Interest in psychotic disorders in childhood is a relatively recent historical phenomenon. As commonly used, the term psychotic can itself be problematic given the changing nature of children’s conceptions of reality (Volkmar, 2000; Volkmar & Tsatsanis, 2002). At one time, the term childhood psychosis was used quite broadly, but more recently the term and its related diagnostic concepts have been more narrowly defined. The prototypic disorder is childhood-onset schizophrenia (COS) but other psychotic conditions, much less well defined, and transient psychotic phenomena are not uncommon (Box 15.1). When present in children, schizophrenia and related psychotic conditions have the potential for serious disruption of development and hence may be more severe than in adulthood (Ordóñez & Gogtay, 2018). This chapter focuses on onset of schizophrenia in childhood and adolescence and, more briefly, children with psychosis resembling schizophrenia but not meeting current DSM-5 criteria for the condition. Psychosis related to mood disorder is discussed in Chapter 14 and delirium and catatonia in Chapter 27.

Before age 5, psychotic phenomena are uncommon and often stress related. Hallucinations may be seen as isolated phenomena and generally are relatively much more prognostically benign. By school age, the presence of psychotic phenomena is more concerning as such symptoms may persist. Transient psychotic phenomena are less worrisome than more persistent or recurrent ones (Ordóñez & Gogtay, 2018). In this age group, potential factors such as drug exposure or other mental disorders should be considered, for example, hallucinations and psychotic phenomena may also be observed in mood disorders. By the end of middle childhood, the presence of persistent psychotic phenomena becomes much more predictive of subsequent schizophrenia (Ordóñez & Gogtay, 2018).

Kraepelin’s description of “dementia praecox” or what would today be termed schizophrenia was a watershed in psychiatric taxonomy. He explicitly noted that in some cases, onset was in childhood and a downward extrapolation of the concept to children (“dementia praecossisma”) quickly followed and severe psychiatric disturbance in childhood became
equated with schizophrenia (Volkmar & Tsatsanis 2002). Kanner’s description of the syndrome of infantile autism was taken, by some, as a description of the first manifestations of schizophrenia in infants and young children. However, with the pioneering work of Kolvin (1971), Rutter (1972), and others, it became clear that autism (see Chapter 7) was quite different from schizophrenia of childhood onset and the latter was, if anything, much less common than autism. As a result, a category for childhood schizophrenia was dropped and infantile autism officially recognized in DSM-III (APA, 1980). Since then, the same criteria for schizophrenia are used in children as well as in young and older adults (the much more typical period of onset). Changes made in DSM-5 (APA, 2013) to the criteria for schizophrenia include elimination of subtypes and catatonia, and there is an emphasis on discussing current state and historical development (i.e., to better understand the progression of the disorder).
The previous emphasis on Kurt Schneider’s “first-rank” symptoms¹ is replaced by a comprehensive assessment of multiple areas of functioning (reality testing, thought organization, negative symptoms, and so forth), and, for children, information can be facilitated by information from parents, teachers, and others.

An assessment covering eight domains of psychopathology—including reality distortion, negative symptoms, thought and action disorganization, cognition impairment, catatonia, and symptoms similar to those found in certain mood disorders, such as whether hallucination or mania is experienced—is recommended to help clinical decision making. In clinical literature, it is frequent to make a distinction between COS with onset between 13 and 18 years and “very early onset schizophrenia” (VEOS) with onset under 13 years (Volkmar & Tsatsanis, 2002).

Several different patterns of onset have been identified. These include a pattern of acute onset, another with gradual onset, and finally a pattern where there is gradual onset followed by acute exacerbation of symptoms. The insidious-onset pattern is more common. Sometimes symptoms emerge in a child who has previous premorbid difficulties.

Auditory hallucinations are reported most frequently (in about 80% of cases). As with adults, the content might include voices, often with persecutory content, commenting about the child. Somatic and visual hallucinations are also observed less frequently (David et al., 2011). The content of the hallucinations is usually less elaborate in children and often reflects age-appropriate concerns, for example, monsters or toys rather than more sexual themes (Volkmar & Tsatsanis, 2002).

Delusions in children with schizophrenia are often less bizarre and systematic than those observed in adults. There may be concerns with thought broadcasting, thought insertion, and thought withdrawal (Caplan et al., 2000). It is important to note that some children have psychotic features but may not exhibit more classic childhood schizophrenia (Ordóñez & Gogtay, 2018). The National Institute of Mental Health group has described these as multidimensionally impaired (see Kumra et al., 1998, and discussion of differential diagnosis later in this chapter).

Although solid epidemiologic data are critically needed, the childhood form of the disorder is probably 50 times less frequent than adult-onset schizophrenia (Ordóñez & Gogtay, 2018; Volkmar & Tsatsanis, 2002). Among school-age children, schizophrenia is relatively rare on the order of 1 in 10,000 for VEOS and around 5% in COS. The rate increases with age so that by late adolescence the condition appears to be somewhat more common in males, particularly with younger children. As with adult schizophrenia, males also appear to have an earlier onset than females. Also, as with adults, there is probably some bias for increased rates in lower socioeconomic status families. Reported rates may be somewhat underestimating the true prevalence of the condition given lack of clinical familiarity and reluctance to make the diagnosis.

Several factors have been associated with risk for COS. These include older paternal age (Tsuchiya et al., 2005) as well as obstetric complications (Cannon et al., 2002) and family history of the conditions (Asarnow et al., 2001). The strength of these associations has been questioned. Higher rates of cytogenetic abnormalities and potential gene associations have led to the identification of associated specific syndromes, for example, the Smith-Magenis
syndrome (17p11.2del), 15q11-q13 deletions/duplications as well as the 22q11 deletion syndrome (22q11DS), and the velocardiofacial syndrome that is associated with a high risk for schizophrenia. Mosaic Turner syndrome has also been noted at a higher than expected rate (see Ordóñez & Gogtay, 2018 for a more extensive discussion). A number of other associations have been noted as well (e.g., Addington et al., 2004; Ambalavanan et al., 2019).