Kraepelin, who introduced the concept of dementia praecox almost a century ago, considered that the causes included both heredity and organic brain damage. He and his colleagues at the Munich Clinic initiated a series of family and twin studies that supported the former while Southard, Professor of Pathology at Boston Psychopathic Hospital, undertook detailed postmortem studies at the time of World War I. Southard1 attributed his neuropathological findings, some of which are now being reconfirmed, to abnormal development of the brain.
Unfortunately, their successors gradually forgot what Kraepelin and Southard knew. Indeed, by the 1960s many authorities denied the existence of genetic and organic causal factors, and the very concept of schizophrenia as a medical illness had fallen into disrepute. Fortunately, improved twin and then adoption studies reemphasized the role of aberrant genes, and the subsequent computerized tomography (CT) scan study of Johnstone et al2 unleashed a plethora of neuroimaging and neuropathological research. Thus, the Kraepelinian view of schizophrenia was rediscovered, and a period of exciting advance began.
The evidence from classical clinical studies for a genetic contribution to schizophrenia is too well known to be repeated here in detail. Suffice it to say that an individual's risk rises the more closely he or she is related to a schizophrenic. The maximum risk is almost 50% for the identical cotwin of a schizophrenic.
Family members, of course, share not only their genes but also many aspects of their physical and cultural environment; these effects can be separated out by adoption studies. Kety and colleagues3 demonstrated that the crucial factor in determining the risk for schizophrenia is not the type of upbringing an adopted child receives from his or her foster parents, but whether a biological parent had the disease.
How genetic is schizophrenia? One way of estimating this for a disease is the correlation in liability in first-degree relatives. The estimate for schizophrenia is .37 to .43, which compares with .27 to .40 for diabetes mellitus, .33 for coronary artery disease, and .24 for hypertension. By comparing these figures, one can see that schizophrenia is genetically determined as much (and probably more) as these other common diseases.
There has been a long, and mainly unproductive, dispute about the mode of inheritance.4 Advocates of a single major locus theory believe that one mutant gene causes nearly all schizophrenia. They have been locked in argument with those who favor a model in which many genes, each of small effect, act additively with environmental factors.5 Both these groups have been pitted against those proposing etiological heterogeneity with various genetic and environmental subtypes.6
Fortunately, these ideological quarrels are beginning to resolve as new data become available. Disorders that follow classical mendelian patterns of inheritance are rare, and common diseases, such as coronary artery disease and major psychoses, now are categorized as multifactorial. However, this does not mean that each individual case is multifactorial. For example, in coronary artery disease, some cases are a direct consequence of the mendelian inheritance of the gene for familial hypercholesterolemia, while others are almost entirely dependent on hazardous behaviors, such as obesity and smoking; in many cases ischemic heart disease results from the additive effect of genetic predisposition and damaging environment. A similar spectrum of cases ranging from the wholly genetic through those with mixed etiology to the totally environmental is likely to be true for schizophrenia.
This conceptual reformulation has encouraged researchers to seek out unusual families in which schizophrenia appears to show mendelian patterns of inheritance. These multiply-affected families are then examined to see whether the disease and a genetic marker occur together (cosegregate) more often than should occur by chance; if this were to be the case, then schizophrenia and the marker would be said to be linked. This would indicate that the disease gene and the marker must be relatively close together on the same chromosome. Consequently, if the chromosomal location of the marker was known, one could deduce that the diseased gene must lie nearby.
The molecular genetic revolution has made available a host of DNA markers, or RFLPs, which span all the human chromosomes. Unfortunately, working systemically through each chromosome, looking for linkage between markers and schizophrenia, would take an enormous amount of time, and researchers did not know where to concentrate their efforts. Fortunately, a clue came when a young man presented to a hospital in Vancouver with schizophrenia plus certain minor physical anomalies suggestive of a chromosome abnormality. His maternal uncle also had both the disease and the anomalies, and cytogenetic analysis showed that both had a translocation of part of chromosome 5. Bassett and colleagues7 reasoned that if having extra genetic material for a segment of chromosome 5 could cause schizophrenia, then perhaps this segment might be the site of a putative schizophrenia gene.
