this subject, see Chapter 15).

Inevitably, the above distinction is arbitrary. For example, major depression occurs in about 50% of patients with Cushing’s disease. Since not all patients with Cushing’s disease suffer from depressive disorder, it follows that variables other than raised plasma cortisol levels are involved. However, organic mood disorders can give clues to aetiology. For example, depressive disorders in patients with Cushing’s disease remit after cortisol levels are restored to normal, a finding that led to the proposal that increased cortisol secretion may play a role in the pathophysiology of major depression (see p. 410). It is also worth noting here that the puerperium, although not an illness, is associated with an increased risk of mood disorders (see p. 424).

Psychological approaches to aetiology

These theories are concerned with the psychological mechanisms by which recent and remote life experiences can lead to depressive disorders. Much of the literature on this subject fails to distinguish adequately between the symptom of depression and the syndrome of depressive disorder. The main approaches to the problem are derived from the ideas of psychoanalysis (see Box 10.2) and cognitive–behavioural theories.

Cognitive theories

Depressed patients characteristically have recurrent and intrusive negative thoughts (‘automatic thoughts’). Beck (1967) proposed that these depressive cognitions reveal negative views of the self, the world, and the future (the depressed patient usually reviews the past in a similar vein). These automatic thoughts appear to persist because of illogical ways of thinking, which Beck called cognitive distortions. These include:

• arbitrary inference (drawing a conclusion when there is no evidence for it and even some evidence against it)

• selective abstraction (focusing on a detail and ignoring more important features of a situation)

• overgeneralization (drawing a general conclusion on the basis of a single incident)

• personalization (relating external events to oneself in an unwarranted way).

Box 10.2 Psychoanalytical theory

• In his seminal paper ‘Mourning and melancholia’, Freud drew attention to the resemblance between the phenomena of mourning and symptoms of depressive disorders, and suggested that their causes might be similar (Freud, 1917).

• Freud suggested that, just as mourning results from loss by death, so melancholia results from loss of other kinds. Since it was apparent that not every depressed patient had suffered an actual loss, it was necessary to postulate a loss of ‘some abstraction’ or internal representation, or in Freud’s terms the loss of an ‘object.’

• Freud pointed out that depressed patients often appear to be critical of themselves, and he proposed that this self-accusation was really a disguised accusation of someone else for whom the patient ‘felt affection.’ In other words, depression was thought to occur when feelings of love and hostility were present at the same time (ambivalence). When a loved ‘object’ is lost, the patient feels despair; at the same time, any hostile feelings attached to this ‘object’ are redirected against the patient him- or herself as self-reproach.

• Freud’s ideas were developed by Melanie Klein (1882–1960), who believed that weaning (loss of the breast) represents a major symbolic loss for the infant, who then feels remorse and guilt about the disappearance of this ‘good object.’ In normal development this ‘depressive anxiety’ leads to attempts at reparation and concern for others. Failure to negotiate this process can result in depressive reactions in the face of future losses.

• John Bowlby (1907–1990) showed that the rearing abilities of the main caregiver play an important role in giving the infant a secure emotional ‘attachment’, which is of critical importance in the development of satisfactory interpersonal relationships. Insecure attachments can increase the risk of various kinds of adult psychopathology, including depression.

For a review of the psychoanalytical model of depression and its application to treatment, see Taylor (2008).

Although most psychiatrists regarded these cognitions as secondary to a primary disturbance of mood, Beck suggested that another set of cognitions precede depression and predispose to it. These cognitions are dysfunctional beliefs such as ‘If I am not perfectly successful then I am a nobody.’ These beliefs are thought to affect the way in which a person responds to stress and adversity. For example, a failure at work is more likely to provoke depression in a person who holds the belief described above. Others have shown that, when depressed, people are more likely to remember unhappy events. This selective recall adds to the person’s depression, making it more difficult to reverse.

One of the major problems for the cognitive model is that although dysfunctional attitudes and beliefs, as well as negatively biased information processing, are easy to demonstrate when patients are depressed, many of these phenomena remit on clinical recovery. For this reason, more recent cognitive theory has suggested that dysfunctional beliefs may be present but difficult to demonstrate (i.e. ‘latent’) until they are activated by a relevant stressor or by a minor change of mood (within the normal range). According to this model a minor (normal) lowering of mood can activate all of the concepts and constructs associated with clinical depression. An interactive cycle of low mood and negative thinking will result (see Figure 10.1; Kuyken et al., 2007).

In fact more recent investigations do suggest that abnormalities in information processing, such as negative biases in facial expression recognition, can be demonstrated in both recovered depressed patients and those at high risk of depression prior to the development of illness (Joorman et al., 2007; Victor et al., 2010). Therefore it is possible that subtle negative biases in information processing may predispose to the development of depression, particularly in the context of psychosocial stress and adverse life events.

Neurobiological approaches to aetiology

Whatever their aetiology, the clinical manifestations of depressive disorders must ultimately be mediated through changes in brain neurochemistry and the circuitry involved in emotional regulation. Biochemical investigations in depressed patients have focused on the monoamine neurotransmitters because monoamine pathways appear to play an important role in the actions of effective antidepressant drugs (see below). Of course the actions of antidepressant drugs may not reverse the cause of depression, but may merely change the expression of symptoms. However, another reason for studying monoamines is that they play an important role in mediating adaptive responses to stressful events, and depressive disorders can be viewed as a failure of these adaptive responses.


Figure 10.1 Cognitive model of how latent dysfunctional assumptions (laid down by early experience) are activated by critical incidents, leading to a vicious cycle of negative thinking and depressed mood. Reproduced from Keith Hawton, Paul. M. Salkovskis, Joan Kirk, and David M. Clark, Cognitive Behaviour Therapy for Psychiatric Problems A Practical Guide © Oxford University Press, 1989, with permission.

Monoamine pathways, particularly those involving noradrenaline and 5-HT, innervate cortical and subcortical brain regions thought to be involved in mood regulation. Recent studies have used structural and functional imaging techniques to elicit changes in the neural circuitry that underpins the expression of clinical affective symptomatology. In hand with this has come the realization that mood disorders, despite their fluctuating and remitting clinical course, are associated with distinct and persistent neuropathological changes in relevant brain regions. Some of these changes are likely to be developmental in origin; this could encompass either genetic inheritance or the consequences of adverse childhood experiences. There is currently much ongoing research effort in both animal and human studies to assess how traumatic experiences in development might affect the subsequent maturation of the hypothalamic–pituitary–adrenal (HPA) axis and thereby influence responses to environmental stress. A final point worth considering is that the prognosis of mood disorders worsens as the number of episodes increases. Although this phenomenon may simply represent an association between severity and poor prognosis, it is also possible that acute episodes of mood disturbance themselves produce neurobiological effects that worsen the subsequent prognosis (a phenomenon that is sometimes known as ‘scarring’).

The monoamine hypothesis

This hypothesis suggests that depressive disorder is due to an abnormality in a monoamine neurotransmitter system at one or more sites in the brain. In its early form, the hypothesis suggested a changed provision of the monoamine; more recent elaborations postulate alterations in receptors as well as in the concentrations or the turnover of the amines. Three monoamine transmitters have been implicated—serotonin (5-hydroxytryptamine, 5-HT), noradrenaline, and dopamine. The latter two neurotransmitters are called catecholamines.

The hypothesis has been tested by observing three kinds of phenomenon:

• the biochemistry of neurotransmitters in patients with mood disorders

• the effects of selective drugs on measurable indices of the function of monoamine systems (usually neuroendocrine indices)

• the pharmacological properties shared by antidepressant drugs.

The biochemical effects of antidepressant drugs are considered in Chapter 19. The present chapter will consider the evidence for abnormalities in monoamine neurotransmitters in untreated depressed patients (see Box 10.3). Much less work of this kind has been carried out in mania. Therefore the present discussion will be largely restricted to depressive disorders, although the dopamine hypothesis of mania will also be outlined.

5-HT function

Plasma trypytophan. The synthesis of 5-HT in the brain depends on the availability of its precursor amino acid, L-tryptophan. Plasma tryptophan levels are decreased in untreated depressed patients, particularly in those with melancholic depression. Studies in healthy subjects have shown that weight loss through dieting can lower plasma tryptophan levels, and this factor appears to explain some, but not all, of the reduction in plasma tryptophan concentration that is seen in depression. Decreases in plasma tryptophan levels may contribute to the impairments that are seen in brain 5-HT function in depressed patients, but are probably not an important causal factor (Cowen, 2008).

Box 10.3 Abnormalities in monoamine neurotransmission in depression


• Decreased plasma tryptophan

• Blunted 5-HT neuroendocrine responses

• Decreased brain 5-HT1A receptor binding (PET)

• Decreased brain 5-HT reuptake sites (SPET and PET)

• Clinical relapse after tryptophan depletion


• Blunted noradrenaline-mediated growth hormone release

• Clinical relapse after AMPT


• Decreased homovanillic acid (HVA) levels in CSF

• Clinical relapse after AMPT

Studies of cerebrospinal fluid. Indirect evidence about 5-HT function in the brains of depressed patients has been sought by examining cerebrospinal fluid (CSF). Numerous studies have been carried out, but overall the data do not suggest that drug-free patients with major depression have a consistent reduction in CSF concentrations of 5-hydroxyindoleacetic acid (5-HIAA), the main metabolite of 5-HT formed in the brain. However, there is more consistent evidence that depressed patients who have made impulsive and more dangerous suicide attempts have low CSF 5-HIAA levels. This finding is not restricted to patients with depression. It has also been reported in, for example, patients with schizophrenia and personality disorder who have a history of aggressive behaviour directed towards themselves or other people. It has been proposed that low levels of CSF 5-HIAA, although not related specifically to depression, may be associated with a tendency of individuals to respond in an impulsive and hostile way to life difficulties (Placidi et al., 2001).

Studies of post-mortem brain. Measurements of 5-HT and 5-HIAA have been made in the brains of depressed patients who have died, usually by suicide. Although this is a more direct test of the monoamine hypothesis, the results are difficult to interpret for two reasons. First, the observed changes may have taken place after death. Secondly, the changes may have been caused before death, but by factors other than the depressive disorder (e.g. by anoxia, or by drugs used in treatment or taken in order to commit suicide). Overall there is little consistent evidence that depressed patients dying from natural causes or suicide have lowered brain concentrations of 5-HT or 5-HIAA. More recent studies have adopted techniques such as receptor autoradiography and mRNA expression. Such studies have suggested that suicide victims have increased expression of 5-HT2A receptors and decreases in serotonin transporters (5-HT reuptake sites) in the prefrontal cortex. However, the data are not consistent (Stockmeier, 2003).

Neurochemical brain imaging studies. Recent developments in brain imaging with selective labelled ligands have enabled the assessment of certain brain 5-HT receptor subtypes in vivo. There is evidence of a widespread modest decrease in 5-HT1A-receptor binding throughout the cortical and subcortical regions (Savitz et al., 2009). There also appear to be reductions in the number of brainstem 5-HT reuptake sitesin depressed individuals, consistent with a decrease in the density of 5-HT cell bodies (Selvaraj et al., 2011).

Neuroendocrine tests. The functional activity of 5-HT systems in the brain has been assessed by giving a substance that stimulates 5-HT function, and by measuring an endocrine response that is controlled by 5-HT pathways, usually the release of prolactin, growth hormone, or cortisol. Neuroendocrine challenge tests have the advantage that they measure an aspect of brain 5-HT function. However, the 5-HT synapses that are involved presumably reside in the hypothalamus, which means that important changes in 5-HT pathways in other brain regions could be missed. A number of drugs have been used to increase brain 5-HT function for the purposes of neuroendocrine challenge.

Studies in unmedicated depressed patients have shown consistent evidence that 5-HT-mediated endocrine responses are blunted in depressed patients. A number of these abnormalities persist into clinical recovery, which suggests that there is persistent dysfunction of some aspects of 5-HT neurotransmission in those at risk of depression (Cowen, 2008).

Tryptophan depletion. Although the findings outlined above provide strong evidence that aspects of brain 5-HT neurotransmission are abnormal in depression, they do not reveal whether these changes are central to pathophysiology or might instead represent some form of epiphenomenon. To assess this, it is necessary to study the psychological consequences of lowering brain 5-HT function in healthy subjects and those at risk of mood disorder.

As mentioned above, the synthesis of brain 5-HT is dependent on the availability in the brain of its amino acid precursor, L-tryptophan. It is possible to produce a transient lowering of plasma tryptophan and brain 5-HT function over a few hours by administering a mixture of amino acids that lacks tryptophan. This procedure is called tryptophan depletion. Tryptophan depletion in subjects with no personal or family history of mood disorder has little measurable effect on mood, and certainly does not produce significant clinical depressive symptomatology. By contrast, unmedicated euthymic patients with a personal history of mood disorder undergo a rapid but temporary depressive relapse when exposed to tryptophan depletion. People with a strong family history of depression but no previous illness show some lowering of mood after tryptophan depletion, but these changes are relatively mild and not of clinical severity. The following conclusions can be drawn.

• Low brain 5-HT function is not sufficient to cause depression, because tryptophan depletion fails to alter mood in those who are not vulnerable to mood disorder.

• Low brain 5-HT function interacts with other vulnerability factors to cause depressive symptoms.

The nature of these other vulnerability factors remains conjectural. It is possible that the 5-HT pathways of vulnerable individuals react abnormally to precursor deficit. Alternatively, there may be pre-existing deficits in the central mood-regulating circuitry which are ‘revealed’ in the presence of low brain 5-HT states. For a review of tryptophan depletion studies in depression, see Ruhe et al. (2007).

