Delineating the profile of autism spectrum disorder characteristics in cornelia de lange and fragile x syndromes.
ABSTRACT Abstract An atypical presentation of autism spectrum disorder is noted in Cornelia de Lange and Fragile X syndromes, but there are few detailed empirical descriptions. Participants in this study were individuals with Cornelia de Lange syndrome (n = 130, M age = 17.19), Fragile X syndrome (n = 182, M age = 16.94), and autism spectrum disorder (n = 142, M age = 15.19), who were comparable on chronological age. Using the Social Communication Questionnaire, the proportion meeting cutoff for autism spectrum disorder and autism was 78.6%, and 45.6%, respectively, in Cornelia de Lange syndrome and 83.6% and 48.6% in Fragile X syndrome. Domain and item analyses indicate differing, atypical autism spectrum disorder profiles in Fragile X and Cornelia de Lange syndromes. A limited association between adaptive behavior and autism spectrum disorder was identified in all groups. The findings have implications for intervention in genetic syndromes and conceptualization of autism spectrum disorder in the wider population.
- [Show abstract] [Hide abstract]
ABSTRACT: Recent research shows that a significant minority of children with Down's syndrome (DS) also meet diagnostic criteria for an autism spectrum disorder (ASD). The present study investigated what proportion of children aged 6–15 years with a confirmed diagnosis of DS in England and Wales display autistic-type behaviours, and explored the characteristics of this group of children. The Social Communication Questionnaire (SCQ) was used to screen for autism characteristics and the Strengths and Difficulties Questionnaire (SDQ) to explore behavioural difficulties. The proportion of children who met the cut-off score for ASD on the SCQ (total score ≥ 15) was 37.7% (95% CI: 33.4–42.0%); for autism (total score ≥ 22) the proportion was 16.5% (95% CI: 13.2–19.8%). Children who met the cut-off for ASD were significantly more likely to be reported as having emotional symptoms, conduct problems and hyperactivity on the SDQ than children who scored well below cut-off (total score < 10). However, the profile of their autism characteristics on the SCQ was atypical compared with individuals with idiopathic ASD. The pervasiveness of ASD in children with DS in England and Wales is substantially higher than in the general population. These children also experience significantly greater behavioural problems than children with DS only. Early detection of autism characteristics is important for appropriate intervention. However, the unusual profile of autism characteristics in this group may affect the recognition of the disorder and hinder the implementation of appropriate interventions. Autism Res 2014, ●●: ●●–●●. © 2014 International Society for Autism Research, Wiley Periodicals, Inc.Autism Research 03/2014; 7(4). · 3.99 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Abstract We evaluated autism spectrum disorder (ASD) characteristics and social behavior in Angelman (AS; n = 19; mean age = 10.35 years), Cornelia de Lange (CdLS; n = 15; mean age = 12.40 years), and Cri du Chat (CdCS, also known as 5 p-syndrome; n = 19; mean age = 8.80 years) syndromes. The proportion of individuals meeting the ASD cutoff on the Social Communication Questionnaire was significantly higher in the AS and CdLS groups than in the CdCS group (p < .01). The groups demonstrated divergent social behavior profiles during social conditions in which adult availability, adult familiarity, and social demand were manipulated. Social enjoyment was significantly heightened in AS, whereas social approaches were heightened in individuals with CdCS. Social motivation, social communication, and enjoyment were significantly lower in CdLS. The findings highlight the importance of detailed observation when evaluating ASD and social behavior in genetic syndromes.American Journal on Intellectual and Developmental Disabilities 07/2013; 118(4):262-83. · 2.08 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Cornelia de Lange syndrome (CdLS) is a cohesinopathy causing delayed growth and limb deficits. Individuals with CdLS have mild to profound intellectual disability and autistic features. This study characterizes the behavioral phenotype of children with CdLS, focusing on autistic features, maladaptive behaviors, and impact of age. Children with CdLS (5–18 years) were administered normed instruments to characterize autism features (Childhood Autism Rating Scale, CARS), maladaptive behaviors (Aberrant Behavior Checklist), and adaptive skills (Vineland Adaptive Behaviors Scales). CdLS features and severity were rated with Diagnostic Criteria for CdLS. Forty-one children with CdLS (23 females, 18 males) were classified as having “no autism” (n = 7; 17.1%), “mild autism” (n = 17; 41.4%), and “severe autism” (n = 17; 41.4%), using CARS scores. Characteristic items were abnormal emotional response, stereotypies, odd object use, rigidity, lack of verbal communication, and low intellectual functioning. Verbal communication deficits and repetitive behaviors were higher compared to sensory, social cognition, and behavior abnormalities (P ≤ 0.0001). Maladaptive behaviors associated with autism traits were stereotypies (P = 0.003), hyperactivity (P = 0.01), and lethargy (P = 0.03). Activities of daily living were significantly affected; socialization adaptive skills were a relative strength. However, with advancing age, both socialization (P < 0.0001) and communication (P = 0.001) domains declined significantly. CdLS is characterized by autistic features, notably excessive repetitive behaviors and expressive language deficits. While other adaptive skills are impacted, socialization adaptive skills are less affected. Advancing age can worsen communication and socialization deficits relative to neurotypical peers. © 2014 Wiley Periodicals, Inc.American Journal of Medical Genetics Part A 04/2014; · 2.30 Impact Factor
ASD in CdLS and FXS
Delineating the profile of Autism Spectrum Disorder
characteristics in Cornelia de Lange and Fragile X
Moss, J., Oliver, C., Nelson, L., Richards, C. and Hall, S.