Sherrington et al8 collected seven extended families (five from Iceland and two from England) in which 31 of the 104 individuals at risk were diagnosed as schizophrenic using DSM-III criteria. Two probes for the crucial segment of chromosome 5 were "blindly" run through the DNA of the affected and unaffected members of these families; evidence highly suggestive of linkage between the markers and schizophrenia was found. This was demonstrated by the finding of a log of the odds (LOD) score of 3.22. An LOD score is the geneticists' way of calculating the odds of linkage. Traditionally, an LOD score of + 3 has been taken as the statistically significant level, indicating a 95% overall probability of linkage. Intriguingly, the LOD score in the study of Sherrington et al rose to 6.49 when not only schizophrenic patients but also individuals with schizotypal personality or even other psychiatric illnesses were considered as affected. Thus, this study suggests that schizophrenia can be inherited in an autosomal dominant manner and that the responsible gene, in some families at least, is located on the proximal long arm of chromosome 5. Furthermore, the phenotypic expression of the mutant gene may be quite variable.
However, other groups have failed to replicate the linkage finding, leaving us with two possibilities. First, schizophrenia may be genetically heterogeneous, with only a minority of cases caused by a gene on chromosome 5. Second, it may be that the diagnostic and penetrance problems inherent in studies of psychiatric disorders make it necessary to demand a much higher LOD score than the usual + 3 before linkage for psychiatric disorders can be regarded as proved.
Another area of impressive advance has been neuroimaging. Over 50 CT scan studies of schizophrenic patients have been carried out. Agreement has been reached that enlargement of the lateral ventricles is observed in many cases,9 although the exact proportion of cases and the extent of the enlargement remains contentious. The increase is not specific to schizophrenia and also is seen in some cases of affective psychosis. Enlargement of the third ventricle and cortical sulci, as well as cerebellar atrophy also have been reported, though less consistently. Ventricular enlargement appears to be present at the onset of schizophrenia and shows no evidence of progression on follow-up after several years.10 This strongly suggests that the changes are antecedent to the onset of the positive syndrome of delusions and hallucinations.
The clinical characteristics that have been found correlated most frequently with structural changes are poor premorbid personality, cognitive impairment, neurological "soft signs," and negative symptoms. Crow11 characterized this constellation as the Type II syndrome. It is true that these correlates have not been universally confirmed but the major reason for this is probably that only a small proportion of the variance in ventricular size is pathogenic. Thus, trying to correlate other variables with ventricular enlargement is rather like trying to associate particular clinical characteristics with low IQ. The fact that some individuals with pathological changes remain within the normal distribution of IQ (or ventricular size) produces too much overlap to enable the universal replication of pathogenic correlations.
The best way out of this dilemma is to study individuals who should have the same ventricular size. Because ventricular size is normally under a high degree of genetic control, monozygotic (MZ) twins discordant for schizophrenia are such a group (Table). A study of discordant MZ twins revealed larger ventricles in the schizophrenic twin than in the nonschizophrenic cotwin. Because MZ twins have identical genotypes, the causes of the differences must be environmental.6 Recently, magnetic resonance imaging has been applied to schizophrenic patients. This has largely confirmed the changes revealed by CT scan. In addition, Suddath and his colleagues12 have used a study of 15 pairs of MZ twins discordant for schizophrenia to demonstrate that the illness is associated with diminished volume of the temporal lobes. The disparity was particularly marked for measurements of gray matter volume in the left temporal region, a finding consonant with other evidence implicating leftsided abnormalities in schizophrenia. For example, in the Maudsley discordant MZ twins, the affected twins had a relative decrease in left cerebral density.
Neuroimaging Studies of Identical !Wins Discordant for Schizophrenia
These neuroimaging studies have provoked a resurgence of work in neuropathology, particularly on the temporal and the frontal lobes. Neuronal loss has been seen in the hippocampus13 and the parahippocampal gyrus together with a striking enlargement of the temporal horn.14 There is little evidence of gliosis, which would be expected if the neuronal loss was a consequence of some adult onset degenerative disorder; its absence suggests either a genetically programmed loss of cells or an environmental factor operating very early in life to impair brain development. In line with this thinking, Jakob and Beckman ' 5 described abnormally positioned neurons in the pre-alpha layer of the parahippocampal gyrus. This latter finding points particularly toward a failure of neuronal migration in utero.
Thus, the structural abnormalities found in the brains of many schizophrenics suggest a neurodevelopmental rather than a degenerative process. Aberrant genes clearly are involved but the identical twin of a schizophrenic has a better than even chance of avoiding the disorder, and most authorities accept that environmental factors must be involved. Indeed, the discordant twin studies described earlier suggest that the enlarged ventricles and left temporal hypoplasia are, to some degree at least, environmentally determined. What are the crucial factors?
