Shorter Oxford Textbook of Psychiatry, 6th Ed.

CHAPTER 11. Schizophrenia


Clinical features

Diagnosis and classification




Social and psychosocial factors

Psychological factors

Current aetiological hypotheses about schizophrenia

Course and prognosis



Discussing schizophrenia with patients and their carers


Of all the major psychiatric syndromes, schizophrenia is the most difficult to define and describe. This partly reflects the fact that, over the past 100 years, widely divergent concepts have been held in different countries and by different psychiatrists. Although there is now a greater consensus, substantial uncertainties remain. Indeed, schizophrenia remains the best example of the fundamental issues with which psychiatry continues to grapple—concepts of disease, classification, and aetiology. Having noted the complexities, we start with an introduction to acute schizophrenia and chronic schizophrenia. The reader should bear in mind that these will be idealized descriptions and comparisons, but it is useful to oversimplify at first before introducing the controversial issues.

Table 11.1 Common symptoms of acute schizophrenia


The predominant clinical features in acute schizophrenia are delusions, hallucinations, and interference with the flow of thoughts (see Table 11.1). These features are often called positive symptoms, and include specific symptoms known as first-rank symptoms, which are given particular weight in the diagnosis. Subtypes of schizophrenia are recognized based upon the relative prominence of the symptoms. Many patients recover from the acute illness, but progression to the chronic syndrome is also common. Its main features are very different, consisting of apathy, lack of drive, slowness, and social withdrawal (see Table 11.2). These features are often called negative symptoms. Once the chronic syndrome is established, few patients recover completely. In both phases of the disorder there may also be cognitive and affective symptoms and other features, as described later. For a review, see Arango and Carpenter (2011).

Table 11.2 Features of chronic schizophrenia


Clinical features

In this section, it is assumed that the reader has read the descriptions of symptoms and signs in Chapters 1 and 3, which include definitions of many of the cardinal features of schizophrenia.

The acute syndrome

The vignette in Box 11.1 illustrates several common features of acute schizophrenia, including prominent persecutory delusions, with accompanying hallucinations, gradual social withdrawal and impaired performance at work, and the odd idea that other people can read one’s thoughts.

In appearance and behaviour some patients with acute schizophrenia are entirely normal. Others seem changed, although not always in a way that would immediately point to psychosis. They may be preoccupied with their health, their appearance, religion, or other intense interests. Social withdrawal may occur—for example, spending a long time in their room, perhaps lying immobile on the bed. Some patients smile or laugh without obvious reason. Some appear to be constantly perplexed, while others are restless and noisy, or show sudden and unexpected variability of behaviour.

The speech often reflects an underlying thought disorder. In the early stages, there is vagueness in the patient’s talk that makes it difficult to grasp their meaning. Some patients have difficulty in dealing with abstract ideas. Other patients become preoccupied with vague pseudo-scientific or mystical ideas. When the thought disorder is more severe, two characteristic kinds of abnormality may occur. Disorders of the form (or stream) of thought include pressure of thought, poverty of thought, thought blocking, and thought withdrawal; some of these constitute first-rank symptoms (see Table 11.3). Thought disorder is reflected in the loosening of association between expressed ideas, and may be detected in illogical thinking (e.g. ‘knight’s move’ thinking) or talking past the point (vorbeireden). In its severest form, the structure and coherence of thinking are lost, so that utterances are jumbled (word salad or verbigeration). Some patients use ordinary words or phrases in unusual ways (metonyms or paraphrases), and a few coin new words (neologisms).

Box 11.1 A vignette of acute schizophrenia

A previously healthy 20-year-old student had been behaving in an odd way. At times he appeared angry and told his friends that he was being persecuted; at other times he was seen to be laughing to himself for no apparent reason. For several months he had seemed increasingly preoccupied with his own thoughts. His academic work had deteriorated. When interviewed, he was restless, suspicious, and exhibited odd mannerisms. He described hearing voices commenting on his actions and abusing him. He believed that the police had conspired with his university teachers to harm his brain and interfere with his thoughts. He also suspected that they could read his thoughts.

Auditory hallucinations are among the most frequent symptoms. They may take the form of noises, music, single words, brief phrases, or whole conversations. They may be unobtrusive, or so severe as to cause great distress. Some voices seem to give commands to the patient. Some patients hear their own thoughts apparently spoken out loud either as they think them (Gedankenlautwerden) or immediately afterwards (echo de la pensée), and some voices discuss the patient in the third person or comment on his actions; these are first-rank symptoms (see Table 11.3). Visual hallucinations are less frequent, and usually occur together with other kinds of hallucination. Tactile, olfactory, gustatory, and somatic hallucinations are reported by some patients. They are often interpreted in a delusional way—for example, hallucinatory sensations in the lower abdomen are attributed to unwanted sexual interference by a persecutor.

Table 11.3 Schneider’s symptoms of the first rank

Hearing thoughts spoken aloud

Third-person hallucinations

Hallucinations in the form of a commentary

Somatic hallucinations

Thought withdrawal or insertion

Thought broadcasting

Delusional perception

Feelings or actions experienced as made or influenced by external agents

Delusions are characteristic. Primary delusions are infrequent and are difficult to identify with certainty. Delusions may originate against a background of so-called primary delusional mood (Wahnstimmung). Persecutory delusions are common, but are not specific to schizophrenia, as they also characterize delusional disorders (see Chapter 12). Less common, but of greater diagnostic value, are delusions of reference and of control, and delusions about the possession of thought. The latter are delusions that thoughts are being inserted into or withdrawn from one’s mind, or ‘broadcast’ to other people.

Insight is almost always impaired. Most patients do not accept that their experiences result from illness, but usually ascribe them to the malevolent actions of other people.

Orientation is usually normal, although this may be difficult to determine if there is florid thought disorder or if the patient is too preoccupied with their psychotic experience to attend to the interviewer’s questions.

Alterations in mood are common and are of three main kinds. First, there may be symptoms of anxiety, depression, irritability, or euphoria. These can be clinically significant, but if such features are sufficiently prominent and sustained, the possibility of schizoaffective disorder or other affective psychosis should be considered. Secondly, there may be blunting (or flattening) of affect—that is, sustained emotional indifference or diminution of emotional response. Thirdly, there may be incongruity of affect, in which the expressed mood is not in keeping with the situation or with the patient’s own feelings.

Finally, we emphasize the variability of the clinical picture. Few patients experience all of the symptoms introduced above (as illustrated by the percentage figures shown in Table 11.1), while others already have features of the ‘chronic’ syndrome at first presentation. Moreover, the overall pattern and duration of features are also taken into account before making a diagnosis. These issues, together with some additional clinical features, are discussed later in the chapter.

The chronic syndrome

Although the positive symptoms of the acute syndrome may persist, the chronic syndrome is characterized by the negative symptoms of underactivity, lack of drive, social withdrawal, and emotional apathy. The vignette in Box 11.2illustrates several of the negative features of what is sometimes called a schizophrenic ‘defect state.’ The most striking feature is diminished volition—that is, a lack of drive and initiative. Left to himself, the patient may remain inactive for long periods, or may engage in aimless and repeated activity. He withdraws from social encounters, and his social behaviour may deteriorate in ways that embarrass other people. Self-care may be poor, and the style of dress and presentation may be careful but somewhat inappropriate. Some patients collect and hoard objects, so that their surroundings become cluttered and dirty. Others break social conventions by talking intimately to strangers or shouting obscenities in public.

Box 11.2 A vignette of chronic schizophrenia

A middle-aged man lives in a supported hostel and attends a sheltered workshop. He spends most of his time alone. He is usually dishevelled and unshaven, and cares for himself only when encouraged to do so by others. His social behaviour seems odd and stilted. His speech is slow, and its content is vague and incoherent. He shows few signs of emotion. For several years this clinical picture has changed little except for brief periods of acute symptoms which are usually related to upsets in the ordered life of the hostel.

Speech is often abnormal, showing evidence of thought disorder of the kinds found in the acute syndrome described above. Affect is generally blunted, but when emotion is shown, it is often incongruous or shallow. Hallucinations and delusions occur, but are by no means universal. They tend to be held with little emotional response. For example, the patient may be convinced that they are being persecuted but show neither fear nor anger.

Various disorders of movement occur, including stereotypies, mannerisms and other catatonic symptoms, and dyskinesias (see below). The latter are common, and are primarily but not entirely due to antipsychotic medication.

As with acute schizophrenia, the symptoms and signs of the chronic illness are variable. At any stage, positive symptoms may recur or become exacerbated; this may be in response to life events, or to discontinuation of medication.

Subtypes of schizophrenia

Schizophrenia is conventionally divided into several subtypes, based upon the predominant clinical features.

• Paranoid schizophrenia is the commonest form. It is characterized by persecutory delusions, often systematized, and by persecutory auditory hallucinations. Thought disorder and affective, catatonic, and negative symptoms are not prominent. Personality is relatively well preserved. It has a later age of onset and a better prognosis than the other subtypes.

• In hebephrenic schizophrenia, also called disorganized schizophrenia, thought disorder and affective symptoms are prominent. The mood is variable, with behaviour often appearing silly and unpredictable. Delusions and hallucinations are fleeting and not systematized. Mannerisms are common. Speech is rambling and incoherent, reflecting the thought disorder. Negative symptoms occur early, and contribute to a poor prognosis.

• In catatonic schizophrenia, the most striking features are motor symptoms, as noted on p. 16, and changes in activity that vary between excitement and stupor. At times the person may appear to be in a dream-like (oneiroid) state. Formerly common, catatonic schizophrenia is now very rare, at least in industrialized countries. Possible reasons for this include a change in the nature of the illness, improvements in treatment, or past misdiagnosis of organic syndromes with catatonic symptoms. It has also been argued that catatonia is a distinct syndrome (Fink et al., 2010).

• Simple schizophrenia is characterized by the insidious development of odd behaviour, social withdrawal, and declining performance at work. Delusions and hallucinations are not evident. Clear schizophrenic symptoms are absent, and there is always the possibility of an earlier acute episode having gone undetected. Given the limited utility of the category and its history of abuse—for example, in the detention of political dissidents in the former Soviet Union (‘sluggish schizophrenia’)—its use should be avoided.

• Undifferentiated schizophrenia is the term used for cases which do not fit readily into any of the above subtypes, or where there are equally prominent features of more than one of them.

Two other categories have been introduced more recently. Residual schizophrenia refers to a stage of chronic schizophrenia when, for at least a year, there have been persistent negative symptoms but no recurrence of positive symptoms. Deficit syndrome describes a subtype of schizophrenia with early, severe, and persistent negative symptoms, and with some other differences in symptomatology and antecedents (Kirkpatrick et al., 2001).

These various subtypes are primarily descriptive, rather than delineating valid sub-syndromes. Other sub-classifications of schizophrenia have been proposed, intended to reflect biologically more valid entities. Two examples are given here.

Type I and type II schizophrenia

Crow (1985) described two syndromes of schizophrenia, based upon a combination of clinical and neurobiological factors. Type I has an acute onset, mainly positive symptoms, and preserved social functioning during remissions. It has a good response to antipsychotic drugs, associated with dopamine overactivity. By contrast, type II has an insidious onset, mainly negative symptoms, and poor outcome and response to antipsychotic drugs, without evidence of dopamine overactivity but with structural brain changes (especially ventricular enlargement). Subsequent research has not strongly supported the biological subtypes and correlations which are predicted by the model. However, it was important as an example of the renewed focus on the neurobiological aspects of schizophrenia which occurred around that time.

Three clinical sub-syndromes

Liddle (1987) studied patients with chronic schizophrenia and, based on clustering of symptoms, described three overlapping clinical syndromes, which he called reality disturbance, disorganization, and psychomotor poverty (see Table 11.4). These findings have largely been confirmed by later studies. Notably, Liddle then linked these symptom clusters to distinct patterns of neuropsycho-logical deficit and to regional cerebral blood flow (Liddle et al., 1992). The most reproducible finding is the link between psychomotor poverty, impaired performance on frontal lobe tasks, and decreased frontal blood flow.

Other aspects of the clinical syndrome

Cognitive features

Despite Kraepelin’s original term dementia praecox, cognitive impairment was for many years a neglected component of schizophrenia, even though it is substantial and significant (Dickinson and Harvey, 2009). Indeed, it has been argued that neuropsychological abnormality is an invariable feature (Wilk et al., 2005). Impairments are seen across all domains of learning and memory, with disproportionate involvement of semantic memory, working memory, and attention. The deficits average 1 to 2 standard deviations below expected performance, and are present at the first episode; IQ is also reduced (Mesholam-Gately and Giuliano, 2009). Cognitive deficits are greater in early-onset schizophrenia (Rajji et al., 2009). Executive function and attention may be the core deficits, in that they are seen even in patients with otherwise intact cognition.

The time course of neuropsychological involvement in schizophrenia is complex and not entirely clear. People who later develop schizophrenia already have reduced IQ scores during childhood (Woodberry et al., 2008). The onset of illness is associated with further cognitive impairments, which may partially improve after resolution of the acute episode. However, most of the cognitive deficits appear to be trait features, largely independent of other symptom domains. In the seventh decade a proportion of patients undergo a substantial decline in cognitive performance and develop dementia, similar in nature to Alzheimer’s disease, but without any detectable neuropathology (Radhakrishnan et al., 2012). Like many other features of schizophrenia, impaired cognitive performance is also observed in attenuated form in unaffected relatives.

Table 11.4 Cerebral and psychological correlates of three sub-syndromes of chronic schizophrenia


Cognitive aspects of schizophrenia are currently being emphasized for several reasons. First, they are a major determinant of poor functional outcome (Green, 2006). Secondly, they are now viewed as potential therapeutic targets (see p. 290). Thirdly, cognitive features are increasingly conceptualized as being central to the disorder and as underlying the psychotic symptoms (O’Connor et al., 2009). Reflecting these developments, methods are being developed to allow routine in schizophrenia cognitive assessments in schizophrenia (e.g. the MATRICS Battery) (Nuechterlein et al., 2008), and the Brief Cognitive Assessment Tool for Schizophrenia (B-CATS) (Hurford et al., 2011).

Depressive symptoms

As noted earlier, and as recognized since the time of Kraepelin, depressive symptoms commonly occur in schizophrenia, at any phase of the illness, even before the florid positive symptoms become apparent. Indeed, they are prominent features of the prodrome (see below). In addition, about 25% of patients subsequently exhibit persistent and significant depression, called post-schizophrenic depression. Comorbid depression in schizophrenia worsens the functional outcome.

There are several reasons why depressive symptoms may be associated with schizophrenia.

• Depression may be an integral part of schizophrenia. This view is supported by the observation that about 50% of patients with acute schizophrenia experience significant depressive symptomatology, which improves as the psychosis remits.

• In the post-psychotic phase, depressive symptoms may be a response to recovery of insight into the nature of the illness and the problems to be faced. Again, this may happen at times, but it does not provide a convincing general explanation.

• Depression may be a side-effect of medication. This is not the only explanation, as depressive symptoms can occur in the absence of antipsychotic drug therapy.

Depressive symptoms in schizophrenia should not be confused with motor side-effects of medication or negative symptoms. These distinctions are important, but can be difficult to make.

Neurological signs

Neurological signs are a neglected clinical feature of schizophrenia (Chan et al., 2010). They are called ‘soft signs’ because they do not localize pathology to a particular tract or nucleus. They include abnormalities in sensory integration, coordination, and sequencing of complex motor acts; catatonic features and dyskinesias may also be considered under this category (Koning et al., 2010). Neurological signs are seen in unmedicated, first-episode patients (and are therefore separate from the effects of antipsychotic drugs in this regard), but are more common in chronic schizophrenia. The presence of neurological signs correlates with cognitive dysfunction, evidence for developmental anomaly, and diffuse brain pathology, and they are thought to be a manifestation of the neurodevelopmental origins of schizophrenia (Picchioni and Dazzan, 2009). A neurological examination that focuses upon the extrapyramidal system should be part of the routine assessment of a patient who presents with schizophrenia.

Olfactory dysfunction

Patients with schizophrenia have deficits in olfactory function, which affect the identification of, sensitivity to, and memory for odours (Turetsky et al., 2006). The deficits are not attributable to medication or smoking. They are predictive of cognitive and negative symptoms, and may contribute to the lack of social drive that is apparent in many patients (Rupp, 2010).

Water intoxication

A few patients with chronic schizophrenia drink excessive amounts of water, thus developing a state of water intoxication. When severe, it may give rise to severe hyponatraemia, cerebral oedema, seizures, coma, and sometimes death. The ‘drive’ to drink water may include unsuitable sources, such as toilet bowls. The reasons for this behaviour are unknown. Possible mechanisms include a response to a delusional belief, or abnormalities affecting hypothalamic regulation of thirst (Goldman, 2009).

Pain insensitivity

A diminished sensitivity to pain in patients with schizophrenia has long been noted by clinicians, and can on occasion be extreme. It is not related to medication (Potvin and Marchand, 2008). Its cause is unknown. Proposed explanations include thalamic lesions, and the possibility that psychotic symptoms render the patient less concerned by, or distracted from, the pain.

Factors that modify the clinical features

The social and cultural background of the patient affects the content of symptoms. For example, religious delusions are less common now than they were a century ago, and have been replaced by delusions concerned with cloning, HIV, or terrorism. Age also seems to modify the picture. In adolescents and young adults, the clinical features often include thought disorder, mood disturbance, passivity phenomena, thought insertion, and withdrawal. With increasing age, paranoid symptomatology is more common, with more organized delusions.

Intelligence also affects the clinical features, and the psychiatrist’s ability to elicit them. Patients with learning disability usually present with a simple clinical picture, sometimes referred to as pfropfschizophrenie. In contrast, highly intelligent people develop complex delusional systems, and are also better able to articulate, or conceal, their experiences.

The amount of social stimulation has a considerable effect. Under-stimulation is thought to increase negative symptoms, whereas over-stimulation precipitates positive symptoms. Modern psychosocial approaches to treatment are designed to avoid both extremes—the under-stimulation associated with institutionalization, and the excess stimulation of high expressed emotion (see p. 285).

