Nitin Gogtay, MD

Although psychotic illnesses are fortunately rare in children, contrary to common belief, psychotic symptoms can be fairly common in very young healthy children. As a symptomatic child grows older, these experiences tend to become much less frequent but, if symptoms persist, they can be indicators of increased risk for more serious mental illness and poorer outcomes later in life.

Given the reasonably common occurrence of hallucinations in otherwise healthy children it is critically important that the symptoms be carefully characterized and a diagnosis such as schizophrenia be made with extreme caution in children. For example, an “imaginary friend,” or hallucinations while transitioning either into asleep (hypnogogic) or wakefulness (hypnopompic) are common and should not be considered problematic in healthy children. The important challenge comes when hallucinations or other psychotic experiences are more pervasive, serious, and pose a diagnostic and treatment challenge.

Although clinicians were already aware by the early 20th century that schizophrenia could manifest very early in childhood, the diagnostic criteria were unclear. This was probably because the term “psychosis” was often loosely used in children to describe a range of symptoms from behavioral disturbances to those on the autism spectrum; a problem that has continued to date. To complicate things further, psychotic disorders in children often have a complex and severe presentation that makes it harder to tease apart the origins of symptoms.

In 1990, the National Institute of Mental Health (NIMH), launched a study on childhood onset schizophrenia (COS).

 This study, the only one of its kind in the world, was prospective and had a comprehensive design that encompassed clinical diagnosis, risk factors, neuropsychological and brain development, and genetic underpinnings. The effort continued for 25 years and generated the largest cohort of well-characterized children with COS in the world. The insights gained were critical both for the understanding of childhood psychotic illnesses as well as for schizophrenia in general and comprise most of the literature in the neurobiology and clinical course of COS.

For COS diagnosis, the NIMH COS study required the onset of psychosis to be before age 13, premorbid IQ of 70 or above, and absence of a neurological disorder. The screening process was extremely thorough: it included inpatient observation with a complete medication washout in most cases. A startling diagnostic observation was that almost 30% of children who met the screening criteria for suspicion of COS (and who were treated as such) had other diagnoses (such as anxiety disorders, mood disorders) after careful observation and medication washout. The diagnosis remained stable at long-term follow up, which highlights the importance of spending the time and effort for diagnostic characterization.

Clinically, COS resembles severe, chronic, poor outcome adult onset schizophrenia (AOS) cases. However, the psychosis of COS is usually severe, pervasive, and remarkably “non-episodic and unremitting” from the initial presentation. A detailed history usually reveals poorer premorbid functioning in social, motor, and language domains, learning disabilities, and disruptive behavior disorders, which suggests widespread brain dysfunction from the get-go. Commonly associated are transient autistic symptoms such as hand flapping and echolalia and upon careful evaluation almost 29% show comorbid autism spectrum disorder.

Careful evaluation of differential diagnosis is critical because, as learned from the imaging analyses from the NIMH COS study, the underlying brain pathology is highly diagnostically specific. Several childhood conditions can manifest with psychotic symptoms and/or deterioration in function. Pediatric mood disorders, such as pediatric bipolar disorder or MDD can present with psychotic symptoms, although they tend to be mood congruent unlike those in COS where the psychosis is much more pervasive.

Autism spectrum disorders and childhood disintegrative disorder have severe impairment in reciprocal communication, social interactions, and odd stereotyped behaviors and thus can be mistaken for the severe psychosis of COS. Behavioral disturbances such as conduct disorder can be associated with hallucinations and psychosis. They may also be an outward manifestation of medical conditions such as diseases due to metabolic errors; neurological illnesses, such as epilepsy; or substance use disorders. A carefully collected patient history and good medical and laboratory work up as well as observations in a neutral setting, and preferably without medications, can usually distinguish these other conditions from the pervasive psychosis of COS.

A group of unique patients who often pose a common diagnostic dilemma is worth special mention. In the NIMH COS study, a subgroup of patients had real but transient psychotic symptoms. The main presentation for these children was severe, daily periods of emotional/mood lability disproportionate to the precipitating triggers. Characteristically these children had difficulty forming relationships and multiple developmental deficits (eg, ADHD, learning disabilities, auditory processing abnormalities).

