Findings from this study support some past research suggesting that children with the 22q11.2 deletion syndrome may exhibit ASDs at a markedly higher rate than that found in the general population. However, although a diagnosis of an ASD was appropriate for approximately 14% of the sample, this proportion represents a substantially lower number than previously suggested (Niklasson et al., 2001
). The explanation for this discrepancy becomes clear when taking the methods of these studies into account. The present study used strict diagnostic criteria for the purpose of screening and classifying children as exhibiting levels of symptoms commensurate to what one might observe in a child with an ASD. Although previous studies also used screening measures, the present study included a more in-depth diagnostic interview to further explore children’s symptoms. Use of these stringent criteria determined that several children whose caregivers had indicated were exhibiting significant levels of autism spectrum symptoms did not meet requirements to receive a diagnosis of an ASD.
Even when applying the strict diagnostic criteria by administering the ADI-R, the findings of the present study appear to conflict with those of Ogilvie et al. (2000)
, who reported that none of the children with autism in their large sample of individuals from multiplex families exhibited a deletion on chromosome 22. There are some possible explanations for the discrepancies in these findings. First, we used the ADI-R (2003, unpublished at time of administration) version of the diagnostic interview, which was revised from the previous versions to improve differentiation between autism and other developmental disorders such as fragile X, particularly in younger children. The revision was also intended to better distinguish between delays in development and developmental deviance, which was important in the current study because of the prevalence of developmental delays. The differences between the ADI (the algorithm from which Ogilvie and colleagues used to confirm a diagnosis of autism in their sample) and the revision may have enabled us to detect autism symptoms in a sample of children already identified as having special needs. Furthermore, Ogilvie and colleagues determined the occurrence of the 22q11.2 deletion in a unique sample of children with autism, limiting their investigation to only individuals from multiplex families. This constraint limits the generalizability of their findings, as multiplex cases of autism may represent a specific genetic subgroup unto themselves. Finally, the 22q11.2 deletion occurs less frequently in the general population than does autism, and thus the Ogilvie et al. (2000)
study may not have had sufficient power to accurately estimate the prevalence of the 22q11.2 deletion in individuals with autism. Although additional studies are required to replicate the findings of the present study, it does appear as if autism spectrum disorders occur more frequently in the population of children with the 22q11.2 deletion than in the general population.
Characteristics of children in the sample who met criteria for autism based on their caregivers’ reports were similar to those reported in the general population of individuals with ASDs. Even when taking into account the disproportionate percentage of males in the sample, the male-female ratio in children who met criteria for autism was 3–4:1, which is consistent with past research (Burd, Severud, Kerbeshian, & Klug, 1999
; Fombonne, 1999
; Steffenburg & Gillberg, 1986
). In addition, although the sample in general tended to be mildly globally developmentally delayed, children who met criteria for ASD tended to have more severe delays, which is in line with past research that has suggested that ASD is often co-morbid with severe to profound cognitive impairments (Fombonne, 1999
; Gillberg et al., 1990
; Lord & Volkar, 2002
; Sigman & Capps, 1997
Reports from caregivers in the present study suggest that perhaps many children who have the 22q11.2 deletion and are exhibiting symptoms of ASDs may not be formally diagnosed with autism. There are several explanations for the possible under-identification of ASDs in this population. The 22q11.2 deletion is often identified at birth, making caregivers immediately aware that their child may not develop typically. Therefore, caregivers may notice unusual or atypical behaviors as their children develop, but they may integrate these behaviors into their conceptualization of the genetic disorder and be less likely to question them than parents of children who were not identified at birth as having a special condition. Anecdotally, some caregivers in the study who completed the ADI-R attributed their children’s behaviors directly to the 22q11.2 deletion rather than to another distinct psychiatric or developmental disorder. Additionally, because of the lack of general knowledge about individuals with the 22q11.2 deletion, many of their caregivers are placed in the unfortunate position of having to educate and inform the professionals who are working with them. Professionals observing and evaluating these children may also integrate autism spectrum-like behaviors into a broader conceptualization of the 22q11.2 deletion rather than identifying them as fitting into a particular diagnostic category.
Another explanation for the under-identification of ASDs in children with the 22q11.2 deletion is that medical procedures and hospital stays in early childhood (which these children often experience) may further obscure caregivers’ perspectives of their children’s behaviors and abilities. For example, losses of skills or regressions in one area of development, common to children with ASDs, were sometimes explained by caregivers in the present study as the result of long and trying hospitalizations. Surgical procedures could also obscure possible delays in language and social responsiveness. Aggressive behaviors may be dismissed by parents who have been close observers of their child’s hardships and who perceive such behavior solely as frustration due to physical inability to speak or ambulate. Finally, in the context of stressful medical care, particularly early in life, symptoms of ASDs may be de-prioritized when caregivers, who are simply thankful that their child is surviving, discuss issues with professionals.
One weakness of the present study is method variance. Due to the nature of the sample, which included families residing in several different geographical regions throughout the United States, direct observation of the children at The Children’s Hospital of Philadelphia during the time frame of the present study was not feasible. Therefore, evaluation was conducted via mail survey and telephone interviews with the caregivers of these children. We attempted to mitigate potential reporter bias by using some measures that were scored by a trained interviewer rather than caregivers themselves. Although this type of interview measure is still influenced by caregiver reports, the number value assigned to a particular behavior is based on concrete descriptions coded by a trained interviewer rather than subjective impressions. Despite use of these measures, however, the results should be interpreted in light of the fact that caregivers provided all of the information.