The prodrome of schizophrenia

In recent years increasing attention has been paid to the prodrome of schizophrenia—that is, the period of time during which psychosis is ‘brewing’, when there are identifiable symptoms but before diagnosable criteria are met, and before the patient typically presents for help (Yung and McGorry, 1996). The focus on the prodrome reflects the fact that in many patients this period may last for several years, and that the longer the duration of untreated psychosis, the worse the outcome appears to be. These considerations have led to a rapid growth in ‘early intervention’, with services designed to detect and then prevent emerging cases of schizophrenia (see Chapter 21). The therapeutic aspects and implications are considered on p. 294.

The prodrome is characterized by insidious and shifting profiles of largely non-specific symptoms, including mild psychotic-like positive and negative symptoms, and also commonly mood and anxiety symptoms. Cognitive and social functioning also deteriorates. Various criteria are available to assess and rate the prodrome, including the Comprehensive Assessment of At Risk Mental State (CAARMS). Importantly, many people who are considered to be prodromal do not progress to overt psychosis; typically, only 20–30% do so during a 2- to 3-year follow-up. This has important implications for treatment interventions. Studies are attempting to identify predictors of conversion (Ruhrmann et al., 2010).

For review of the schizophrenia prodrome, see Addington and Lewis (2011).

Diagnosis and classification

This section is concerned mainly with the diagnostic criteria for and classification of schizophrenia as specified in DSM-IV and ICD-10. However, the current approach—and its problems—can be understood better with some knowledge of the historical perspective.

Historical development of ideas about schizophrenia

The development of ideas about schizophrenia is discussed in Box 11.3. To a large extent, they mirror the development of ideas about psychiatric illness in general, reflecting the central position in psychiatry held by schizophrenia during the last century. For a review, see Andreasen (2009).

In the 1960s it was noticed that there were wide divergences in the use of the term schizophrenia, and marked differences in diagnostic practice. This was unsurprising, given the multiple views and traditions summarized in Box 11.3. For example, in the UK and continental Europe, psychiatrists generally employed Schneider’s first-rank symptoms to identify a narrowly delineated group of cases. In the USA, however, interest in psychodynamic processes led to diagnosis on the basis of mental mechanisms, and to the inclusion of a much wider group of cases. First-admission rates for schizophrenia were also much higher in the USA than in the UK.

These discrepancies prompted the two major cross-national studies of diagnostic practice, discussed in Chapter 2 (see p. 29), which proved highly influential not just for schizophrenia but also for the development of diagnostic criteria and classificatory systems, namely the US–UK Diagnostic Project (Cooper et al., 1972), and the International Pilot Study of Schizophrenia (IPSS) (World Health Organization, 1973). These findings led to a consensus that agreed diagnostic criteria were required, and the development of standardized methods by which these could be defined and identified, culminating in the current ICD-10 and DSM-IV criteria for schizophrenia.

In summary, a wide range of views about schizophrenia have been held over the past century since Kraepelin and Bleuler established the concept. With the advent of current classificatory systems, the term now refers to a syndrome which can be diagnosed reliably, which is essential for rational clinical practice. However, as noted in Chapter 3, reliability is not in itself sufficient. For example, use of Schneider’s first-rank symptoms leads to high reliability in diagnosis but has no prognostic value. Moreover, because the cause of schizophrenia is still largely unknown, the syndrome remains of uncertain validity. Until this fundamental question is answered, there will continue to be dispute as to its most important features, diagnostic boundaries, and internal subdivisions.

Box 11.3 Development of schizophrenia concepts and terminologies

In the nineteenth century, one view was that all serious mental disorders were expressions of a single entity, which Griesinger called Einheitpsychose (unitary psychosis). The alternative view, put forward by Morel in France, was that mental disorders could be separated and classified. Morel searched for specific entities, and argued for a classification based on cause, symptoms, and outcome. In 1852 he gave the name démence précoce to a disorder which he described as starting in adolescence and leading first to withdrawal, odd mannerisms, and self-neglect, and eventually to intellectual deterioration. A few years later, Kahlbaum (1863) described the syndrome of catatonia, and Hecker (1871) wrote an account of a condition he called hebephrenia.

Emil Kraepelin (1855–1926) derived his ideas from studying the course and outcome of the disorder. His observations led him to argue against the idea of a single psychosis, and to propose a division into dementia praecoxand manic–depressive psychosis. This grouping brought together as subclasses of dementia praecox the previously separate entities of hebephrenia and catatonia. (His adoption of the word dementia emphasizes the prominence he attributed to the cognitive impairments of the disorder.) Kraepelin’s description of dementia praecox appeared for the first time in 1893, in the fourth edition of his textbook, and the account was expanded in subsequent editions. He described the illness as occurring in clear consciousness, and consisting of ‘a series of states, the common characteristic of which is a peculiar destruction of the internal connections of the psychic personality’ (Kraepelin, 1919). Kraepelin originally divided dementia praecox into three subtypes (catatonic, hebephrenic, and paranoid), and later added a fourth subtype (simple). He separated the condition that he named paraphrenia(Box 12.2, p. 300) from dementia praecox on the grounds that it started in middle life and seemed to be free from the changes in emotion and volition found in dementia praecox. It is commonly held that Kraepelin regarded dementia praecox as invariably progressing to chronic deterioration. However, he reported that, in his series of cases, 13% recovered completely (although some relapsed later) and 17% were ultimately able to live and work without difficulty.

Eugen Bleuler (1857–1959) based his work on that of Kraepelin, and in his own book wrote ‘the whole idea of dementia praecox originates with Kraepelin’ (Bleuler, 1911). He also acknowledged the help of his younger colleague, Carl Jung, in trying to apply some of Freud’s ideas to dementia praecox. Compared with Kraepelin, Bleuler was concerned less with prognosis and more with the mechanisms of symptom formation. Bleuler proposed the name schizophrenia to denote a ‘splitting’ of psychic functions, which he considered to be of central importance. He believed in a distinction between fundamentaland accessory symptoms. Fundamental symptoms included disturbances of associations, changes in emotional reactions, and autism (withdrawal from reality into an inner world of fantasy). It is interesting that, in Bleuler’s view, some of the most frequent and striking symptoms were accessory (secondary)—for example, hallucinations, delusions, catatonia, and abnormal behaviours. Bleuler was interested in the psychological study of his cases, but did not rule out the possibility of a neuropathological cause of schizophrenia. Since Bleuler was preoccupied more with psycho-pathological mechanisms than with symptoms themselves, his approach to diagnosis was less precise than that of Kraepelin. In addition, Bleuler took a more optimistic view of the outcome, but still held that one should not ‘speak of cure but of far-reaching improvement.’ He also wrote: ‘as yet I have never released a schizophrenic in whom I could not still see distinct signs of the disease, indeed there are very few in whom one would have to search for such signs’ (Bleuler, 1911).

Kurt Schneider (1887–1967)—not to be confused with Carl Schneider, who categorized types of thought disorder—tried to make the diagnosis more reliable by identifying a group of symptoms characteristic of schizophrenia and rarely found in other disorders. Schneiderian first-rank symptoms (see Table 11.3) have a major role in the contemporary diagnosis of schizophrenia. However, unlike Bleuler’s fundamental symptoms, Schneider’s symptoms were not supposed to have any central psychopathological role. Moreover, he noted that they were neither necessary nor sufficient for the diagnosis:

Among the many abnormal modes of experience that occur in schizophrenia, there are some which we put in the first rank of importance, not because we think of them as basic disturbances, but because they have this special value in helping us to determine the diagnosis of schizophrenia. When any one of these modes of experience is undeniably present and no basic somatic illness can be found, we may make the diagnosis of schizophrenia. … Symptoms of first rank importance do not always have to be present for a diagnosis to be made.

(Schneider, 1959)

Several German psychiatrists tried to define subgroupings within schizophrenia. Karl Kleist, a pupil of the neurologist Wernicke, looked for associations between brain pathology and different subtypes of psychotic illness. He accepted Kraepelin’s main diagnostic framework, but used careful clinical observation in an attempt to distinguish various subdivisions within schizophrenia and other atypical disorders. His attempt to match these subtypes to specific kinds of brain pathology was ingenious but unsuccessful.

Leonhard continued this approach of careful clinical observation, and published a complicated classification that distinguishes schizophrenia from the ‘cycloid’ psychoses—a group of non-affective psychoses of good outcome (Leonhard, 1957). He also divided schizophrenia into two groups. The first group is characterized by a progressive course, and is divided into catatonias, hebephrenias, and paraphrenias. Leonhard gave this group a name which is often translated as systematic. The second group, referred to as non-systematic, is divided into affect-laden paraphrenia, schizophasia, and periodic catatonia. Affect-laden paraphrenia is characterized by paranoid delusions and the expression of strong emotion about their content. In schizophasia, speech is grossly disordered and difficult to understand. Periodic catatonia is a condition with regular remissions. During an episode, akinetic symptoms are sometimes interrupted by hyperkinetic symptoms.

Scandinavian psychiatrists were influenced by Jaspers’ distinction between process schizophrenia and reactive psychoses. In the late 1930s, Langfeldt, using follow-up data on patients in Oslo, proposed a distinction between true schizophrenia, which had a poor prognosis, and schizophreniform states, which had a good prognosis. True schizophrenia was defined narrowly and was similar to Kraepelin’s dementia praecox. Schizophreniform states were described as often precipitated by stress and accompanied by confusional and affective symptoms. Langfeldt’s distinction between cases with good and bad prognoses has been influential, but research has not found that his criteria predict prognosis accurately. According to modern diagnostic criteria, most of Langfeldt’s schizophreniform states would be classified as mood disorders. Note also that in DSM-IV the term schizophreniform disorder is used in a different way.

In Denmark and Norway, cases of psychosis arising after stressful events have received much attention. The terms reactive psychosis or psychogenic psychosis are commonly applied to conditions which appear to be precipitated by stress, are to some extent understandable in their symptoms, and have a good prognosis. In current diagnostic schemes such disorders would be classified as brief psychotic disorder or schizophreniform disorder.

Classification of schizophrenia in DSM-IV and ICD-10

The classification of schizophrenia and schizophrenia-like disorders in DSM-IV and ICD-10 is outlined and compared in Table 11.5. The classifications can be seen to be broadly similar.


In this classification, schizophrenia is defined in terms of the symptoms in the acute phase and also of the course, for which the requirement is continuous signs of disturbance for at least 6 months (see Table 11.6). The acute symptoms (criterion A) include delusions, hallucinations, disorganized speech or behaviour, and negative symptoms. At least two of these symptoms must have been present for a period of 1 month (unless successful treatment has occurred). Where subjects have major disturbances of mood occurring concurrently with the acute-phase symptoms, a diagnosis of a schizoaffective disorder or mood disorder with psychotic features should be made.

A further criterion for the diagnosis of schizophrenia is that the patient must have exhibited deficiencies in their expected level of occupational or social functioning since the onset of the disorder. As noted above, for the diagnosis to be made there must have been at least 6 months of continuous disturbance; this can include prodromal and residual periods when acute-phase symptoms, as described above, are not evident. Patients whose symptom duration does not meet this 6-month period criterion will be classified as suffering from schizophreniform disorder or brief psychotic disorder. Other related disorders include delusional disorder (see Chapter 12) and psychotic disorders not otherwise classified (see below).

In DSM-IV, schizophrenia is divided into a number of subtypes, defined by the predominant symptomatology at the time of evaluation. After at least 1 year has elapsed since the onset of active-phase symptoms, DSM-IV also allows the disorder to be classified by longitudinal course, and specification of the pattern of relapse and remission.

Table 11.5 Classification of schizophrenia and schizophrenia-like disorders in ICD-10 and DSM-IV*


Table 11.6 Criteria for schizophrenia in DSM-IV

A. Characteristic symptoms of the active phase

Two (or more) of the following, each present for a significant portion of time during a 1-month period (or less if successfully treated)

1. Delusions

2. Hallucinations

3. Disorganized speech (e.g. frequent derailment or incoherence)

4. Grossly disorganized or catatonic behaviour

5. Negative symptoms (i.e. affective flattening, alogia, or avolition)

B. Social/occupational dysfunction

For a significant portion of the time since the onset of the disturbance, one or more major areas of functioning such as work, interpersonal relations, or self-care are markedly below the level achieved prior to the onset (or when the onset is in childhood or adolescence, failure to achieve expected level of interpersonal, academic, or occupational achievement)

C. Duration

Continuous signs of the disturbance persist for at least 6 months. This 6-month period must include at least 1 month of symptoms (or less if successfully treated) that meet criterion A (i.e. active-phase symptoms) and may include periods of prodromal or residual symptoms; the signs of the disturbance may be manifested by only negative symptoms or two or more symptoms listed in criterion A present in an attenuated form (e.g. odd beliefs, unusual perceptual experiences)

D. Schizoaffective and mood disorder exclusion

Schizoaffective disorder and mood disorder with psychotic features have been ruled out because either (1) no major depressive, manic, or mixed episodes have occurred concurrently with the active-phase symptoms, or (2) if mood episodes have occurred during active-phase symptoms, their total duration has been brief relative to the duration of the active and residual periods

E. Substance/general medical condition exclusion

The disturbance is not due to the direct physiological effects of a substance (e.g. drug of abuse, or medication) or a general medical condition

F. Relationship to a pervasive developmental disorder

If there is a history of autistic disorder or another pervasive developmental disorder, the additional diagnosis of schizophrenia is made only if prominent delusions or hallucinations are also present for at least 1 month (or less if successfully treated)


The ICD-10 criteria for schizophrenia (see Table 11.7) place more reliance than DSM-IV on first-rank symptoms, and requires a duration of only 1 month, excluding prodromal symptoms. There are several clinical subtypes, which differ from DSM-IV mainly in the inclusion of categories of simple schizophrenia and post-schizophrenic depression. As in DSM-IV, cases in which symptoms of schizophrenia are accompanied by prominent mood disturbances are classified as schizoaffective disorders.

Table 11.7 Criteria for schizophrenia in ICD–10

The normal requirement for a diagnosis of schizophrenia is that a minimum of one very clear symptom (and usually two or more if less clear-cut) belonging to any one of the groups listed as (a)–(d) below, or symptoms from at least two of the groups referred to as (e)–(h), should have been clearly present for most of the time during a period of 1 month or more

(a) Thought echo, thought insertion or withdrawal, and thought broadcasting

(b) Delusions of control, influence, or passivity, clearly referred to body or limb movements or specific thoughts, actions, or sensations; delusional perception

(c) Hallucinatory voices giving a running commentary on the patient’s behaviour, or discussing the patient among themselves, or other types of hallucinatory voices coming from some part of the body

(d) Persistent delusions of other kinds that are culturally inappropriate and completely impossible

(e) Persistent hallucinations in any modality, when accompanied either by fleeting or half-formed delusions without clear affective content, or by persistent overvalued ideas, or when occurring every day for weeks or months on end

(f) Breaks or interpolations in the train of thought, resulting in incoherence or irrelevant speech, or neologisms

(g) Catatonic behaviour, such as excitement, posturing, or waxy flexibility, negativism, mutism, and stupor

(h) ‘Negative’ symptoms such as marked apathy, paucity of speech, and blunting or incongruity of emotional responses, usually resulting in social withdrawal and lowering of social performance; it must be clear that these are not due to depression or to neuroleptic medication

(i) A significant and consistent change in the overall quality of some aspects of personal behaviour, manifested as loss of interest, aimlessness, idleness, a self-absorbed attitude, and social withdrawal

Certain aspects of personality are associated with schizophrenia, clinically and genetically. ICD-10 includes schizotypal disorder among schizophrenic disorders, whereas in DSM-IV this condition is classified as a personality disorder. Schizotypal disorder is associated with social isolation and restriction of affect; in addition, there are perceptual distortions, with disorders of thinking and speech and striking eccentricity or oddness of behaviour.

Similarly to DSM-IV, ICD-10 classifies delusional disorders separately from schizophrenia (see Chapter 12). In addition, a group of acute and transient psychotic disorders are recognized that have an acute onset and complete recovery within 2–3 months (see below).

Summary of the differences between DSM-IV and ICD-10

• ICD-10 places greater weight on Schneider’s first-rank symptoms. DSM-IV emphasizes course and functional impairment.

• ICD-10 requires a duration of illness of 1 month, whereas DSM-IV requires a duration of 6 months.

• Schizotypal disorder is included in ICD-10, but is categorized as a personality disorder in DSM-IV.

• ICD-10 includes some additional subtypes, namely simple schizophrenia and post-schizophrenic depression.

• Disorganized schizophrenia in DSM-IV is called hebe-phrenic schizophrenia in ICD-10.

Schizophrenia-like disorders

Whatever definition of schizophrenia is adopted, there will be cases that resemble schizophrenia in some respects and yet do not meet the criteria for diagnosis. In DSM-IV and ICD-10, these disorders are divided into four groups:

1. delusional disorders (paranoid psychoses)

2. brief disorders

3. disorders accompanied by prominent affective symptoms

4. disorders without all of the symptoms required for schizophrenia.

Delusional disorders are discussed in Chapter 12. The latter three groups are discussed here.

Brief disorders

DSM-IV uses the term brief psychotic disorder to refer to a syndrome characterized by at least one of the acute-phase positive symptoms shown in Table 11.6. The disorder lasts for at least 1 day but not more than 1 month, by which time full recovery has occurred. The disorder may or may not follow a stressor, but psychoses induced by the direct physiological effects of drugs or medical illness are excluded. Schizophreniform disorder is a syndrome similar to schizophrenia (meeting criterion A) which has lasted for more than 1 month (and so cannot be classified as brief psychotic disorder), but for less than the 6 months required for a diagnosis of schizophrenia to be made. Social and occupational dysfunction are not required in order to make the diagnosis.