A careful interview revealed lack of formal thought disorder and the psychotic symptoms were transient at best. For lack of anything better, the DSM diagnosis was multidimensionally impaired (MDI). The MDI cohort, not uncommon and often mislabeled, appeared to have a distinct long-term clinical course, with none progressing to schizophrenia.

While considering the differential diagnosis for schizophrenia in a child with psychosis, it is important to recognize that COS could co-exist with many comorbid psychiatric disorders. Because symptom manifestations can also be part of (or masked by) the symptoms of schizophrenia, a diagnosis of an independent Axis I condition can often be ignored. The most frequent comorbid diagnoses in the NIMH COS cohort were depression (54%), followed by OCD (21%), generalized anxiety disorder (15%), and ADHD (15%). The rate of any one anxiety disorder (generalized anxiety disorder, OCD, separation anxiety, PTSD, panic disorder combined) at screening was 42%. Furthermore, comorbid diagnoses (particularly depression) correlated with poorer overall functioning and continued comorbidity, which suggests either a refractory nature of these conditions, or a close association with core schizophrenia pathology. Importantly, there were no significant associations between comorbid diagnoses and IQ, familiarity, medication status, premorbid functioning, or age of psychosis onset.

The NIMH COS study spent significant effort in establishing the neurobiological continuity of COS and more typical AOS. All published neurobiological risk factors were examined carefully in patients with COS. As expected, COS showed a similar yet exaggerated pattern of risk factors that may have reflected greater impairment in early brain development, but also supported the continuity between COS and AOS. These observations included higher rates of early developmental abnormalities (language, social, and motor domains), larger effect sizes for smooth pursuit eye movement abnormalities, and higher incidence of familial schizophrenia spectrum disorders.

Neuropsychological function in COS warrants special mention as it is commonly evaluated in clinical practice. This has been well studied in the NIMH COS cohort as well as by some other groups studying COS or adolescent onset schizophrenia. Analyses on the NIMH COS cohort showed a clear decline in IQ post-onset of psychosis, but more importantly, this drop in cognitive function did not continue even after 8 years. The incidence of neuropsychological deficits for COS were not significantly higher than for AOS per se, but healthy siblings of children with COS also shared some neuropsychological deficits, suggesting a more salient genetic/ familial link in COS.

Prospective brain imaging studies in COS patients and their siblings have contributed to the understanding of abnormal brain development in COS and have also provided critical insights into brain development in schizoprehnia in general.

 Most of these studies have originated from the NIMH COS cohort.

In addition to replicating the brain abnormalities seen in AOS (eg, large ventricular volumes, smaller gray matter volumes), the finer brain mapping studies in COS showed that there is profound gray matter loss in COS during adolescence that mimics the normal cortical gray matter maturation pattern, which suggests that the healthy pruning process in normal maturation may be exaggerated in COS. This gray matter loss, however, stabilizes over time and the pattern merges with that seen for AOS, thus establishing biological continuity between the two.

Young healthy COS siblings share this gray matter deficit pattern during their younger years, but the deficits normalize by late adolescence, which suggests that the gray matter loss in COS is an age-dependent trait marker. More importantly, the profound gray matter abnormalities are diagnostically specific, unrelated to medication exposure, and appear proportional to the degree or severity of psychosis. The abnormalities are not limited either to cortex or only to gray matter as white matter and deeper cortical structures also share the deficits and the deficit pattern. Connectivity analyses using structural, functional, and diffusion tensor imaging, show specific neurocircuitry abnormalities in circuits relevant to psychosis.

As was expected, COS shows more salient genetic underpinnings. The patients in the NIMH COS study had a significantly higher incidence of velocardiofacial syndrome (with spontaneous 22q11.2 deletion) than AOS patients or than any clinical population. Patients with COS also had a higher incidence of cytogenetic abnormalities as well as other common and rare variant alleles. None of the genetic abnormalities have explained the profound brain abnormalities, but the obviously higher salience suggests that continued effort will eventually provide valuable insights.

