Congenital risk factors for schizophrenia

Technological advances in medicine often lead to a better understanding of disease processes. Computerized neuroimaging is an advance that has had a more profound impact than most. In schizophrenia, it has shown brain changes of at least a quantitative nature to exist in a significant minority of patients. Until recently, these changes were seen as evidence for an underlying neurodegenerative process: the dementia of dementia praecox. After all, the ventricular enlargement and cortical sulcal widening looked like cerebral atrophy and, from the beginning (Johnstone et al. 1976), a correlation between the radiological changes and degree of cognitive impairment in schizophrenic patients was noted. However, the expected correlation between length of illness and degree of ventricular dilatation stubbornly refused to emerge (Weinberger et al. 1979; Andreasen et al. 1982; Owens et al. 1985). Furthermore, changes of a similar degree were found in young patients only weeks or months after the onset of their first episode of schizophrenia (Turner et al. 1986). Most recently, follow-up computed tomographic studies have shown that the ventricular enlargement does not appear to progress over periods of up to seven years (Nasrallah et al. 1986; Illowsky et al. 1988; Reveley et al. 1988). These findings effectively rule out a progressive brain atrophy and have forced a reappraisal of the nature of such brain changes in schizophrenia. The changes are static, present at and, by inference, before the onset of the illness. Since major neurological events are seldom found to coincide with the onset of schizophrenia, the possibility arises that these changes represent long-standing non-progressive sequelae of much earlier events, events which could themselves be remote precursors of the illness. Consistent with this hypothesis is the handful of reports of sporadic cases of schizophrenia and similar psychoses in conjunction with a variety of gross brain lesions which are more obviously congenital: aqueduct stenosis (Reveley & Reveley, 1983), corpus callosum agenesis (Lewis et al. 1988), cerebral hamartomata (Taylor, 1975) and arteriovenous malformations (Aleem & Knesevich, 1987), perinatal ischaemic encephalopathy (Lewis, 1987), porencephaly (Owens et al. 1980), and septal (Lewis & Mezey, 1985) and arachnoid cysts (Kuhnley et al. 1981). This view is also supported by recent neuropathological studies of the brains of schizophrenic patients. Limbic abnormalities found independently by three groups include unusual gyral patterns and reduced cortical thickness in the hippocampal region (Bogerts et al. 1985; Brown et al. 1986; Jakob & Beckmann, 1986). The reported absence of gliosis accompanying these lesions would only be expected if these were the results of insult in infancy or before (Kolb et al. 1983; Bogerts, 1988). This line of reasoning has quickly become elaborated, drawing on advances in other fields which have also reaped the benefits of new technology, in particular, neonatal medicine and developmental neuroanatomy. There now exist the rudiments of a working hypothesis about the nature of underlying mechanisms. Straws in the wind (Mednick, 1970; Fish, 1977; Bellak, 1979) preceded the recent emergence of fairly well-developed speculations, more or less independently, from several groups (Schulsinger et al. 1984; Murray et al. 1985) about what might broadly be called a 'neurodevelopmental' approach to schizophrenia (Murray & Lewis, 1987).