It has been known for many years that a history of adverse events during fetal and neonatal life is obtained more frequently regarding schizophrenic patients than normal subjects. At first these reports were dismissed on the basis that a mother might retrospectively exaggerate adverse obstetric events in the child who became schizophrenic. Subsequent studies that relied on contemporary birth records came to the same conclusion, but it was still difficult to conceive of how obstetric hazard might be associated with the occurrence of schizophrenia some two decades or so later.
Neurodevelopmental Model of Schizophrenia
However, we now know that obstetric complications (the term used to cover both pre- and perinatal events) are noted particularly in schizophrenics without a positive family history for psychosis.16 This reciprocal relationship implies that obstetric complications may augment or even substitute for genetic predisposition. Furthermore, a number of studies have shown that obstetric complications predict ventricular enlargement in adult schizophrenics.17
Pediatricians have shown that a variety of pre- and perinatal events can cause hypoxic ischemic damage to the neonate who can develop intraventricular hemorrhage and perventricular damage as a result. As yet such children have not been followed to adulthood, but in childhood they show persistent ventricular enlargement, behavioral deviance, and cognitive abnormalities, as well as soft neurological signs - characteristics that are found in many schizophrenics.
It is easy to see how obstetric complications might impair neuronal proliferation and migration. However, in the last decade it has become apparent that the immature brain consists of a large excess of neurons and that during development a large proportion of these neurons die and the axons of the remainder thin out. This process eliminates early errors of connection. Destruction of an area of neurons reduces the fall-out rate of adjacent neurons. Thus early brain damage can lead not only to cell loss but also to the persistence of immature patterns of cells and anomalous connections in the maturing nervous system.
Murray et al14,18 therefore suggest that neural dysplasia results in premorbid cognitive deficits and abnormal personality, as well as negative symptoms and abnormal CT scans, ie, the Type II syndrome (Figure). Maturational brain changes in adolescence, possibility myelination or synaptic pruning, then reveal the immature neuronal circuitry; this in turn leads to misinterpretation of stimuli and the consequent onset of hallucinations and delusions (the Type I syndrome). Weinberger19 also emphasizes the importance of brain maturational changes in adolescence.
Mednick and his colleagues20 consider that the mutant gene or genes that predispose to schizophrenia do so by causing a type of neuronal migration particularly vulnerable to pre- or perinatal damage. These researchers have shown that individuals who were in their second trimester of fetal life during the influenza epidemic in Finland in autumn 1957 had an increased risk of later schizophrenia.
The neurodevelopmental model of schizophrenia holds out the hope of explaining several of the hitherto curious epiphenomena of the disease. For example, schizophrenics are more likely to be born (by some 5% to 15%) in the late winter months. Could it be that this is a consequence of some seasonal difference in the prevalence of viral infections or other perinatal hazard? Second, why do males have an earlier onset of schizophrenia than females? It is well known that males are more prone than females to neurodevelopmental damage, and indeed neurodevelopmental disorders such as autism are much more common in males than in females. Perhaps the earlier onset of schizophrenia in males is caused by their greater Hability to neurodevelopmental damage. Indirect support for this notion comes from the evidence that concordance rates for schizophrenia are lower in male identical twins than in female identical twins.21 A recent study of Goldstein et al22 using data from the large Iowa 500 study also has indicated that the morbid risk is lower in the first-degree relatives of male than female schizophrenic probands. These factors may indicate that a smaller proportion of male schizophrenia is genetically determined.
It will be evident to the reader that we have not mentioned the role of psychosocial factors. There is no doubt that adverse life events frequently precede florid psychotic symptoms. A recent large World Health Organization study examined the occurrence of independent fife events in 386 psychotic patients; an increase occurred in the 3 weeks before onset.23 However, this does not prove that adverse life events can cause schizophrenia; most authorities believe that they precipitate breakdown in individuals already predisposed. A similar explanation probably underlies the evidence (see "Family Factors in Schizophrenia" by J. Leff, BSc, MD, FRCPsych, MRCP, pp 542-547) that excessive criticism, hostility, and emotional overinvolvement on the part of close family members predicts more frequent relapse in their schizophrenic relative.
However, research in the last few years has demonstrated clearly abnormalities in the brains of schizophrenics, particularly enlarged ventricles and hypoplasia of the temporal lobes. The mutant gene(s) and adverse environmental factors that cause these appear to operate by impairing neural development in fetal or neonatal life. Thus, schizophrenia is best regarded as a congenital disorder.