Cerebra Centre for Neurodevelopmental Disorders,
School of Psychology,
University of Birmingham
Please use this reference when citing this work:
Moss, J., Oliver, C., Nelson, L., Richards, C. & Hall, S. S. (2013). Delineating the profile of Autism
Spectrum Disorder characteristics in Cornelia de Lange and Fragile X syndromes. American Journal
on Intellectual and Developmental Disabilities, 118, 55-73.
The Cerebra Centre for Neurodevelopmental Disorders,
School of Psychology, University of Birmingham, Edgbaston, Birmingham, B15 2TT
ASD in CdLS and FXS
We are grateful to the Cornelia de Lange Syndrome Foundation, UK & Ireland, the Fragile X
Society, UK, the National Autistic Society and Research Autism UK for their help with participant
recruitment, to Ms (anonymised for blind review) for her help with data collection and data entry and
to all the families who participated in the study. Work conducted in this study was funded by a PhD
studentship from the Cornelia de Lange Syndrome Foundation, UK & Ireland (data collection for
CdLS and FXS groups), a PhD studentship from Research Autism UK (data collection for ASD
reference group) and by Cerebra.
ASD in CdLS and FXS
An atypical presentation of Autism Spectrum Disorder (ASD) is noted in Cornelia de Lange (CdLS)
and Fragile X syndromes (FXS) but there are few detailed empirical descriptions. Participants were
individuals with CdLS (N=130; mean age=17.19), FXS (N=182; mean age=16.94) and ASD
(N=142; mean age=15.19), who were comparable on chronological age. Using the Social
Communication Questionnaire, the proportion meeting cut off for ASD and autism was 78.6% and
45.6% in CdLS respectively and 83.6% and 48.6% in FXS. Domain and item analyses indicate
differing, atypical ASD profiles in FXS and CdLS. A limited association between adaptive behavior
and ASD was identified in all groups. The findings have implications for intervention in genetic
syndromes and conceptualization of ASD in the wider population.
Key words: Autism Spectrum Disorder, Cornelia de Lange syndrome, Fragile X syndrome,
ASD in CdLS and FXS
The last decade has seen increasing research interest into the association between a range of
neurodevelopmental disorders of genetic origin (genetic syndromes) and the presence of autism
spectrum disorder (ASD) and/or ASD symptomatology. The number of genetic syndromes in which
such an association has now been described is growing and includes (amongst others): Angelman,
Cohen, Williams, Fragile X, Rett, Cornelia de Lange, Down, 22q11 deletion and Prader Willi
syndromes and Tuberous Sclerosis Complex (for reviews see Fombonne, 1999; Gillberg & Coleman,
2000; Moss & Howlin, 2009; Moss, Oliver & Howlin, 2011). It has been suggested that describing
ASD symptomatology in these syndromes may be central to furthering our understanding of the
aetiological pathways of ASD (Persico & Bourgeron, 2006).
Skuse (2007) argues that the increasingly apparent association between particular genetic syndromes
and ASD symptomatology is likely to be accounted for by compromised cognitive capacity.
However, the degree to which intellectual disability (ID) accounts for the presence of ASD-like
symptomatology in these syndrome groups is varied (see Moss & Howlin, 2009) and has rarely been
explicitly evaluated in these populations. Further criticism of research conducted in this field
highlights that the focus of assessment for which the association between ASD and a particular
syndrome is evaluated is important. While the proportion of individuals meeting cut off criteria may
indicate a strong association, evaluation of specific ASD symptomatology, using domain and item
level analyses, has demonstrated unusual or atypical profiles of ASD characteristics in many of these
populations. Atypical profiles are evident in Fragile X, Cornelia de Lange, Angelman and Rett
syndromes (Hall, Lightbody, Hirt, Rezvani, & Reiss, 2010; Moss, Oliver, Wilkie, Berg, Kaur, &
Cornish, 2008; Mount, Hastings, Reilly, Cass, & Charman, 2003; Trillingsgaard & Østergaard,
2004). Moreover, there is likely to be a great deal of within syndrome heterogeneity with regard to
ASD symptomatology. While a syndrome group, as whole, may show elevated levels of these
characteristics, they are rarely evident, to the same degree, in all individuals with the syndrome. In
Down syndrome (for example), it has been reported that those who meet criteria for ASD form a
distinct sub-group of individuals who show different characteristics of ID and behavior relative to
individuals with Down syndrome who do not meet criteria (Capone et al., 2005; Carter et al., 2007;
Molloy et al., 2009; paper in press - anonymised for blind review). Thus, when evaluating ASD in
genetic syndromes it may be important to consider which characteristics distinguish those individuals
who score highly on measures of ASD from those who do not.
