Editorial: Do Different Neurogenetic Disorders Impart Different Profiles of Psychiatric Risk?

The best-studied examples of genetically defined developmental disorders, such as Down syndrome (trisomy 21) and velocardiofacial syndrome (del22q11), have been known since before the genomic era and were initially recognized as distinct syndromes based on their own unique constellation of dysmorphic and multisystem features. For example, Down syndrome is characterized by the co-occurrence of several dysmorphic features, including a flattened facial profile, slanted palpebral fissures, protruding tongue, and transverse palmar crease, with accompanying hypotonia, cardiac issues, and developmental delay.¹ None of these features in isolation is specific to Down syndrome, and all features are not present in all cases, but the co-occurrence of multiple features from this set is a specific and sensitive marker for the presence of trisomy 21. To what extent might similar principles apply to the patterning of cognitive and behavioral features across different neurogenetic syndromes?

The idea that different genetically defined neurodevelopmental disorders may have not only have distinctive physical presentations but also distinctive behavioral presentations is encapsulated in the term behavioral phenotype, which was first coined by Nyhan² in reference to self-injurious behaviors in Lesch-Nyhan syndrome. Other classic examples of proposed behavioral phenotypes include relatively simple motor or appetitive features, such as handwringing in Rett syndrome or hyperphagia in Prader-Willi syndrome, as well as more complex behavioral styles, such as hypersociality in Williams syndrome. This latter example points toward the possibility that different neurogenetic disorders may be associated with different phenotypic signatures in higher-order domains of cognition and affective processing, which in turn raises the question of whether different neurogenetic syndromes may also be associated with significantly different profiles of psychiatric risk. The answer to this question, which is considered by Glasson et al.³ in this issue of the Journal, not only has major clinical implications but is also relevant for how we think about genetic mechanisms of risk for psychopathology more broadly. For instance, if profiles of psychiatric risk are indistinguishable across different genetic syndromes, this would suggest that knowledge of the specific genes impacted by a given genetic disorder will have limited utility for tailored prediction of clinical risk at the individual level. A lack of clear variegation in psychiatric risk across distinct neurogenetic disorders would also argue for cross-disorder convergence on a common pathway that increases risk for psychopathology in a nonspecific manner—with the specific profile of psychopathology observed in any individual carrier being determined by factors outside the primary genetic lesion.

Despite the importance of these questions, very few studies have compared rates of psychopathology across a range of different neurogenetic syndromes using the same set of standardized measures. This is precisely the approach taken by Glasson et al.³ in their article through a meta-analysis of 39 publications that collectively administered standardized instruments such as the Child Behavior Checklist (CBCL)⁴ or the Developmental Behavior Checklist (DBC) to 4,039 individuals spanning the following 10 different neurogenetic syndromes: Down, velocardiofacial, fragile X, Williams, Prader-Willi, Smith-Magenis, Noonan, Cornelia de Lange, Mowat-Wilson, and Angleman syndromes. Psychopathology was characterized by diagnostic rates as well as the proportion of patients with suprathreshold scores on dimensional symptom scales. Focusing on Total, Internalizing, and Externalizing problem scales from the CBCL, Glasson et al.³ found evidence for both qualitatively and quantitatively distinct risks for psychopathology across different neurogenetic syndromes. Specifically, five of the genetic syndromes (where suitable data were available) could be arranged in order or increasing overall psychopathology, with Down syndrome showing the smallest proportion of individuals with CBCL Total Problem scores above threshold (approximately 12% versus general population reference of 14%), Prader-Willi syndrome showing the largest proportion (> 70%), and other syndromes (del2q11, fragile X, Williams) occupying intermediate positions (lower to higher). Importantly, this ordering of syndromes was not the same for all CBCL subscales, indicating some dissociable in the profile of subscale scores across different syndromes. Moreover, the observed ordering of syndromes by psychopathology was not congruent with reported differences across syndromes in mean IQ.³ In combination with the variable ordering of disorders across CBCL subscales, this finding indicates that cross-syndrome variation in risk for psychopathology is unlikely to simply reflect cross-syndrome variation on the degree of overall cognitive impairment. Collectively, the study by Glasson et al.³ provides a useful reference and a systematic integration of existing studies into rates of psychopathology among youths impacted by different neurogenetic syndromes. This effort represents a much-needed early step forward on the long journey toward more fully characterizing how genetically defined neurodevelopmental disorders modulate risk for mental disorders. As this journey progresses, it will be important to keep some of the following considerations in mind.

