In contrast to the dementias and the affective illnesses, psychoses characterised mainly by paranoid delusions and hallucinations, but without evidence, during initial assessment and later course, of structural cerebral changes or of an affective psychosis with paranoid symptoms, are diagnosed in only 1-2% of psychogeriatric patients, and have, understandably, engaged only a small number of researchers. Twenty of them were joined by two researchers in less focused aspects of schizophrenia at a meeting of the International Late Onset Schizophrenia Group. With two members from the host country (UK), attenders came from Australia, Canada, Denmark, France, India, Japan, Spain, Switzerland and the USA. Their papers and a consensus statement are presented in this book, which appeared less than 1 year after the meeting, a cause of congratulations to both editors and contributors. The volume contains all that is known or is being discussed on the subject of late-life psychoses and their management. It is required reading for both researchers and clinicians in psychogeriatrics.

Readers will find up-to-date accounts of symptomatology and valuable hints on the management of the first meeting with patients to bring about acceptance and compliance with drug treatment. The mode of action of antipsychotic drugs is elucidated and the atypical ones recommended as less likely to produce troublesome and sometimes irreversible side-effects. Epidemiological studies have confirmed the far greater prevalence of these late paranoid psychoses in women, and this leads a Canadian worker to consider an antipsychotic action of oestrogens and the possible role of future drugs modulating oestrogen receptors. Spanish clinical scientists advocate the role of non-biological, psychological treatments together with drug therapy for patients in whom drug treatment alone had produced little or no symptom relief.

More than half of the book is mainly of theoretical interest, debating whether these late paranoid psychoses belong among the schizophrenias or are diseases sui generis. Roth's descriptive concept of late paraphrenia is attacked, especially by the Swiss contingent, who describe it as both unnecessary and confusing. Their early- and late-onset cases of schizophrenia only differed owing to age influences. Declaring an interest, this reviewer in his 1966 monograph Persistent Persecutory States of the Elderly reported that one-third of his patients had presented with a few paranoid delusions, one-third had more widely spread delusions and related hallucinations and one-third was set apart by the presence of Schneider's first-rank symptoms for the diagnosis of schizophrenia. I left the question of whether all three conditions were schizophrenias open until future workers had unravelled the biological bases of these illnesses. I therefore welcomed the straight-forward declaration by the eminent American schizophrenia researcher, Nancy C. Andreasen: “I don't believe in late onset schizophrenia”. She thinks that recent work makes a convincing case for schizophrenia being a neurodevelopmental disorder and that its symptoms are the result of neural misconnections. However, at an older age these misconnections could not possibly be developmental, but due only to degenerative processes. What Emil Kraepelin and Eugen Bleuler regarded as secondary symptoms (delusions and hallucinations) are produced, but not what they considered primary symptoms, such as formal thought disorder and affective flattening, ambivalence and avolia. This absence of primary symptoms had, in fact, been found by all workers, although a few had thought them doubtfully present on long-term follow-up. In their contributions, the American editor, Rabins, and his British colleague, Howard, also accept the neural misconnection theory, but other chapters report that so far no differences between early- and late-onset cases could be found by neuroimaging or neuropsychological examinations.

Thus, in its consensus statement, the group agrees that for purposes of future research, cases arising between the ages of 40 and 60 years should be called late-onset schizophrenia. Cases with onset after the age of 60 should usually be called very late-onset schizophrenia-like psychoses. A further version of this consensus statement is in press with the American Journal of Psychiatry.