ASD in CdLS and FXS
These findings highlight two important methodological points. Firstly, that the association between
ASD symptomatology and a given syndrome should be conducted at a number of levels of
assessment in order to describe accurately the profile of association and secondly, that without
comparison to individuals with idiopathic ASD it is difficult to determine whether these profiles can
and should be considered to be ASD-like.
Two syndromes, both with previously reported heightened association with ASD, are the focus of the
current study; Cornelia de Lange syndrome (CdLS) and Fragile X syndrome (FXS). Previous studies
have demonstrated that both of these syndromes have a strong, but likely atypical association with
ASD. Reported prevalence rates of ASD are similar in both groups and some of the areas of
atypicality appear to be shared, making them interesting groups for comparison.
CdLS is caused by a deletion in the NIP-BL gene on chromosome 5 (locus 5p13) in 20 to 50% of
cases (Gillis et al., 2004; Krantz et al., 2004; Miyake et al., 2005; Tonkin, Wang, Lisgo, Bamshad, &
Strachan, 2004). Additional mutations on the SMC3 gene on chromosome 10 (Deardorff et al., 2007)
and X linked SMC1 gene (Musio et al., 2006) are reported to account for 5% of cases. Intellectual
disability ranges from mild (11%) to profound (50%; Oliver, Arron, Sloneem & Hall, 2008).
Prevalence estimates of ASD in CdLS are fairly consistent ranging from 50 – 67% (Basile, Villa,
Selicorni, & Molteni, 2007; Berney, Ireland, & Burn, 1999; Bhyuian et al., 2006; Moss et al., 2008;
manuscript in press – anonymised for blind review; Oliver, Arron, Sloneem, & Hall, 2008; Oliver,
Berg, Moss, Arron, & Burbidge, 2011). Studies that have employed matched contrast groups have
demonstrated that the occurrence of ASD symptomatology in CdLS is not entirely accounted for by
the degree of ID that is characteristic of the syndrome (Moss et al., 2008; Moss, Howlin, Magiati &
Oliver, 2012; Oliver et al., 2008).
Social impairment in CdLS is reported to be characterized by selective mutism, extreme shyness and
social anxiety (Collis, Oliver, & Moss, 2006; Goodban, 1993; Moss et al., 2008; Richards, Moss,
O’Farrell, Kaur, & Oliver, 2009). Repetitive behaviors are common in individuals with CdLS,
particularly lining up and tidying up behaviors (Moss et al., 2009). However, repetitive behavior is
not considered to be the most prominent aspect of the triad of impairments associated with the
syndrome (Oliver et al., 2011; Moss et al., 2012). Comparison of individuals with CdLS to those
with idiopathic ASD (matched for chronological age and nonverbal skills) confirm that while there
may appear to be broad similarities between these two groups, subtle differences in specific areas of
communication and social interaction are evident (Moss et al., 2012).
ASD in CdLS and FXS
FXS is the most common cause of inherited intellectual disability, occurring in 1 in 3,600 males and
1 in 8,000 females (Cornish et al., 2008). It results from an excess of CGG trinucleotide repeats on
the FMR1 (Fragile X Mental Retardation- 1) gene at location Xq27-3 (Dykens, Hodapp & Finucane,
2000). Intellectual disability is reported to be within the mild to severe range in males, while
females typically show a mild level of ID (Cornish, Turk & Hagerman, 2008). Reported prevalence
rates of ASD in FXS vary widely from 0 to 60%; although the most consistent estimates from studies
range from 21% to 50% (see Moss & Howlin, 2009 and Moss et al., 2011 for reviews). Studies
evaluating the profile of ASD phenomenology in FXS have described social anxiety, extreme
shyness and gaze avoidance, alongside seemingly preserved emotion sensitivity and willingness to
interact (Cornish, Turk, & Levitas, 2007; Hall, deBernardis, & Reiss, 2006; Lesniak-Karpiak,
Mazzocco, & Ross, 2003; Roberts, Weisenfeld, Hatton, Heath, & Kaufmann, 2007; Turk & Cornish,
1998). Similarly to individuals with CdLS, this presentation is somewhat different to the social
impairments that are characteristic of individuals with idiopathic autism. While some studies have
reported a strong association between ASD and ID in FXS (Demark et al., 2003; Dissanayake et al.,
2009; Kaufmann et al., 2004; Lewis et al., 2006; Loesch et al., 2007), ASD has also been identified
in individuals with pre-mutation FXS who have mild intellectual impairments or an IQ in the normal
range (Hagerman et al., 2005).