Noradrenaline function

Metabolism and receptors. There is no consistent evidence that brain or CSF concentrations of noradrenaline or its major metabolite 3-methoxy-4-hydroxy-phenylethylene glycol (MHPG) are altered in depressed patients (Anand and Charney, 2000). As with 5-HT receptors, noradrena-line receptors in the brain can be divided into a number of subclasses. There is some evidence that depressed patients who die from suicide have increased expression of α2-adrenoceptor binding in some brain regions (Escriba et al., 2004).

Neuroendocrine tests. Increasing brain noradrenaline function elevates plasma concentrations of adrenocorticotropic hormone (ACTH), cortisol, and growth hormone. There is fairly consistent evidence that the growth hormoneresponse to both the noradrenaline reuptake inhibitor desipramine and the noradrenaline-receptor agonist clonidine is blunted in patients with melancholic depression. Clonidine acts directly on postsynaptic α2-adrenoceptors in the hypothalamus to increase plasma growth hormone, and therefore the blunted response in depressed patients suggests a decreased responsivity of hypothalamic postsynaptic α2-adrenoceptors (Anand and Charney, 2000). Clearly this finding appears to be inconsistent with the increased expression of α2-adrenoceptors in cortical regions noted above.

Catecholamine depletion. It is possible to lower the synthesis of catecholamines by inhibiting the enzyme tyrosine hydroxylase, which catalyses the conversion of the amino acid tyrosine to L-DOPA, a precursor of both noradrenaline and dopamine. The drug that is used to achieve this effect is α-methyl-para-tyrosine (AMPT). In healthy subjects, AMPT produces sedation but not significant depressive symptoms. However, as with tryptophan depletion, when administered to recovered depressed patients off drug treatment it causes a striking clinical relapse in depressive symptomatology (Ruhe et al., 2007). This could be mediated by diminished function of either dopamine or noradrenaline, or by combined inhibition of both of these neurotransmitters.

These findings suggest that subjects at risk of mood disorder are vulnerable to decreases in both 5-HT and catecholamine neurotransmission. This is consistent with the clinical evidence that drugs that act selectively on noradrenaline or 5-HT pathways are effective antidepressant treatments.

Dopamine function

Depression. The function of dopamine in depression has been less well studied than that of 5-HT or noradrena-line, but there are a number of reasons for believing that dopamine neurons may be involved in the pathophysiology of the depressed state.

• Dopamine neurons in the mesolimbic system play a key role in incentive behaviour and reward. These processes are disrupted in depression, particularly melancholic states.

• Antidepressant treatments in animals increase the expression of dopamine receptors in the part of the mesolimbic system called the nucleus accumbens.

There is some evidence which suggests that dopamine function may be abnormal in depression. For example, CSF levels of the dopamine metabolite homovanillic acid (HVA) are consistently low in depressed patients, and some brain imaging studies in depressed patients have found increased binding of dopamine D2/D3 receptors in striatal regions. However, this may be restricted to patients with psychomotor retardation. There is also limited evidence for regional reductions in the density of dopamine D1 receptors (Hirvonen et al., 2008; Cannon et al., 2009).

These findings, taken together with the effect of AMPT in causing relapse in recovered depressed patients, suggest that impaired dopamine function may play a role in the manifestation of the depressive syndrome and in the effects of antidepressant drug treatment.

Mania. An influential hypothesis links excessive dopamine activity to the pathophysiology of schizophrenia (see p. 278). It has also been proposed that manic states may be attributable to dopamine overactivity. Studies of dopamine metabolism and function have provided little direct evidence for this suggestion, but mania can be provoked by dopamine agonists such as bromocriptine and pramipexole, and the euphoriant and arousing effects of psychostimulants such as amphetamine and cocaine are well known. Finally, dopamine-receptor-antagonist drugs such as haloperidol are useful in the treatment of mania.

A study in recovered bipolar patients showed that they had a greater psychological response to intravenous amphetamine but no greater release of presynaptic dopamine as measured by the technique of raclopride displacement (see p. 106). This suggests that there may be a heightened responsivity to increased dopamine neurotransmission in patients at risk of mania, but this is unlikely to be due to increased presynaptic dopamine release.

For a review of the role of dopamine in mania, see Cousins et al. (2009).

Role of monoamines

There is now good evidence that unmedicated depressed patients have abnormalities in various aspects of monoamine function. However, these abnormalities vary in extent from one case to another, and the changes are not large and are not sufficiently sensitive to be diagnostic. Some abnormalities may also persist into clinical recovery, suggesting that they are related to vulnerability to illness rather than the acute depressive state.

The most convincing studies that show a key role for monoamines in the pathophysiology of depression are the 5-HT and catecholamine depletion paradigms. This is because lowering of 5-HT and noradrenaline and dopamine function is sufficient to cause clinical depression in those at risk by virtue of a previous illness. It is unclear what neurobiological changes produce this psychological vulnerability to monoamine depletion, but it seems likely that they are at least in part consequences of previous episodes of illness and perhaps their treatment.

Amino acid neurotransmitters

Developments in magnetic resonance spectroscopy (MRS) have facilitated measurement of the amino acid neurotransmitters GABA and glutamate in patients with mood disorders. Overall there is evidence for decreased levels of glutamate in the anterior brain regions in depressed patients. Brain levels of GABA are also lowered in depression, although for technical reasons much of this work involves measurement in the occipital cortex. GABA concentrations may also be lower in patients with panic disorder, consistent with the known role of GABA in anxiety (Sanacora, 2010).

There is growing interest in the role of glutamate in mood disorders, partly because drugs with glutamatergic properties might be helpful in treatment. For example, the NMDA-receptor antagonist ketaminecan produce rapid and striking, albeit transient, relief of symptomatology in patients with treatment-refractory depression (Zarate et al., 2010).

Endocrine abnormalities

Abnormalities in endocrine function may be important in aetiology for two reasons.

• Some disorders of endocrine function are followed by mood disorders more often than would be expected by chance, suggesting a causative relationship.

• Endocrine abnormalities found in depressive disorder indicate that there may be a disorder of the hypothalamic centres that control the endocrine system.

Endocrine pathology and depression

About 50% of patients with Cushing’s syndrome suffer from major depression, which usually remits when the cortisol hypersecretion is corrected. Depression also occurs in Addison’s disease, hypothyroidism, and hyperparathyroidism. Endocrine changes may account for depressive disorders that occur premenstrually, during the menopause, and after childbirth. These clinical associations are discussed further in Chapter 15.

Hypothalamic–pituitary–adrenal (HPA) axis

Much research effort has been concerned with abnormalities in the control of cortisol in depressive disorders. In about 50% of patients whose depressive disorder is at least moderately severe, plasma cortisol secretion is increasedthroughout the 24-hour cycle. This increase in cortisol secretion is associated with enlargement of the adrenal gland and increased cortisol response to corticotropin (ACTH) challenge.

When studying depressed patients, use has been made of the dexamethasone suppression test, which suppresses cortisol levels via inhibition of ACTH release at pituitary level. About 50% of depressed inpatients do not show the normal suppression of cortisol secretion induced by administering 1 mg of the synthetic corticosteroid dexamethasone, an agent which suppresses ACTH via interaction with specific glucocorticoid receptors.

Dexamethasone non-suppression is more common in depressed patients with melancholia, but it has not been reliably linked with any more specific psychopathological feature. However, abnormalities in the dexamethasone suppression test are not confined to mood disorders; they have also been reported in mania, chronic schizophrenia, and dementia. This lack of diagnostic specificity diminished early hopes that dexamethasone non-suppression could be used as a diagnostic marker of melancholic depression.

The diminished responsivity to dexamethasone that is seen in depressed patients led to the glucocorticoid-receptor hypothesis of depression, whereby dysfunction of the HPA axis and the resulting depressive syndrome are linked to genetic or acquired defects of glucocorticoid receptors. This has coincided with findings from animal experimental studies that different classes of antidepressant medication increase expression of glucocorticoid receptors. Therefore one therapeutic mechanism of anti-depressant drug action may be to normalize excessive HPA axis activity via increased ability of the glucocorticoid receptors to provide feedback regulation (Pariante and Lightman, 2008).

In general, HPA axis changes in depressed patients have been regarded as state abnormalities—that is, they remit when the patient recovers. However, there is some evidence that changes in HPA axis function may persist in recovered depressed subjects. This suggests that some vulnerable individuals may have fairly enduring abnormalities is HPA axis regulation (Cowen, 2010). In experimental animal studies, early adverse experiences produce long-standing changes in HPA axis regulation, indicating a possible neurobiological mechanism whereby childhood trauma could be translated into increased vulnerability to mood disorder. Recent studies have confirmed that adults who were abused as children have altered HPA responses to stress (Carpenter et al., 2009).

Corticotropin-releasing hormone (CRH) and depression. In addition to its effects on cortisol secretion, CRH may play a more direct role in the aetiology of depression. It is well established that CRH has a neurotransmitter role in limbic regions of the brain where it is involved in regulating biochemical and behavioural responses to stress. Administration of CRH to animals produces changes in neuroendocrine regulation, sleep, and appetite that parallel those found in depressed patients. Furthermore, CRH levels may be increased in the CSF of depressed patients. Therefore it is possible that hypersecretion of CRHcould be involved in the pathophysiology of the depressed state, and non-peptide antagonists of CRH receptors may have value as antidepressant agents. For a review of this subject, see Pariante and Lightman (2008).

Thyroid function

Circulating plasma levels of free thyroxine appear to be normal in depressed patients, but levels of free tri-iodothyronine may be decreased. About 25% of depressed patients have a blunted thyrotropin-stimulating hormone (TSH)response to intravenous thyrotropin-releasing hormone (TRH). This abnormality is not specific to depression, as it is also found in alcoholism and panic disorder. Like CRH, TRH has a role in brain neurotransmission and is found in brain neurons co-localized with classical monoamine neurotransmitters such as 5-HT. Therefore it is possible that the abnormalities in thyroid function which are found in depressed patients may be associated with changes in central TRH regulation (Bunevicius and Prange, 2010).

Depression and the immune system

There is growing evidence that patients with depression manifest a variety of disturbances of immune function. Earlier studies found decreases in the cellular immune responses of lymphocytes in depressed patients, but more recent research has produced evidence of immune activation, with increases in particular in the release of certain cytokines (see Box 10.4). Cytokines are known to provoke HPA axis activity, and therefore changes in immune regulation may play a role in HPA axis dysfunction in depression.

Box 10.4 Immune changes in depression

• Lowered proliferative responses of lymphocytes to mitogens

• Lowered natural killer cell activity

• Increases in positive acute phase proteins

• Increases in cytokine levels (e.g. IL-6 and TNFα)

• Induction of indoleamine 2,3-dioxygenase

It is also possible that the changes seen in immune activity are secondary to other depressive features (e.g. lowered food intake and diminished self-care). However, the fact that medical administration of some cytokines (e.g. interferon and tumour necrosis factor) can cause significant depressive symptoms suggests that, in some situations, changes in immune function may have a more direct role in provoking mood disorders. Current formulations propose that cytokines can induce expression of the tryptophan-metabolizing enzyme indoleamine 2, 3-dioxygenase. This lowers tryptophan levels, which could put vulnerable individuals at risk of depression. For reviews of this subject, see Blume et al. (2011) and Maes (2011).

Sleep changes in depression

Disturbed sleep is characteristic of depression. Recordings of the sleep EEG (polysomnogram) have shown a number of abnormalities in sleep architecture in patients with major depression, including the following:

• impaired sleep continuity and duration

• decreased deep sleep (stages 3 and 4)

• decreased latency to the onset of rapid eye movement (REM) sleep

• an increase in the proportion of REM sleep in the early part of the night.

Decreased REM sleep latency is of interest in relation to aetiology because there is some evidence that it may persist in recovered depressed patients and indicate a vulnerability to relapse. A further link between REM sleep and depression is that many (but not all) effective antidepressant drugs decrease REM sleep time and the latency to its onset. In addition, both total sleep deprivation and selective REM sleep deprivation can produce a temporary alleviation of mood in depressed patients. For a review, see Steiger and Kimura (2010).

The neurochemical mechanism that links changes in REM sleep to mood is not known. The abnormalities in REM sleep in depressed patients could be attributable to excessive sensitivity of muscarinic cholinergic receptors(Modell and Lauer, 2007).

Brain imaging in mood disorder

Structural brain imaging

Changes in brain volume. Computerized tomography (CT) and magnetic resonance imaging (MRI) studies have found a number of abnormalities in patients with mood disorders, particularly in those with more severe and chronic disorder. The most consistent findings are:

• enlarged lateral ventricles (predominantly in elderly subjects with late-onset depression)

• decreased hippocampal volume (more consistently in unipolar than in bipolar subjects)

• decreased volume of basal ganglia structures (in unipolar but not bipolar subjects)

• decreased grey matter volume in the subgenual pre-frontal cortex (in unipolar and bipolar subjects).

The origin of these structural abnormalities is unclear, but might be related to the cellular neuropathological abnormalities that are increasingly described in depression (see Box 10.5). The neurotrophic hypothesis of depression suggests that stress (perhaps aided by cortisol hypersecretion) can lead to atrophy and death of neurons and down-regulation of adult neurogenesis, particularly in the hippocampus (Duman, 2009). Conceivably processes of this kind could lead to the kind of structural deficits that are listed above. In elderly subjects with late-onset depression, structural brain changes may be attributable to cardiovascular disease (see below and Drevets et al., 2009).

Box 10.5 Some cellular pathological abnormalities in patients with mood disorders

1. Decreased glial cell numbers (anterior cingulate cortex)

2. Decreased neuronal size and density (anterior cingulate, prefrontal cortex)

3. Decreased synaptic markers (prefrontal cortex)

From Harrison (2002).