1. Southard EE. On the topographical distribution of cortex lesions and anomalies in dementia praecox with some account of their functional significance. Am J Insanity. 1914/15; 71:383-403.
2. Johnstone EC, Crow TJ, Frith CD, Husband J, Kreel L. Cerebral ventricular size and cognitive impairment in chronic schizophrenia. Lancet. 1976; ii:924-926.
3. Kety SS. Mental illness in the biological and adoptive relatives of schizophrenic adoptees. Am J Psychiatry. 1983; 140:720-727.
4. Murray RM, Reveley AM, McGuffin P. Genetic vulnerability to schizophrenia. Psychiatr Clin North Am. 1986; 9:3-15.
5. Gottesman II, Shields J. Schizophrenia: The Epigenetic Puzzle. Cambridge, England: Cambridge University Press; 1982.
6. Murray RM, Lewis S, Reveley AM. Toward an etiological classification of schizophrenia. Lancet. 1985; i:1023-1026.
7. Bassett A, McGill vray B, Jones B, Pantzar J. Partial trisomy chromosome 5 cosegregating with schizophrenia. Lancet. 1988; i : 799-801.
8. Sherrington R, Brynjjolfsson J, Petursson H, et al. Localization of a susceptibility locus for schizophrenia on chromosome 5. Nature. 1988; 336:164-167.
9. Shelton RC, Weinberger DR. X-ray computerized tomography studies in schizophrenia: a review and synthesis. In: Nasrallah HA1 Weinberger DR, (eds): The Neurology of Schizophrenias. Amsterdam, Netherlands: Elsevier; 1986:207-250.
10. Illowsky BP, Juliano DM, Bigelow LB, Weinberger DR. Stability of CT scan findings in schizophrenia: results of an 8 year follow-up study. J Neurol Neurosurg Psychiatry. 1988; 51:209-213.
11. Crow TJ. The two-syndrome concept: origins and current status. SchizophrBull. 1985; 11:471-485.
12. Suddath RL, Christison G, Torrey EF, Casanova M, Weinberger DR. Quantitative MRI in twin pairs discordant for schizophrenia. Schizophrenia Research. 1989; 2:129.
13. Falkai P, Bogerts B. Cell loss in the hippocampus of schizophrenics. Eur Arch Psychiatry Neurol Sci. 1986; 236:154-161.
14. Brown R1 Colter N, Corsellis JAN, et al. Postmortem evidence of structural brain changes in schizophrenia. Arch Gen Psychiatry. 1986; 43:36-42.
15. Jakob H, Beckman H. Prenatal developmental disturbances in the limbic allocortex in schizophrenics. J Neural Transm. 1986; 65:303-26.
16. Lewis SW, Murray, RM. Obstetric complications, neurodevelopmental deviance and risk of schizophrenia. J Psychiatr Res. 1988; 21:473-421.
17. Owen MJ, Lewis SW, Murray RM. Obstetric complications and schizophrenia: a CT scan study. Psychol Med. 1989; 18:332-340.
18. Murray RM, Lewis SW, Owen MJ. Foerster A. Neurodevelopmental origins of dementia praecox. In: Bebbington P, McGuffin P, eds. Schizophrenia: The Major Issue. London, England: Heinemann; 1988.
19. Weinberger DR. Implications of normal brain development for the pathogenesis of schizophrenia. Arch Gen Psychiatry. 1987; 44:660-669.
20. MednickSA,MachonRA,HuttunenMO, Bonnet D. Adult schizophrenia following prenatal exposure to an influenza epidemic. Arch Gen Psychiatry. 1988; 45:189-192.
21. Kringlen E. Contributions of genetic studies on schizophrenia In: Hafner H, Gattaz WF, Janzarick W. Search for the Causes of Schizophrenia. Berlin, West Germany: Springer-Verlag; 1988.
22. Goldstein JM. Gender and schizophrenia. Schizophrenia Research, 1989; 2:9.
23. Day R, Nielsen JA, Korlen A, et al. Stressful life events precedingthe acute onset of schizophrenia: a cross-national study from the World Health Organization. Cult Med Psychiatry. 1987; 11:123-205.
Neuroimaging Studies of Identical !Wins Discordant for Schizophrenia
Neurodevelopmental Model of Schizophrenia