The following is a highly selective overview of what has been written about schizophrenia. Because of space considerations, the author has emphasised recently published material. Should we put so many widely disparate presentations under the one label of schizophrenia? Attempts at discovering a cause or causes for schizophrenia must consider the probability that the broader syndrome may represent a heterogeneous group of disorders.

Is schizophrenia an end-stage syndrome where certain symptoms are shared by all and which is reached by a slow decompensation of personality? Earlier writers were conscious of this when they described process or nuclear schizophrenia.

Neuro-psycho-social aspects

Patients with schizophrenia have difficulties interpreting social cues from the faces of other persons and those with positive symptoms may be impaired in recognising even basic facial cues. Adolescents at ultra-high risk of psychosis attribute anger to neutral faces and patients with paranoid schizophrenia react excessively to neutral social scenes but not to negative social scenes. A positron emission tomography (PET) study involving facial emotion recognition revealed that right-handed non-acute patients with schizophrenia showed left amygdalar hyperactivation in both emotional and control tasks and the right amygdala showed no differential activation in any of the tasks: schizophrenia is associated with non-task-specific amygdalar hyperactivation during a continuous emotional and non-emotional task compared to healthy controls. A recent project used an affective face recognition model and functional magnetic resonance imaging (fMRI) and discovered that patients with schizophrenia performed slowly; comparison subjects recruited expected cortical areas more than did patients, and more severely symptomatic patients showed relatively decreased recruitment; increased symptoms correlated with excess amygdalar and orbitofrontal cortex response to threatening faces; and comparison subjects showed a negative relationship between amygdala and cortical areas involved in cognition, while patients exhibited weakening of this relationship; in other words, there is limbic over-responsiveness and diminished cortical facial recognition memory response. 

Patients with schizophrenia may be impaired in their ability to cope with normal levels of negative emotion. Expressed emotion (EE) is a measure of criticality, hostility, intrusiveness and over-involvement among people with whom patients have significant contact. Four times as many patients from high EE homes as those from low EE homes relapse during the first nine months at home following discharge after treatment of an acute episode. Young, single men, who are living with parents, are at very high risk from this type of ambient tension. The perception of high versus low EE varies with culture. Measured EE may be less reliable than relatives’ attribution of responsibility for symptoms: low EE relatives may see the condition as being beyond the patient’s control whereas high EE relatives may blame the patient. Parents of children and adolescents with schizophrenia may manifest less EE than the parents of adults with schizophrenia. Not all studies equate EE with relapse rates.

People with schizophrenia may do better if they live in suburbia, if they are of high socio-economic status, if they live in rural developing countries, or if they like their relatives. Stress may arise from outside the family, eg. in occupational or interpersonal arenas. Coping mechanisms (such as problem-solving) and medication may prevent the effects of stress reaching the non-specific symptomatic stage. High environmental tension may keep the patient near the threshold for relapse, so that minor stresses (including an increase in EE) may precipitate florid symptoms. There is evidence of an increase in EE levels in the three weeks before the onset of an episode of schizophrenia. Antipsychotic drugs reduce relapse rates in the presence of high EE, but only relatively so. A greater number of independent (not brought on by illness) life events are probably required to induce relapse in adequately medicated patients. If relatives can be trained to recognise non-specific symptoms, medication dose may be increased pending psychiatric review. Relapse rates may be reduced by educating the family about schizophrenia and by conducting group sessions for community-based carers. While many clinicians focus on improving the interpersonal coping skills of patients, a focus on the family may improve outcomes. Research has tended to concentrate on the stresses inherent in family life for the patient but such patients can cause stress in their relatives, ie. it is debatable whether parental EE is a trait or a reaction. Relatives may dislike the idea of EE because it sounds as if they are being blamed. It is therefore important to emphasise positive aspects of EE reduction rather than negative aspects of high EE. EE levels may fall when carers feel less burdened and when there is less contact with the patient.

Life events lead to stress, which in turn, via physiological arousal, leads to non-specific symptoms like insomnia, and then relapse. Slow habituation to arousal is associated with delusional thinking and hallucination-proneness.

It has been suggested that reduced cortical volume in schizophrenia is due to reduced social interaction, a phenomenon reported in animals, but such theorising ignores direction of causation. On the other hand, in 2005, a study found that whole brain volume in schizophrenia was of genetic origin, while lateral ventricular size was related to the individual-specific environment. A 2010 systematic review warned that antipsychotics may cause some of the MRI findings attributed to schizophrenia, and this was supported by a later prospective MRI study of first-episode schizophrenia. 