White matter hyperintensities. Hyperintense MRI signals can be detected in a number of regions in both normal ageing subjects and older patients with major depression. The usual sites are in the deep white matter and periventricular white matter. In major depression, increased deep white matter hyperintensities are associated with the following:

• late onset of depressive disorder

• greater illness severity and poorer treatment response

• apathy, psychomotor slowness, and retardation

• the presence of vascular risk factors.

It has been proposed that major depression with these clinical and radiological features is likely to be caused by vascular disease, which presumably impairs functioning in the pathways involved in mood regulation (Drevets et al., 2009). There are also reports that white matter hyperintensities can be found in both old and young patients with bipolar disorder. However, the significance of these abnormalities is not clear at present.

Cerebral blood flow and metabolism

Cerebral blood flow can be measured in a number of ways—for example, with single-photon emission tomography (SPET), positron emission tomography (PET), or functional magnetic resonance imaging (fMRI). PET can also be used to measure cerebral metabolism. Cerebral blood flow and metabolism are normally highly correlated.

Numerous studies have examined both cerebral metabolism and blood flow in groups of depressed patients. The findings have often been contradictory, and there are many methodological factors, such as patient selection, drug status, and imaging techniques, that may account for the discrepant findings. Nevertheless, there is some consensus that in depressed patients there is evidence of altered cerebral blood flow and metabolism in the following regions:

• the prefrontal cortex (orbitofrontal, dorsolateral, and dorsomedial cortex)

• the anterior cingulate cortex, particularly the subgenual region

• the amygdala and thalamus

• the caudate nucleus and ventral striatum.

Taken together, the abnormalities in functional brain imaging in depression support a circuitry model in which mood disorders are associated with abnormal interactions between several brain regions, rather than a major abnormality in a single structure. The circuitry that is implicated involves regions of the medial prefrontal cortex, anterior cingulate cortex, amygdala, ventral striatum, thalamus, and hypothalamus (Price and Drevets, 2010). Some tentative correlations between these brain regions and clinical depressive features are listed in Box 10.6.

Neuropsychological changes in mood disorder

Patients with acute depression and mania show poor performance on several measures of neuropsychological function. Impairment is typically seen over a wide range of neuropsychological domains, including attention, learning, memory, and executive function. There is disagreement as to whether these defects are best regarded as global and diffuse, or whether there may be some selectivity with regard to the changes that are seen. However, some authors have suggested that deficits in executive function may be particularly prominent, which would be consistent with the abnormalities seen in prefrontal perfusion in imaging studies.

Most of the cognitive impairments resolve as the mood disorder remits, but a growing literature attests to the persistence of cognitive defects, albeit less striking, in euthymic patients with recurrent depression and bipolar disorder (Chamberlain and Sahakian, 2004; Clark et al., 2009). Such deficits are particularly apparent in elderly patients, in whom they have been found to correlate with decreases in hippocampal volume.

Box 10.6 Some neuropsychological correlates of altered central perfusion and metabolism in depressed patients

Dorsolateral and dorsomedial prefrontal cortex

• Cognitive dysfunction (particularly executive dysfunction) and cognitive slowness

Orbital and lateral prefrontal cortex

• Abnormal emotional processing

• Perseverative thinking

Anterior cingulate

• Impaired attentional processes

• Abnormal emotional processing


• Abnormal emotional processing

Ventral striatum

• Impaired incentive behaviour

• Psychomotor disturbances

As noted above, depression is clinically associated with negative biases in the processing of emotional information. Behavioural studies have shown that such biases affect attention to emotional stimuli as well as emotional perception, memory and appraisal (Clark et al., 2009). There is growing understanding of the neural circuitry that underpins emotional processing, in particular the way that limbic brain regions interact with regulatory cortical areas. In patients with depression, functional imaging studies have shown that presentation of negative emotional stimuli (e.g. fearful faces) is associated with overactivity of limbic regions, whereas there is a decrease in activity in the prefrontal regulatory areas. These changes are reversed by antidepressant medication, but are still apparent in recovered patients who have been withdrawn from medication. This suggests that exaggerated limbic processing of aversive material may be a trait marker of vulnerability to depression (Victor et al., 2010). Interestingly, patients with mania show exaggerated amygdala responses to positive emotional stimuli, although it is unclear whether this might be a trait abnormality (Bermpohl et al., 2009).


The predisposition to develop mania and severe depressive disorders has an important genetic contribution. The levels at which genetic factors operate are not entirely clear. They could act via effects on the regulation of neural responses to emotion, or through more remote factors such as temperament and the psychological response to stressful life events. Adverse early experiences such as parental conflict, or abuse of various kinds, may play a part in shaping features of personality which in turn determine whether, in adulthood, individuals are able to access emotional support to help to buffer the stressful effects of social adversity. In addition, early life experiences could programme the HPA axis to respond to stress in a way that might predispose to the development of mood disorder.

The precipitating causes of mood disorders are stressful life events and certain kinds of physical illness. Some progress has been made in discovering the types of event that provoke depression, and in quantifying their stressful qualities. Such studies show that loss can be an important precipitant, but not the only one. The effects of particular events may be modified by a number of background factors that may make a person more vulnerable—for example, caring for several small children without help, and being socially isolated. As noted in the preceding paragraph, the impact of potentially stressful events also depends on early life experience, personality factors, and probably genetic inheritance.

Two kinds of pathophysiological mechanism have been proposed to explain how precipitating events lead to the phenomena observed in depressive disorders. The first mechanism is psychological and the second is neurobio-logical. The two sets of mechanism are not mutually exclusive, and as the neural basis of emotional processing becomes better understood, the two approaches are converging in a way that promises a richer understanding of pathophysiology. In particular it seems that abnormalities in the neural processing of emotion can account for the emotional biases that are characteristic of depression, and can probably account for much of the subjective symptomatology. The relevant brain regions receive a strong innervation from monoamine neurons, and manipulation of 5-HT and noradrenaline neurotransmission is capable of altering emotional processing at both a behavioural level and a neural level (Harmer et al., 2009). This might account for the role of monoamine changes in the pathophysiology of depression, and the ability of monoamine manipulation to produce symptomatic relief in some depressed patients.

Using sophisticated statistical techniques it is possible to derive quantitative estimates of the roles of different risk factors in the development of depressive disorders. For example, from a prospective study of 1942 female twin pairs, Kendler et al. (2002) calculated that about 50% of the susceptibility to an episode of major depression is attributable to the following factors (in order of relative importance):

• recent stressful life events and difficulties

• adolescent risk factors (neuroticism, early-onset anxiety, and conduct disorder)

• genetic risk

• past history of major depression.

The same study found that adverse childhood experiences also played a significant role in the risk of subsequent depression, but this was expressed indirectly through an increased risk of life events and difficulties and diminished social support. Interestingly, some of the impact of genetic factors was mediated through an increased risk of early adverse experiences. This suggests that part of the genetic risk associated with depression is expressed though greater likelihood of exposure to an adverse family environment. This in turn could be linked to temperamental factors in both parents and children.

Taken together, the findings of recent studies show that major depression is a disorder with important genetic, environmental, and interpersonal determinants. These factors do not interact in a simple additive manner, but modify each other in both direct and indirect ways. Although this formulation precludes the use of simple models to explain the aetiology of major depression, it does correspond more closely with clinical experience. In addition, it suggests that a number of different kinds of intervention could be useful for decreasing the liability of individuals to develop mood disorders.

Course and prognosis

When considering course and prognosis it is necessary to deal with bipolar and unipolar disorders separately. However, it should be remembered that any sample of patients with unipolar depression will contain a proportion of patients whose bipolar illness has not yet declared itself. It has been estimated that about 10% of patients who present with a depressive disorder will eventually have a manic illness. Such patients are more likely to have a family history of mania or to show brief, mild manic mood swings during antidepressant treatment. From recent hospital and population-based studies, the following general conclusions can be derived (Angst, 2009).

Bipolar disorders

• The age of onset of bipolar disorder is typically about 21 years in hospital studies, but earlier (about 17 years) in community surveys. Late-onset bipolar disorder is rare, and may be precipitated by organic brain disease.

• Bipolar disorder usually begins as depression, with the first manic episode manifesting about 5 years later.

• The average length of a manic episode (treated or untreated) is about 6 months.

• At least 90% of patients with mania experience further episodes of major mood disturbance.

• Over a 25-year follow-up, on average bipolar patients experience about 10 further episodes of major mood disturbance.

• The interval between episodes becomes progressively shorter with both age and the number of episodes.

• Over long-term follow-up, patients with bipolar disorder experience mood-related symptomatology of varying severity for about one-third of the time. In both bipolar I and bipolar II disorder this most commonly takes the form of depressive symptoms.

• Nearly all bipolar patients recover from acute episodes, but the long-term prognosis is rather poor. For example, less than 20% of bipolar patients achieve a period of 5 years of clinical stability with good social and occupational performance. Patients with bipolar II disorder have a somewhat better outcome.

Unipolar depressive disorders (recurrent major depression)

• The age of onset of unipolar disorders varies widely, but is generally agreed to be later on average than for bipolar cases. There is a fairly uniform incidence of onset of cases throughout the lifespan, but the relative contribution of different aetiological factors probably varies between early- and late-onset cases.

• The average length of a depressive episode is about 6 months, but around 25% of patients have episodes that last for more than 1 year, and around 10–20% develop a chronic unremitting course.

• About 80% of patients with major depression will experience further episodes (i.e. have recurrent major depression).

• Over a 25-year follow-up, patients with recurrent major depression experience on average about four further episodes.

• As with bipolar patients, the interval between episodes becomes progressively shorter.

• About 50% of depressed patients do not achieve complete symptom remission between episodes, and experience continuing subsyndromal depressive symptomatology of fluctuating severity.

• The longer-term prognosis of recurrent major depression may be a little better in some ways than that of bipolar disorder, but is still modest. For example, only about 25% of patients with recurrent depression achieve a period of 5 years of clinical stability with good social and occupational performance.


Dysthymia is by definition a chronic disorder that lasts for many years. Despite this, about 50% of outpatients may be expected to show a clinical recovery over a 5-year follow-up. Over the lifespan, some patients with dysthymia develop major depression (so-called double depression), while some patients who originally present with major depression subside into dysthymia. The development of mania is rare.

Minor mood disorder

Most information is available for minor depression, which shows a recurrence rate similar to that of major depression. Patients who meet the criteria for minor depression at one point in follow-up may then subsequently be diagnosed as suffering from major depression, and vice versa. Thus, in a similar manner to dysthymia, minor depression is likely to be a risk factor for major depression, and may also be a residual state following remission of major depression. Overall, the longitudinal data suggest that major and minor depression and dysthymia are not distinct conditions, but represent part of a spectrum of depressive disorders (Angst, 2009).

Mortality of mood disorders

Mortality is significantly increased in patients with mood disorders, largely, although not exclusively, due to suicide. The standardized mortality ratio in mood disorders is about twice that found in the general population. Apart from suicide, excess deaths are due to accidents, cardiovascular disease, and comorbid substance misuse. Epidemiological studies suggest that treatment lowers the mortality in patients with mood disorders (Angst et al., 2002).

Rates of suicide in patients with mood disorders are at least 15 times higher than those in the general population, and tend to be higher in unipolar than in bipolar disorder (Angst, 2009). Longer-term follow-up of patients with depression has yielded differing rates of lifetime risk of suicide. In those with severe illnesses who have been treated as inpatients, the risk may be as high as 15%. However, in community samples the risk is lower. The proportion of patients with mood disorders who die by suicide decreases as the period of follow-up increases, presumably because mortality from natural causes becomes more significant. However, it is also possible that the risk of suicide is highest during the early stages of the illness.

Prognostic factors

The best predictor of the future course is the history of previous episodes. Not surprisingly, the risk of recurrence is much higher in individuals with a history of several previous episodes. Other factors that predict a higher risk of future episodes include the following:

• incomplete symptomatic remission

• bipolar disorder

• early age of onset

• poor social support

• poor physical health

• comorbid substance misuse

• comorbid personality disorder.

The various risk factors, particularly previous pattern of recurrence and the extent of current remission, have important implications for the use of longer-term maintenance treatments (see below). In many patients, mood disorders are best conceptualized as chronic relapsing conditions that require an integrated long-term treatment approach.

The acute treatment of depression

This section is concerned with the efficacy of various forms of treatment in the acute management of depression. Details of treatment with drugs and ECT are provided in Chapter 19, which should be consulted before reading this section. Advice on the selection of treatments and the day-to-day care of patients is given in the section on management (see p. 244).

Antidepressant drugs

Antidepressant drugs are effective in the acute treatment of major depression. The largest effects relative to placebo are seen in patients with major depression whose symptoms are of at least moderate severity. Short-term response rates in controlled trials are about 50% for patients on active treatment, and about 30% for those on placebo; the number needed to treat (NNT) is 5 (Anderson et al., 2008). In terms of efficacy there is little to choose between the various antidepressants, although some are better than others in certain defined situations.

Similar clinical response rates are seen in dysthymia, where again several classes of antidepressant drugs, including tricyclic antidepressants and SSRIs, have shown therapeutic efficacy. A meta-analysis by de Lima et al. (1999) showed a response rate in patients receiving active treatment of 55%, compared with 30% in those on placebo (NNT = 4). Antidepressants do not appear to be more effective than placebo in the treatment of minor depression (Barbui and Cipriani, 2011).