In ICD-10, the grouping is acute and transient psychotic disorder. These disorders are of acute onset, and complete recovery within 2–3 months is the rule. The disorder may or may not be precipitated by a stressful life event. The category is then subdivided into several overlapping and somewhat confusing subtypes. In the first two, acute polymorphic psychotic disorders, with or without symptoms of schizophrenia, hallucinations, delusions, and perceptual disturbance are obvious, but change rapidly in nature and extent. There are often accompanying changes in mood and motor behaviour. Bouffée délirante and cycloid psychosis (see Box 11.3) are given as synonyms for these categories. A third subtype, acute schizophrenia-like psychotic episode, is a non-committal term for cases that meet the symptom criteria for schizophrenia but last for less than a month. Residual cases that do not fit these subtypes of acute psychosis are called other acute psychotic episodes. Overall, the ICD-10 category identifies a heterogeneous group of disorders in terms of outcome (Singh et al., 2004).

Schizophrenia-like disorders with prominent affective symptoms

Some patients have a more or less equal mixture of schizophrenic and affective symptoms. As mentioned earlier, such patients are classified under schizoaffective disorder in both DSM-IV and ICD-10.

The term schizoaffective disorder has been used in several distinct ways. It was first applied by Kasanin (1994/1933) to a small group of young patients with severe mental disorders characterized by a very sudden onset in a setting of marked emotional turmoil. The psychosis lasted a few weeks and was followed by recovery.

The current definitions differ substantially from this description. DSM-IV requires that there should have been an uninterrupted period of illness during which, at some time, there is either a major depressive episode, a manic episode, or a mixed episode concurrent with symptoms that meet criterion A for schizophrenia. During this continuous episode of illness, the acute-phase psychotic symptoms must have been present for at least 2 weeks in the absence of prominent mood symptoms (or the diagnosis would be a mood disorder with psychotic features). However, the episode of mood disturbance must have been present for a substantial part of the illness. The definition of schizoaffective disorder in ICD-10 is similar. It also specifies that the diagnosis should only be made when both definite schizophrenic and definite affective symptoms are equally prominent and present simultaneously, or within a few days of each other. (This is an important point. The label ‘schizoaffective’ should not be applied just because a patient has an isolated symptom or two consistent with both diagnoses, or because the assessment has not been sufficiently detailed to identify the primary diagnosis.) ICD-10 classifies schizoaffective disorder according to whether the mood disturbance is depressive, manic, or mixed. In DSM-IV schizoaffective disorder is specified as either depressive type or bipolar.

Family studies have shown that the more recent diagnostic concepts of schizoaffective disorder delineate a syndrome in which first-degree relatives have an increased risk of both mood disorders and schizophrenia. In addition, the outcome of schizoaffective disorder is generally thought to be better than that for schizophrenia, with negative symptoms rarely developing (Tsuang et al., 2000).

As noted above, it is not uncommon for patients with schizophrenia to develop depression as the symptoms of acute psychosis subside. This is recognized in ICD-10 as post-schizophrenic depression, where prominent depressive symptoms have been present for at least 2 weeks while some symptoms of schizophrenia (either positive or negative) still remain.

Persistent disorders without all of the required symptoms for schizophrenia

A difficult problem is presented by cases that have longstanding schizophrenia-like symptoms but which do not fully meet the diagnostic criteria. There are four groups.

1. Patients who have exhibited the full clinical picture of schizophrenia in the past, but who no longer have all of the symptoms required to make the diagnosis. These cases are classified as residual schizophrenia in both DSM-IV and ICD-10.

2. People who from an early age have behaved oddly and shown features seen in schizophrenia—for example, ideas of reference, persecutory beliefs, and unusual types of thinking. When long-standing, these disorders can be classified as personality disorders in DSM-IV (schizotypal personality disorder), or with schizophrenia in ICD-10 (schizotypal disorder). Because of a suggested close relationship to schizophrenia, these disorders have also been called latent schizophrenia, or part of the schizophrenia spectrum.

3. People with social withdrawal, lack of initiative, odd behaviour, and blunting of emotion, in whom positive psychotic symptoms are never known to have occurred. A variety of terms may be applicable, including simple schizophrenia (see p. 258), schizoid personality disorder (see p. 145), and Asperger’s syndrome (see p. 653). (We advise caution in using the terms simple schizophrenia or latent schizophrenia. Both of these lack utility—since treatment is rarely indicated or wanted—and represent speculative diagnostic labels for what many would consider simply to be eccentricity.)

4. People with stable, persistent delusions but without other features of schizophrenia (see Chapter 12).


Many patients with schizophrenia also meet the criteria for another psychiatric diagnosis, complicating diagnosis and treatment, and worsening the outcome. The example of depression (which occurs in 50% of cases) has already been mentioned. The estimated lifetime prevalence of any substance abuse is 47%, and of alcohol abuse is 21%. Psychiatric disorders often comorbid with schizophrenia include post-traumatic stress disorder (29%), obsessive–compulsive disorder (23%), and panic disorder (15%). For a review, see Buckley et al. (2009).

Differential diagnosis

We have already described how current classifications include schizophrenia-like disorders as well as schizophrenia, but the boundary between them is blurred and to some extent arbitrary. Similarly, there is no clear distinction between disorders that are considered to be variants of schizophrenia, and some which are viewed as being part of its differential diagnosis. For example, delusional disorders are often included in both categories, while schizotypal disorder is classified as a personality disorder in DSM-IV but with schizophrenia in ICD-10. Such difficulties reflect the unknown validity of the syndrome(s), and are unlikely to be resolved until the classification is based upon aetiology or other empirically validated markers.

With these caveats in mind, current diagnostic practice requires schizophrenia to be distinguished from a number of other disorders, which are listed below.

• Organic syndromes. Acute schizophrenia can be mistaken for delirium, especially if there is pronounced thought disorder and a rapidly fluctuating affect and mental state. Careful observation is needed for clouding of consciousness, disorientation, and other features of delirium (see Chapter 13). Schizophrenia-like disorders can also occur, in clear consciousness, in a range of neurological and medical disorders. These conditions are referred to as a psychotic disorder secondary to a general medical condition (DSM-IV), organic delusional disorder (ICD-10), or simply as secondary schizophrenia. Classic examples include temporal lobe epilepsy (complex partial seizures), general paralysis of the insane, and metachromatic leukodystrophy (see Chapter 13, and Hyde and Ron, 2011). Visual, olfactory, and gustatory hallucinations are said to be suggestive of an organic disorder. The occurrence of organic schizophrenia-like disorders emphasizes the importance of a careful medical history and physical examination (and investigations, if indicated) in all such patients. One study in London found that 10 out of 268 cases (3.7%) of ‘first-episode schizophrenia’ had an organic cause other than substance misuse (Johnstone et al., 1987).

• Drug-induced states (including substance misuse). Certain prescribed drugs, particularly steroids and dopamine agonists, can cause florid psychotic states. Psychoactive substance misuse, particularly with psychostimulants or phencyclidine, and also alcohol, should always be considered in the presentation of schizophrenia-like psychoses. Urine or hair testing can be helpful in diagnosis, as is the temporal association between drug use and symptoms. However, the high prevalence of recreational drug use in young adults means that a clear distinction between a drug-induced psychosis and schizophrenia is not always possible.

• Mood disorder with psychotic features. The distinction of schizophrenia (and schizoaffective disorder) from affective psychosis depends on the degree and persistence of the mood disorder, the relationship of any hallucinations or delusions to the prevailing mood, and the nature of the symptoms in any previous episodes. The distinction from mania in young people can be particularly difficult, and sometimes the diagnosis can only be clarified by longer-term follow-up. A family history of mood disorder may be a useful pointer.

• Delusional disorders. These disorders (see Chapter 12) are characterized by chronic, systematized paranoid delusions, but lack other symptoms of schizophrenia, and many areas of the mental state are unremarkable.

• Personality disorder. Differential diagnosis from personality disorder, especially of the paranoid, schizoid, or schizotypal forms, can be difficult when insidious changes are reported in a young person, or paranoid ideas are present. Prolonged observation may be required to detect genuine symptoms of psychosis, and the additional features indicative of schizophrenia. Some patients with borderline personality disorder also exhibit transient psychotic symptoms, although the presence of affective instability and other features means that there should rarely be diagnostic confusion with schizophrenia.


Schizophrenia is a disorder with a low incidence and a high prevalence. Estimates of both have varied markedly, reflecting the methodology and diagnostic criteria used, as well as genuine differences between populations. Recent reviews and meta-analyses have helped to clarify the estimates and their confidence intervals. For reviews, see McGrath et al. (2008) and Jablensky et al. (2011).


The annual incidence using current diagnostic criteria is 0.16–1.00 per 1000 population using a broad definition; for more restrictive diagnoses, the incidence is about two to three times lower (Jablensky et al., 2011). A systematic review provided an estimated mean incidence of 0.24 per 1000, but an estimated median incidence of 0.15 per 1000. This difference reflects a skew, with many estimates in the upper tail of the distribution, which in turn results from studies that sampled from within populations with a high incidence (e.g. migrant groups, as discussed below). These recent analyses strongly question the belief, largely arising from the World Health Organization Ten Country Study (Jablensky et al., 1992), that schizophrenia has a similar incidence (and prevalence) in all populations (McGrath et al., 2008). There may also be variations in incidence related to latitude, and urban–rural differences (see p. 274).

There has been debate as to whether the incidence of schizophrenia is falling in industrialized countries, stimulated by a Scottish study (Eagles and Whalley 1985). However, studies vary in their findings. In any event, whether observed declines represent a true fall in inception rate, as opposed to changes in diagnostic criteria or their implementation, remains unclear (Jablensky et al., 2011).


Systematic reviews indicate a lifetime prevalence of about 4 per 1000, and a lifetime morbid risk of 7.2 per 1000 (McGrath et al., 2008). These are median values because the data are skewed; the means are higher, at 5.5 and 11.9 per 1000, respectively. The latter figure is more consistent with the frequent statement that about 1 in 100 people develop schizophrenia, but the value of 7 in 1000 is statistically more appropriate. As with incidence, prevalence estimates are higher in men than in women, in some migrant groups, and at higher latitudes.

Age at onset

Schizophrenia can begin at any stage of life, from childhood to old age. The usual onset is between 15 and 54 years, most commonly in the mid-twenties; more detailed analyses suggest that there are two peaks, one at 20 years of age, and a second peak at 33 years. Gender differences in age of onset may explain these observations, as men have a single peak in their early twenties, whereas women have a broader range of age of onset, and a second peak in their fourth decade.

Late-onset schizophrenia is discussed in Chapter 18.


In addition to gender differences in age of onset, meta-analyses show that schizophrenia is more common in men than in women, with a male:female ratio of 1.4. The difference is more marked for severe cases, and is not due to differences in referral, identification, or age of onset. It has been attributed to the neuroprotective effects of oestrogen, but the evidence for this is poor. Gender differences in genetic and epigenetic risk factors may also be relevant. For a general review of gender issues in schizophrenia and its management, see Grossman et al. (2008) and Abel et al. (2010).


Whether patients with schizophrenia show decreased fertility has been a controversial topic, not least because of its implications for debate about whether schizophrenia is a ‘disease’, and with regard to its genetic basis. A recent meta-analysis showed a substantial reduction in fertility in people with schizophrenia (about 40% of that expected), which was greater in men than in women (Bundy et al., 2011). These figures are difficult to interpret as being an intrinsic feature of schizophrenia, since opportunities for patients to have children have historically been limited by institutionalization, as well as by the amenorrhoea and sexual dysfunction caused by antipsychotic drugs. On the other hand, the unaffected siblings of patients also show a slight reduction in fertility (Bundy et al., 2011).

Other aspects of schizophrenia epidemiology

The course and outcome of schizophrenia are discussed later in the chapter, and other epidemiological aspects relevant to aetiology are considered in the next section.



Views about the aetiology of schizophrenia have been inextricably linked with the controversies regarding its nature and classification (discussed earlier in this chapter and in Chapter 3).

Schizophrenia therefore exemplifies the whole range of biological, psychological, and social factors considered to be important in psychiatric causation (see Chapter 5 and Table 11.8), and the methods which have been applied to try to identify them. This section summarizes the current knowledge and theories in each domain, and also mentions some outdated but influential views. It adopts a broad definition of aetiology, and includes pathogenesis and pathophysiological findings.

Key aspects of the present consensus regarding the aetiology of schizophrenia can be summarized as follows. The most important influence is genetic, with about 80% of the risk of schizophrenia being inherited. The mode of inheritance is complex, and the genes—some of which have recently been identified—act as risk factors, not determinants of illness. A number of environmental factors contribute, too, many of which appear to act prenatally, and which interact with the genetic predisposition. Together, these and subsequent risk factors lead to a neurodevelopmental disturbance that renders the individual vulnerable to the later emergence of symptoms, and that manifests itself premorbidly in a range of behavioural, cognitive, and neuroanatomical features. In schizophrenia, there are slight but robust differences in brain structure and function, supporting the view that the syndrome is a disorder of brain connectivity. Acute psychosis is associated with excess dopamine neurotransmission, which may be secondary to abnormalities of the glutamate system. Psychosocial factors have been relatively neglected recently, but significantly influence the onset and course of illness. Finally, it is emphasized that there are few certainties about the aetiology of schizophrenia, and even where the facts are robust, their interpretation often remains unclear.

For contemporary reviews of the aetiology and conceptualization of schizophrenia, see Tandon et al. (2008), van Os and Kapur (2009), and Insel (2010).


Family, twin, and adoption studies have all been applied extensively in schizophrenia, and cumulatively provide irrefutable evidence for a major genetic contribution to the syndrome (Riley and Kendler, 2011). Such studies set the context for the current wave of molecular studies designed to identify the causative genes and genetic abnormalities.

Table 11.8 Examples of aetiological factors and theories in schizophrenia


Family studies

The first systematic family study was conducted in Kraepelin’s department by Ernst Rudin, who showed that the rate of dementia praecox was higher among the siblings of probands than in the general population. Later research used better criteria for proband selection and diagnosis, and yielded estimates of an average lifetime risk of around 5–10% among first-degree relatives of people with schizophrenia, compared with 0.2–0.6% among first-degree relatives of controls (see Table 11.9). Taken together, the family studies provide clear evidence of a familial aetiology, but do not distinguish between genetic effects and the influence of the family environment.

Table 11.9 Approximate lifetime risk of developing schizophrenia for relatives of a proband with schizophrenia


Family studies can also be used to determine whether the liability to schizophrenia and other disorders (e.g. bipolar disorder) is transmitted independently, which should be the observed pattern if the two disorders are separate syndromes with differing aetiology. However, the findings from the various studies have been contradictory, with some researchers arguing for separate familial transmission of mood disorders and schizophrenia, while others suggest that what is transmitted is a shared vulnerability to psychosis. Two broad conclusions can be drawn (see Table 11.10 and Kendler and Diehl, 1993). First, the risk of schizophrenia, schizoaffective disorder, and schizotypal and paranoid personality disorders is increased in first-degree relatives of patients with schizophrenia. Secondly, the risk of both schizophrenia and mood disorder is increased in first-degree relatives of patients with schizoaffective disorder.

Twin studies

Twin studies were introduced on p. 99. They have been of considerable importance in research on schizophrenia, providing unequivocal evidence of the heritability of this disorder.

The first substantial twin study was conducted in Munich by Hans Luxenberger in the 1920s. He found concordance in 11 of his 19 monozygotic (MZ) pairs and in none of his 13 dizygotic (DZ) pairs. Subsequent investigations, although differing with regard to rates, all agree that concordance is several-fold higher in MZ than in DZ twins. Representative figures for concordance are 40–50% for MZ pairs and about 10% for DZ pairs (Cardno and Gottesman, 2000).

Table 11.10 Lifetime risk of developing psychiatric disorder in first-degree relatives of probands with schizophrenia, schizoaffective disorder, and controls


Modern twin studies also produce estimates of heritability (the proportion of liability to schizophrenia in the population which can be attributed to genes), and can separate environmental factors into those that are unique to the individual and those which are shared with others. A meta-analysis (Sullivan et al., 2003) thereby confirmed the substantial heritability of schizophrenia (81%; 95% confidence interval, 73–90%). These data unambiguously show that inheritance (i.e. genes) contribute the majority of the risk for schizophrenia. However, there are caveats about the interpretation of heritability figures. For example, the estimates make some assumptions about the ‘genetic architecture’ (i.e. how the genetic factors operate) and include gene–environment interactions. The meta-analysis also showed, less predictably, that most of the environmental contribution comes from shared rather than individual-specific influences (11%; 95% confidence interval, 3–19%). The identity of these shared environmental factors is unknown.

It is worth noting that, among discordant MZ twins, the risk of schizophrenia is increased equally in children of the unaffected and the affected co-twin. This indicates that the unaffected co-twin indeed had the same genetic susceptibility to developing schizophrenia as the affected twin, but for some reason did not express the phenotype. This is most probably due to environmental protective factors, or chance (‘stochastic processes’), affecting the penetrance or expression of the genetic predisposition. Moreover, unaffected identical co-twins do exhibit some mild features of schizophrenia (in terms of symptoms and biological findings) which fall well short of being diagnostically significant, but which are greater than those seen where neither member of a twin pair has schizophrenia. This probably reflects a partial expression of the risk genotype.

Adoption studies

Heston (1966) studied 47 adults who had been born to mothers with schizophrenia and separated from them within 3 days of birth. As children they had been brought up in a variety of circumstances, although not by the mother’s family. At the time of the study their mean age was 36 years. Heston compared them with controls matched for circumstances of upbringing, but whose mothers had not suffered from schizophrenia. Among the offspring of the affected mothers, five were diagnosed as having schizophrenia, compared with none of the controls. The rate for schizophrenia among the adopted-away children was comparable with that among children with a schizophrenic parent who remained with their biological family.

