Using his father’s term for dementia praecox, ‘schizophrenia’, Manfred Bleuler¹⁹ described individuals with ‘late-onset schizophrenia’ as those with onset after age 40 exhibiting symptoms similar to those with an earlier onset of the disorder and no evidence of brain disease. Very few of these patients had onset after the age of 60. This classification was adopted by most subsequent German authors⁹.

In the UK during this period, however, the classification of late- onset psychotic disorders took a somewhat different path. Studying a group of patients with onset after age 60, Roth and Morrissey²⁰ described a syndrome of paranoid delusions and hallucinations in the context of preserved intellect, personality and affect. Because of the phenomenological similarity to Kraepelin’s ‘paraphrenia’ and due to its late onset, Roth and colleagues termed this disorder ‘late paraphrenia’^21,22, a name that was designed to encompass all late- onset, non-affective, non-organic psychoses in which paranoid symptoms were prominent. Thus, the term was both broader than late-onset schizophrenia, in that it encompassed late-onset delusional disorder, and more restrictive, in that it did not include non-paranoid forms of late-onset psychosis. Post²³ developed a different descriptive system. He divided late-onset (after age 50) psychoses into paranoid hallucinosis, schizophreniform syndrome, and schizophrenic syndrome. Based on a three-year follow-up, however, he concluded that these three diseases were actually a continuum of the same disorder with slightly different symptom profiles.

European debates and developments were slow to influence the classification system used in the USA. The first Diagnostic and Statistical Manual of Mental Disorders (DSM-I)²⁴ used the term ‘involutional psychotic reaction’, which encompassed both paranoid ideation and depression in older patients. This amalgam of affective and psychotic symptoms in the elderly was split in the second edition (DSM-II)²⁵ in favour of ‘involutional paranoid state (involutional paraphrenia)’ and ‘involutional melancholia’. The former disorder, like Roth’s late paraphrenia, was characterized by ‘delusion formation with onset in the involutional period. The absence of conspicuous thought disorders typical of schizophrenia distinguishes it from that group’²⁵. Schizophrenia could be diagnosed in individuals with any age of onset.

As European psychiatrists began to study patients with late para- phrenia more systematically, new classification systems in the USA were restricting the diagnosis of late-onset psychosis. One of the five Feighner Research Criteria²⁶ for schizophrenia was age of onset before age 40. In the third edition of the DSM (DSM-III)²⁷, a diagnosis of schizophrenia could not be made if the onset of symptoms was after age 45. Late-onset psychosis that involved persistent per- secutory delusions with prominent hallucinations could be given a diagnosis of ‘paranoid disorder’. This classification system was in stark contrast to both earlier RDC (research diagnostic criteria)²⁸ and to the ninth version of the International Classification of Diseases (ICD-9)²⁹, neither of which imposed age-of-onset restrictions for schizophrenia. The ICD-9 also allowed for a diagnosis of paraphrenia at any age. The revised version of DSM-III (DSM-III-R)³⁰ rectified the omission of late-onset schizophrenia by providing a separate diagnostic category for those diagnosed with schizophrenia after age 45. In the most recent versions of the DSM (DSM-IV-TR)⁴ and the ICD (ICD-10)³, no special categories exist for late-onset psychoses, although schizophrenia may be diagnosed at any age.

It is clear from this historical review that there has been little consensus regarding the classification of late-onset, non-affective, non- organic psychoses. Two opposing lines of thought have pulled the terminology in different directions. On the one hand, some authors have preferred to emphasize the similarity of late-onset psychoses to the corresponding early onset disorders. This has resulted either in the use of terms such as ‘late-onset schizophrenia’ and ‘late-onset delusional disorder’ or has prompted a move toward ignoring age of onset altogether in classification (e.g. DSM-IV-TR⁴, ICD-10³). On the other hand, some members of the psychiatry community (mainly in Europe) have preferred to emphasize differences between the phenomenology of late- and early onset psychosis and thus have tended to use distinct terminology, such as ‘paraphrenia’ or ‘late paraphrenia’³¹.

Thus, questions remain about which terminology would optimally serve the clinical and research communities. These questions have become particularly timely with the development of DSM-V^32,33