Tricyclic antidepressants

Tricyclic antidepressants have been extensively compared with placebo in both inpatients and outpatients with major depression. In all but the most severely depressed patients, tricyclic antidepressants are clearly more effective than placebo (Morris and Beck, 1974). There is little evidence that tricyclic antidepressant drugs differ from one another in clinical efficacy, but they do differ in terms of their side-effect profile (see Chapter 19). Lofepramine is relatively safe in overdose. Among the other classes of antidepressant drug, none is more effective than the tricyclics, although individual patients may show a preferential response to other compounds (see below).

Selective serotonin reuptake inhibitors (SSRIs) and serotonin and noradrenaline reuptake inhibitors (SNRIs)

Selective serotonin reuptake inhibitors (SSRIs) have undergone extensive trials both against placebo and against comparator antidepressants. There is good evidence that they are as effective as tricyclic antidepressants in the broad range of depressed patients, although they may be less effective in hospitalized depressed patients. Venlafaxine, a serotonin and noradrenaline reuptake inhibitor (SNRI), also appears to be slightly more effective than SSRIs in patients with more severe depressive states. SSRIs are more effective than tricyclic antidepressants (with the exception of clomipramine) where depression occurs in association with obsessive–compulsive disorder (Anderson et al., 2008). Duloxetine is another SNRI, but current evidence does not suggest that it is more effective than SSRIs. A network meta-analysis suggested that escitalopram and sertraline are the most effective of the SSRIs, but it has been disputed whether these differences are of clinical relevance (Gartlehner et al., 2008; Cipriani and Farukawa, 2009).

Tolerance relative to tricyclics

In short-term clinical trials, compared with tricyclic antidepressants, SSRIs are associated with lower dropout rates due to side-effects, but the differences are modest (relative risk of dropout due to side-effects = 0.73, NNT = 33). However, the differences in favour of SSRIs increase in routine clinical situations, particularly when the duration of treatment exceeds a few weeks. Clearly, their relative safety in overdose gives SSRIs an advantage in certain clinical situations; venlafaxine is less safe in overdose than SSRIs (Cheeta et al., 2004). The relative toxicity of duloxetine in overdose has not been clarified.

Monoamine oxidase inhibitors (MAOIs)

The efficacy of monoamine oxidase inhibitors (MAOIs) in the treatment of major depression (particularly with melancholic features) has been a matter of controversy. However, placebo-controlled trials have shown that MAOIs are effective antidepressants and of equal therapeutic activity to tricyclic antidepressants for moderate to severe depressive disorders.

MAOIs are liable to cause dangerous reactions with other drugs and some foods, and for this reason they are not recommended as first-line antidepressant drugs. However, controlled trials have shown that for the following conditions MAOIs carry an advantage over tri-cyclic antidepressants:

• atypical depression

• anergic bipolar depression (characterized by fatigue, retardation, increased appetite, and sleep)

• depression that is resistant to tricyclic antidepressants and SSRIs.

The reversible type-A MAOI moclobemide has the advantage of not requiring adherence to a low-tyramine diet. However, it can still produce hazardous interactions with other drugs (see p. 542). Controlled trials have shown that moclobemide is more effective than placebo in the treatment of uncomplicated major depression. However, it is doubtful whether moclobemide at standard doses is as effective as conventional MAOIs for atypical and tricyclic-resistant depression (Cowen, 2005).

Combination of MAOIs and tricyclic antidepressants

It has been reported that, in cases which are resistant to treatment with either a tricyclic or an MAOI, a combination of the two drugs is more effective than either drug given alone in corresponding dosage. This claim has not been proved; indeed, two clinical trials have failed to confirm it (Cowen, 2005). However, it can be argued that neither of those trials was concerned with patients known to be resistant to single drugs (the patient group to whom combined treatment is most often given).

Other antidepressants

A wide variety of other antidepressant drugs are now available (see Chapter 19 and Table 10.10), all of which have established efficacy relative to placebo. The main differences between the various preparations are in side-effect profile. Apart from reboxetine, anticholinergic-type side-effects are uncommon. Trazodone and mirtazapine are sedating. All of these agents are safer than tricyclic antidepressants in overdose. Agomelatine may have a different mechanism of action, involving the activation of melatonin receptors (Stahl, 2007).


Lithium as a sole treatment

This section is concerned only with lithium as a treatment for depressive disorders. The use of lithium in the treatment of mania and in the maintenance treatment of mood disorders is considered later. Placebo-controlled trials suggest that lithium may have some antidepressant efficacy in bipolar depression, but its effects in unipolar depression as a sole treatment are not established (Goodwin, 2009).

Table 10.10 Clinical characteristics of some antidepressant drugs


Lithium in combination with antidepressants

Despite its limited utility as a sole drug treatment for depression, lithium can produce useful therapeutic effects when added to antidepressant medication in treatment-resistant patients (lithium augmentation). In a meta-analysis, Crossley and Bauer (2007) found that about 40% of depressed patients responded to lithium augmentation of their antidepressant regimen, compared with about 15% of patients who were given placebo (NNT = 4). However, in the pragmatic STAR*D study the effects of lithium augmentation were disappointing in terms of low response and high dropout rates (Nierenberg et al., 2006).

Although some studies have reported a rapid amelioration of the depressed state within as little as 48 hours after the addition of lithium, the more usual pattern of response is a gradual resolution of symptoms over 2 to 3 weeks. Unipolar depressed patients seem to respond as well as bipolar patients, and so far there are no reliable clinical or biochemical predictors that identify depressed patients who are likely to respond to lithium augmentation. The effects of lithium augmentation do not appear to be restricted to any specific class of antidepressants.


Anticonvulsants such as carbamazepine, valproate, and lamotrigine are useful in the management of bipolar disorder, and in these circumstances can prevent episodes of major depression. Whether these agents also have acute antidepressant efficacy in unipolar depression is unclear, but lamotrigine has been shown to have antidepressant effects in placebo-controlled trials in bipolar depressed patients, particularly those with higher levels of symptomatology (Geddes et al., 2009). It is possible that valproate may have a similar effect (Smith et al., 2010).

Atypical antipsychotic drugs

Antipsychotic drugs are often combined with antidepressant drugs in the treatment of patients with depressive psychosis, and there is also evidence that atypical antipsychotic agents can be of benefit when combined with antidepressants in non-psychotically depressed patients who have failed to respond to antidepressant treatment alone (see below). Quetiapine also appears to possess anti-depressant activity when given as monotherapy in patients with unipolar and bipolar depression, and has a useful role in the latter condition (see below, and Ravindran et al., 2010).

Electroconvulsive therapy (ECT)

This treatment is described in Chapter 19, where its unwanted effects are also considered. The present section is concerned with evidence about the therapeutic effects of ECT for patients with depressive disorders.

Comparison with simulated ECT

Six double-blind controlled trials have compared the efficacy of ECT and simulated ECT (anaesthesia with electrode application but no passage of current) in patients with major depression. Five of these studies found ECT to be more effective than the simulation. In the study which did not find the full procedure to be more effective, unilateral low-dose ECT was used, a procedure which is considered on other grounds to be relatively ineffective. The overall response rate is about 70% for ECT and 40% for simulated treatment (NNT = 3–4) (UK ECT Review Group, 2003; see also Chapter 19).

Comparison with other treatments

Several studies have compared depressed inpatients receiving ECT with those receiving antidepressant drugs. In a total of nine comparisons with tricyclic antidepressants, ECT was therapeutically more effective in six studies and equally effective in the remaining three. In five comparisons with MAOIs, ECT was superior in each trial and worked more quickly. These data suggest that in severely depressed inpatients, ECT is probably superior to antidepressant drug treatment, at least in the short term (UK ECT Review Group, 2003; see also Box 21.12).

Indications for ECT

Clinicians generally agree that the therapeutic effects of ECT are greatest in severe depressive disorders, especially those in which there is marked weight loss, early-morning waking, retardation, and delusions. From the trials comparing full ECT with simulated ECT, it appears that delusions and (to a lesser extent) retardation are the features that distinguish patients who respond to full ECT from those who respond to placebo (UK ECT Review Group, 2003).

Other studies have established that patients with depressive psychosis respond better to ECT than to tricyclic antidepressants or antipsychotic drugs given alone. However, combined treatment with antidepressants and antipsychotic drugs may be about as effective as ECT, although no direct comparisons have been made (Cowen, 2005). Another point of practical importance is that ECT may often prove effective in depressed patients who have not responded to full trials of medication, whether or not psychotic features are present. However, in such patients relapse rates are high (Kellner et al., 2006).

Psychological treatment

All depressed patients, whatever other treatment they may be receiving, require psychotherapy in a general sense, which provides education, reassurance, and encouragement. These measures (‘clinical management’) can provide some symptomatic relief, and can also increase the likelihood that pessimistic patients will adhere to specific treatments. Education and reassurance should also be given to the patient’s partner, other close family members, and other people involved in their care.

The psychological treatments used for depressive disorders can be divided into the following categories:

• supportive psychotherapy

• cognitive–behaviour therapy

• interpersonal psychotherapy

• marital therapy

• dynamic psychotherapy.

These psychotherapies can be employed as alternatives to antidepressant medication, or as adjuncts. Psychotherapies have been less well evaluated than antidepressant medication in major depression, but the most extensive evidence base is for cognitive–behaviour therapy. In general, psychotherapies are somewhat more effective than treatment as usual in the management of mild to moderate depression, particularly in primary care. In this setting, however, structured treatments such as cognitive–behaviour therapy do not appear to be superior to other therapies such as interpersonal therapy. This suggests that factors common to all psychological treatments are likely to be important in mediating the therapeutic effect (see p. 574). In randomized trials, psychotherapies usually perform as well as drug treatment in moderately depressed patients (National Institute for Health and Clinical Excellence, 2009a).

Supportive psychotherapy and problem solving

Supportive psychotherapy goes beyond clinical management in focusing on the identification and resolution of current life difficulties, and in using the patient’s strengths and available coping resources. A development of this approach is problem solving, in which the therapist and the patient identify the main problems of concern and devise feasible step-by-step ways of tackling them. Randomized trials suggest that problem-solving treatment is better than treatment as usual in moderately depressed patients in primary care. Its efficacy relative to drug treatment and other psychotherapies is unclear (National Institute for Health and Clinical Excellence, 2009a).

Cognitive–behaviour therapy

For depressive disorder, the essential aim of cognitive–behaviour therapy is to help the patient to modify their ways of thinking and acting in relation to life situations and depressive symptoms (for further information, see p. 585). There have been numerous studies of cognitive–behaviour therapy in acute major depression, which have been reviewed recently (National Institute for Health and Clinical Excellence, 2009a). Current conclusions can be summarized as follows:

• There is strong evidence that cognitive–behaviour therapy is superior to a waiting-list control in relieving depressive symptomatology.

• Cognitive–behaviour therapy is not generally superior to other structured psychological treatments, such as behavioural activation and interpersonal therapy.

• Cognitive–behaviour therapy is as effective as pharmacological treatment in moderately depressed outpatients.

• Combined cognitive–behaviour therapy and pharmacological treatment is better than pharmacological treatment alone.

The opinion of many clinicians is that cognitive–behaviour therapy is not effective as a sole treatment for patients with severe depression, but this view does not have clear support from trial evidence (Driessen et al., 2010). Issues such as what is meant by ‘severity’ in different settings, as well as the problem of generalizing from randomized trials to everyday practice, are likely to be implicated in this controversy. In addition, therapist expertise may be a critical factor in the delivery of effective cognitive–behaviour therapy (DeRubeis et al., 2005).

Interpersonal psychotherapy

Interpersonal therapy is a systematic and standardized treatment approach to personal relationships and life problems (see p. 576). It has been less studied than cognitive–behaviour therapy in depression, but seems to be as effective. The National Institute for Health and Clinical Excellence (2009a) concluded that in depressed patients, interpersonal therapy:

• is more effective than placebo with clinical management and GP treatment as usual

• is as effective as antidepressant medication

• is more effective when combined with antidepressants than when given alone. However, it is not clear whether the combination of interpersonal therapy and anti-depressants is better than antidepressants alone for the treatment of acute depression.

Couple therapy

Couple therapy can be given to depressed patients for whom interactions with a partner appear to have contributed to causing or maintaining the depressive disorder. The aim of the intervention is to understand the nature of these interactions and modify them so that the relationship becomes more mutually supportive. There are few randomized trials available, but the evidence suggests that couple therapy is significantly more effective than a waiting-list control and as effective as cognitive–behaviour therapy. There are insufficient data to indicate how couple therapy compares with antidepressant medication (National Institute for Health and Clinical Excellence, 2009a). In practice, antidepressant treatment and couple therapy are often used together, but the combination has not been well evaluated.

Dynamic psychotherapy

Dynamic psychotherapy has a different aim from the treatments described so far, which is to resolve underlying personal conflicts and attendant life difficulties that are believed to be causing or maintaining the depressive disorder. Opinions differ about its value. There have been few attempts to evaluate treatment, and they have mainly been concerned with brief dynamic psychotherapy in groups. The results for the treatment of depression suggest that dynamic psychotherapy may be less effective than cognitive–behaviour therapy or antidepressant medication (National Institute for Health and Clinical Excellence, 2009a). However, Taylor (2008) has argued that consideration of a wider range of evidence suggests that psychodynamic therapies can be useful in depression, particularly in the longer term.

Other treatments

Sleep deprivation

Several studies suggest that, in some depressive disorders, rapid short-term changes in mood can be brought about by keeping patients awake overnight. The alleviation of depressed mood after total sleep deprivation is nearly always temporary; it disappears after the next night’s sleep or even during a daytime nap after the night of sleep deprivation. Although the antidepressant effect of sleep deprivation is of great theoretical interest, its brevity makes it unpractical. However, there are reports that sleep deprivation can be used to hasten the onset of effect of antidepressant drugs, and also that some pharmacological manipulations can prolong the effect of sleep deprivation (Hemmeter et al., 2010).