Childhood-onset schizophrenia is a devastating disorder that is treatment-resistant. It is important to realize that the NIMH sample may have a referral bias in that patients who responded to treatment early were unlikely to be referred to the NIMH. Thus, in the absence of evidence-based guidelines to specifically treat children with schizophrenia, treatment guidance for schizophrenia in general should be followed. This includes atypical (and/or typical) antipsychotic medications as first-line treatment and clozapine as the final resort treatment after at least two failed trials of adequate antipsychotic treatment with two agents. The only two randomized controlled trials (RCTs) established superiority of typical antipsychotics over placebo in COS, and a single trial demonstrated superiority of clozapine over haloperidol. Subsequently, results from a relatively small double-blind RCT showed that clozapine was better than olanzapine in alleviating negative symptoms. These data and the observations made during the NIMH study have led to the understanding that clozapine is the best option for this extremely ill cohort.

Unfortunately, as anticipated, clozapine is associated with more adverse effects, including risk of agranulocytosis, significant weight gain, akathisia, enuresis, tachycardia, orthostatic hypotension, and seizures. Managing the adverse effects when treating with clozapine is a major challenge where the observations from the NIMH COS study provide critical insights.

The most significant adverse effect is neutropenia that can lead to agranulocytosis. The FDA requires weekly blood monitoring with neutropenia, often resulting in discontinuation of clozapine with potentially disastrous clinical worsening. This can be successfully managed by the addition of lithium.

Weight gain is pronounced with clozapine particularly during childhood. Along with significant psychoeducation for life-style changes, the antidiabetic medication metformin (which improves peripheral insulin sensitivity) showed some success when started with clozapine.

Akathisia, seen rarely in adults taking clozapine, is surprisingly common in children and can manifest as worsening of psychotic symptoms or agitation in children. This often results in dose increments that further worsens the akathisia. Akathisia is responsive to adjunctive beta blocker (eg, propranolol) treatment.

Cardiac adverse effects such as orthostatic hypotension and tachycardia need regular monitoring and can be managed by adjusting the dose, fluid intake, activity, or medications such as beta blockers. Incontinence is not uncommon with clozapine but although it is responsive to vasopressin, it also tends to be get better with longer treatment. Drooling also gets better with time, but children need to be taught to manage it daily.

Finally, clozapine can lower the seizure threshold and can induce seizures. Regular EEG monitoring and prophylactic antiepileptic medication such as gabapentin has helped manage this adverse effect. Metabolic changes from chronic clozapine use (such as hyperlipidemia, hyperglycemia) should be monitored regularly so that appropriate treatment can be initiated. Clozapine may be the best medication available to treat COS symptoms and improve functionality, but careful monitoring and managing adverse effects is a challenge.

Calkins ME, Moore TM, Merikangas KR, et al. The psychosis spectrum in a young U.S. community sample: findings from the Philadelphia Neurodevelopmental Cohort. 

Nicolson R, Rapoport JL. Childhood-onset schizophrenia: rare but worth studying. 

Ordonez AE, Driver DI, Orloff M, et al. Clozapine treatment in pediatric onset schizophrenia: lessons from the NIMH experience. 

Dr Gogtay is Director, Office of Clinical Research, National Institute of Mental Health, Rockville, MD.

 Rapoport JL, Gogtay N. Childhood onset schizophrenia: support for a progressive neurodevelopmental disorder. 

 Ordonez AE, Gogtay N. Childhood onset schizophrenia and other early onset psychotic disorders. Martin A, Block MH, Volkmar, FR, Eds. 

Lewis’s Child and Adolescent Psychiatry: A Comphrehensive Textbook

, 5th ed. Philadelphia, PA: Walters Kluwer Health; Chapter 5.3.

 Ordonez AE, Gogtay N. Phenomenology and neurobiology of childhood onset schizophrenia. 

 Gogtay N. Cortical brain development in schizophrenia: Insights from neuroimaging studies in childhood-onset schizophrenia. 

 Moran ME, Hulshoff Pol H, Gogtay N. A family affair: brain abnormalities in siblings of patients with schizophrenia. 

Articles

Although psychotic illnesses are fortunately rare in children, contrary to common belief, psychotic symptoms can be fairly common in very young healthy children.

Nitin Gogtay, MD

Articles

Neurobiology and Clinical Management of Childhood Onset Schizophrenia