First, the accuracy with which we can profile rates of psychopathology in youths with neurogenetic syndromes is fundamentally tied to our ability to fully (or, more realistically, randomly) ascertain all carriers. Such nonbiased ascertainment is very hard to achieve outside selected countries with well-developed national registries that allow unbiased ascertainment of carriers from a population-based sampling frame with complete linkage to medical records (eg, iPSYCH; Lundbeck Foundation Initiative for Integrative Psychiatric Research, Aarhus, Denmark, https://ipsych.dk/). To date, most studies of psychopathology in neurogenetic disorders (including those considered by Glasson et al.³) are based on case-control study designs using clinically ascertained patient cohorts, which are likely to show lower rates of psychopathology than unidentified carriers. This concern is well illustrated by comparing the extremely high early estimates for schizophrenia risk in del22q11 (velocardiofacial) syndrome from clinically selected cohorts (eg, approximately 30-fold increase in risk⁵), with the far lower risk estimates that have come from National Registry data on psychiatric diagnoses in patient cohorts from population-based sampling frames (eg, approximately 6-fold increase in risk⁶). Because sources of ascertainment bias are likely to vary across different neurogenetic disorders (eg, disorders with versus disorders without physical signs and symptoms), study designs that cannot remove or otherwise control for ascertainment bias may induce apparent variation in psychiatric presentation across different genetic disorders.

Second, the ability to detect and fully characterize dissociable profiles of psychiatric risk across neurogenetic syndromes is also heavily influenced by the granularity with which symptoms are profiled in each disorder and whether there are similar measures across all the disorders to be compared. To take an example from the field of developmental neuropsychology, impairments in general cognitive ability are shared across many neurogenetic syndromes, but there are highly reproducible differences between neurogenetic syndromes in the degree to which cognitive impairments impact verbal versus nonverbal subcomponents of general cognitive ability.⁷ Thus, to conduct a robust test for variegated psychiatric risk across different neurogenetic disorders, it is not only important to ensure that we are measuring psychopathology in a detailed manner, but also that the details we capture in our measurement instruments are sampling subcomponents of psychopathology that may be differentially impacted by different genetic disorders. If one can measure only how many wheels a vehicle has, a car can look like a tractor, but the same is true if one’s measurement scale is a highly detailed coding of paint color.

Third, in asking whether different neurogenetic disorders impart different profiles of psychiatric risk, it is important to consider what degree of difference would matter, which in turn relates to the envisaged scientific and/or clinical goal in mind. For example, the proportion of variance explained by neurogenetic subtype may be relatively low across several dimensions of psychopathology that are being considered individually,⁸ with it still being possible to distinguish neurogenetic subtypes by their multivariate score profiles or to identify differential contribution of different psychopathology domains to caregiver strain in each neurogenic subtype.

Finally, and regardless of the aforementioned challenges, there is a pressing need to better characterize and understand the difficulties experienced by individuals affected by genetically defined neurodevelopmental disorders and their primary caregivers. Each of these disorders is a clinically impactful condition in its own right, above and beyond the special potential of such disorders to serve naturally occurring models of genetic risk in humans. However, there remains a widespread shortage of clinicians with specialized training in the assessment and care of those complex mental health issues and co-occurring cognitive impairments that often arise in genetically defined neurodevelopmental disorders.⁹