Bright light treatment

Over 50% of patients with recurrent winter depression respond to bright light treatment (about 10 000 lux). Treatment is usually given for an hour or two in the morning, but the timing of light treatment is not always critical, and evening light or even midday exposure can be effective. The duration of exposure usually needs to be 1–2 hours (see p. 564).

Designing placebo-controlled trials of bright light for winter depression presents problems, because most patients are aware before treatment that bright light is believed to be the important therapeutic ingredient. Within this limitation, most studies have found that dim light is less effective than bright light. The usual onset of the antidepressant effect of bright light is within 2 to 5 days, but longer periods of treatment appear to be needed in some patients. Patients with ‘atypical’ depressive features such as overeating and oversleeping appear to respond best. To avoid relapse, light treatment usually needs to be maintained until the usual time of natural remission, in the early spring (see Eagles, 2004 and p. 564).

Some studies have shown that bright light treatment may also be effective in non-seasonal depression—for example, in elderly people with depression, where circa-dian rhythm disturbances may be involved in pathophysiology (Lieverse et al., 2011). However, the durability of such an effect has not been established.

The acute treatment of mania

The treatment of mania presents a formidable clinical challenge which often tests the management skills of the psychiatric team. Drug treatment plays a pivotal role in the management of mania, and has the aim of reducing physical and mental overactivity, improving features of psychosis, and preventing deterioration in health due to exhaustion, sleep deprivation, and poor fluid intake. It is worth noting that before the advent of modern drug treatment the mortality of mania in the hospital setting was over 20%; nearly 50% of these patients died from exhaustion (Derby, 1933).


Antipsychotic drugs

Typical antipsychotic drugs

Several randomized controlled trials have shown the efficacy of chlorpromazine and haloperidol in treating mania, whether or not patients have clear psychotic features. However, the use of typical antipsychotic drugs has limitations. Manic patients often receive high doses, and may be particularly susceptible to extrapyramidal side-effects. In addition, conventional antipsychotic drugs do not protect against the depressive downswings that can follow resolution of a manic illness.

Atypical antipsychotic drugs

Because of their improved tolerability profile, atypical antipsychotic agents are being increasingly used in the treatment of mania. Placebo-controlled trials support the efficacy of aripiprazole, olanzapine, quetiapine, and risperidone in patients with mania (Smith et al., 2007). However, it should be noted that more severely ill patients may not be included in placebo-controlled trials, which could limit the generalizability of these findings.

Mood stabilizers


Five placebo-controlled trials have shown that lithium is effective in the acute treatment of mania (Smith et al., 2007). In the only study which compared lithium with both placebo and an active comparator, Bowden et al. (1994) found a response rate of 49% for lithium, 48% for valproate, and 25% for placebo. In general, lithium is as effective as antipsychotic medication, but its onset of action is slower. Moreover, in highly active states, antipsychotic medication may be preferable (Goodwin, 2009). Prominent depressive symptoms and psychotic features predict a poorer response to lithium alone, as does a rapid cycling disorder.


Assessment of the efficacy of carbamazepine in acute mania has been limited by problems with regard to study design. These include small numbers, the use of adjunctive medications, and mixed diagnostic groupings. However, more recent placebo-controlled trials of an extended-release form of carbamazepine found a clinically significantly anti-manic effect (Weisler et al., 2006). Carbamazepine induces drug-metabolizing enzymes in the liver, which can lead to lower plasma levels of concomitantly administered drugs.


Several randomized controlled trials have shown that valproate possesses anti-manic activity that is greater than that of placebo and equivalent to that of lithium (Smith et al., 2007). Valproate may be more effective than lithium in the subgroup of patients with prominent dysphoric symptoms and rapid cycling.

One advantage of valproate is that it has an earlier onset of activity than other mood stabilizers. For example, with ‘valproate loading’ (20 mg/kg/day) the onset of the anti-manic effect can occur as early as within 1 to 4 days. This is probably because the tolerability of valproate allows rapid dose escalation, whereas lithium and carbamazepine generally need to be introduced more gradually (Goodwin, 2009).


Benzodiazepines are useful adjuncts in the treatment of mania, because they can rapidly diminish overactivity and restore sleep. They have been used as sole therapy in the treatment of mania, but this carries a risk of disinhibition. Their most useful role is as an adjunct to mood stabilizers, because the latter drugs can take several days to become effective. Benzodiazepines can also be given in combination with antipsychotic drugs, because this lowers the doses of antipsychotic drug that are required to calm agitated and overactive patients. The benzodiazepines that are usually employed are the high-potency agents, lorazepam and clonazepam. Use should be on an ‘as needed’ basis and for as short a time as possible, in order to minimize the risk of tolerance and dependence (Goodwin, 2009).

Electroconvulsive therapy

ECT has been widely used to treat mania, although there have been only two prospective controlled trials in the modern era. In one of these trials, bilateral ECT was found to be superior to lithium (Small et al., 1988). In the other study, unilateral and bilateral ECT were compared with a combination of lithium and haloperidol in patients who had not responded to antipsychotic drugs alone. The response rate of these patients to ECT (13 of 22) was greater than that to combined drug treatment (none of five). The efficacy of unilateral and bilateral ECT did not differ (Mukherjee et al., 1988).

Retrospective investigations have shown ECT to be effective in acute mania, with the overall response rate being about 80%. Many of these patients had been unresponsive to medication. ECT is also effective in mixed affective states (Valenti et al., 2008). It is often given at shorter intervals in the treatment of mania than in the treatment of depression, but there is no evidence that this regimen is necessary or that it speeds up the treatment response. It is unclear whether bilateral electrode placement is better than unilateral placement in manic patients.

The longer-term treatment of mood disorders

Follow-up studies have shown that mood disorders often recur and that, if left untreated, they have a rather poor long-term prognosis. For this reason there is now increasing emphasis on long-term management.

Prevention of relapse and recurrence

Strictly speaking, the term relapse refers to the worsening of symptoms after an initial improvement during the treatment of a single episode of mood disorder, whereas recurrence refers to a new episode after a period of complete recovery. Treatment to prevent relapse should be called continuation treatment, and treatment to prevent recurrence should be called prophylactic or maintenance treatment. In practice, however, it is not always easy to maintain the distinction between these two kinds of treatment, because a therapy may be given initially to prevent relapse, and may then be used to prevent recurrence.

Drug treatment of unipolar depression

Continuation treatment

It is now well established that stopping antidepressants soon after a treatment response has been obtained is associated with a high risk of relapse. About one-third of patients who are withdrawn from medication will relapse during the next year, with the majority of the relapses occurring in the first 6 months. Placebo-controlled studies of the role of continuation therapy have reached the following conclusions (Anderson et al., 2008):

• Continuing antidepressant treatment for 6 months past the point of remission halves the relapse rate.

• Treatment should be at the originally effective dose of medication if possible.

• In patients who are at low risk of further episodes, continuation of antidepressant treatment for longer than 6 months confers little extra benefit except in the elderly, where continuation therapy for 12 months is more appropriate.

Maintenance treatment

Controlled studies involving patients with recurrent depression (usually defined as at least three episodes over the last 5 years) have shown that maintenance antidepressant treatment can substantially reduce relapse rates. For example, in a 3-year study of 128 patients, Frank et al. (1990) found a relapse rate of 22% in patients taking imipramine, compared with 78% in patients treated with placebo. The effects of longer-term maintenance treatment were confirmed in a systematic review (Geddes et al., 2003), where over a period of 1 to 2 years of continued antidepressant treatment the relapse rate was lowered from 41% on placebo to 18% on active medication (see Figure 10.2). During longer-term treatment, it is common practice to lower the dose of medication, especially that of tricyclic antidepressants, to a ‘maintenance’ level, but this may decrease the prophylactic efficacy unless it is carried out carefully, with close monitoring. The general rule appears to be ‘What gets you well, keeps you well’ (Anderson et al., 2008).

Lithium carbonate has also been used in the prevention of recurrent unipolar depression, but while some patients show a clear response, the overall evidence for its efficacy is less robust than for the prevention of bipolar disorder (Anderson et al., 2008). However, lithium may be a useful treatment for depressed patients who have shown any sign of elevated mood, because (unlike antidepressant drugs) it also prevents mania (see below).

Drug treatment of bipolar disorder

Continuation treatment

It is a common clinical experience that too rapid a reduction in doses of drug treatment can lead to the sudden recrudescence of an apparently treated manic disorder. Since the average length of a manic episode is about 6 months, it seems prudent to continue some form of medication for at least this period. There are several studies involving atypical antipsychotic drugs which show that, compared with placebo, continuation treatment decreases the risk of manic relapse (Goodwin, 2009). Patients who have been severely ill may well be taking both a mood stabilizer and an antipsychotic agent. Because of the adverse effects of antipsychotic drugs, it will often be appropriate to slowly withdraw this form of treatment first.

Maintenance treatment

Lithium. There is substantial evidence for the efficacy of lithium in the maintenance treatment of recurrent mood disturbances in patients with bipolar disorders. In a systematic review, Geddes et al. (2004) found that lithium was significantly more effective than placebo in preventing relapses of all mood disorder (mean relative risk, 0.65; 95% CI, 0.5–0.84). However, although lithium was clearly effective in preventing manic relapse, its effect in preventing depressive episodes was more equivocal (mean relative risk, 0.72; 95% CI, 0.49–1.07).


Figure 10.2 Meta-analysis of rates of depressive relapse in patients randomly allocated to receive continuation treatment for on average about 1 year with antidepressant medication or placebo. Reprinted from The Lancet, 361 (9358), John R Geddes, Stuart M Carney, Christina Davies, Toshiaki A Furukawa, David J Kupfer, Ellen Frank, and Guy M Goodwin, Relapse prevention with antidepressant drug treatment in depressive disorders: a systematic review, pp. 653-661, Copyright (2003), with permission from Elsevier.

In general, about 50% of bipolar patients who are treated with lithium respond well. The following are predictors of a relatively poorer response to lithium maintenance treatment:

• rapid-cycling disorders or chronic depression

• mixed affective states

• alcohol and drug misuse

• mood-incongruent psychotic features.

There is also evidence that use of lithium in patients with recurrent mood disorders is associated with a significant reduction in mortality from suicide. This effect is most often seen in dedicated lithium clinics, and the mechanism and specificity of the association require further study. Nevertheless, it is clinically important that the carefully supervised use of lithium is associated with a lowered risk of suicidal behaviour. This effect is not necessarily shared by other mood stabilizers (Cipriani et al., 2005b).

Carbamazepine. Although the number of patients studied in randomized controlled studies is relatively few, carbamazepine appears to have efficacy in the prophylaxis of bipolar disorder. However, it appears to be less effective than lithium in ‘classical’ bipolar illness (Goodwin, 2009). Therefore carbamazepine can be considered in patients who respond poorly to lithium, particularly those with rapid-cycling disorders. It can be given alone or in combination with lithium.

Valproate. This is increasingly used in the treatment of acute mania, and therefore evidence of its longer-term maintenance effects is needed. Bowden et al. (2003) randomized 372 patients to 1 year of maintenance treatment with valproate, lithium, and placebo. There were no significant differences between the three treatment groups on the primary outcome measure of time to recurrence of any mood disorder. However, the number of patients who left the study because of the occurrence of a mood episode was significantly less among those taking valproate compared with those on placebo. Patients who were taking valproate experienced more tremor, alopecia, and weight gain than those on placebo. In a recent pragmatic study in bipolar patients, the relapse rate on lithium (59%) was significantly less than that with valproate (69%) over 24 months. The combination of lithium and valproate was associated with the lowest relapse rate (54%) (BALANCE Study Group, 2010).

Lamotrigine. Two prospective placebo-controlled trials have shown that lamotrigine has prophylactic effects in patients with bipolar illness. Although it may have some modest benefit in the prevention of mania, lamotrigine has a clearer prophylactic effect against depression. Its profile of activity therefore contrasts with those of lithium and valproate, which are more effective in preventing mania (Goodwin, 2009). However, lamotrigine is not licensed for the treatment of mood disorder in the UK.

Antipsychotic drugs. Patients with bipolar disorder have sometimes been maintained on typical antipsychotic drugs, usually given in addition to mood-stabilizing agents. It is usually wise to minimize this form of treatment where possible, because conventional antipsychotic drugs do not protect against depression, and may be more liable to cause tardive dyskinesia in bipolar patients.

Atypical antipsychotic drugs are also being employed in the longer-term maintenance treatment of bipolar disorder, either as a sole treatment or as an adjunct to mood stabilizers. There is evidence from randomized studies for the efficacy of olanzapine and quetiapine in this respect in the prevention of both depression and mania, whereas aripiprazole is effective in the prevention of mania but not depression (Goodwin, 2009). However, a common design of these studies is to continue antipsychotic treatment in patients who have responded to short-term, open-label treatment, and to compare this with placebo substitution. This sample, enriched for antipsychotic treatment responders, may favour active treatment unduly, and therefore the findings may not be generally applicable to other clinical situations (Cipriani et al., 2010).