Further evidence came from a series of Danish studies that started in the 1960s. In one major project (Kety et al., 1975), two groups of adoptees were identified—a group of 33 adoptees who had schizophrenia, and a matched group who were free from schizophrenia. Rates of disorder were compared in the biological and adoptive families of the two groups of adoptees. The rate for schizophrenia was higher among the biological relatives of the adoptees with schizophrenia than among the relatives of the controls, a finding which supports the genetic hypothesis. Furthermore, the rate for schizophrenia was not increased among couples who adopted the affected children, which suggests that environmental factors were not of substantial importance. Follow-up studies using a national sample of Danish adoptees confirmed that biological first-degree relatives of patients with schizophrenia have an approximately tenfold increased risk of suffering from schizophrenia or a related (‘spectrum’) disorder (Kety et al., 1994).

The data from adoption studies thus strongly support the view that genetic factors explain the familial clustering of schizophrenia. However, it should be noted that they cannot control for the prenatal environment, which other studies suggest is important. Nor can they rule out an interaction between environmental causes in the adoptive family and genetic predisposition; indeed, Finnish data show that adoptees at high genetic risk of schizophrenia are more sensitive to adverse upbringing (Tienari et al., 2004).

The mode of inheritance

The frequency profile of schizophrenia among people with different degrees of genetic proximity to the proband does not fit any simple Mendelian pattern, as would be expected if the disorder was caused by a single major gene (Gottesman, 1991). Instead, the evidence suggests that schizophrenia arises from the cumulative effect of several genes, as a so-called complex or non-Mendelian disorder. The liability to schizophrenia lies along a continuum in the population, and is expressed when a certain threshold of genetic susceptibility is exceeded. No genes are either necessary or sufficient, and they act as risk factors, not determinants. Within this model, a range of genetic mechanisms and subtypes are possible—for example, concerning the nature of the genetic abnormalities themselves (e.g. single-nucleotide polymorphisms and copy-number variations), whether the gene variants are common or rare in the population, and whether a given gene acts independently or interactively with other genes (epistasis). For review, see Owen et al. (2010) and Bassett et al. (2010).

What is heritable?

Identification of the genes responsible for a condition is much easier if the boundaries of the inherited phenotype are known. However, this is a circular argument, as it is ultimately the identification of causative genes that, whenever possible, defines diseases. The uncertain boundaries between schizophrenia and schizophrenia spectrum disorders and mood disorders have already been mentioned. Recent molecular genetic studies strongly support the view that schizophrenia and bipolar disorder are not distinct disorders, but lie along a genetic spectrum (Owen et al., 2007), and have also suggested that other aspects of the genetic predisposition to schizophrenia are shared with autism and with learning disability (Bassett et al., 2010).

Another possibility is that it is particular components of schizophrenia, or other features associated with it, which are actually being inherited. Much evidence suggests that there are such features, called endophenotypes or intermediate phenotypes (see p. 104), which are more closely related to the genes, and therefore more aetiologically valid. A range of heritable endophenotypes of schizophrenia are known, including evoked potentials, working memory, brain structure, and eye-tracking dysfunction. Whether any of these components become part of a redefined syndrome or syndromes of schizophrenia will depend upon confirmation of their genetic basis, and demonstration that they have reliability and utility (e.g. in predicting outcome or therapeutic responsiveness).

Schizophrenia susceptibility genes

Despite the high heritability of the disorder, and the large number of studies that have been conducted, it has proved difficult to identify schizophrenia genes. This difficulty probably reflects the various issues already mentioned, notably the uncertainty as to the phenotype being inherited, and the existence of multiple genes of small effect which are thus hard to detect. However, significant progress has been made in the past few years. This is due to a combination of methods which were introduced in Chapter 5. Each has contributed to the current state of knowledge, although at present not all of the findings from each approach can readily be integrated.

A recent meta-analysis of 32 genome-wide linkage studies identified several chromosomal regions (loci) which, by reasonably stringent statistical criteria, show linkage to schizophrenia (Ng et al., 2009), namely regions of chromosome 1, 2q, 3q, 4q, 5p, 8p, and 10q (p and q refer to the short and long arms, respectively). This suggests that a gene (or genes) contributing to schizophrenia risk is located at each of these regions, although both false-positive and false-negative findings are likely.

Several approaches have been taken to finding schizophrenia genes themselves, each of which has had some successes.

First, testing for genetic association with schizophrenia of single-nucleotide polymorphisms (SNPs) in genes within these chromosomal loci led to the discovery of several schizophrenia genes, notably neuregulin, dysbindin, and G72.

Secondly, chromosomal abnormalities that confer an increased risk of schizophrenia were used as a rationale for searching for genes in that region. The best example is velocardiofacial syndrome (VCFS, also known as di George syndrome), which is caused by deletion of one copy of chromosome 22q11 (hence its alternative name, 22q11 hemideletion syndrome). It is a relatively common cytogenetic anomaly, occurring in 1 in 4000 live births, and causes a range of physical abnormalities and cognitive impairment (Murphy, 2002). Of relevance here, it is also associated with psychosis (either schizophrenia-like or affective) in about 30% of individuals. Thus even though VCFS is a rare cause of schizophrenia overall, 22q11 is implicated as a locus for schizophrenia genes in general. The resulting focus on genes within this region has identified several which may be involved in schizophrenia risk (Karayiorgou et al., 2010). Another example in this category is disrupted in schizophrenia 1 (DISC1), a gene on chromosome 1 which was identified after extensive study of a large Scottish family in which a translocation between chromosomes 1 and 11 is linked to a high incidence of schizophrenia (and other psychiatric disorders) (Porteous et al., 2006).

Thirdly, ‘candidate genes’ (i.e. those implicated on the basis of biological theories of, or findings in, schizophrenia) have been studied. This approach has not proved very successful, but has provided some evidence for the involvement of dopamine D2 and D3 receptors.

Fourthly, genome-wide association studies (GWAS) (see p. 104) have been used. These have produced the best statistical evidence of all the approaches to date, with several genes, notably for miR137, ZNF804A, and the major histocompatibility complex (MHC) locus. However, for various reasons these studies may also miss important genes and genetic abnormalities (Mitchell and Porteous, 2011).

Finally, analysis of copy-number variation (CNV) (see p. 101) has been used. There is now good evidence that schizophrenia is associated with an increased number of CNVs, especially deletions (see Table 11.11). (VCFS, mentioned earlier, can be regarded as a large CNV.) The CNVs can be anywhere in the genome, but genes involved in neurodevelopment and synaptic function may be especially important. CNVs can either be inherited, or arise de novo in an individual (Levinson et al., 2011). It is unclear how important, overall, CNVs will prove to be compared with other forms of genetic variation (e.g. polymorphisms) in the causation of schizophrenia (Bassett et al., 2010).

Table 11.11 summarizes the current evidence, but it is important to note that this is ‘work in progress.’ Some of the findings may not be replicated, and other genes will undoubtedly emerge. There is also a considerable divergence of opinion as to how the current data should be interpreted (Kim et al., 2011). At present it is not possible to state how many susceptibility genes for schizophrenia exist, nor is their relative importance known.

Crow’s lateralization hypothesis

A very different view of the aetiology of schizophrenia has been proposed by Crow (2002). He argues that schizophrenia is due to a single gene, which is also responsible for two unique properties of the human brain, namely cerebral asymmetry and language. No other genes, or environmental factors, are required. For various reasons, the gene is postulated to reside at a particular part of the sex chromosomes, and its abnormality is one of its epigenetic regulation, not its DNA sequence. Protocadherin 11XY gene is the leading candidate. This theory is ingenious but is not widely supported (O’Donovan et al., 2008).

The biology of schizophrenia genes

Efforts are under way to understand the function of the genes implicated in schizophrenia, and how this effect can be explained biologically. One proposal is that many of the genes may converge upon the development, functioning, and plasticity of synapses, particularly glutamatergic ones (Harrison and Weinberger, 2005). The molecular basis with regard to how and why the variants increase schizophrenia is unknown, in part because, with few exceptions, several (or many) different schizophrenia-associated variants have been reported in each gene, and it is not known which of them is the causal one. Nevertheless, at present it is thought that the genetic variants affect the regulation and expression of the gene in a way that is deleterious for brain development or functioning. Interestingly, forms of each gene (‘isoforms’) which are expressed preferentially in the brain (rather than more widely in the body), and which are enriched in fetal life, seem to be particularly implicated (Kleinman et al., 2011).

To complicate matters, there is increasing evidence that schizophrenia risk genes do not operate independently, but interact with each other (epistasis) in biologically meaningful ways. For example, a recent study of the neuregulin signalling pathway found only weak evidence for individual genes being associated with schizophrenia, but individuals who had risk variants in neuregulin 1, and its receptor, and a downstream protein, had a markedly increased risk of schizophrenia (Nicodemus et al., 2010). Genes may also affect an individual’s vulnerability to environmental risk factors, as discussed below.

Table 11.11 Susceptibility loci and genes for schizophrenia


Environmental risk factors

Despite the prominent genetic component, the twin studies discussed earlier, along with many epidemiological and other studies, show that environmental factors are also important in the aetiology of schizophrenia. A range of factors, especially pre- and perinatal ones, have been identified. This section describes the main biological factors; social influences are covered in a subsequent section. For reviews of environmental risk factors, see van Os et al. (2010) and McGrath and Murray (2011).

Obstetric complications

Rates of schizophrenia are increased in individuals who experienced obstetric complications, compared with their unaffected siblings or normal controls. Meta-analyses suggest an odds ratio of about 2 (Geddes et al., 1999; Cannon et al. 2002). A range of exposures are associated with increased risk, including antepartum haemorrhage, diabetes, low birth weight, asphyxia, and Rhesus incompatibility. Obstetric complications may be more relevant in individuals who have a genetic predisposition to schizophrenia (McNeil et al., 2000), a view which is supported by recent findings of interactions between schizophrenia risk genes and birth complications (Nicodemus et al., 2008).

The association of obstetric complications with schizophrenia has several possible explanations. They might be directly causal (e.g. via fetal hypoxia), or a reflection of pre-existing fetal abnormality, or even a reflection of maternal characteristics (e.g. a mother’s antenatal health behaviour).

Maternal influenza and other infections

Several studies have suggested that fetuses exposed during the second trimester to influenza, especially the 1957 influenza A2 pandemic, have an increased risk of schizophrenia. Some biological support for this association was provided by animal studies which showed that prenatal influenza affects fetal brain development. Additional support emerged with the demonstration that serological evidence of influenza infection during early pregnancy was associated with a sevenfold increase in risk of schizophrenia in the offspring (Brown et al., 2004). However, many other studies, and a meta-analysis of the 1957 pandemic, have yielded negative results (Selten et al., 2010), and a relationship between influenza and schizophrenia has been hard to establish.

Several other maternal (and childhood) infections have also been associated with schizophrenia, including toxoplasmosis, herpes simplex virus 2, and rubella. For a review of this subject, see Brown and Derkits (2010).

Maternal malnutrition

Children born to mothers who experienced famine early in their pregnancy have an increased risk of schizophrenia, with an odds ratio of about 2. This was shown initially in children born after the Dutch ‘Hunger Winter’ of 1944, and subsequently in two large Chinese populations that were exposed to famines in the 1960s. It is not known whether the cause is related to general malnutrition or lack of specific micronutrients, nor is it clear whether less extreme fluctuations in nutritional status during pregnancy are relevant to schizophrenia risk. For a review, see Brown and Susser (2008).

Winter birth

Schizophrenia is slightly more frequent among people born in the winter than among those born in the summer (Davies et al., 2003). It has been shown in both the northern and southern hemispheres, and becomes more prominent at higher latitudes. Winter birth may be more common in patients without a family history of schizophrenia. The explanation for the winter birth effect is unknown. It has been linked to the prevalence of influenza earlier in the winter, and to sunshine and vitamin D levels around the time of birth, but could also be related to the time of conception, via seasonal fluctuations in the genetic make-up of gametes.

Paternal age

A replicated finding is that schizophrenia is associated with paternal age, especially in those without a family history of psychosis. One large study found that the risk increased by almost 50% for each 10-year increase in paternal age (Sipos et al., 2004). The favoured explanation has been that the frequency of de novo genetic mutations in sperm is age related. However, a recent study shows that the phenomenon is only observed for a man’s first-born child, and not for subsequent ones (Petersen et al., 2011), which makes this explanation unlikely. Instead, the paternal age effect may be related to how personality (or relationship factors) modify the age at which a man first fathers a child.

Child development

Parnas et al. (1982) reported a study of 207 children of mothers with schizophrenia who were first assessed when they were 8–12 years of age and then again, as adults, 18 years later, by which time 13 individuals had developed schizophrenia and 29 had ‘borderline schizophrenia.’ Of the measures that were used on the first occasion, those that predicted schizophrenia were poor rapport at interview, being socially isolated from peers, disciplinary problems mentioned in school reports, and parental reports that the person had been passive as a baby.

In a national birth cohort of more than 16 000 children who were prospectively studied over a 16-year period, those who developed schizophrenia could be distinguished at the age of 11 years, if not earlier, by greater hostility towards adults, and by speech and reading difficulties (Done et al., 1994). This difference was apparent compared to those who grew up to develop neurotic illness, as well as those who remained well. In a similar study, Jones et al. (1994) found that children who eventually developed schizophrenia showed delayed milestones and speech problems, together with lower education test scores and less social play. A graded relationship between delayed milestones (e.g. age at walking) and schizophrenia has been confirmed in several subsequent cohorts. Finally, Walker and colleagues have shown that the behaviour of children (including motor acts in infants), as recorded in home movies and videotapes, is also associated with later schizophrenia (Schiffman et al., 2004).

Overall, there is good evidence that individuals who will develop schizophrenia show increased rates of intellectual and motor dysfunction and poor social competence in childhood. However, it is unclear how specific these changes are, or how they may be related to the subsequent development of the illness. Most of the children who were destined to develop schizophrenia were not considered clinically abnormal at the time, and many children who perform poorly on these indices do not develop schizophrenia. Nevertheless, the findings support an early neurodevelopmental contribution to schizophrenia (see p. 282).

Substance use

As discussed earlier, the high prevalence of drug and alcohol use in patients with schizophrenia is well established, but whether substance misuse plays a causal role in schizophrenia is more controversial. If substance use is considered to have directly caused or precipitated the psychosis, it is diagnosed as such, and not as schizophrenia. However, there is evidence that prior use of some drugs, particularly cannabis, is associated with an increased risk of later developing schizophrenia. Andreasson et al. (1987) followed up over 45 000 Swedish conscripts for 15 years, and found that the relative risk of developing schizophrenia was 2.5 times higher in subjects who used cannabis, with a sixfold increase in risk for heavy users. These data could be interpreted in two ways—first, that cannabis misuse is indeed a risk factor for schizophrenia, and secondly, that those predisposed to develop the illness (or who are in the prodrome) tend to misuse cannabis. Recent studies support a genuine causal contribution, with early (and heavy) use of cannabis conferring the greatest excess risk (Moore et al., 2007). Moreover, the risk is greater in those predisposed to psychosis for other reasons, including genetic factors (Caspi et al., 2005; van Winkel et al., 2011). However, the magnitude of the cannabis risk for psychosis, and its implications for public health policy, remain controversial (Hall and Degenhardt, 2011) (see also Chapter 17).

Other risk factors

Many other risk factors have been reported to be associated with schizophrenia. For example, earlier studies suggested that head injury is a risk factor for schizophrenia and other psychoses. However, the evidence for this is weak (David and Prince, 2005). For a review, see McGrath and Murray (2011).


Schizophrenia is at the forefront of attempts to understand psychiatric disorders in terms of alterations in brain structure and function, utilizing the whole range of contemporary neuroscientific concepts and techniques. For reviews of this subject, see Harrison (2009) and van Os and Kapur (2009).

Structural brain changes

Whether there is a neuropathology associated with schizophrenia has been a matter of debate for over a century. The search began with Alois Alzheimer, who spent a decade studying the brains of patients with dementia praecox before he reported the case of presenile dementia with which his name is associated. The failure of Alzheimer and others to identify reliable, characteristic brain changes was central to the view of schizophrenia as a functional disorder rather than an organic one. However, evidence has accrued over the past 30 years which disproves the null hypothesis that there are no structural differences in the brain in schizophrenia, although the details and interpretation of the neuropathology remain poorly understood, and the findings are not clinically useful in the diagnosis of individual patients. The main findings are summarized in Table 11.12.

Structural imaging

In a landmark study, Johnstone et al. (1976) used the novel technique of computerized tomography (CT), and found significantly larger ventricles in 17 elderly hospitalized patients with schizophrenia than in 8 healthy controls. Lateral ventricular enlargement in schizophrenia had been reported decades previously using pneumoencephalography, but it was the study by Johnstone and colleagues which stimulated renewed interest. A large number of subsequent imaging studies, mostly using MRI, have confirmed and extended those findings.

A meta-analysis of MRI studies concluded that there are reliable and significant differences in the volumes of the whole brain, lateral and third ventricles, and hippocampus in individuals with schizophrenia (Wright et al., 2000). The ventricular enlargement, and probably also the other changes, shows a unimodal distribution in patients, indicating that the abnormality is not confined to an ‘organic’ subtype of schizophrenia. There is no clear association with gender. The differences are present in first-episode and unmedicated patients. However, the changes are modest (e.g. the difference in brain volume is about 3%), and with significant overlap between patients and controls. Furthermore, they are not diagnostically specific—for example, some findings are also seen in bipolar disorder (Arnone et al., 2009). There are few clear clinico-pathological associations, although particular structural alterations have been associated with specific symptoms and cognitive deficits—for example, thought disorder has been associated with smaller superior temporal gyri.

Table 11.12 Summary of structural brain changes in schizophrenia

Brain imaging

• Decreased brain volume

• Enlarged lateral ventricles

• Enlarged third ventricle

• Smaller medial temporal lobes

• Decreased cortical grey matter

• Altered gyrification


• Decreased brain weight

• Absence of neurodegenerative changes and of gliosis

• Decreased synaptic markers

• Decreased oligodendroglia

• Decreased markers of some interneurons

• Smaller pyramidal neurons in some areas

• Fewer thalamic neurons

As well as measuring volume, MRI methods are now available for assessing white matter pathways and connectivity between brain regions (e.g. diffusion tensor imaging). A range of abnormalities have been reported in schizophrenia, especially affecting the left frontal and temporal lobes (Ellison-Wright and Bullmore, 2009). Other MRI techniques have also identified differences in cortical thickness, surface area, and sulcogyral patterns.