Post-traumatic stress disorder appears common in patients with schizophrenia and other severe mental disorders, prevalences ranging from 3.8-43% (mostly 29-43%) being reported (compared to 8-12% in the general population). Immigrants and ethnic minorities are thought to be at increased risk for schizophrenia.

It appears that African-Caribbean and black Africans in England are at especially high risk for both schizophrenia and mania. Rates of mental hospital admission for immigrants to England and Wales tend to be higher for females. Many authors point to lower admission rates for schizophrenia in the countries of origin, but this is open to many interpretations, such as ‘the dysfunctional or sick emigrate’, ‘emigration makes you sick’ or ‘host countries mistakenly see them as being ill’. Some studies show an excess of schizophrenia in second-generation immigrants but not in the actual immigrants, although others show an excess in first- and second-generation immigrants and in nationals with a history of foreign residence, but such findings are difficult to explain. Also, immigrants may have a better prognosis than the indigenous population of the host country, probably because the illnesses are precipitated by social difficulties. A 2005 meta-analysis found a mean weighted relative risk for schizophrenia among first- and second-generation migrants of 2.7 and 4.5, respectively; the risk was increased if migrants came from developing countries or if they came from countries with a black majority. Recently, Coid and co-workers conducted an inner-east London population-based two-year epidemiological study of first-episode psychosis in people aged 18-64 years: black and minority ethnic subgroups all had an increased incidence of affective and non-affective psychoses compared to white British people; only black Caribbean second-generation individuals had a significantly increased risk compared to first-generation counterparts; Asian women, but not men, of both first-generation and second-generation, were at increased risk for psychoses compared to white British people. A 2008 study examined first-episode psychosis cases and community controls in two English cities over three years: cases were more socially disadvantaged and isolated, even when they confined the sample to affective diagnosis, a short prodrome and short duration of untreated psychosis; the greater the disadvantage the more likely was psychosis to occur; and there were similar patterns in white British and black Caribbean groups, although the black Caribbean group was more disadvantaged.

A very recent meta-analysis of the risk for psychotic disorders among first- and second-generation immigrants found that an increased risk of schizophrenia and related disorders among immigrants persisted into the second generation and there was no significant difference between generations but there were such differences according to ethno-racial status and host country, ie. the risk is mediated by social context. Bhugra and Bhui suggested in 2010 that life in Britain can cause distress in African-Caribbean people by effects on self-esteem, identity and prospects; ‘schizophrenia’ may have different causes in African-Caribbean people and white British people. These authors do not ignore genetic factors but suggest different weightings for these and psychosocial risk factors. Ethnic density refers to the number of people of one’s own ethnic origin in the immediate environment. Evidence suggests that when there are few such people one is more prone to psychosis.

Fifteen per cent of people in the general population report delusions and hallucinations, although clinical psychosis is much less prevalent. The importance of these symptoms is being debated, opinion varying from normal variant to psychosis precursor. A study published in 2009 found an association in 12-year-olds between having definite symptoms and maternal infection during pregnancy, maternal diabetes, need for resuscitation and five-minute Apgar score. A 2010 study found a one-year prevalence of auditory vocal hallucinations in seven- and eight-year-olds of 9%; 15% of these were reported to suffer significantly and to behave problematically; rural children had a higher prevalence but urban children were more functionally affected by the experience; and there was little evidence of a role for developmental variables. Possible precursors of schizophrenia such as psychotic-like experiences may be more common among the children of African-Caribbean origin in the UK than among their white peers.

Psychodynamic and family theories

Freud, in 1911, published his analysis of Daniel Schreber, the presiding judge of the Dresden appeal court. This analysis (based only on Schreber’s autobiography and the report of a Dr Weber) led Freud to conjecture that Schreber was unable to accept that he was a homosexual. ‘I love him’ was changed subconsciously to ‘I hate him’ but, because this was unacceptable to the superego, the latter was altered to ‘It is not I who hates him but he who hates me’ and ‘I am persecuted by him’, ie. denial with projection. In fact, Schreber may have suffered from paranoid schizophrenia or even encephalitis lethargica.

Melanie Klein placed the origin of schizophrenia in infancy, the paranoid-schizoid position. There was a developmental failure to understand that others can possess a number of characteristics, such as ‘good’ and ‘bad’, simultaneously. Fromm-Reichmann coined the term ‘schizophrenogenic mother’ in 1948. This notorious term has fallen from favour. Egeland and Sroufe, in the 1980s, felt the schizophrenic mother may be unable to offer secure attachment for her child, with resultant poor bonding, social incompetence and problem-solving difficulties in offspring. Recently, and surprisingly in view of the foregoing, Pawlby and colleagues found that mothers with schizophrenia on a mother-and-baby unit communicated quite well with their babies. However, Suvisaari and co-workers reported that the children of mothers with psychotic disorders are at increased risk of several adverse outcomes, including early death.