Cognitive therapy

Unipolar depression

As noted above, there is some evidence that cognitive therapy given during an acute phase of depression leads to a more sustained improvement in depressive symptomatology and lessens the risk of subsequent relapse (Hollon et al., 2005; National Institute for Health and Clinical Excellence, 2009a). There is also growing interest in the use of continuation and maintenance treatment with cognitive therapy, particularly in patients who have residual depressive symptomatology and are therefore at increased risk of relapse. For example, Paykel et al. (1999) studied 158 patients who experienced significant residual symptoms after treatment of an episode of major depression. All of the patients received clinical management and continuation treatment with antidepressant medication, and half of them also received 16 sessions of cognitive therapy. Over the next 16 months the relapse rate in the patients who received cognitive therapy was 29%, compared with 47% in the group who received clinical management only. The benefit of cognitive–behaviour therapy persisted, with diminishing effect, for a further 2 years (Paykel et al., 2005).

Mindfulness-based cognitive therapy (MBCT) integrates cognitive–behaviour therapy with meditation techniques designed to lower stress by facilitating acceptance and self-compassion. There is evidence that MBCT lowers the risk of relapse in patients with recurrent depression, although further trials controlling for the non-specific effects of this complex treatment are needed (Chiesa and Serretti, 2011).

Bipolar disorder

Psychotherapy has been less studied in bipolar patients, but cognitive techniques may be valuable in helping patients to accept their illness and the need for medical treatment. Some randomized studies have shown that cognitive techniques combined with education about early signs of relapse reduced the rates of relapse in bipolar patients (Lam et al., 2005). However, a large pragmatic trial showed no benefit using this approach (Scott et al., 2006). There is better evidence for group psychoeducation, where a 6-month course of treatment was of benefit in preventing in relapse over a 5-year follow-up period (Colom et al., 2009). The best results were shown in people with fewer episodes of illness, which suggests that the intervention should be applied as early as possible in the course of the illness.

Interpersonal therapy

In patients with recurrent depressive episodes, interpersonal therapy given once a month delayed, but did not prevent, depressive relapse compared with standard clinical management. In the same study, interpersonal therapy was less effective than imipramine maintenance treatment, and the combination of psychotherapy and imipramine was no better than imipramine alone (Frank et al., 1990). The effect of combined interpersonal therapy and medication has also been studied in older depressed patients. A benefit of the combined treatment over medication alone was seen with nortriptyline but not with paroxetine (Reynolds et al., 2006). These findings suggest that in terms of recurrence prevention, interpersonal therapy is less effective than continued medication, while the benefit of combined treatment over medication alone is equivocal.

The assessment of depressive disorders

The aims of the various steps of assessment are as follows:

• to decide whether the diagnosis is depressive disorder

• to judge the severity of the disorder, including the risk of suicide

• to form an opinion about the causes

• to assess the patient’s social resources

• to gauge the effect of the disorder on other people.

Diagnosis depends on thorough history taking and examination of the physical and mental state, and has been discussed earlier in this chapter. Particular care should be taken not to overlook a depressive disorder in patients who do not complain spontaneously of being depressed (‘masked depression’). It is equally important not to diagnose a depressive disorder simply on the grounds of prominent depressive symptoms, as the latter could be part of another disorder, such as a general medical condition. It should also be remembered that certain drugs—both legal and illegal—can induce depression (see p. 384 and Chapter 17).

The history of previous mood disturbance is important in assessment. Some patients will have had recurrent episodes of mood disorder. A history of these episodes often provides clues to the probable course of the current disorder and its response to treatment. It is particularly important to ask about possible previous episodes of mania and hypomania, even if these were mild and short-lived. If there is a history of mania, the mood disorder is bipolar. Interviews with relatives and close friends may help to establish whether such episodes have occurred.

The severity of the disorder is judged from the symptoms. Considerable severity is indicated by ‘biological’ symptoms, hallucinations, and delusions, particularly the latter two. It is also important to assess how the depressive disorder has reduced the patient’s capacity to work or to engage in family life and social activities. In this assessment, the duration and course of the condition should be taken into account as well as the severity of the present symptoms. Not only does the length of the history affect the prognosis, but also it gives an indication of the patient’s capacity to tolerate further distress. A long-continued disorder, even if not severe, can bring the patient to the point of desperation. The risk of suicide must be judged in every case (the methods of assessment are described on p. 427).

Aetiology is assessed next, with reference to precipitating, predisposing, and maintaining factors. No attempt need be made to allocate the syndrome to an exclusively ‘endogenous’ or ‘reactive’ category. Instead, the importance of all relevant risk factors should be evaluated in every case.

Provoking causes may be psychological and social (the ‘life events’ discussed earlier in this chapter), or they may be physical illness and its treatment. In assessing such cases, it is good practice to enquire routinely about the patient’s work, finances, family life, social activities, general living conditions, and physical health. Problems in these areas may be recent and acute, or may take the form of chronic background difficulties such as prolonged marital conflict, problems with children, and financial hardship.

The patient’s social resources are considered next. Enquiries should cover family, friends, and work. A loving family can help to support the patient through a period of depressive disorder by providing company, encouraging them when they have lost confidence, and guiding them into suitable activities. For some patients, work is a valuable social resource, providing distraction and comradeship. For others it is a source of stress. A careful assessment is needed in each case.

The effects of the disorder on other people must be considered carefully. The most obvious problems arise when a severely depressed patient is the mother of young children who depend on her. This clinical observation has been confirmed in objective studies which have demonstrated that depressive disorder in either parent is associated with emotional disorder in the children (Ramchandani et al., 2005). It is important to consider whether the patient could endanger other people by remaining at work (e.g. as a bus driver). If there are depressive delusions, it is necessary to consider what would happen if the patient were to act on them. For example, severely depressed parents may occasionally kill their children because they believe them to be doomed to suffer if they remain alive.

The management of depressive disorders

This section starts by considering the management of patients with a depressive disorder of moderate or greater severity. The first question concerns the level of care and supervision that may be required. The answer depends on the severity of the disorder and the quality of the patient’s social resources. When judging severity, particular attention should be paid to the risk of suicide (or any risk to the life or welfare of family members, particularly dependent children) and to any failure to eat or drink that might endanger the patient’s life. Provided that these risks are absent, most patients with a supportive family can be treated at home, even if they are severely depressed. Patients who live alone, or whose families cannot care for them during the day, may need intensive community treatment or day-patient care. Inpatient treatment can be considered for patients who are not improving with these measures, or where safety issues raise particular concern. Involvement of the family wherever possible is also likely to improve the outcome (see Box 10.7).

The next question is whether the patient should continue to work. If the disorder is of moderate severity, work can provide a valuable distraction from depressive thoughts, and be a source of companionship. If the disorder is more severe, retardation, poor concentration, and lack of drive are likely to impair performance at work, and such failure may add to the patient’s feelings of hopelessness. In severe disorders, there may be dangers to other people if the patient remains at their job.

The need for antidepressant drug treatment should be considered next. This treatment is indicated for most patients with a major depressive syndrome of at least moderate severity, and particularly those with melancholic symptoms. Other indications include a family history or personal history of depression, particularly if there has been a clear response to drug treatment. Dysthymia is also an indication for antidepressant medication, but the role of antidepressants in milder depressive conditions is not established. However, where a patient with a clear history of major depression develops symptoms of mild depression, antidepressant medication can be considered. Guidelines for a stepped-care approach to the management of depression have been developed by the National Institute for Health and Clinical Excellence (2009a) (see Box 10.8).

Choice and use of antidepressant drug

Several kinds of antidepressant drug treatment are available, and the choice should be made in collaboration with the patient, with particular consideration of the likely side-effects. These are fully described in Chapter 19.

Box 10.7 What patients and families want to know

Particularly for patients with a first episode of moderate to severe depression, questions such as the following are important:

1. What is wrong with me?

2. Can I recover?

3. What treatment do I need?

4. What can I do to improve the situation?

5. Can my family be helped?

The answers to these questions require a good knowledge of the patient and their circumstances. Patients will understand and appreciate that a careful practitioner will need to gather information before giving definitive answers to important questions. However, some general advice is given below.

• Depression is regarded as a medical condition and not any fault of the individual concerned. Feelings of guilt are common and are seen as part of the condition.

• The prognosis for individual episodes of depression is good, although people with depression are naturally pessimistic about the future.

• Treatment consists of a comprehensive package with psychological support and medication if indicated. The treatment plan will be developed as a collaborative exercise between the patient and the clinical team, and the patient’s rights and choices will be respected.

• Part of the treatment plan will involve a full discussion about the most appropriate level of occupational and social activity for the current circumstances. This will be reviewed frequently to allow adjustment to the patient’s changing clinical state.

• It is very useful for the family to be fully involved in discussions about the nature of depression and the treatment plan, as their support can be invaluable. Conversely, critical comments can be demoralizing for depressed people. One of the aims of family involvement is to help to improve mutual communication and support.

Box 10.8 Guidelines for the management of depression

1. Patients with short-lived mild depression who may recover quickly without treatment should be offered an early review (‘active monitoring’).

2. Antidepressants are not recommended for the treatment of mild depression.

3. Patients with persistent mild depression should be recommended a guided self-help programme based on cognitive–behaviour therapy. Group cognitive–behaviour therapy is an alternative. An exercise programme and sleep hygiene can also be recommended.

4. For patients with persistent mild depressive symptoms that do not respond to these measures, consider drug treatment with an SSRI or a higher-intensity psychological treatment (cognitive–behaviour therapy, behavioural activation, interpersonal therapy, or couple therapy).

5. Patients who present with moderate or severe depression should be treated with a combination of antidepressant medication and a high-intensity psychological intervention.

6. Patients who respond to antidepressant medication should continue treatment for at least 6 months. Patients at high risk of relapse should be advised to continue antidepressant treatment for 2 years.

7. Consider cognitive–behaviour therapy for patients who have relapsed despite antidepressant treatment, or mindfulness-based cognitive therapy for patients who are well but who have experienced three or more previous episodes of depression.

From the National Institute for Health and Clinical Excellence (2009a).

• SSRIs are the usual first choice in moderately depressed patients, unless the patient considers that they have done well previously with a different agent. Lofepramine, a tricyclic antidepressant that is relatively safe in overdose, is another non-sedating compound with a different side-effect profile (see Table 10.8).

• Mirtazapine may be considered if the patient needs concomitant sedation.

• Tricyclic antidepressants such as amitriptyline may be the first choice for patients with severe depression, and particularly for depressed inpatients. For severely ill patients who are unable to tolerate tricyclic antidepressants, or where there are medical contraindications, venlafaxine is an alternative, although it is somewhat less well tolerated than SSRIs. Both amitriptyline and venlafaxine are less safe in overdose than the other compounds listed here.

The dosage of these drugs, the precautions to be observed when using them, and the instructions to be given to patients are described in Chapter 19. Here it is necessary only to emphasize again the importance of explaining to the patient that, although side-effects will appear quickly, the therapeutic effect is likely to be delayed for 2 weeks, with maximum improvement occurring over 6–12 weeks. During this time the patient should be seen regularly to provide suitable clinical management (see above); those with more severe disorders may need to be seen every 2 or 3 days, and other patients once a week. It is important to make sure that the drugs are taken in the prescribed dose. The patient should be warned about the effects of taking alcohol. They should be advised about driving, particularly that they should not drive while experiencing sedative side-effects or any other effects that might impair their performance in an emergency.

The use of ECT

ECT will very rarely be part of the first-line treatment of depression, and will usually be considered only for patients who have already been admitted to hospital. The only indication for ECT as a first measure is the need to bring about improvement as rapidly as possible. In practice this applies to two main groups of patients:

• those who refuse to drink enough fluid to maintain an adequate output of urine (including the rare cases of depressive stupor)

• those who present a highly dangerous suicidal risk.

Occasionally, ECT is indicated for a patient who is suffering such extreme distress that the most rapid form of treatment is deemed justifiable. Such cases are rare. It should be remembered that, with the exception of patients who are unresponsive to antidepressant drugs, the effects of ECT differ from those of antidepressant drugs in terms of greater rapidity of action, rather than the final therapeutic result. In patients with depressive psychosis, ECT is considerably more effective than an antidepressant given alone, but probably about the same therapeutic effect can be achieved, albeit more slowly, if a combination of an antidepressant drug and an antipsychotic drug is used.


The need for suitable activity should be considered for every patient. Depressed patients give up activities and withdraw from other people. As a result they become deprived of social stimulation and rewarding experiences, and their original feelings of depression are increased. It is important to make sure that the patient is occupied adequately, although they should not be pushed into activities in which they are likely to fail because of slowness or poor concentration. Thus there is a fairly narrow range of activity that is appropriate for the individual depressed patient, and the range changes as the illness runs its course. If the patient remains at home, it is important to discuss with the relatives how much they should be encouraged to do each day. For inpatients, this question should be decided by the clinical team.


The appropriate kind of psychological treatment should also be decided in each individual case. As noted earlier, all depressed patients require support, encouragement, and a thorough explanation that they are suffering from illness and not moral failure. Similar counselling of partners and other family members is often useful.

The use of one of the more specific psychological treatments discussed earlier should also be considered. These treatments can be used as the sole therapy for patients with mild to moderate depression without melancholic features, particularly if the patient prefers not to take drug therapy (National Institute for Health and Clinical Excellence, 2009a). The kind of psychological treatment that is used depends largely on the availability of a suitably trained therapist and the preference of the patient, although the more structured therapies (such as interpersonal therapy and cognitive–behaviour therapy) have a greater evidence base with regard to the treatment of moderate to severe depression. The therapeutic response to antidepressant drugs is usually faster than that to psychotherapy.

Current recommendations are that in patients with moderate to severe depression it is helpful to add cognitive–behaviour therapy to antidepressant medication (National Institute for Health and Clinical Excellence, 2009a). In practical terms, some time is usually needed to arrange cognitive–behaviour therapy, and in any event some initial improvement produced by antidepressant medication may enable the patient to make more use of psychological treatment. It is also important to tackle any psychosocial problems that are making an important contribution to the depression. In particular, couple therapy can be a helpful addition in depressed patients where problems with a partner are playing a role in maintaining the disorder.