Longitudinal studies have attempted to identify the course of structural brain changes in schizophrenia. The results are complex and controversial. Some changes are present before the onset of symptoms, while others emerge during the first episode (Smieskova et al., 2010). Thereafter there is some progression (Olabi et al., 2011), but the timing, magnitude, location, and cause remain unclear (Weinberger and McClure, 2002). For example, the extent to which the volume reductions are attributable to medication, and whether there are differences between atypical and typical antipsychotics in this respect, has not been established (Ho et al., 2011).

First-degree relatives of patients with schizophrenia show changes in regional brain volumes that are intermediate between those of patients and unrelated healthy controls, which suggests that some of the brain abnormalities are associated with the genetic predisposition to the disorder (Boos et al., 2007).


Post-mortem neuropathological studies have sought to explain the cellular and molecular basis for the neuroim-aging findings (see Table 11.12). For a review of this subject, see Harrison (1999) and Dorph-Petersen and Lewis (2011). A few findings deserve mention here.

• Brain weight is decreased by 2–3%.

• There is no evidence of any neurodegenerative processes. This supports a neurodevelopmental origin of the pathology, and also argues that any progression of pathology during the illness, as noted in some MRI studies, is not neurotoxic or degenerative in nature.

• The main positive histological findings are alterations in markers of synapses and dendrites, and in specific populations of neurons and glia. These changes, which are present in the cerebral cortex, hippocampus, and thalamus, suggest differences in synaptic connections, and have contributed to the ‘aberrant connectivity’ hypothesis of schizophrenia.

• An abnormal position or clustering of some neurons, notably in the entorhinal cortex and in the subcortical white matter, has been reported in several studies. Such findings are strongly suggestive of a prenatal abnormality in neuronal migration.

• Studies of gene expression have shown differences in families of genes consistent with the involvement of both neuronal and oligodendroglial cells in schizophrenia.

Note that virtually all of the patients whose brains were studied post mortem had been on medication. Therefore treatment may have contributed to the findings; indeed, some experimental data in non-human primates support this possibility.

Functional brain imaging

As described in Chapter 5, a number of imaging techniques, particularly positron emission tomography (PET), single-photon emission tomography (SPET), and functional magnetic resonance imaging (fMRI), have been used to assess patterns of brain activity in schizophrenia. Study designs usually compare groups of patients with groups of controls, either at rest, or during performance of different kinds of cognitive task. For a review, see Meyer-Lindenberg (2010).

Cerebral blood flow

The first study was conducted by Ingvar and Franzen (1974), who used injections of radioactive xenon, and found decreased perfusion of the frontal cortex compared with the posterior regions in chronic, medicated patients with schizophrenia. This ‘hypofrontality’ has subsequently been considered to be a feature of schizophrenia and, although many studies have given negative results, was confirmed in a meta-analysis (Hill et al., 2004). However, stronger relationships with hypofrontality are seen if the phase of illness and symptom profile are taken into account (e.g. the association with psychomotor poverty) (see Table 11.4).

BOLD signal on fMRI

Altered frontal activity is also apparent in fMRI studies of schizophrenia, particularly in the prefrontal cortex during performance of working memory tasks, such as the Wisconsin Card Sorting Test and the N-back task. The profile of changes is complex (Minzenberg et al., 2009). When patients are matched with controls for performance, patients require more frontal cortex activation to achieve the same level of performance as controls, which suggests that there is a reduced ‘efficiency’ of cortical processing (Callicott et al., 2003). Such abnormalities have been related to theories of aberrant connectivity and dopamine in schizophrenia (see below).

Other fMRI studies have looked at the cerebral correlates of specific symptoms—for example, the cortical regions that are activated during auditory hallucinations (Jardri et al., 2011).

Neurophysiological findings

Neurophysiological function in schizophrenia has been examined using a range of methods. For a review of this subject, see Winterer and McCarley (2011).


The electroencephalogram (EEG) in schizophrenia generally shows increased amounts of theta activity, fast activity, and paroxysmal activity. The significance of these findings is not known. Recently, more sophisticated combinations of EEG and event-related potential analyses have been used. This work shows that there is a decreased synchronization or coherence of electrical activity in the prefrontal cortex in schizophrenia, which suggests ‘noisy’ or inefficient cortical processing (Winterer and Weinberger, 2004).

Sensory evoked potentials: P300 and P50

The P300 response is an auditory evoked potential which occurs 300 milliseconds after a subject has identified a target stimulus embedded in a series of irrelevant stimuli. The response provides a measure of auditory information processing. In patients with schizophrenia, and a proportion of their first-degree relatives, the amplitude of the P300 wave is reduced.

Abnormalities in another evoked potential, the P50 wave, have also been reported in patients with schizophrenia and their relatives. Moreover, molecular genetic studies in families with schizophrenia have shown that P50 deficits are linked to the gene coding for a subunit of the nicotinic cholinergic receptor. This suggests that alterations in cholinergic neurotransmission could underlie abnormalities in information processing in schizophrenia. A meta-analysis has confirmed the presence of P300 and P50 deficits in schizophrenia (Bramon et al., 2004).

Eye tracking

Over 50% of patients with schizophrenia show defective performance in tests of eye tracking. A similar abnormality affects about 25% of first-degree relatives. Holzman (2000) proposed that eye-tracking dysfunction is an electrophysiological endophenotype of schizophrenia.

Neurochemical findings

Many neurotransmitter receptors, enzymes, and metabolites have been measured in schizophrenia, and a wide range of abnormalities have been described. Efforts are also being made to link post-mortem findings with functional imaging studies in vivo, and to relate the neuro-chemistry to genes (Lisman et al., 2008). However, for many of the findings that will be described here, their role in the aetiology of schizophrenia is still unclear, and is difficult to distinguish from alterations which are part of the brain’s adaptive or maladaptive response to the illness, including the effects of medication.


Two early lines of research converged and implicated dopamine as having a key role in schizophrenia. The first concerned the effects of amphetamine which, among other actions, releases dopamine at central synapses. Repeated use of amphetamine at high doses can induce a disorder similar to acute schizophrenia in some normal individuals, and acute amphetamine administration worsens psychotic symptoms in people with schizophrenia. The second approach starts from the finding that all antipsychotic drugs are dopamine-receptor antagonists, and that their affinity at dopamine D2 receptors is the property that correlates best with their clinical potency. These observations led to the well-known ‘dopamine hypothesis’ of schizophrenia, which emerged during the early 1970s and has persisted. For an early review, see Seeman et al. (1987).

Despite the prominence and longevity of the theory, evidence that dopamine neurotransmission is abnormal in schizophrenia was hard to obtain, in part because of the difficulty in distinguishing drug effects from disease effects in medicated patients. However, the use of PET and SPET techniques to image dopamine receptors and dopamine synthesis in the brain has now provided good evidence to support the presumed ‘hyperdopaminergia’ in acute schizophrenia. There is a slight increase in dopamine D2 receptors, but the main finding is that the release of dopamine, and the occupancy of D2 receptors by dopamine, are markedly increased. Moreover, the extent of this increase predicts the response to antipsychotic medication. The abnormality develops during the prodrome and is not seen in patients during remission, which suggests that excess dopamine release is directly related to acute psychotic symptoms—the ‘wind of the psychotic fire’ (Laruelle and Abi-Dargham, 1999). A recent study suggests that, contrary to expectations, the excess dopamine function occurs primarily in the associative and not the ventral striatum (i.e. in the mesocortical rather than the mesolimbic dopamine pathway). For a review, see Guillin et al. (2007) and Howes and Kapur (2009).

Other phases and aspects of schizophrenia may be associated with different abnormalities in dopamine function (Howes and Kapur, 2009). In particular, various lines of evidence suggest that there is deficient dopaminergic activation of the prefrontal cortex, relevant to the enduring cognitive deficits of the illness. There are plausible pathways and mechanisms to link these opposing dopaminergic abnormalities.

The mechanisms of dopaminergic involvement in schizophrenia are unclear. There may be a genetic component, but current theories emphasize dysregulation secondary to glutamatergic and developmental abnormalities, as outlined below. A phenomenological approach has been proposed by Kapur (2003). He argues that the known roles of dopamine in cognition and behaviour suggest that it mediates the ‘salience’ of external events and internal representations, and that in schizophrenia the dopamine abnormalities lead the patient to misattribute stimuli and their meaning. Delusions and hallucinations are viewed as the consequence of these experiences and the patient’s attempt to make sense of them.


The amino-acid neurotransmitter glutamate is now second only to dopamine in neurochemical theories of schizophrenia. For a review, see Krystal and Moghaddam (2011). The key finding which triggered this interest was that antagonists of the N-methyl-D-aspartate (NMDA) type of glutamate receptor (e.g. phencyclidine and ketamine) can induce a schizophrenia-like psychosis (Javitt and Zukin, 1991). This finding was complemented by several other lines of evidence (see Table 11.13). For example, pro-glutamatergic drugs, including positive NMDA-receptor modulators, may have antipsychotic efficacy.

The origins of glutamatergic involvement in schizophrenia are not known. The ‘NMDA receptor hypofunction’ model postulates a developmental abnormality in the receptor (Olney and Farber, 1995). There may be a genetic component (see p. 273), and an autoimmune subgroup due to anti-NDMA receptor antibodies (Lennox and Vincent, 2012).

Glutamate and dopamine are intricately linked, and their dysfunction in schizophrenia is also thought to reflect shared factors. Glutamate is generally considered to be the more primary or ‘trait’ abnormality, with the dopamine abnormality downstream and ‘state’ related. Alternatively, there may be differential involvement of glutamate and dopamine in different patients (Coyle, 2006; Stone et al., 2007).

Table 11.13 Evidence for glutamatergic involvement in schizophrenia

NMDA-receptor antagonists induce or exacerbate schizophrenia symptoms

NMDA-receptor co-agonists reduce some symptoms

Alterations in glutamatergic molecules (e.g. receptors) in schizophrenia brain

Altered levels of glutamate in schizophrenia brain shown using spectroscopy

Schizophrenia risk genes affect the glutamate system

Animals with modifications of glutamate genes or glutamate function show schizophrenia-like phenotypes

Gamma-aminobutyric acid (GABA)

GABA is the major inhibitory transmitter in the brain, and it is implicated in schizophrenia for several reasons (Krystal and Moghaddam, 2011). First, there is a decrease in its synthetic enzyme, glutamic acid decarboxylase (GAD), in the cerebral cortex. Secondly, the density of a particular GABAergic neuron type (those which contain parvalbumin) and their synaptic terminals are reduced. Thirdly, there are alterations in the expression of GABAA receptors. Taken together, these findings indicate impairment of GABA signalling in schizophrenia, and have led to therapeutic approaches (Lewis and Sweet, 2009).

Serotonin (5-HT)

A possible role of serotonin (5-HT) in schizophrenia has long been considered, because the hallucinogen, lysergic acid diethylamide (LSD), is an agonist at 5-HT2 receptors. Other evidence which suggests such a role is that 5-HT2-receptor antagonism may contribute to the atypical profile of some antipsychotics, and alterations in 5-HT receptor densities in schizophrenia, notably a reduction in the number of frontal cortex 5-HT2A receptors and an increase in 5-HT1A receptors. The possible involvement of 5-HT in schizophrenia has been related to its roles in neurodevelopment and the many interactions between 5-HT, dopamine, and glutamate (Carlsson et al., 2001).

Social and psychosocial factors

In recent decades, social and psychosocial factors in schizophrenia have been neglected relative to neurobiological ones. However, recent studies highlight their importance, and emphasize that there are major interactions between biological and social aspects of the disorder. For a review, see van Os et al. (2010) and Bebbington and Kuipers (2011).

Occupation and social class

Schizophrenia is over-represented among people of lower social class. In Chicago, Hollingshead and Redlich (1958) found both the incidence and the prevalence of schizophrenia to be highest in the lowest socio-economic groups. At first, these findings were thought to be of aetiological significance, but they could be a consequence of schizophrenia. For instance, Goldberg and Morrison (1963) found that people with schizophrenia were of lower social status than their fathers, and that this was usually because they had changed status after the onset of the illness. However, Castle et al. (1993) found that, compared with controls, patients with schizophrenia were more likely to have been born into socially deprived households. Those authors proposed that some environmental factor of aetiological importance was more likely to affect individuals of lower socio-economic status.

Place of residence

Faris and Dunham (1939) studied the place of residence of mentally ill people in Chicago, and found that schizophrenics were over-represented in the disadvantaged inner-city areas. This distribution has been confirmed in other cities, and it has been suggested that unsatisfactory living conditions can cause schizophrenia. These findings have often been ascribed to the occupational and social decline described above, or to a search for social isolation by people who are about to develop schizophrenia.

However, recent data suggest that schizophrenia is associated with place of birth and upbringing, findings which cannot be explained as a consequence of illness. Specifically, population-based studies in several countries show that urban birth is associated with an increased risk of schizophrenia. Larger cities carry a higher odds ratio than small towns or suburban areas (Pedersen and Mortensen, 2001). The cause of the association remains unclear, and may relate to social deprivation, migration, infections, stress, or interactions between genetic vulnerability and urban environments (Peen and Dekker, 2004).

Migration and ethnicity

High rates of schizophrenia have been reported among migrants. In a study of Norwegians who had migrated to Minnesota, Ødegaard (1932) found that the inception rate for schizophrenia was twice that of Norwegians living in Norway. The high rates have been attributed mainly to a disproportionate migration of people who are unsettled because they are becoming mentally ill. The effects of a new environment may also play a part in provoking illness in predisposed individuals. Thus ‘social selection’ and ‘social causation’ may both contribute to an excess of schizophrenia among migrants.

A meta-analysis of 18 studies has confirmed that migration is a risk factor for schizophrenia, and that this cannot be explained solely by selection (Cantor-Grae and Selten, 2005). The relative risk for migrants is 2.7 (95% confidence interval, 2.3–3.2), with a higher risk for the children of migrants (4.5, 95% confidence interval, 1.5–13.1). Risks were also higher for migrants from developing countries.

A particularly controversial aspect of migration and schizophrenia concerns the Afro-Caribbean population in the UK. This group has a strikingly increased incidence of schizophrenia, particularly in the ‘second generation’, who were born in the UK. The relative risk for all psychosis ascertained in the multi-centre first-onset AESOP study was 1.5 for the Asian population and 6.7 for the Afro-Caribbean population, compared with the white population in the UK (Fearon et al., 2006). The relative risk increased to 9.1 for the Afro-Caribbean group when the diagnosis was restricted to ‘narrow schizophrenia.’ Concern about these figures is compounded by the increased rates of compulsory admission among the black population. However, rates in the Caribbean are not appreciably increased, even among the families of origin of ill migrants (Bhugra et al., 1996).

There has been concern that the high rates represent misdiagnosis because of poor diagnostic practice by UK psychiatrists, and even because of ‘institutional racism’ (Singh and Burns, 2006). However, there is now strong evidence that this is not the case, and that the raised incidence is a real phenomenon (Morgan et al., 2010). Several studies have compared the diagnoses in Afro-Caribbean and white patients using structured research instruments, and found them to be equivalent (Singh and Burns, 2006). The most rigorous test has been the AESOP study in which, despite the raised rates, symptom profiles were broadly the same across the ethnic groups, and the association with known risk factors was also similar. Tellingly, an as yet unpublished 10-year follow-up of over 300 subjects from the AESOP sample has found identical clinical and social outcomes in the ethnic groups. Had the diagnostic threshold been lower in the Afro-Caribbean patients, the outcome would have been predicted to be substantially better.

Experiential factors must play a significant part in these different rates. Studies in London (Boydell et al., 2001) and in The Hague in the Netherlands (Veling et al., 2008) have confirmed that the high incidence in ethnic-minority patients varies according to where they grew up in the host country. Controlling for known risk factors (e.g. cannabis use and deprivation), the rates are still more than doubled when these individuals grow up in areas where they are members of a small minority as opposed to areas where they are a relatively large minority. How this differential ‘ethnic density’ interacts with rates of schizophrenia is unclear, but it is unlikely simply to be an association.

Social isolation

People with schizophrenia often live alone, unmarried, and with few friends. The findings reviewed above suggest that this pattern of isolation may begin before the illness, sometimes in early childhood. Of course, some patients may choose to isolate themselves as a way of decreasing social stimulation. In this sense social isolation, although undesirable in a general sense, is not a straightforward stressor.

Life events and difficulties

Life events and difficulties have often been proposed as precipitants of schizophrenia, but few satisfactory studies have been carried out. In one of the most convincing studies, Brown and Birley (1968) used a standardized procedure to collect information from 50 patients newly admitted with a precisely datable first onset or relapse of schizophrenia. By comparison with a control group, the rate of ‘independent’ events in those with schizophrenia was increased during the 3 weeks before the onset of the acute symptoms. When the events (which included moving house, starting or losing a job, and domestic crises) were compared with events preceding depression, neurosis, and suicide attempts, they were found to be nonspecific. Paykel (1978) calculated that experiencing a life event doubles the risk of developing schizophrenia during the subsequent 6 months. Others have confirmed these findings for both first episodes and relapse of schizophrenia. In patients with chronic schizophrenia, the level of symptoms over time correlates with life events. However, there is little evidence that patients with schizophrenia experience more life events than the general population.

For a review, see Bebbington and Kuipers (2011).

Childhood trauma

A high prevalence of childhood trauma and abuse has been reported in individuals who develop schizophrenia (Morgan and Fisher, 2007). It is unclear whether this is a causal relationship.