Lidz, extrapolating from uncontrolled psychoanalytic studies of a few wealthy families of schizophrenic patients, suggested there were two types of abnormal families: marital skew (one parent dominant, usually the mother, and the other partner yields to her whims) and marital schism (parents hold contrary views, leaving the child with divided loyalties). Bateson wrote about double bind during the 1950s: overt instructions from the parent are contradicted by covert instructions so that the child can only give ambiguous and meaningless responses. A 2008 study examining responses of male adolescents to questions posed by a draft board found an increased risk for non-affective/schizophrenic psychoses in those reporting poor family functioning; again, this does not inform about direction of causality. Studies suggest an excess of various forms of victimisation (particularly sex abuse) in the past histories of adult psychotic patients in private British households but there are methodological flaws in this research.

Living in cities

In the 19th century, urban life was often thought to be protective against insanity. In a 1987 Nottingham study, people with schizophrenia were concentrated in poor central city areas, while a more varied distribution was found for affective psychosis. The former may seek out the low levels of social demands for performance and the relative anonymity of city centres, while the latter benefit from periods of normality, the manic drive or other unknown associated factors. A 2004 Swedish study reported that the incidence of first-admission psychosis and depression rose with increasing levels of urbanisation, although Danish evidence suggested that urban-rural differences in schizophrenia risk were unrelated to exposures that became more common in urban areas over time. In 2003 Murray suggested that the increased risk for onset of schizophrenia in cities is due to social isolation. An association between vagrancy and schizophrenia has been noted in various parts of the world. The ‘drift hypothesis’ holds that patients’ fortunes decline as a result of illness, they fall down the social scale into poverty. Downward social drift has also been demonstrated for bipolar disorder.

The ‘breeder’ hypothesis holds that impoverished city centres are aetiologically related to schizophrenia. One criticism of the breeder theory is that parents may have moved to cities because of their genetic predisposition to psychosis.

Both poverty and illness may lead to drift. Rented accommodation tends to be cheaper in less attractive, run-down areas. There was a significant movement of patients with schizophrenia from outer to inner London areas during the period 1986-1991; those inner-London schizophrenic patients who were relatively mobile were more likely to be male, have prominent hallucinations, and have no contact with a GP. Kelly et al, in 2009, found that the incidence of schizophrenia in males was higher in Irish urban than rural areas whereas the incidence of affective psychosis was lower in urban than rural areas for both sexes.

Danish workers found that living in an urban area and a higher degree of urbanisation increased risk for schizophrenia. Both level of urbanicity and familial liability independently and synergistically seem to increase risk for psychosis. 

Fertility and mortality

All psychiatric disorders carry an increased risk of premature death, the highest risk, from natural and unnatural causes, being associated with substance abuse and eating disorders. Risk of premature death from unnatural causes is especially pronounced in the functional disorders, particularly schizophrenia and major depression. Deaths from natural causes are especially increased in organic mental disorders, intellectual disability and epilepsy. People with schizophrenia have a reduced life expectancy of about 20% (males more than females) and a reduced reproduction rate, although fertility may be improving, partly because more patients live outside institutions. A Swedish study found that the number of people with schizophrenia whose bodies were not discovered for some time following death increased in keeping with the decline in bed availability during 1952-2005 in Malmo. Among people with schizophrenia in rural China, all-cause mortality and suicide rates were greater in males than in females in a 10-year cohort study. Male gender, earlier age at onset, older present age, longer illness duration, physical illness, inability to work, and no history of treatment were independent predictors of increased mortality. Finnish studies suggest that antipsychotic treatment, especially clozapine, is associated with lower mortality compared with the non-use of antipsychotics.

In a 2010 report on the 25-year mortality of a community-based schizophrenia cohort mortality risk was two to three times that of the general population, mostly due to natural causes (unnatural deaths tended to occur during the first five years of the study), and the authors suggested that cardiovascular mortality may have increased compared to the general population – the steepest rise in cardiovascular mortality appeared to coincide with the introduction of the newer antipsychotic drugs, although the authors felt unable to form any firm cause-effect conclusions.