If the depressive disorder is severe, too much discussion of problems at an early stage is likely to increase the patient’s feelings of hopelessness. Therapy directed toward self-examination is particularly likely to make the disorder worse. During intervals between acute episodes, such therapy may be given to patients who have recurrent depressive disorders that are largely caused by their ways of reacting to life events.

Failure to respond to initial treatment

If a depressive disorder does not respond within a reasonable time to a chosen combination of antidepressant drugs, graded activity, and psychological treatment, the plan should be reviewed. The first step is to check again that the patient has been taking medication as prescribed. If not, the reasons for this should be sought. The patient may be convinced that no treatment can help, or may find the side-effects unpleasant. The diagnosis should also be reviewed carefully, and a check made that important stressful life events or continuing difficulties have not been overlooked.

If this enquiry is unrevealing, antidepressants should be continued at an increased dose if possible. In general, SSRIs do not have clear dose–response relationships, although some patients respond to higher doses, particularly if a partial response has been observed at a standard dose. Venlafaxine is probably more effective at higher doses in patients with treatment-resistant depression (Cowen, 2005). If it becomes clear that the patient is not improving, a number of further steps can be taken (see Box 10.9). There is not a fixed order in which treatments should be offered. Prescribing decisions should be made in collaboration with the individual patient, taking into account factors such as specific symptomatology, how far the condition may have shown a partial response, and the side-effect profile of the various treatment options.

Box 10.9 Some pharmacological treatments for resistant depression

• Increase antidepressant to the maximum tolerated dose; if the patient has depressive psychosis, add an antipsychotic drug; try a different class of anti-depressant drug, including venlafaxine and tricyclic antidepressants.

• Try an antidepressant combination (e.g. an SSRI or venlafaxine with mirtazapine).

• Add lithium to antidepressant drug treatment.

• Add an atypical antipsychotic drug to an SSRI or venlafaxine.

• Add tri-iodothyronine to tricyclic antidepressant treatment.

• MAOIs (can be usefully combined with lithium).

• ECT.

Change in antidepressant drug treatment

If a patient does not respond to one antidepressant, the first step is usually to stop the first medication and try another. Most published studies of this approach have studied patients in an open sequential way; clearly this cannot control for the placebo effect or the possibility of spontaneous remission. Overall, however, there is reasonable evidence that switching to a second antidepressant can produce benefit in about 50% of patients who are unresponsive to an initial medication trial (Anderson et al., 2008).

If a patient has not responded to one kind of anti-depressant, it would seem sensible to switch to an anti-depressant that has a different pharmacological profile. However, it must be acknowledged that open studies have shown equally good response rates when patients who failed to respond to one SSRI were switched to another. A meta-analysis of randomized studies suggested that switching from an SSRI to a different class of drug (bupropion, mirtazapine, or venlafaxine) was marginally better in terms of remission rate (SSRI, 23.5% vs. non-SSRI, 28%; NNT = 22) than switching to a second SSRI, but this advantage is of doubtful clinical significance (Papakostas et al., 2008). Switching from a drug with serotonergic properties to another serotonergic compound should be carried out cautiously because of the risk of serotonin toxicity. In practice this means that the first compound should be fully withdrawn if at all possible, and the second one started at a half-dose or less. In the case of fluoxetine, because of its long duration of action, at least 1 week should elapse before starting a second serotonergic agent. However, when switching between agents with different pharmacological properties (e.g. from citalopram to mirtazapine or reboxetine), cross-tapering can be employed. Detailed instructions are provided by the Maudsley Prescribing Guidelines (Taylor et al., 2009).

Both amitriptyline and venlafaxine appear to be slightly more effective than SSRIs; these drugs are therefore worth trying at some point in the management of patients who are unresponsive to initial medication trials (Anderson et al., 2008). Similar comments apply to the use of conventional MAOIs, which have a clearly beneficial effect in some patients with resistant depression (Nolen et al., 1988). However, the drug and food restrictions associated with MAOIs mean that it is unusual to use them early in drug-resistant depression unless a patient has responded well to them in the past.

Augmentation of antidepressant drug treatment

When switching antidepressant preparations, one problem is that withdrawal of the first compound may not be straightforward. Patients may have gained some small benefit from the treatment (e.g. improved sleep or reduced tension), and this benefit may be lost. Also, if the first medication is stopped quickly, withdrawal symptoms may result (see Chapter 19). However, if the dose is reduced gradually, the changeover in medication may be protracted and may not be easily tolerated by a depressed and despairing patient.

For this reason, in patients who are unresponsive or partly responsive to first-line medication it may be more appropriate to add a second compound to the antidepressant—so-called augmentation therapy. A disadvantage of augmentation therapy is the increased risk of adverse effects due to drug interaction.

Lithium augmentation

We have already seen that data from randomized trials indicate that lithium can be effective in the treatment of resistant depression. Provided that there are no contraindications, the addition of lithium to antidepressant drug treatment is usually safe and well tolerated. About 50% of depressed patients will show a useful response over 1 to 3 weeks. Lithium can be added to any primary antidepressant medication with good effect, although combination with SSRIs and venlafaxine should be undertaken with caution because of the risk of 5-HT neurotoxicity (see p. 236). In the latter case it is appropriate to start lithium at a low dose (200 mg daily) and increase by a maximum of 200 mg weekly. In augmentation treatment, the aim should be to obtain a lithium concentration within the range used for prophylactic purposes (0.5–0.8 mmol/l); the lower end of the range is appropriate for combination with serotonergic antidepressants. It has been suggested that the combination of lithium with MAOIs or clomipramine (sometimes supplemented with tryptophan) can be effective in depressed patients who are otherwise refractory to treatment (Cowen, 2005).

Antipsychotic drugs

As noted above, in patients with psychotic depression it is usually best to prescribe a combination treatment of antidepressant and antipsychotic medication (Anderson et al., 2008; Wijkstra et al., 2010). In non-psychotic resistant depression, conventional antipsychotic drugs are of little value except, at low doses, for ameliorating agitation and distress. However, there is good evidence that some atypical antipsychotic drugs may have anti-depressant effects when used in combination with SSRIs in patients with non-psychotic depression. In a meta-analysis of over 3000 patients, Nelson and Papakostas (2009) found that, relative to placebo, the addition of atypical antipsychotic drugs (aripiprazole, olanzapine, quetiapine, and risperidone) was significantly more likely to result in clinical remission (odds ratio, 2.0; 95% CI, 1.69–2.37; NNT = 9).

The use of atypical antipsychotics to augment SSRIs employs lower doses than would be used to treat schizophrenia, perhaps because the key pharmacological mechanism involves 5-HT2 receptor rather than dopamine D2receptor blockade (see p. 523). Despite this, olanzapine and quetiapine, even at low doses, can cause troublesome sedation and weight gain, while concomitant use of risperidone also produces some degree of weight gain together with hyperprolactinaemia. Aripiprazole is less likely to cause metabolic side-effects, but is associated with agitation and restlessness. In randomized studies of SSRI augmentation, the discontinuation rate due to adverse effects was significantly greater with atypical antipsychotics than with placebo (odds ratio, 3.9; 95% CI, 2.68–5.72; NNH = 17) (Nelson and Papakostas, 2009). This suggests that although atypical antipsychotic drug addition can be useful in SSRI-resistant depression, the clinical utility of this approach may be limited by its tolerability.

Tri-iodothyronine (T3)

Some open and controlled studies have suggested that the addition of tri-iodothyronine (T3) in doses of 20–40 µg daily to ineffective tricyclic antidepressant treatment can bring about a useful clinical response. However, a meta-analysis of four published randomized trials which assessed the efficacy of T3 addition to ineffective tricyclic antidepressant treatment was of borderline significance (Aronson et al., 1996). The evidence for the efficacy of T3 augmentation of newer antidepressant drugs, such as SSRIs, is inconsistent (Cooper-Kazaz and Lerer, 2008). Despite this, some depressed patients do appear to be helped by the addition of T3, which can start at a dose of 10 µg daily and be increased to 20 µg after 1 week, if tolerance is good. At this dose the side-effects are mild, but tremor and sweating can occur. It is prudent not to use T3 addition in patients with evidence of cardiovascular disease.

Combination treatment with antidepressants

Combination strategies aim to supplement the antidepressant effect of an ineffective or partially effective medication with another antidepressant agent. This approach can therefore be considered an augmentation strategy, although if the patient’s condition remits it may be unclear whether the response is due to the combined effect of the two antidepressants or to the second agent acting alone.

The pharmacological rationale of combination treatment is the use of two agents to produce a broader spectrum of activity on monoamine pathways than either agent could produce alone. In practice this means that SSRIs or venlafaxine are usually combined with nora-drenergic-promoting agents such as tricyclic antidepressants, reboxetine, or mirtazapine (which increases noradrenaline function through blockade of auto-inhibitory α2-adrenoceptors).

The evidence for any of these strategies is limited, although they are endorsed in case series and expert reviews. The best evidence is probably for the addition of mirtazapine or its predecessor, mianserin, to ineffective SSRI treatment, although a large randomized trial in patients with resistant depression failed to find a benefit of combined sertraline and mianserin over mianserin alone (Anderson et al., 2008). Combination of a tricyclic antidepressant with SSRIs must be approached with caution because of the risk of elevation of tricyclic levels (see p. 537).

The use of ECT in resistant depression

If severe depression persists, ECT should be considered. ECT is often beneficial to patients who have not responded to antidepressant drugs, and it is probably more effective than drugs in the most severe depressions, particularly when psychotic features are present. Whether medication resistance lowers the response to ECT has been debated. However, a meta-analysis of observational studies suggested that the overall response rate to ECT in patients who failed pharmacotherapy (48%) was significantly lower than the rate in those who did not (65%) (Heijnen et al., 2010).

It is important to note that, among patients who have not responded to full-dose antidepressant medication, the relapse rate in the year after ECT may be as high as 50%. This may be because patients are often continued on the same antidepressant medication to which they previously failed to respond. In these circumstances it seems reasonable to use an antidepressant of a different class, or lithium, for continuation treatment after ECT, but there is currently no unequivocal evidence that this procedure lowers the relapse rate. However, a randomized study of 200 patients found post-ECT prophylaxis with a combination of lithium and nortriptyline to be as effective as maintenance ECT in sustaining remission in the 6 months following a successful course of treatment. Despite this, just over 50% of the patients in each group relapsed (Kellner et al., 2006).

Other important measures

Continuing support is essential for patients who do not respond to treatment. Lack of improvement increases the pessimism experienced by the depressed patient. Therefore it is important to give reassurance that depressive disorders have a good chance of recovering eventually, whether or not treatment speeds recovery. Meanwhile, if the patient is not too depressed, any problems that have contributed to their depressed state should be discussed further. Techniques derived from cognitive therapy can be used to help the patient to limit distorted thinking and judgement. In resistant cases, it is particularly important to watch carefully for suicidal intentions. For patients with severe symptoms and long-standing refractory illness, other treatment approaches such as deep brain stimulation and neurosurgery can be considered. These are discussed in Chapter 19.

Prevention of relapse and recurrence in depression

Drug treatment

Continuation therapy

After recovery, the patient should be followed up for several months by the psychiatric team or general practitioner. If the recovery appears to have been brought about by an antidepressant drug, that drug should usually be continued for about 6 months and then gradually withdrawn over a period of several weeks. If residual symptoms are still present, it is safer not to withdraw medication. At follow-up interviews, a careful watch should be kept for signs of discontinuation reactions or relapse. It is helpful to discuss with the patient the possible early signs of relapse, and to develop a plan of action should any of these signs appear. If the patient is in agreement, it is helpful to involve the relatives in this plan.

Maintenance treatment

Major depression is often recurrent, and long-term maintenance treatment may need to be considered. The risk factors for recurrence have been noted above (see p. 234). In addition, the clinician will need to take into account the following individual factors:

• the likely impact of a recurrence on the patient’s life

• the previous response to drug treatment

• the patient’s view of long-term drug treatment.

It is estimated that, among patients who have had three episodes of major depression, the likelihood of another episode is 90%. The usual recommendation is that maintenance drug treatment should be considered if a patient has had two previous episodes of depression within a 5-year period, particularly if there is a family history of recurrent major depression, or personal and social factors predictive of recurrence (Anderson et al., 2008). Persistent residual symptoms are also an indication for maintenance treatment.

Choice of medication

For most patients, the choice of antidepressant will be derived from their response in the acute or continuation phase of treatment. In this case, the same medication can be continued, if possible at the same dose. As noted earlier, newer antidepressant drugs are better tolerated in the longer term, and may be preferred for maintenance treatment. If a change needs to be made because of adverse effects (e.g. sexual dysfunction with SSRIs), an alternative choice should be determined on the basis of side-effect profile.

Lithium can also be effective for long-term maintenance treatment of recurrent depression in some patients. Its adverse-effect profile and the need for plasma monitoring mean that it will not be a first choice for most patients, but its use should be considered if there is any history of manic episodes, even if these have been mild and transient. In addition, lithium will sometimes be effective in patients with recurrent depression who do not respond to maintenance treatment with antidepressants.

Other measures

If a depressive disorder was related to self-imposed stressors, such as overwork or complicated social relationships, the patient should be encouraged to change to a lifestyle that is less likely to lead to further illness. These readjustments may be helped by psychotherapy, which may be individual, marital, or group therapy. As noted above, cognitive–behaviour therapy appears to be helpful in lowering the risk of relapse, particularly in patients with residual depressive symptomatology. There is also growing interest in the role of mindfulness-based cognitive therapy. Community nurses and nurse therapists can play a valuable role in delivering treatment of this kind. General practitioners can play a key role in the long-term monitoring of patients, and should always be involved in treatment planning.