Another potential limitation of the present study is the use of the ADI-R to detect ASD within this sample of children with an established genetic diagnosis. The ADI-R has been revised to discriminate ASD from other genetic disorders (e.g., Fragile X). It is possible, however, that the estimated prevalence of ASD in this sample of children with 22q11.2 deletion may have been somewhat inflated due to a measurement artifact. That is, the ADI-R may not be able to reliably discriminate behaviors that characterize children with ASD from behaviors of children with 22q11.2 deletion. This potential weakness in the ADI-R, which certainly applies to other measures such as the Vineland Adaptive Behavior Scales, highlights the importance of using multiple methods to carefully and thoroughly assess ASD, particularly in children who have genetic diagnoses.
Certain characteristics of the sample may limit the generalizability of these findings. There is a possibility of ascertainment bias, in that all children described in the present study had experienced manifestations of the 22q11.2 deletion that were apparent enough to warrant genetic testing. This characteristic suggests that they may have been more seriously affected than other children who have the deletion but have not experienced the structural anomalies such as congenital heart defects or the other health problems such as hypocalcemia associated with the deletion. It is also possible that parents of children who were more seriously affected by the 22q11.2 deletion were more motivated to respond to the invitation to participate than parents of children with fewer difficulties. As a result, the findings may represent an overestimate of the prevalence of ASD in the population of children with 22q11.2 deletion. Moreover, participating families had the time and resources to allow their children to travel to The Children’s Hospital of Philadelphia, in some cases from considerable distances, to participate in the larger study. They were further able to take the time to participate in the current study over the telephone, which may mean that they represented a slightly more economically privileged group. Given these sample characteristics, findings should be interpreted with caution, as this sample may not be fully representative of all individuals with the 22q11.2 chromosomal deletion.
Implications and Future Directions
This study represents one of the initial steps in determining the prevalence of ASDs in individuals with the 22q11.2 deletion. Future studies should employ direct observation measures, such as the Autism Diagnostic Observation Schedule (ADOS; Lord et al., 2000
), to evaluate these children for symptoms of autism. Although it would be difficult to include a geographically diverse sample, a secondary study might utilize a face-to-face administration of the ADI-R. A replication of the results would not only support the findings, but would also support the use of the ADI-R through telephone administration. Replication would broaden the scope of this tool and its availability for broad-based studies by supporting its use through a telephone interview, rather than requiring a face to face interview. Others have already presented findings supporting this method of administration with a screening version of the ADI-R (Vrancic et al., 2002
Diagnosis is crucial to our understanding of disease. Through this investigation and future ones, we are obtaining a broader understanding of the very meaning of autism and the 22q11.2 deletion syndrome. Such research provides better understanding of co-morbid conditions and fleshes out our understanding of the genetic components of these disorders. Future research must continue to relate overlapping disorders while simultaneously providing more definitive definitions of ASDs. Possibly, ASDs overlap with other disorders, such as the 22q11.2 deletion, in only a few specific subgroups. Subsequent studies should explore this possibility in order to increase our knowledge of potential genetic subtypes of autism (Bassett & Chow, 1999
; cf. Ogilvie et al., 2000
). Further research focusing on linkages between specific genes and ASDs is greatly needed, as phenotype research on family characteristics and personality types remains imprecise and relies upon clinical interviews that require further testing and standardization (e.g., M-PAS, FHI; Bolton et al., 1994
; Folstein et al., 1999
; Piven, Palmer, Jacobi, Childress, & Arndt, 1997
; Szatmari et al
Findings from the present study may directly impact current practices and medical care for children with a 22q11.2 deletion. In this study, hospital stays and previous diagnoses might have delayed the detection of ASDs. Parents may not be have been prepared to recognize behaviors that did not confirm the primary diagnosis or may have ignored these behaviors as the result of the trauma of their child’s hospitalization. These factors suggest that certain situations, such as prematurity or regular hospital stays, require careful attention and consideration of missed diagnoses. Future studies should investigate methods to prevent biased interpretations of symptoms in the context of the 22q11.2 deletion. In addition, it might be fruitful to provide all children with a 22q11.2 deletion with more stringent developmental assessments that specifically rule out or confirm ASDs.
Given the host of difficulties faced by many children with a 22q11.2 deletion, such as chronic medical conditions, learning disabilities, and other psychiatric issues, some may question the incremental value to families of formally diagnosing an ASD. However, the existence of empirically supported treatments and interventions for individuals with autism means that children who are diagnosed can have access to early intervention and ongoing special services that can improve social, behavioral, and language functioning (Goldstein, 2002
; Horner, Carr, Strain, Todd, & Reed, 2002
; McConnell, 2002
). Diagnosis provides understanding to parents and informs their attributions about their children’s behaviors. Rather than treat behaviors related to the ASD as naughty, parents can find therapeutic ways to mitigate behavioral difficulties. In addition, diagnosis prepares parents for areas of difficulty that develop in the later years of autism, including depression among higher functioning individuals (Ghaziuddin, Ghaziuddin, & Greden, 2002
; Volkmar et al., 1999
) and the need for residential living and preparation for sheltered occupations (Frith, 1991
). An accurate diagnosis of ASD may expedite acquisition of services and be particularly empowering for families of children with a 22q11.2 deletion, who are too often faced with informing medical and educational professionals about this little known but relatively common genetic disorder, about which we still have so much to learn.