Despite concerns about fertility/mortality, schizophrenia still occurs at a similar rate to that in the past half-century. Inbreeding does not seem to affect the incidence of schizophrenia. Hyperprolactinaemia due to medication, especially high doses of the typical agents, may inhibit ovulation. Married status may delay onset of illness, especially in men. Perhaps less vulnerable people marry and delay breakdown for longer. Never-married persons of either sex are likely to be admitted earlier in life than those who are married. Schizophrenic males do not marry and reproduce as much as other age-matched males, although they may have more children when they do marry than do married schizophrenic females, and affected females and males have decreased rates of marriage and fertility. A Swedish birth cohort study found that, relative to the general population, people with schizophrenia had fewer children and grandchildren (partly due to lower marriage rates), their unaffected siblings had no more children than the population norm, there was a trend for offspring of schizophrenic patients to have more children, and patients with affective psychosis and their relatives resembled the general population regarding fertility measures. It appears that either something in the environment is aetiologically important or that new mutations keep the disorder going. Another possibility is that schizophrenia confers a biological advantage on the sufferer, although the evidence for this is weak. Reduced fertility among Finnish schizophrenic patients is not compensated for by higher fertility among their siblings. Alternative theories involve mutation-selection balance and a role for quantitative traits.

Season of birth

There is an excess of schizophrenic births in winter and spring from the northern but not the southern hemisphere, suggesting a possible viral aetiology and the possible role of obstetric problems. The season-of-birth effect is small with differences of only 5-10% from expected rates, but it may be important in some cases. 

In 1989 O’Callaghan, whose passing was sadly noted recently, observed an excess of spring births in the general Irish population and he found that schizophrenic patients with a family history of schizophrenia followed the same pattern, but that schizophrenic patients with a negative family history had an excess of winter births. A 1970s study of Norwegian psychiatric inpatients born 1866-1939 found an excess of winter births, although the trend was less obvious in better-off patients. Retrospective Scottish work suggested a cold autumn might lead to the birth of people prone to developing schizophrenia during the following spring! Two studies reported an association between summer births and deficit schizophrenia. One group described a large increase in broadly and narrowly defined schizophrenia, mainly in younger people, in part of Camberwell, London; it was unclear if migration played a role. Finnish work on birth cohorts suggests a move from rural to urban births over time, but there are still clusters that suggest possible genetic isolation. 

Genetics

The lifetime risk for schizophrenia is about 1% in the general population. Schizophrenia often runs in families and the risk increases with the number of affected relatives and with severity of illness in the proband.

The genetic input in schizophrenia is probably a graded matter in the general population. Murray and Dean have suggested that liability to psychosis, like hypertension, is a continuous trait in the population with schizophrenic patients lying at one extreme of this continuum. Liability to schizophrenia is most likely polygenic, ie. additive effect of a number of minor genes, ie. quantitative trait loci (QTLs). A QTL is the genomic region containing naturally occurring allelic variations affecting a quantitative phenotype. Interplay between genes and bidirectional gene-environment interactions complicate matters. Many loci have been highlighted as possibly being important in schizophrenia. Copy number variation (CNV) refers to a segment of DNA for which copy-number differences have been found by comparing at least two genomes. Humans, who are diploid, usually have two copies of each chromosomal region, one per chromosome. This may vary because of deletion or duplication. CNVs may have a role in schizophrenia.

As in haemophilia, up to one-third of schizophrenia patients have no known family history of schizophrenia. A 1991 project found that paranoid schizophrenic patients had a lower likelihood of having affected first-degree relatives than did other subtypes as well as having a lower concordance ratio between identical (DZ) and non-identical (MZ) twins. Non-familial cases may have an environmental aetiology because they are more likely to have cerebral abnormalities, eg. ventriculomegaly. A meta-analysis of 12 twin studies found the point estimate of heritability in liability to schizophrenia to be 81%; there was evidence for environmental influences as well (joint estimate of 11%). 

Tim Crow postulated a continuum of psychosis and suggested that we have failed to genetically demarcate schizophrenia from affective disorder. A 1980s report described a case of MZ twins in which one had manic-depressive psychosis and the other had schizophrenia. Other authors have also reported overlap between affective disorders and schizophrenia in the families of patients with schizophrenia. Kendler et al, working in Roscommon during the 1990s, found a familial tendency to develop a wide variety of psychotic syndromes rather than schizophrenia per se, but a vulnerability to develop depressive and manic affective illness was found to be somewhat more specific. Anglin and others recently found that schizophrenia patients with a family history of affective illness performed relatively well on measures of IQ and executive function, perhaps suggesting the existence of a distinct subgroup that is biologically nearer to severe affective disorder than to schizophrenia. Lately, a very large Swedish study found evidence that schizophrenia and bipolar disorder partly share a common genetic aetiology: 63% of the comorbidity between disorders was due to additive genetic effects common to these two conditions. Also, a meta-analysis of family studies of probands with schizophrenia and bipolar disorder found that they co-aggregated in families.