The assessment of mania

In the assessment of mania, the steps are the same as those for depressive disorders, as outlined above, and can be summarized as follows.

• Decide the diagnosis.

• Assess the severity of the disorder.

• Form an opinion about the causes.

• Assess the patient’s social resources.

• Judge the effects on other people.

Diagnosis depends on a careful history and examination. Whenever possible, the history should be taken from relatives as well as from the patient, because the latter may not recognize the extent of their abnormal behaviour. Differential diagnosis has been discussed earlier in this chapter. It is always important to remember that mildly disinhibited behaviour can result from frontal lobe lesions (e.g. tumours) as well as from mania. There should be a urine screen for illegal substances.

Severity is judged next. For this purpose it is essential to interview an informant as well as the patient. Manic patients may exert some self-control during an interview with a doctor, and may then behave in a disinhibited and grandiose way immediately afterwards. At an early stage of mania, the doctor can easily be misled and may lose the opportunity to persuade the patient to enter hospital before causing him- or herself long-term difficulties (e.g. due to ill-judged decisions or unjustified extravagance). Where possible, it is important to identify any life events that may have provoked the onset of manic illness. Some manic episodes follow physical illness, treatment with drugs (especially steroids or antidepressants), or operations. Sleep deprivation may trigger mania in some susceptible individuals.

The patient’s resources and the effect of the illness on other people are assessed in the ways described above for depressive disorders. Even the most supportive family will find it extremely difficult to care for a manic patient at home for more than a few days unless the disorder is exceptionally mild. The patient’s responsibilities with regard to the care of dependent children, or at work, should always be considered carefully.

The management of manic patients

Acute episodes

The first decision is whether to admit the patient to hospital. In all but the mildest cases, admission is nearly always advisable, in order to protect the patient from the consequences of their own behaviour. If the disorder is not too severe, the patient will usually agree to enter hospital after some persuasion. If the disorder is more severe, compulsory admission is likely to be needed.

Clinical management

Development of a therapeutic alliance with the manic patient is an important goal of treatment, although lack of insight into the condition and anger at involuntary detention can make this a testing and time-consuming process. It is best to use an understanding, yet firm, approach that minimizes confrontation. For example, it is often possible to avoid an argument by taking advantage of the manic patient’s easy distractibility; instead of refusing demands, it is better to delay until the patient’s attention switches to another topic that they can be encouraged to pursue. In hospital, nursing staff play a key role in this kind of management, where they attempt to provide manic patients with a low-stimulus and safe environment that limits confrontations and ill-judged sexual liaisons.

Supportive, reality-orientated psychotherapy is an important element of treatment, and may need to be extended to the patient’s partner and family, who often will have experienced great strain during the manic episode. Educational sessions are important when the patient is more settled. Patients may need much practical help in order to limit the financial, legal, and occupational repercussions of the illness.


Medication plays an important role in the acute management of mania. However, all current drug treatments have limitations in terms of their efficacy and tolerability. In the UK, antipsychotic drugs are generally used as primary agents in the management of acute mania, with mood stabilizers being reserved for longer-term prophylaxis or where the initial response to antipsychotic drugs is unsatisfactory. For patients who are not already taking mood stabilizers, the drug management can be summarized as follows:

• Begin treatment with an atypical antipsychotic drug (a modest dose of haloperidol is an alternative).

• Where necessary, use adjunctive high-potency benzodiazepines (e.g. lorazepam or clonazepam) to avoid the need for high-dose antipsychotic therapy, and to ensure satisfactory sleep.

• Add a mood stabilizer if the response to this treatment is unsatisfactory (use lithium for ‘classical’ mania, and valproate for mania with prominent dysphoric or mixed states or where lithium is contraindicated).

If it is decided to use a mood stabilizer as the initial treatment, the choice lies between lithium and valproate. Valproate is somewhat easier to use and has a faster onset of action. As before, benzodiazepines can be used as an adjunct to reduce overactivity and permit sleep, while antipsychotic agents can be reserved for patients who are unresponsive to these measures. If a patient presents with a manic episode and is already taking a mood stabilizer, the first step is to optimize the mood-stabilizer treatment in terms of dosing and adherence. Concomitant treatment with antipsychotic drugs will then probably be needed; addition of a second mood-stabilizing drug is an alternative (Goodwin, 2009).

Progress of a manic episode can be judged not only by the mental state and general behaviour, but also by the pattern of sleep and the regaining of any weight lost during the illness. As progress continues, antipsychotic drug treatment is reduced gradually. It is important not to discontinue the drug too soon, otherwise there may be a relapse and a return of all the original problems of management.

ECT was used to treat mania before antipsychotic drugs were introduced, but evidence about its effectiveness is still limited (see above). It is appropriate to consider the treatment for the unusual patients who do not respond to drug treatment even when this is given at maximum doses. In such cases, clinical experience suggests that a relatively short course of ECT may often be followed by a significant reduction in symptoms sufficient to allow treatment to be continued with drugs. ECT may also be helpful for mixed affective states in which depressive symptoms are prominent.

Whatever treatment is adopted, a careful watch should be kept for symptoms of depressive disorder. It should be remembered that transient but profound depressive mood change, accompanied by depressive ideas, is common among manic patients (see p. 212). The clinical picture may also change rapidly to a sustained depressive disorder. If either change happens, the patient may develop suicidal ideas. A sustained change to a depressive syndrome is likely to require additional treatment unless the disorder is mild.

Treatment of bipolar depression

Depressive episodes are common in bipolar illness and are associated with the majority of the illness burden in both bipolar 1 and bipolar II disorder (Angst, 2009). However, treatment of depression in the context of a bipolar disorder can be problematic because standard antidepressant treatments have a number of disadvantages, which include:

• an apparently lower efficacy than in unipolar depression

• a risk of inducing mania

• a risk of inducing rapid cycling.

Lithium can be a useful antidepressant in bipolar patients and does not cause mania. However, it is ineffective in a significant proportion of cases, and many bipolar patients become depressed despite taking lithium. In these circumstances, the use of antidepressants is common, but the evidence for their effectiveness relative to placebo is equivocal (Sidor and MacQueen, 2011). The treatment of depression in bipolar I disorder with antidepressant medication should generally include a mood stabilizer because this probably reduces the risk of manic upswing. Whether the same should apply to bipolar II disorder is not yet clear (Amsterdam and Shults, 2010). Overall, the evidence suggests that in bipolar patients the risk of switching from depression to mania is less with SSRIs than with tricyclic antidepressants, and probably also less than with venlafaxine (Goodwin, 2009). If a patient with bipolar depression appears to respond to antidepressant medication, consideration should be given to discontinuing the antidepressant after a few weeks’ remission, because of the equivocal evidence for long-term benefit of continued treatment (Ghaemi et al., 2010).

There is growing evidence for the short- and long-term efficacy of quetiapine in the treatment of bipolar depression, and its effects may be greater than those of SSRIs and lithium (McElroy et al., 2010; Young et al., 2010). Whether this efficacy extends to other atypical antipsychotic drugs is uncertain, and there is speculation that the efficacy of quetiapine in these circumstances is due to the formation of norquetiapine, an active metabolite of quetiapine with noradrenaline-reuptake-blocking properties (Goodwin, 2009). There is no evidence that quetia-pine used as a monotherapy causes mania or mixed states. Lamotrigine may also have a role in the acute management of bipolar depression (Geddes et al., 2009), and is probably free of the risk of inducing mania or rapid cycling.

For a review of the treatment of bipolar depression, see Baldessarini et al. (2010).

Prevention of relapse and recurrence

Continuation treatment

Since the duration of an untreated manic episode can be several months (see above), some form of continuation treatment, as in the management of depression, is advisable for at least 6 months. As a guide, treatment should not be withdrawn finally until the patient has been asymptomatic for at least 8 weeks. Withdrawal of lithium should be particularly cautious (e.g. by not more than about 100 mg a week), because sudden discontination can lead to ‘rebound’ mania (Goodwin, 2009).

Maintenance treatment


We have already seen that the majority of patients with bipolar disorder will experience recurrence, suffering both depressive and manic episodes, although the pattern in individual patients is variable. For patients who have had two or more episodes of illness in less than 5 years, particularly where the illnesses have proved personally disruptive or hazardous, longer-term maintenance treatment should be considered. This will usually involve long-term treatment with a mood stabilizer.

People with bipolar disorder are often reluctant to consider long-term treatment with medication. There are usually a number of reasons for this, and it necessary to spend time in discussion with patients so that their point of view can be fully appreciated and understood. An excellent personal account has been provided by Jamison (1997). The reasons for reluctance to consider such treatment may include:

• difficulties in accepting a diagnosis of a lifelong condition and the need for maintenance treatment with medicines; issues of stigma are usually important

• a belief that it should be possible to control mood without medication, and that the use of medication for this purpose is a sign of personal weakness

• a fear that medication will blunt the patient’s emotional life, and sometimes specifically that mood-stabilizing drugs will decrease feelings of joy and creativity. If the latter feelings are associated with hypomania, it is indeed possible that they will be lessened by medication. For some individuals this can represent a significant loss, which needs to be balanced against the prevention of disabling episodes of depression.

In the UK, lithium is still regarded as the first choice of mood stabilizer, although in the USA valproate is the most popular choice because of its better tolerability. However, the evidence for its long-term prophylactic efficacy is less complete. In addition, there is evidence that the supervised use of lithium is associated with a decreased risk of suicidal behaviour in bipolar illness.

The practical aspects of using lithium to prevent further episodes are discussed on p. 552. There are two aspects of maintenance that require emphasis. First, the patient should be seen and their plasma lithium levels and renal and thyroid function assessed at regular intervals. Secondly, some patients stop lithium of their own accord because they fear that the drug may have harmful long-term effects, or because it makes them feel ‘flat.’ The risk associated with stopping lithium suddenly is that it may well lead to relapse. Therefore patients should be advised to discontinue lithium slowly under supervision.

With carefully supervised follow-up, the likelihood of relapse can be substantially reduced by maintenance treatment, although lesser degrees of mood change often continue. These mood changes may require adjunctive antidepressant or antipsychotic drug treatment. For patients who do not show a good response to lithium, alternative mood-stabilizing drugs can be used. It is possible to use combinations of mood stabilizers, although the risk of adverse reactions is increased. When using combinations of mood stabilizers it is helpful to bear in mind that lamotrigine appears to be more effective in preventing depression than in preventing mania, while the reverse is true for lithium and valproate. Quetiapine appears to be effective in preventing both depression and mania, although its longer-term use is associated with weight gain and metabolic abnormalities (Vieta et al., 2008, and see p. 523). Current guidelines stress the importance of regular clinical and biochemical monitoring of the physical health of bipolar patients, as well as monitoring of their psychological status (Goodwin, 2009).

Rapid-cycling disorder. Some patients with bipolar disorder develop rapid-cycling mood disorder (see p. 212). Sometimes this disorder is apparently provoked by anti-depressant treatment, in which case the best plan is to try to withdraw the antidepressant drug. Rapid cycling often requires treatment with combinations of mood stabilizers, and atypical antipsychotic drugs can also be helpful. In particularly resistant cases, the use of high-dose thyroid hormone has been advocated. For a review, see Datta and Cleare (2009).

Psychosocial approaches

Bipolar disorder is often a lifelong, disruptive illness which carries a substantial morbidity. There are several areas of psychosocial function where problems may arise:

• adjustment to the diagnosis, and the need for lifestyle limitations

Box 10.10 Psychological approaches to bipolar disorder

• Advice about lifestyle (regular social and sleep routines, avoidance of illegal drugs).

• Identification and avoidance of triggers for relapse (e.g. sleep deprivation, substance misuse).

• Identification of early subjective signs of relapse (e.g. feeling driven, sleeping badly), with contingency plans for action.

• Education about the importance of medication, and discussion of sensitivity to side-effects and active measures to reduce these.

• interpersonal and relationship difficulties, and occupational problems

• misuse of both illegal and legal substances

• problems relating to concordance with medication.

Psychological treatment approaches designed to help with these important problems employ a number of theoretical approaches. However, they have a number of features in common which centre around education about the illness and enhancing self-management (Goodwin, 2009) (see Box 10.10).

Although there are some useful drug therapies for bipolar illness, the effectiveness of treatment in clinical practice is often disappointing. The use of the psychosocial measures outlined above should enhance the overall effectiveness of treatment plans, and randomized trials of certain approaches, such as psychoeducational programmes, are promising in this respect (Goodwin, 2009).

Further reading

Ghaemi SN (2008). Mood Disorders. Lippincott, Williams and Wilkins, Philadelphia, PA. (A well-written practical guide to the clinical assessment and management of mood disorders.)

Kraepelin E (1921). Manic-Depressive Insanity and Paranoia (translated by RM Barclay). Churchill Livingstone, Edinburgh [reprinted in 1976 by Arno Press, New York]. pp. 1–164. (The classical account. Of particular interest is the description of untreated mania, which is seldom seen today.)

Jamison KR (1997). An Unquiet Mind. Picador, London. (A superb autobiographical account of the experience of bipolar illness and its treatment.)

Lewis AJ (1934). Melancholia: a clinical survey of depressive states. Journal of Mental Science80, 277–378. [Reprinted in Lewis AJ (1967). Inquiries in Psychiatry. Routledge and Kegan Paul, London, pp. 30–117]. (This landmark study contains a detailed account of the clinical picture of depressive disorder.)