Psychological factors

Personality factors

Several early writers, including Bleuler (1911), commented on the frequency of abnormalities of personality preceding the onset of schizophrenia. Kretschmer (1936) proposed that both personality and schizophrenia were related to the asthenic type of body build. He suggested a continuous variation between normal personality, schizoid personality, and schizophrenia. He regarded schizoid personality as a partial expression of the psychological abnormalities that are manifested in their full form in schizophrenia. Such ideas must be treated with caution, as it is difficult to distinguish between premorbid personality and the prodromal phase of slowly developing illness. However, Kretschmer’s ideas are similar to the current concept of the schizophrenia spectrum, in which schizophrenia represents the most severe end of a continuum. Taken together, the findings suggest that abnormal personality features are not uncommon among people who later develop schizophrenia, and among their first-degree relatives. However, many people with schizophrenia have no obvious disorder of personality before the onset of the illness, and only a minority of people with schizotypal or schizoid personalities develop schizophrenia.

Neuropsychological factors

As noted above, it is now well established that patients with schizophrenia have widespread cognitive deficits, particularly in tasks that involve learning and memory. Here we consider psychological theories of schizophrenia.

A well-known neuropsychological model of schizophrenia is that of Gray et al. (1991), who proposed that positive symptoms arise from a failure to integrate stored memories with current stimuli. The theory was unusual at the time, in attempting to explain psychological features in terms of limbic circuits and dopamine projections. The aberrant salience model of Kapur (2003), mentioned earlier, shares some features with Gray’s model.

Frith (1996) argued that in schizophrenia there is a breakdown in the internal representation of mental events. For example, a failure to monitor and identify one’s own willed intentions might give rise to the idea that thoughts and actions arise from external sources, resulting in delusions of control.

The concept of social cognition—the role of cognitive factors in the impaired social functioning that occurs in schizophrenia—has become prominent recently. Abnormalities of emotional processing, face recognition, and theory of mind are postulated to be important, and to result from the involvement of neural circuits implicating the frontal cortex and amgydala. There is good evidence for theory of mind impairment (Bora et al., 2009), and also that social cognition interacts with other aspects of cognitive impairment (mentioned earlier) to predict functional outcome (Fett et al., 2011).

Dynamic and interpersonal factors

Psychodynamic theories of schizophrenia are largely of historical interest (see Box 11.4). However, they remain important both because they focus attention on interpersonal and developmental aspects, and because some carers and health professionals continue to regard them as relevant. For a review, see Jackson and Cawley (1992).

Box 11.4 Psychodynamic and family theories of aetiology

Psychodynamic theories

Freud’s theory of schizophrenia was stated most clearly in his 1911 analysis of the Judge Schreber case and in his 1914 paper ‘On narcissism: an introduction.’ According to Freud, in the first stage, libido was withdrawn from external objects and attached to the ego. The result was exaggerated self-importance. Since the withdrawal of libido made the external world meaningless, the patient attempted to restore meaning by developing abnormal beliefs. Because of libidinal withdrawal, the patient could not form a transference and therefore could not be treated by psychoanalysis. Although Freud developed his general ideas considerably after 1914, he elaborated his original theory of schizophrenia but did not replace it.

Melanie Klein (1952) believed that schizophrenia originated in infancy. In the ‘paranoid schizoid position’, the infant dealt with innate aggressive impulses by splitting both his own ego and his representation of his mother into two incompatible parts, one wholly bad and the other wholly good. Only later did the child realize that the same person could be good at one time and bad at another. Failure to pass through this stage adequately was the basis for the later development of schizophrenia.

The family as a cause of schizophrenia

Two kinds of theory have been proposed about the family as a cause of schizophrenia, namely deviant role relationships (associated with Fromm-Reichmann and Lidz) and disordered communication(associated with Bateson and Wynne). The different role of the family in influencing the course of schizophrenia is discussed later.

The concept of the ‘schizophrenogenic’ mother was suggested by the analyst Fromm-Reichmann in 1948. Lidz and his colleagues used intensive psychoanalytical methods to study the families of 17 patients with schizophrenia, of whom 14 were in social classes I or II (Lidz et al, 1965). Two types of abnormal family pattern were reported:

• marital skew, in which one parent yielded to the other’s (usually the mother’s) eccentricities, which dominated the family

• marital schism, in which the parents maintained contrary views, and consequently the child had divided loyalties. It was suggested that these abnormalities were the cause rather than the result of the schizophrenia.

Investigations by other clinicians have not confirmed these findings. Even if they were confirmed, the abnormalities in the parents could be an expression of genetic causes, or secondary to the disorder in the patient. These and other speculations about the causative role of family relationships have had the unfortunate consequence of inducing unjustified guilt in parents.

Research on disordered communication in families originated from the idea of the double bind (Bateson et al, 1956). A double bind is said to occur when an instruction is given overtly, but is contradicted by a second more covert instruction. For example, a mother may overtly tell her child to come to her, while conveying by her manner and tone of voice that she rejects him. Double binds are only pathogenic within families that implicitly prohibit the acknowledgement of this contradiction. They are, after all, far from uncommon. According to Bateson, double binds leave the child able to make only ambiguous or meaningless responses, and schizophrenia develops when this process persists. The theory is ingenious but not supported by evidence.

Studies of communication in families of patients with schizophrenia gave rather conflicting results (Singer and Wynne, 1965; Hirsch and Leff, 1975). However, it is worth noting that an association between abnormal social communication in parents and schizophrenic illness in children may in fact be a consequence of a shared genetic inheritance—for example, because the incidence of schizotypal disorder (characterized by abnormalities of thought and speech) is raised in first-degree relatives of patients with schizophrenia.

Current aetiological hypotheses about schizophrenia

Several models of schizophrenia (e.g. concerning genes, dopamine, and aberrant salience) have already been mentioned. Here we note three other current hypotheses, which have emerged from or contributed to the aetiological considerations discussed above.

Neurodevelopmental hypothesis

The leading hypothesis is that schizophrenia is a disorder of neurodevelopment. Original suggestions of this kind were made by Clouston in 1892 and by others, including Kraepelin, at the turn of the last century, but the current view can be traced to Murray and Lewis (1987) and Weinberger (1987). The hypothesis is simple—schizophrenia results from an abnormality in some aspect of brain development. This is implicitly contrasted with the alternative—that schizophrenia is a neurodegenerative disorder (McClure and Lieberman, 2003). Specific forms of the neurodevelopmental hypothesis implicate the prenatal period, adolescence, or an interaction between the two. Evidence in favour of the model has accrued from several sources (Lewis and Levitt, 2002) and is summarized in Table 11.14. Many of the factors have already been discussed. The absence of gliosis (the proliferation and hypertrophy of astrocytes) is important, as it argues for a prenatal (second-trimester) timing of the neuropathology.

There are no serious challenges to the neurodevelop-mental hypothesis as a pathogenic model of schizophrenia. This partly reflects its vagueness, which allows most findings to be interpreted as being consistent with it. Its other weaknesses include a failure to readily explain the onset or course of the disorder (Broome et al., 2005). For a recent review, see Weinberger and Levitt (2011).

Aberrant connectivity

Eugen Bleuler conceptualized the fundamental symptoms of schizophrenia as reflecting ‘psychic splitting’ or a failure of integration of mental functions. This view is currently framed in terms of abnormal connectivity, whereby the activity of different brain regions or circuits is aberrant in schizophrenia (Andreasen, 1999; Stephan et al., 2009). Some variants of this model also involve a structural component (i.e. changes in the ‘wiring’ of the brain), consistent with the structural imaging and neuro-pathology findings mentioned earlier. The latest models focus on brain oscillations and network activity—that is, the coordinated firing of ensembles of neurons between brain areas, thought to underlie cognition and other brain operations (Woo et al., 2010). Most connectivity models of schizophrenia assume that the abnormality arises developmentally, but this need not be the case.

Stress-vulnerability model

This model (also called the stress-diathesis model) incorporates development, but emphasizes the interaction of the early events (whether genetic, biological, or social) with later stressors (e.g. substance use, life events). The early factors confer vulnerability to schizophrenia, whereas the later ones explain its onset and course (Nuechterlein and Dawson, 1984). The emphasis on interactions between aetiological factors is also appropriate, as an increasing number of studies, including many of those mentioned above, indicate that environmental factors act primarily upon individuals who are already at risk, for genetic or other reasons (van Os et al., 2010). Future studies will need to be very large if such interactions are to be reliably identified.

Table 11.14 Findings that support the neurodevelopmental hypothesis of schizophrenia

Structural brain differences present at or before illness onset

Limited progression of structural brain changes after onset

Cognitive and social impairments in childhood

Neuropathological changes without gliosis

‘Soft’ neurological signs at presentation

Minor physical anomalies and aberrant dermatoglyphics

Pre- and perinatal risk factors

Increased frequency of large cavum septum pellucidum

Schizophrenia risk genes affect neurodevelopment

Animal models show delayed effects of early brain lesions

Course and prognosis

Kraepelin initially believed that dementia praecox had an invariably poor outcome, although he later reported that in the long term, 17% of his patients were socially well-adjusted. Although it remains generally agreed that the outcome of schizophrenia is worse than that of most psychiatric disorders, it is difficult to draw clear conclusions.

An important long-term study was carried out by Manfred Bleuler (1974), son of Eugen Bleuler, who personally followed up 208 patients who had been admitted to hospital in Switzerland between 1942 and 1943. Twenty years after admission, 20% of these patients had a complete remission of symptoms, 35% had a good outcome in terms of social adjustment, and 24% were severely disturbed. When full recovery had occurred, this was usually within the first 2 years. When the illness was recurrent, each subsequent episode usually resembled the first one in its clinical features. Bleuler’s diagnostic criteria were narrow, and his findings suggest that the traditional view of schizophrenia as a generally progressive and disabling condition must be reconsidered. Nevertheless, 10% of his patients suffered an illness of such severity that they required long-term sheltered care.

Bleuler’s conclusions are broadly supported by Luc Ciompi’s larger but less detailed study of long-term outcome in Lausanne (Ciompi, 1980). The study was based on the well-kept records of 1642 patients diagnosed as having schizophrenia from the beginning of the century to 1962. The average follow-up was 37 years. One-third of the patients were found to have a good or fair social outcome. Symptoms often became less severe in the later years of life. More recent studies are in general agreement with these findings. A 15- and 25-year follow-up study of patient cohorts in 15 countries found that one-sixth of the patients had achieved full recovery. Importantly, late recovery was seen in a significant minority, challenging the therapeutic pessimism that such improvements are very rare (Harrison et al., 2001).

Whether the outcome of schizophrenia has improved since Kraepelin’s era remains unclear. A meta-analysis found a ‘recovery rate’ of 35% prior to 1955, and 49% for the period 1956–1985, a difference that may reflect the use of antipsychotic drugs. However, the rate fell back to 36% between 1986 and 1992 (Hegarty et al., 1994). Any apparent time trends must be interpreted very cautiously because of changes in diagnostic practice and outcome measures during the twentieth century.

For a review of the course and clinical outcome of schizophrenia, see an der Heiden and Häfner (2011). Kooyman and Walsh (2011) have reviewed the societal outcomes.

Physical health, mortality, and suicide in schizophrenia

All studies with prolonged follow-up report an increased mortality in patients with schizophrenia. In a meta-analysis of 36 000 patients, Harris and Barraclough (1998) found that the risk of death from all causes was increased by 1.6-fold. More recent systematic reviews have cited even higher figures, and some suggestion that the ‘mortality gap’ is increasing (Saha et al., 2007). About 60% of the excess mortality is accounted for by unnatural causes, mainly suicide, and the rest from natural causes, especially cardiovascular disease, presumably reflecting the high rates of smoking, as well as poor diets, sedentary lifestyle, and higher rates of obesity, type 2 diabetes, and other physical illnesses (Leucht et al., 2007). Medication, especially atypical antipsychotics, may contribute significantly to this excess mortality, due to metabolic syndrome and other adverse effects. However, whether medication overall lowers or increases mortality in schizophrenia remains controversial (De Hert et al., 2010). As an aside, mortality from some illnesses, including some cancers, is lower than expected in individuals with schizophrenia, and in their relatives (Catts et al., 2008). The mechanisms involved are not known.

The lifetime risk of suicide in schizophrenia is usually quoted to be around 10% (partly based on the study by Harris and Barraclough, 1998). However, overall, the evidence is for a lower rate (around 5%) (Hor and Taylor, 2010). The highest risk occurs soon after diagnosis (Dutta et al., 2010).

Predictors of outcome

The outcome of schizophrenia is remarkably heterogeneous and largely unpredictable (an der Heiden and Häfner, 2011). For example, in the International Pilot Study of Schizophrenia (World Health Organization, 1973), tests were made of the predictive value of several sets of criteria based on symptoms. All of them proved largely unsuccessful in predicting outcome at 2 or 5 years. The best predictors of poor outcome appear to be the criteria used for diagnosis in DSM-IV, probably because they stipulate that the syndrome should have been present for 6 months before the diagnosis can be made.

Poor outcome in schizophrenia is associated with younger age of onset, male sex, poor premorbid functioning, and persistence of negative symptoms. There is also evidence that the duration of psychotic symptoms prior to treatment correlates with time to remission and level of remission (Marshall et al., 2005). Although reluctance to seek treatment could be associated with other predictors of poor outcome, it is also possible that early treatment, either with drugs or with psychological therapies, may improve the prognosis (see p. 294).

Some of the predictors that have emerged from the various studies are listed in Table 11.15, and may be taken as a moderately useful guide. However, it is wise to be cautious when asked to predict the outcome of individual cases.

So far this discussion has been concerned with factors that operate before or at the onset of schizophrenia. An account will now be given of factors that exert their effect after the illness has become established.

Geographical variation in course

As noted above, international studies suggest that the incidence of schizophrenia is broadly similar in different countries. However, the course and outcome are not. In the International Pilot Study of Schizophrenia (World Health Organization, 1973), the 2-year outcome was better in India, Colombia, and Nigeria than in the other centres. This finding could not be explained by any recorded differences in the initial characteristics of the patients. The possibility of selection bias remains. For example, in these three countries, patients with acute illness may be more likely to be taken to hospital than patients with illness of insidious onset. However, later studies designed to overcome these objections also found a more favourable course of illness in less developed countries. The major difference in outcome was that patients in developing countries were more likely to achieve a complete remission than those in developed countries, who tended to be impaired by continual residual symptoms.

Table 11.15 Factors that predict a poor outcome of schizophrenia

Early onset

Insidious onset

Long duration of untreated psychosis

Previous psychiatric history

Hebephrenic subtype

Negative symptoms

Enlarged lateral ventricles

Male gender

Single, separated, widowed, or divorced

Poor psychosexual adjustment

Abnormal previous personality

Poor work record

Social isolation

Poor compliance

Substance misuse

The better outcome of schizophrenia in developing countries is unexplained. It does not appear to be due to differences in levels of expressed emotion; a 15- and 25-year follow-up study confirmed the stability of geographical differences in course, and suggested that a difference in the severity and type of illness at first presentation is an important contributory factor (Harrison et al., 2001).

Life events

As explained above, some patients experience an excess of life events in the weeks before the onset of acute symptoms of schizophrenia. This applies not only to first illnesses but also to relapses (Bebbington and Kuipers, 2011). In addition, patients with increased numbers of life events experience a more symptomatic course.

Social stimulation

In the 1940s and 1950s, clinicians recognized that among schizophrenics living in institutions, many clinical features were associated with an unstimulating environment. Wing and Brown (1970) investigated patients at three mental hospitals. One was a traditional institution, another had an active rehabilitation programme, and the third had a reputation for progressive policies and short admissions. The research team devised a measure of ‘poverty of the social milieu’ which took into account little contact with the outside world, few personal possessions, lack of constructive occupation, and pessimistic expectations on the part of ward staff. Poverty of social milieu was found to be closely related to three aspects of the patients’ clinical condition, namely social withdrawal, blunting of affect, and poverty of speech. The causal significance of these social conditions was strongly supported by a further survey 4 years later; improvements had taken place in the environment of the hospitals, and these changes were accompanied by corresponding improvements in the three aspects of the patients’ clinical state.

Although an under-stimulating hospital environment is harmful, an over-stimulating environment can precipitate positive symptoms and lead to relapse. Since factors in the hospital environment play an important part in determining the prognosis, it seems likely that similar factors are important to patients living in the community.

Family life and expressed emotion

Brown et al. (1958) found that, on discharge from hospital, patients with schizophrenia who were returning to their families generally had a worse prognosis than those entering hostels. Brown et al. (1962) found that relapse rates were greater in families where the patient’s relatives showed high expressed emotion (‘high EE’) by making critical comments, expressing hostility, and showing signs of emotional over-involvement. In such families the risk of relapse was greater if the patients were in contact with their close relatives for more than 35 hours a week. These findings were confirmed and extended when Leff and Vaughn (1981) investigated the interaction between expressed emotion in relatives and life events during the 3 months before relapse. The onset of illness was associated either with a high level of expressed emotion or with an independent life event.

Vaughn and Leff (1976) also suggested an association between expressed emotion in relatives and the patient’s response to antipsychotic medication. Among patients who were spending more than 35 hours a week in contact with relatives who showed high expressed emotion, the relapse rate was 92% for those not taking antipsychotics and only 53% for those who were. Patients who were taking antipsychotic drugs and spending less than 35 hours in contact with relatives who showed high expressed emotion had a 15% relapse rate.

Further studies (Leff et al., 1985) strongly suggested that high expressed emotion has a causal role. A total of 24 families were selected in which a schizophrenic patient had extended contact with relatives showing high expressed emotion. All of the patients were on medication. Half of the families were randomly assigned to routine outpatient care. The other half took part in a programme that included education about schizophrenia, relatives’ groups, and family sessions for relatives and patients. The relapse rate was significantly lower in this group than in the controls.

Expressed emotion has now been widely investigated as a predictor of psychotic relapse, and overall it appears that patients living in families with high levels of expressed emotion have a two- to threefold increased risk. An associated finding is that relatives with low levels of expressed emotion are more likely to regard schizophrenia and its associated behaviour as an illness that is beyond the control of the patient, and the nature of these causal attributions is also a predictor of outcome (Barrowclough et al., 1994).