Tim Crow put forward his pseudo-autosomal theory in the 1980s. He locates the genetic mechanism for psychoses at the tips of the short arms of the sex chromosomes and said that language is a human facility made possible by lateralisation, in turn made possible by material from the X-chromosome moving to the Y-chromosome (recombination during male meiosis) and thus escaping inactivation. Attention has been focused on the region of X-Y homology and specifically one gene, protocadherin XY (protocadherin is a cell adhesion molecule involved in axonal guidance) which, he suggests, explains, at least partly, language and psychosis. 

Burns suggested that schizophrenia is a by-product of social brain evolution in humans, with problems in the ‘social brain’, ie. frontotemoral and frontoparietal networks; ‘mirror neurones’ in the social brain allow us to read the intentions of other people. Burns suggests that the basic problem in schizophrenia relates to a sense of detachment and disembodiment from the social self and social world. This concept received support from work with twins discordant for schizophrenia. However, most morphometric work has reported bilateral alterations, anatomical asymmetries are found in other primates, and Crow failed to address the more generalised deficits found in schizophrenia. Nevertheless, there is fMRI evidence to suggest that increased language activity of the right hemisphere may play a role in precipitating psychosis. And recent research found that non-right-handed people have higher schizotypy that reflects a higher incidence of bilateral language lateralisation, potentially underlying loosening of associations.

The notion of a defective filtering of environmental stimuli dates to Broadbent in the 1950s. Hemsley applied defective selective attention theory to negative symptoms: patients learn to cope with information overload by slowed thinking, poverty of speech and flat affect, apathy and withdrawal. Nicotine transiently normalises the P50 deficit in schizophrenia and past and present smokers may have a delayed onset of schizophrenia compared to those who never smoked. The α7-nicotinic receptor gene may be involved in an auditory sensory gating defect reported in schizophrenics: linkage of impaired P50 inhibition was reported to a locus adjacent to the α7-nicotinic receptor gene on 15q. Do schizophrenic patients smoke to improve filtering of environmental stimuli? 

Functional MRI work is consistent with nicotinic receptor mediation of neuronal dysfunction in schizophrenia. According to some workers cigarette smoking may improve visuospatial working memory and attention in schizophrenic patients, perhaps via nicotinic receptors. The idea that schizophrenic subjects smoke tobacco in order to improve their cognitive functioning is as yet unresolved. There is some evidence that DMXB-A, a partial α7-nicotinic agonist, improves cognition in people with schizophrenia.

The velo-cardio-facial syndrome (VCFS – cardiac defects, abnormal facies, thymic hypoplasia, cleft palate, hypoglycaemia and the involvement of chromosome 22) is due to small hemizygous chromosomal deletions in chromosome 22q11.2. This is associated with an increased incidence of schizophrenia-like and bipolar disorder-like manifestations. The gene coding for COMT is located within the small region on chromosome 22 that is deleted in VCFS. 

If paternal half-siblings (same father but different mothers) are separated early in life we can claim to have individuals who share neither intrauterine nor early maternal influences. Kety et al, in the 1970s, found that 13% of paternal half-siblings of schizophrenic patients were diagnosed as having this disorder compared with 1.6% (one case) of 64 half-siblings of controls. Such findings favour genetic transmission over other explanations.

‘Anticipation’ refers to increasing severity and earlier age of onset of a condition over successive generations; a minority of cases of schizophrenia may have this characteristic; it was hypothesised to be associated with unstable coding trinucleotide (triplet) repeat expansions; work has been done on polyglutamine-expanded tracts, encoded by CAG repeats, in protein from lymphoblasts. These findings are reminiscent of those in Huntington’s disease research. Further evidence is required.

There is some evidence that an older father (and possibly an older mother) at the time of conception may be associated with increased risk for developing schizophrenia in offspring. A 2009 study defined an older father as someone over 35 years of age. According to Petersen et al, 2011, risk for schizophrenia in offspring is a function of father’s age at birth of his first child and not of his age at the time of the birth of later children, and they argue that if de novo mutation in the father underlies the risk for schizophrenia (in offspring) then the risk associated with paternal age should increase with the number of children, a theory not supported by their data. Advanced father’s age was more important than the mother’s age as a risk factor for schizophrenia in offspring in a recent study from Iran.

See part two of this series.

See part three of this series.

Declaration of interest: none