The Differential Diagnosis of Childhood- and Young Adult-Onset Disorders That Include Psychosis

We identified 62 congenital disorders that may present from childhood through middle-age and include psychosis ( Table 1 ). Prominent associated neurological and physical sign groups are listed in Table 2 . Forty-four disorders (71%) have prominent associated neurological features that facilitate differential diagnosis. Seventeen disorders (27%) have readily recognizable unique phenotypes. Forty-five disorders (73%) may present without mental retardation. Fifty-three disorders (86%) have characteristic laboratory features. Fifty-three disorders (86%) have known genetic loci or several different etiologies with known loci (e.g., autism), and three disorders (5%) have loci yet unknown. Five disorders (8%) were due to chromosomal nondisjunction. Sixteen disorders (26%) have estimated prevalence of 1/10,000 or greater ( Table 3 ).

A neuropsychiatric diagnostic approach that takes relative prevalence into account is suggested to simplify a differential diagnosis that includes a large number of rare disorders. The diagnosis of unusual disorders that can present with psychosis receives little if any attention in psychiatry or neurology residency training. Although most of these disorders are unlikely to present to psychiatrists or neurologists more than occasionally, the failure of physicians to recognize them may lead to unnecessary delay in diagnosis, misdiagnosis, or missed opportunities to offer genetic counseling. Furthermore, failure to recommend appropriate treatment or application of inappropriate treatment may lead to adverse outcomes. A substantial minority of these disorders have prominent unique phenotypes that are readily recognizable at a distance, which we have called “doorway diagnoses.” These are listed in Table 4 . Fifty-five percent of disorders identified have subtypes that present without mental retardation or easily recognized phenotypes and are thus easily missed ( Table 5 ).

It would be cost-prohibitive to order a laboratory and neurodiagnostic workup to rule out all of the possible causes of psychotic symptoms, even if that workup were reserved only for those disorders whose natural history is atypical for major axis I psychotic disorders. A coherent neuropsychiatric approach, such as the one presented here, improves cost management by providing a probability-guided, examination-based approach to focus the diagnostic workup.

Identification by literature review of disorders that have been reported to include psychotic symptoms is a first step and may be diagnostically useful, but presents a number of confounding variables that should be addressed. With the exception of the more common disorders, it would be difficult to assemble large enough populations with a given diagnosis over sufficient time to carefully study their psychotic symptoms. Whether the psychotic symptoms are directly caused by the disorder, are a common behavioral response to environmental or medical stress, or are merely coincidental cannot be determined from these reports. Despite the inclusion of misleading terms such as “schizophreniform,” “schizophrenia-like,” or “manic” in many case reports, the histories of many of the reported disorders do not appear to be consistent with axis I disorders such as schizophrenia, schizoaffective disorder, bipolar mania, or major depression with psychotic features, despite phenotypic similarity at the time of presentation. Few congenital disorders have been described that closely resemble schizophrenia. Metachromatic leukodystrophy, which has both negative and positive signs and symptoms and may begin within the typical age range of schizophrenia onset, is one such disorder. Velocardiofacial syndrome, which typically progresses from mood disorder to schizophrenia-like disorder and may account for up to 6% of childhood-onset schizophrenia, is another.

Another potential limitation of this review arises from reliance upon the most common presentations to determine the major associated neuropsychiatric signs. For example, although neurofibromatosis type 1 certainly may be a cause of seizures, seizures are estimated to occur only in 7% of affected individuals; therefore, neurofibromatosis type 1 was not included in the seizure group.

Studying neuropsychiatric disorders of known etiology that include psychosis may ultimately lead to research aimed at understanding the etiology of psychotic symptoms in axis I disorders. With recent improvements in DNA analysis, genes have been identified for the majority of these disorders, many of which are causative for the disorder and not merely associated with one subtype or susceptible population. As genotyping becomes less cost-prohibitive and more commonly available, neuropsychiatrists will need to develop increased sophistication with the utilization of genetic tests in the differential diagnosis of psychosis.

As consultants frequently called upon to evaluate atypical presentations of psychosis, neuropsychiatrists should be aware of congenital disorders that can present with psychosis, however rarely. To minimize missed diagnoses, both general psychiatrists and neuropsychiatrists should become familiar with the phenotypes of the 17 “doorway diagnoses” we have listed ( Table 4 ). We recommend a differential diagnostic approach based on estimated prevalence of the disorders and their most prominent associated neuropsychiatric and physical features, to facilitate appropriate diagnosis in a systematic and cost-effective fashion.

Poster presented at the American Neuropsychiatric Association Annual Meeting in Tucson, Ariz., February 18–20, 2007.