Effects of schizophrenia on the family

With the increasing care of patients in the community rather than in hospital, difficulties have arisen for some families. Relatives of patients with schizophrenia describe two main kinds of problems. The first kind relates to social withdrawal. Patients tend not to interact with other family members; they seem slow, lack conversation, have few interests, and neglect themselves. The second kind of problem relates to more obviously disturbed or socially inappropriate behaviour, and threats of violence. These problems are likely to play a part in the development of high expressed emotion in relatives, which in turn may worsen the situation.

Relatives of patients with schizophrenia may feel anxious, depressed, guilty, or bewildered. They may be understandably uncertain about how to deal with difficult and odd behaviour. Further difficulties can arise from differences in opinion between family members, and from a lack of understanding and sympathy among neighbours and friends. The effects on the lives of relatives and families are often serious. Unfortunately, community services and support for patients with schizophrenia and their relatives are often less than adequate. Improving help, advice, and services to carers of patients is a priority of the UK National Service Framework for Mental Health (Department of Health, 1999).

Working with patients with schizophrenia and their carers is discussed further in the section on management, and in Chapter 21.


The treatment of schizophrenia is concerned with both the acute illness and chronic disability. In general, the best results are obtained by combining antipsychotic drugs and certain psychosocial treatments. This section is concerned with evidence for the efficacy of the individual treatments. The section on management deals with the use of these treatments in clinical practice. Pharmacological aspects of antipsychotic drugs are discussed in Chapter 19.

Pharmacological treatments

For recent guidelines, see Buchanan et al. (2010) and Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology (2011) (and see also Table 11.16).

Acute schizophrenia

The effectiveness of antipsychotic medication in the treatment of acute schizophrenia is well established. The National Institute of Mental Health Collaborative Project (Cole et al., 1964) compared chlorpromazine, fluphenazine and thioridazine with placebo. Around 75% of the patients who received antipsychotic treatment for 6 weeks improved, whereas around 50% of those who received placebo worsened. In an early meta-analysis, Thornley et al. (1997) reviewed 42 randomized acute trials of chlorpromazine. Relative to placebo, chlorpromazine was more likely to lead to global clinical improvement at 8 weeks (55% vs. 37%, NNT = 6). More recently, Leucht et al. (2009a) undertook a meta-analysis of 38 trials of atypical antipsychotics and showed the same NNT (i.e. 6) and an effect size of 0.5 compared with placebo.

Table 11.16 Key recommendations in the pharmacological treatment of schizophrenia

Acute episode

• Initiate antipsychotic medication at lower end of the licensed dose range.

• Do not use loading doses (‘rapid neuroleptization’). If sedation is needed, use benzodiazepines.

• The choice of drug should be based on patient preference, effects of previous treatments, relative liability of the drug to cause serious side-effects (especially extrapyramidal and metabolic syndromes), and relevant medical history.

• Titrate dose within licensed range, monitoring for effects and side-effects.

• Aim to achieve optimum dose with good adherence for 4 weeks.

• Record the indications for medication, the anticipated benefits and time course, and discussions with patient and carers.

Maintenance and relapse prevention

• Continue medication, using the same principles as for the acute episode.

• Ensure that dose, duration, and adherence are adequate before switching drug.

• Drug withdrawal should be gradual, and the mental state should be monitored.

• Continuous treatment is more effective than intermittent treatment.

• Monitor adherence regularly.

• Consider depot formulations, especially if adherence is a problem.

Adapted from Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology (2011).

Antipsychotics do not have a delayed onset of action, as is sometimes believed (Agid et al., 2003). Indeed, improvement in psychotic symptoms, as well as sedation, can be detected within 24 hours (Kapur et al., 2005). Although a meaningful therapeutic response takes longer to develop, the bulk of the improvement occurs within 1 month (Agid et al., 2003). Indeed, if there has not been a 20–25% improvement in symptom score after 2 weeks, the chances of a later response to that drug and dose are small (Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology, 2011).

Drug treatment has most effect on the positive symptoms of schizophrenia, and least (or no) effect on negative and cognitive symptoms. This has implications for the management of schizophrenia, as will be discussed later in this chapter.

Treatment after the acute phase

Since the original demonstration by Pasmanick et al. (1964), many controlled trials have shown the effectiveness of continued antipsychotic therapy in preventing relapse. In a systematic review, Thornley et al. (2003) showed that chlorpromazine reduces the risk of relapse over 6–24 months (relative risk = 0.65, NNT = 3) and improves global functioning (relative risk = 0.76, NNT = 7). Intermittent treatment, where antipsychotics are given at the first sign of relapse, is less effective than continuous prophylaxis.

Some patients with chronic schizophrenia do not respond even to long-term medication, whereas other patients remain well without drugs. Unfortunately, there has been no success in predicting which patients benefit from longer-term drug treatment, and there is no good evidence as to how long such prophylaxis should continue. The general consensus is that even long periods of remission do not protect patients from relapse once medication is withdrawn (Cunningham Owens and Johnstone, 2009). Whenever antipsychotic treatment is discontinued, withdrawal should be carried out slowly, and with careful monitoring of the patient’s mental state for several months afterwards.

There is a widespread clinical impression that depot injections are more successful than continued oral medication in preventing relapse, presumably due to improved compliance. A meta-analysis has confirmed this impression, with a relative risk reduction of about 30% (Leucht et al., 2011). However, a large recent trial of long-acting risperidone versus oral antipsychotics did not show superiority over a 2-year follow-up (Rosenheck et al., 2011).

Differences between antipsychotic drugs

There are no clear differences in efficacy between one typical antipsychotic and another, nor—with the exception of clozapine—are there any obvious differences in efficacy between one atypical antipsychotic and another. Whether there are any such differences between typical and atypical antipsychotics has been more controversial. Meta-analyses of acute treatment trials differ in their conclusions as to whether atypical antipsychotics have greater efficacy or greater tolerability and, if so, the size and clinical relevance of the benefit. For example, Geddes et al. (2000) concluded that any benefits were insignificant, and two subsequent large pragmatic (and non-industry-funded) trials—the CATIE study in the USA (Lieberman et al., 2005) and the CuTLASS study in the UK (Jones et al., 2006) similarly found minimal if any differences. However, in an updated meta-analysis, Leucht et al. (2009b) reported a statistically significant (but still limited) greater efficacy for olanzapine, risperi-done, amisulpride (and clozapine) compared with typical antipsychotics. There is agreement across meta-analyses that extrapyramidal side-effects are less common with the newer agents, but this benefit is set against the greater risks of metabolic syndrome (see p. 527).

A similar conclusion can be drawn with regard to the comparative efficacy of antipsychotics in the prevention of relapse in schizophrenia (Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology, 2011). That is, apart from clozapine, there are no robust differences in efficacy between antipsychotics in the maintenance treatment of schizophrenia.

Reflecting these conclusions, current guidelines do not advocate one class over another, but emphasize the individual drug differences in side-effects and tolerability (see below), and the need to discuss the choice of medication with the patient (National Institute for Health and Clinical Excellence, 2009b; Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology, 2011). These issues are described in the section on management.

Treatment-resistant schizophrenia

About 30% of patients do not respond to antipsychotics, or are intolerant of them. The only proven drug intervention for this group is clozapine, which is effective in about one-third of such patients, as shown by the key trial by Kane et al. (1988), and confirmed in subsequent studies and meta-analyses (Leucht et al., 2009a,b). There is no evidence that any other atypical antipsychotic shares this greater efficacy. Clozapine may also have benefits with regard to suicide risk, aggression, and substance misuse (Farooq and Taylor, 2011). The use of clozapine is hampered by the need for regular blood monitoring and the risks of agranulocytosis, as well as by other side-effects, notably weight gain, sedation, and hypersalivation. Nevertheless, clozapine should be recommended to all patients who do not respond to or cannot tolerate at least two other antipsychotics. In practice, only a minority of suitable patients have a trial of the drug, and often this does not occur until many years into the illness (Farooq and Taylor, 2011).

There is little evidence-based guidance on the appropriate drug treatment of patients who are unresponsive to, or unable to take, clozapine. There are two common augmentation strategies for which there is some evidence. The first is to add an antipsychotic that has a high affinity for the dopamine D2 receptor, such as, amisul-pride; there is also emerging evidence for aripiprazole (Taylor et al., 2012). This is the only situation in which antipsychotic polypharmacy is currently well justified (Correll et al., 2009). A trial of at least 3 months may be necessary. The other strategy is to add a mood stabilizer, with the strongest evidence being for lamotrigine (Tiihonen et al., 2009b). Occasionally, ECT is used (see below).

For a review of pharmacological approaches to treatment-resistant schizophrenia, see Van Sant and Buckley (2011).

Antidepressants and mood stabilizers

As has already been explained, symptoms of depression commonly occur in schizophrenia. However, there have been few satisfactory clinical trials (Whitehead et al., 2003).

Siris et al. (1987) conducted a placebo-controlled trial of imipramine in schizophrenic patients receiving fluphenazine in whom a depressive episode became apparent following resolution of the psychotic symptoms. Imipramine was significantly more effective than placebo, improving depressed mood but not altering psychotic symptoms. A second study by the same group confirmed the beneficial effect of imipramine. In patients in whom depressive symptoms coexist with an active psychosis, antidepressants do not appear to be helpful. Newer antidepressants, notably the SSRIs, are now widely used to treat depressive symptoms in schizophrenia, but have not been rigorously evaluated.

One practical issue with regard to clinical practice and trial design is the difficulty in distinguishing between depressive and negative symptoms. In this regard, a recent meta-analysis reported that antidepressant augmentation has some efficacy in treating negative symptoms (Singh et al., 2010).

Other than for clozapine augmentation, the value of mood stabilizers in treating schizophrenia is uncertain. There is no evidence that they have an antipsychotic effect, and the beneficial effects that have been seen in some trials could be due to treatment of affective symptoms. For example, in the Northwick Park ‘Functional’ Psychosis Study, lithium decreased elevated mood, whatever the diagnosis, but had no significant effect on positive or negative symptoms (Johnstone et al., 1988). A recent systematic review confirmed that the benefits of lithium in schizophrenia occur in patients with affective symptoms (or with schizoaffective disorder) (Leucht et al., 2004).

Electroconvulsive therapy (ECT)

In the treatment of schizophrenia, the traditional indications for ECT are catatonic stupor and severe comorbid depressive symptoms. In both situations, the clinical impression is that the effects of ECT can be dramatic. There is also some evidence that ECT is rapidly effective in acute episodes, and may enhance the response to antipsychotics in treatment-resistant cases (Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology, 2011). Overall, however, the evidence base is weak (Tharyan and Adams, 2002), and ECT is seldom used to treat schizophrenia in the UK, although it remains more widely used in some other countries.

Psychosocial approaches

For a discussion of recent psychological and psychosocial treatment guidelines, see Dixon et al. (2010).

The development of community-based services has led to an increasing emphasis on psychosocial interventions in the treatment of schizophrenia. These interventions are of several different kinds, but share the following aims:

• enhancement of interpersonal and social functioning, including promotion of independent living in the community

• attenuation of symptom severity and associated comorbidity (e.g. substance misuse)

• improvement in cognition.

There is evidence in support of a range of psychosocial interventions (see Table 11.17).

Family therapy

Family therapy, both formal and informal, is often employed at various stages of treatment. The most systematically evaluated intervention is that designed to decrease expressed emotion in family members. The procedure is usually combined with education about the illness and its consequences, together with practical advice on management (see Table 11.18). A recent meta-analysis confirmed that such interventions lower rates of relapse (NNT = 7) and improve medication compliance (NNT = 6). However, the studies showed a wide range of outcomes and may overestimate the treatment effect (Pharoah et al., 2010). Psychoeducation and family therapy are currently advocated in the management of schizophrenia for patients who live with, or are in close contact with, their family (National Institute for Health and Clinical Excellence, 2009b).

Table 11.17 Current psychosocial interventions for schizophrenia

Family therapy (psychoeducation)

Cognitive–behaviour therapy

Cognitive remediation

Social skills training

Supported employment

ILLness management skills

Integrated treatment for comorbid substance misuse

Assertive community treatment

Cognitive–behaviour therapy

The use of cognitive–behaviour therapy (CBT) in schizophrenia is based on the rationale that positive psychotic symptoms are amenable to structured reasoning and behavioural modification. In patients with delusional beliefs, for example, individual ideas are traced back to their origin and alternative explanations are explored. However, direct confrontation is avoided. Similarly, it may be possible to modify a patient’s beliefs about the omnipotence and origin of auditory hallucinations, with a resulting decrease in the distress that accompanies the experience and, perhaps, in the intensity of the hallucinations.

Table 11.18 Elements of family intervention in schizophrenia

Education about schizophrenia

Improving communication

Lowering expressed emotion

Expanding social networks

Adjusting expectations

Reducing number of hours of daily contact

Several trials have reported that cognitive–behaviour therapy is effective in schizophrenia, following Tarrier et al. (1998), who found that it was more effective than supportive counselling in decreasing positive symptoms. Turkington et al. (2004) concluded that cognitive–behaviour therapy has a robust effect on positive, negative, and overall symptoms in schizophrenia, and could be delivered as a brief intervention by community psychiatric nurses. Current guidelines recommend cognitive–behaviour therapy as part of an integrated treatment package for schizophrenia (National Institute for Health and Clinical Excellence, 2009b). However, there have also been studies with negative findings, and other reviews which have come to much more cautious conclusions, especially regarding its use in routine clinical settings. Indeed, a recent systematic review which excluded poor-quality trials concluded that cognitive–behaviour therapy has no specific benefits in psychosis (Lynch et al., 2010).

Social skills training and illness self-management

Social skills training uses a variety of approaches to teach the patient complex interpersonal skills, including behavioural rehearsal, feedback, and training. Skills training may be combined with illness self-management, in which the patient learns to adjust their own medication and organize their life so as to minimize troublesome symptoms. The results of these interventions are generally positive, but concerns remain about whether these gains are maintained and whether benefit is restricted to patients who have a good prognosis. A meta-analysis concluded that social skills training does not produce reliable benefits, and that it was not recommended for routine use (Pilling et al., 2002).

Treatment of cognitive impairments

As noted above, cognitive impairments are important determinants of poor outcome in schizophrenia. Both pharmacological and psychological approaches are now being used to try to ameliorate them.

Antipsychotics have minimal beneficial effects on cognitive symptoms (Mishara and Goldberg, 2004), and high doses, especially of drugs with anticholinergic or sedative properties, can impair performance. Suggestions that atypical antipsychotics may produce slightly greater improvements have not been substantiated. Various drugs are being tested or developed with the aim of improving cognition in schizophrenia, independent of any antipsychotic actions. As yet, however, no convincing clinical data have emerged.

A range of psychological treatments for cognitive deficits have also been trialled. Most of these rely on mental exercises and training to improve performance through practice. One of the most extensively tested is cognitive remediation therapy. The results, including the conclusions of meta-analyses, have been variable, and this therapy is not supported in current schizophrenia treatment guidelines. However, an enlarged meta-analysis recently provided good evidence for a sustained, small to medium beneficial effect of cognitive remediation therapy on cognitive outcomes, especially when combined with rehabilitation (Wykes et al., 2011).

Dynamic psychotherapy

In the past, dynamic psychotherapy was quite commonly used in the treatment of schizophrenia (more so in the USA than in the UK). Evidence from clinical trials is sparse, but does not support the use of this kind of psychological treatment. May (1968) found that psychotherapy had little benefit, notwithstanding the fact that the treatment was short and provided by relatively inexperienced psychiatrists. Apart from the lack of convincing evidence that individual psychotherapy is effective in schizophrenia, there are concerns that the treatment may cause overstimulation and relapse.

Interaction of drug and psychosocial treatments

Psychosocial treatments are usually given in conjunction with antipsychotic medication. It is believed that the combination improves outcome, although the data are limited.

It is also unclear to what extent successful psychosocial treatment works because it improves adherence with medication. Compliance therapy, a psychological intervention designed specifically to improve adherence, was effective in an initial trial (Kemp et al., 1996) but not in a replication, and it is not recommended in current guidelines. However, a recent trial with positive findings has led to a call for this to be revised (David, 2010).


Successful management of schizophrenia depends on first establishing a good relationship with the patient. This can be difficult because of the nature of the illness, but with skill and patience, progress can usually be made. It is important to make plans that are realistic, and that are acceptable to the patient and their carers. It is important to explain the benefits of medication, but also to discuss its limitations and its side-effects. Finally, novel treatments and modes of service delivery are likely to have a substantial impact on the management of schizophrenia over the next few years, and it is important to keep up to date with developments in this respect.

Table 11.19 summarizes the key elements of the management of schizophrenia. For a general review, see Cunningham Owens and Johnstone (2009). For a review of the pharmacological aspects, see Barnes and the Schizophrenia Consensus Group of the British Association for Psychopharmacology (2011), and for an account of the psychological and social aspects, see Dixon et al. (2010).

Table 11.19 Components and principles of the management of schizophrenia

Therapeutic partnership with patient and carers

Integrated, multidisciplinary working, involving primary and secondary care

Antipsychotic drugs for treatment and prophylaxis

Trial of clozapine for all patients who meet the criteria

Cognitive–behavioural therapy

Family interventions

Early interventions

Cognitive remediation

Assertive outreach for vulnerable patients

Regular assessment of needs

Crisis resolution and home treatment teams as alternatives to admission

Maintain realistic therapeutic optimism

The acute illness

Admission to hospital for assessment and treatment is usually needed for first episodes of schizophrenia and acute relapses. With adequate resources, home treatment is possible and has some advantages, including patient preference (Killaspy et al., 2006). This is reflected in the development of crisis teams which aim to treat and support acutely ill patients at home, although the evidence that such teams reduce admissions is relatively weak (Glover et al., 2006). Hospital admission does confer some advantages. It allows a thorough assessment, provides a safe environment, and gives a period of relief to the family, who will often have experienced considerable distress during the prodrome and emergence of the illness.

A drug-free observation period is a counsel of perfection, and is rarely attempted now. The main difficulty in establishing the diagnosis is often to elicit all of the symptoms from a withdrawn or suspicious patient. This procedure may require several interviews, as well as information from relatives or close friends and careful observations by nursing staff.

While the diagnosis is being established, a social assessment should be carried out. This includes a history of the patient’s personality, level of functioning, work record, accommodation, and leisure pursuits, and the attitudes to the patient of relatives and any close friends. The current social functioning and capabilities can be assessed by other members of the multidisciplinary team, notably occupational therapists, social workers, and nurses.

Antipsychotic medication is the mainstay of the treatment of schizophrenia and other psychoses. It will usually be initiated soon after the diagnosis is made—or even while this remains uncertain—and will often continue for several years. Therefore decisions about the use of medication, and appropriate involvement of the patient and their carer, are a key aspect of management. During the acute phase, the priority is to suppress the positive symptoms and control agitation, without producing unnecessary side-effects, notably acute dystonias. To achieve this, a combination of a low or moderate dose of an antipsychotic with a benzodiazepine is useful (e.g. haloperidol 2–10 mg daily or risperidone 2–6 mg daily, plus lorazepam 2 mg 4-hourly as required). Haloperidol and lorazepam can both be given intramuscularly if necessary. Alternatively, a more sedative antipsychotic, such as olanzapine 10–20 mg daily, or chlorpromazine 200–400 mg daily, can be used. Oral medication is always preferable, although occasional intramuscular doses may be needed for patients who exhibit acutely disturbed behaviour and are unwilling to comply with oral treatment. If there are doubts about whether the patient is swallowing tablets, the drug can be given as syrup or as an orally dispersible tablet. Short-acting depot preparations (of zuclopenthixol acetate or olanzapine) are available, but are not recommended in patients whose tolerance of antipsychotics has not been determined. Anticholinergic drugs should be prescribed if extrapyramidal side-effects are troublesome, but should not be given routinely or for sustained periods. As noted earlier, the acute symptoms of excitement, restlessness, irritability, and insomnia can be expected to improve rapidly after instituting treatment. Delusions, hallucinations, and other psychotic symptoms also begin to respond during this time, with improvement developing further over a period of several weeks.

As soon as the patient’s mental state permits, it is essential to explain the need for drug treatment and to obtain their full consent, especially in view of the risk of serious and long-term side-effects. If the patient is judged to be unable to make decisions of this nature, the appropriate legal safeguards under the Mental Health Act and its code of practice should be employed. The patient’s family should be involved as far as possible in all decisions about treatment as a matter of principle, and because they will have a role in encouraging compliance. The use of patient advocates is also helpful.

By the time that meaningful symptomatic improvement has occurred, the team should have formulated a provisional plan for continuing care, based upon the social assessment that has been carried out. Although it is difficult to predict the long-term prognosis at this stage, a judgement has to be made about the likely immediate outcome, based on the degree and speed of response to treatment, the history of previous episodes, and the factors listed in Table 11.15. The aim is to decide the nature and extent of aftercare required, and to make realistic plans accordingly.

Choice of antipsychotic drug

The evidence regarding the efficacy of antipsychotic drugs was discussed above, and the broad similarity between drugs and classes was noted. Thus, in practice, other factors determine the choice of drug. The key factors are the patient’s past drug response and preference, the availability of different preparations, and the anticipated side-effects, which do vary markedly between drugs. Important side-effects to consider when choosing an antipsychotic include the following:

• Extrapyramidal side-effects. The reduced liability of atypical antipsychotics to cause these side-effects is the one advantage on which there is agreement across all of the meta-analyses. There is also increasing evidence that the risk of tardive dyskinesia is lower than with typical antipsychotics (Correll and Schenk, 2008).

• Hyperprolactinaemia induced by antipsychotics with a high dopamine D2-receptor affinity. This includes some atypical antipsychotics (e.g. amisulpride, risperi-done), as well as typical ones. Other atypical antipsychotics, especially aripiprazole, are ‘prolactin-sparing.’ Raised prolactin levels are important in both sexes, but especially in women, not only because of the immediate effects on menstrual function and galactorrhoea, but because of concerns about increased risks of malignancy and osteoporosis (Wieck and Haddad, 2004). Both the extrapyramidal and hyperprolactinaemic effects of antipsychotics are dose related, emerging at doses (and dopamine D2-receptor occupancies) higher than that required for the antipsychotic effect (Kapur et al., 1999). Their occurrence thus implies that the dose is already too high for that patient; if the therapeutic response is not adequate, the antipsychotic should be changed to one with less propensity to cause these side-effects.

• Sedation. Some antipsychotics are very sedative, especially those with a high affinity for histamine and muscarinic receptors (e.g. chlorpromazine, olanzapine, and clozapine). This property is often desirable in the acute situation to control agitation, but may be problematic in the longer term, impairing normal levels of functioning.

• Weight gain is an important side-effect of many typical and atypical antipsychotics, notably those with hista-mine and 5-HT2C-receptor blocking effects, and associated in part with genetic variation at the 5-HT2C receptor. For example, 3 months of treatment with olanzapine increases weight on average by about 5 kg. Of the atypical antipsychotics, amisulpride, aripiprazole, and risperidone have less of a propensity to increase weight.

• Type 2 diabetes mellitus is of increasing concern. It is unclear to what extent it is actually caused by antipsychotics, rather than being related to illness factors (e.g. shared genetic predisposition, lifestyle). Furthermore, it is not clear whether some antipsychotics are worse in this regard than others, although the data suggest that the risks are greatest with clozapine and olanzapine (e.g. Smith et al., 2008). In view of the current uncertainties, regular monitoring for diabetes should become a routine part of prescribing of any antipsychotic, together with measurement of body weight and blood lipids (American Diabetes Association et al., 2004).

• Cardiovascular risk factors. Antipsychotics have been associated with increased risks of cardiovascular and cerebrovascular events, in part via their influence on body weight and insulin resistance, and dyslipidaemia. Some drugs also prolong the QTc interval on the electrocardiogram, which is a risk factor for cardiac arrhythmia. An antipsychotic with a low tendency to produce these changes should therefore be chosen in patients with any cardiac or vascular risk factors. This includes the elderly, in whom antipsychotics should be prescribed with great caution (see Chapter 18).

Taking these factors into account, the clinician should become thoroughly familiar with a few antipsychotics, including both typical and atypical agents, drugs with different side-effect profiles, and parenteral and depot preparations, and should have a knowledge of the interactions of antipsychotics with other drug classes.

Aftercare of ‘good prognosis’ patients

After a first episode of schizophrenia, patients who have responded well, who do not have residual impairments, and who are judged to have a good immediate prognosis have two principal needs. The first is to continue medication, at the smallest effective dose. There is surprisingly little evidence as to how long this should persist; most clinicians would recommend 1–2 years. Withdrawal should be undertaken gradually, with regular monitoring for a recurrence of symptoms. The second goal is to regain and maintain the best possible level of social functioning, which can involve supporting realistic structured activity, whether it be attending a day centre, education, or a job. Advice should be given about avoiding obviously stressful events. The patient should be seen regularly as an outpatient or at home by a member of the community mental health team (often a psychiatric nurse), until a few months after medication has been stopped and symptoms have remained quiescent. Thereafter, a cautiously optimistic prognosis can be given, but the patient and their family should be warned to seek medical help quickly if there is any suggestion of the condition returning. The patient’s GP has an important role to play in this phase of treatment.

Aftercare of ‘poor prognosis’ patients without major impairments

In the majority of cases, further relapse will be considered likely and continuing care will be required.

Sustained use of antipsychotic medication, where appropriate by depot injection, is an important component of treatment. The dosage, which should be the minimum required to suppress symptoms, can be determined by cautiously varying its size and frequency while observing the patient’s clinical state. If the response is partial at the maximum tolerated dose of the first antipsychotic, switch to another drug, perhaps with a different receptor profile. As noted earlier, clozapine should be considered in all patients who are treatment resistant or treatment intolerant, taking into account its potential risks. If available, cognitive–behaviour therapy directed at the residual positive symptoms may be a useful adjunct (see p. 289).

The follow-up plan should include regular review of the patient’s mental state and medication and their social and occupational functioning. It may be helpful for advice to be given about stressful situations and family problems. Specific interventions may be needed to help the family to understand the patient’s illness, to have realistic expectations of what the patient can accomplish, and to decrease their emotional involvement where appropriate. Work of this nature can be undertaken by community psychiatric nurses, social workers, psychiatrists, and GPs working together. Some patients understandably tend to withdraw from treatment and follow-up after recovery. For those who are judged to be at continuing risk of relapse, assertive community treatment has a useful role to play (see Chapter 21) (Burns and Firn, 2002).

Many patients fail to return to their full premorbid level of functioning and activities, despite resolution of their positive symptoms. This may be due to negative symptoms, depressive symptoms, cognitive impairments, or the side-effects of medication. These possibilities should all be considered, and, where possible, appropriate action taken.

Patients with a poor prognosis and significant impairments

Patients who have prominent negative symptoms, poor social adjustment, and cognitive and behavioural defects characteristic of chronic schizophrenia require more elaborate and indefinite aftercare. The components of a community service for chronic schizophrenia are described in Chapter 21.

Maintenance drug therapy continues to play an important part, but the main emphasis is on a programme of rehabilitation, psychosocial interventions, and social support, tailored to the needs of the individual patient. An occupational therapist and clinical psychologist can make valuable contributions. The following specific issues require attention.

• Living accommodation and circumstances. Will the patient be able to live alone or will they return to live with their parents or other relatives? Will accommodation need to be arranged?

• Occupational activities. Will the patient be able to undertake some form of paid employment or voluntary work?

• Activities of everyday living. Patients often need help to re-learn and implement basic living skills such as cooking, budgeting, shopping, housework, and personal hygiene.

• If available, cognitive remediation, social skills training, or related therapies mentioned earlier which are aimed at improving cognition and social functioning.

• Optimization of the efficacy and tolerability of medication, while recognizing that any further response beyond that achieved earlier in the illness is unlikely.

If a patient has persistent abnormalities of behaviour, particular attention needs to be given to the problems of their carers. Relatives may be helped by joining a carers’ group and meeting others who have learned to deal with similar problems. Advice is needed about the best ways of responding to abnormal behaviour, and about the expectations that they should have of the patient. Such advice is often best given by community psychiatric nurses. Social workers also have a long tradition of advising and helping relatives.

Other aspects of management

Care plans and the Care Programme Approach

Care plans are important in the management of psychiatric disorders, largely because of the need for coordinated multidisciplinary working, risk management, and clear documentation of responsibilities for their delivery. In the UK this responsibility has been formalized in the Care Programme Approach (see Chapter 21). This approach is particularly appropriate for those with schizophrenia, who often require complex and long-term support. The essential elements of a care plan are to identify and record:

• a systematic assessment of health and social needs

• a treatment plan agreed by the relevant staff, the patient, and the relatives

• the allocation of a key worker whose task is to maintain contact with the patient, monitor their progress, and ensure that the treatment programme is being delivered

• regular review of the patient’s progress and needs.

Early intervention

As noted earlier in the chapter, there is growing interest in the idea that early treatment of schizophrenia may improve the long-term prognosis, based on a number of lines of evidence.

• In conventional practice, patients have often been psychotic for many months before the diagnosis is made and treatment is initiated. Patients with a longer duration of untreated psychosis have a poorer longer-term prognosis (Marshall et al., 2005).

• The treatment response becomes slower and less frequent during subsequent psychotic episodes.

• Much of the deterioration in social function takes place during the first 2 years after diagnosis.

These observations imply that it may be worth attempting to identify patients with psychosis as early as possible, so that effective treatment can be instituted. Specific ‘Early Intervention in Psychosis Service’ (EIS) teams were mandated in the UK in the 2000 NHS Plan, and are endorsed in the current guidelines (National Institute for Health and Clinical Excellence, 2009b). These teams are described in Chapter 21. The EIS approach has also been extended to trials of treatment in individuals thought to be in the prodrome of illness but who do not have overt psychotic symptoms, and even to those who are well but at ‘ultra-high risk’ of developing schizophrenia (e.g. individuals who are children of parents with schizophrenia, and who have behavioural oddities). However, such early interventions are still mainly restricted to ‘demonstration’ or research teams.

Despite the widespread introduction of EIS teams, the evidence that they are effective remains limited and actively debated. There is no convincing evidence that early intervention reduces the duration of untreated psychosis (Lloyd-Evans et al., 2011). There is some evidence that low-dose medication and psychological approaches based on cognitive–behaviour therapy may decrease the conversion to psychosis, but the effectiveness, feasibility, and desirability of such interventions remain controversial (Bosanac et al., 2009; McGorry et al., 2010).

Substance misuse

Patients who misuse substances tend to have a worse course symptomatically and be more difficult to engage in treatment. In addition, substance misuse is associated with increased risks of homelessness, violent behaviour, and poor general health. Comprehensive outpatient and assertive community programmes that address both substance misuse and psychotic illness in an integrated way have been proposed for this group. Cognitive–behaviour therapy methods can reduce substance misuse, but do not clearly improve outcomes (Barrowclough et al., 2010).

The violent patient

Overactivity and disturbances of behaviour are common in schizophrenia, and although major violence towards others is not common, it does occur. For example, Humphreys and Johnstone (1992) found that, in a group of 253 patients with a first schizophrenic episode, 52 patients behaved in a way that was threatening to others. In about 50% of the patients the behaviour was directly attributable to psychotic symptoms, usually delusions. Comorbid substance misuse is probably the main factor implicated in the increased risk of violence (Fazel et al., 2009a). However, patients with schizophrenia are much more at risk of being victims of violence than of being perpetrators. Careful follow-up and close community supervision may reduce the likelihood of victimization.

General management for the potentially violent patient is the same as that for any other patient with schizophrenia, although a compulsory order is more likely to be required. Medication is often needed to bring disturbed behaviour under immediate control, but much can be done by providing a calm, reassuring, and consistent environment in which provocation is avoided. Threats of violence should be taken seriously, especially if there is a history of such behaviour, whether or not the patient was ill at the time. The danger usually resolves as acute symptoms are brought under control, but a few patients pose a continuing threat. The management of violence is considered further in Chapter 24.

Suicide risk

We have already seen that the rate of suicide is markedly increased in schizophrenia. As in other psychiatric disorders, demographic risk factors apply, such as male gender, social isolation, unemployment, and a previous episode of suicidal behaviour (Hawton et al., 2005). Other risk factors include the following:

• comorbid depression and hopelessness

• frequent exacerbations of psychosis

• good premorbid functioning

• being in the early stage of the disorder

• during hospitalization or shortly after discharge.

Patients with pronounced negative symptoms are less likely to make suicide attempts. Although suicide may be associated with active positive symptoms such as delusions of control or auditory hallucinations, more often it occurs in those with previously high levels of educational attainment who are relatively well at the time. Such patients often have depressive symptoms in the context of high premorbid expectations of themselves and dismay at how their illness has damaged their future prospects.

Lowering the risk of suicide in schizophrenia involves being aware of and managing the relevant risk factors, and treating depressive symptoms vigorously. Various data, including a randomized trial, suggest that clozapine treatment may significantly reduce suicide risk (Melt-zer et al., 2003), but the size and mechanism of this effect require further study.

Discussing schizophrenia with patients and their carers

There is probably more stigma and misunderstanding associated with schizophrenia than with any other psychiatric diagnosis. It is therefore particularly important that the term is used carefully and sensitively, and all patients and their carers should have regular opportunities to ask questions, express their views, and raise any concerns that they may have. The involvement of the family in all decisions about treatment and management has already been mentioned.

There should be discussion about the nature of the illness, identification of any misapprehensions, and an explanation of the current meaning of the term schizophrenia. It is worth discussing what is known about the aetiology, in part to ensure that misconceptions do not lead to unnecessary guilt. For example, some parents worry, incorrectly, that their behaviour may have caused the illness. Mention of its heritability may lead to questions about the risks of schizophrenia in other relatives or future generations. These can be answered clearly using the figures in Table 11.9, acknowledging the increased risk while emphasizing the probability that unaffected relatives will remain well. If, after discussion, the family continue to have very different concepts of schizophrenia from those of the clinical team, such views should be respected.

The prospects for recovery should be portrayed realistically but optimistically, given that some people wrongly believe schizophrenia is untreatable and inevitably follows an unremitting or deteriorating course. The key role of medication, together with psychosocial factors (e.g. family environment, substance misuse), in influencing outcome should be emphasized. Note that the family may have very different views about the meaning of recovery. The ‘Recovery Approach’ (sometimes referred to as the ‘Recovery Movement’) has achieved considerable prominence and influence in UK policy (Slade, 2009). Originating in the USA and well established in New Zealand, it is a relatively loose collection of ideas and principles that emphasize the priority of the patient’s aspirations and the goal of living life as fully as possible with the illness. Recovery advocates often criticize psychiatrists for having too pessimistic a view of the prognosis in schizophrenia, and insist on the need for optimism.

A further area in which there may be a difference in emphasis between psychiatrist and carer or patient is the focus upon the content rather than the nature of the patient’s beliefs (McCabe et al., 2002). That is, the psychiatrist is usually more concerned as to whether a thought is or is not delusional, or whether a perception is or is not a hallucination, whereas the patient wishes to discuss the thought or percept itself and its meaning for them. The psychiatrist should give the patient or carer adequate time to voice these concerns, while also explaining why it is necessary to ask detailed questions about the nature of the experiences.

Other than noting these broad themes, the psychiatrist must be prepared to adapt to the circumstances of each case. The full involvement of the patient and their carers should be an integral part of the management of schizophrenia. It is reasonable to assume that establishing and maintaining a genuine therapeutic alliance in this way will increase the likelihood that the patient maintains contact with psychiatric services, is adherent with treatment, and thereby has an improved prognosis.

Further reading

Bleuler E (1911). Dementia Praecox or the Group of Schizophrenias (English edition 1950). International University Press, New York.

Kraepelin E (1919). Dementia Praecox and Paraphrenia. Churchill Livingstone, Edinburgh.

Weinberger DR and Harrison PJ (2011). Schizophrenia, 3rd edn. Wiley-Blackwell, Oxford. (A definitive and up-to-date reference on all aspects of schizophrenia.)