In the current study we evaluated the prevalence and profile of ASD symptomatology in individuals
with CdLS and FXS in comparison to a reference group of individuals with idiopathic ASD. We also
considered the role of adaptive behavior in the association between these syndrome groups and ASD
related characteristics. Using an ASD screening tool; the Social Communication Questionnaire
(SCQ; Rutter, Bailey & Lord, 2003) to conduct fine-grained analysis at cut off, total, domain and
item level scores and comparison to individuals with idiopathic ASD, we aimed to answer the
following research questions:
1. What is the prevalence of ASD symptomatology (according to the SCQ) in individuals with
CdLS and FXS?
2. What is the nature of ASD symptomatology within these syndrome groups relative to
individuals with idiopathic ASD?
3. When individuals with CdLS and FXS meet the cut off for ASD on the SCQ, do they so for
the same reasons as those with idiopathic ASD and how does the profile of ASD
symptomatology differ to that of the broader syndrome group?
ASD in CdLS and FXS
4. Using scores on a measure of self help skills, what role do these adaptive behaviors play in
the association between CdLS/FXS and the severity ASD symptomatology?
Recruitment of CdLS and FXS groups
This study was conducted as part of a larger project evaluating behavioral phenotypes in a number of
neurodevelopmental disorders (anonymised for blind review 2011; anonymised for blind review et
al., 2010; anonymised for blind review et al., 2009; anonymised for blind review et al., 2011).
Parents and carers of 376 individuals with CdLS who had either taken part in previous studies
conducted by the research team (anonymised for blind review, 2002; anonymised for blind review,
2009; anonymised for blind review, 2009) or were members of the CdLS Foundation (UK and
Ireland) were invited to take part in the study (for more detailed information regarding recruitment
methods see anonymised for blind review et al., 2008; anonymised for blind review et al., 2009;
anonymised for blind review et al., 2009). Parents and carers of 432 individuals with FXS were
invited to take part, via the Fragile X Society (male membership of over five years only).
A total of 125 parents and carers of individuals with CdLS (33.24% return rate) and 198 males with
FXS (45.83% return rate) responded to the invitation to take part. Information regarding the
diagnosis of genetic syndromes was obtained in order to establish the validity of diagnosis.
Participants were excluded from the study if they did not have a specific diagnosis from a General
Practitioner, Clinical Geneticist or Pediatrician, were under the age of four years or if a large
proportion (more than 25% of items on the Social Communication Questionnaire; Rutter et al., 2003)
of information was missing. Twenty two participants with CdLS and 21 participants with FXS were
excluded on the basis of these criteria.
ASD reference group:
Participants in the ASD reference group were selected from a larger database of 288 individuals who
had been recruited via the National Autistic Society and related branches as part of a previous study
(anonymised for blind review et al., in press). Participants were selected into the ASD reference
group if they had a confirmed diagnosis from a Pediatrician, G.P., Psychiatrist, Clinical Psychologist
or Educational Psychologist and scored above the cut off for ASD on the Social Communication
Questionnaire (SCQ; Rutter et al., 2003). Individuals under the age of nine years were removed from
ASD in CdLS and FXS
the sample in order to ensure that the ASD reference group was well matched to the FXS and CdLS
groups on chronological age.
In order to validate the nature of this group, SCQ data from the ASD reference group was compared
to that of the normative sample reported in the SCQ manual (Rutter, Bailey, Lord, & Berument,
2003). Odds ratio analysis at item level, using 99% confidence intervals confirmed no significant
differences between the ASD reference group and the normative SCQ sample on 31 out of 40 SCQ
items. The ASD reference group was more likely to score as ‘impaired’ on six SCQ items including
four algorithm items: ‘circumscribed interests’, ‘unusual sensory interests’, ‘social chat’ and ‘eye
gaze’ and two non-algorithm items: ‘self-injurious behavior’ and ‘unusual attachments’. The ASD
reference group was less likely to score as ‘impaired’ on ‘range of facial expression’. Overall, these
findings indicate that the ASD reference group employed in the current study is broadly similar to
the normative sample reported in the SCQ manual and may, in fact, show greater severity of
impairments in some areas, thus confirming the validity of this reference group.
+++Insert Table 1 here+++
A total of 103 participants (43 males; 60 females) with CdLS (mean age =17.19; SD=8.81; range =
4yrs to 40yrs), 177 participants (all males) with FXS (mean age =16.94; SD = 8.82; range = 6yrs to
47yrs) and 142 participants (120 males; 21 females) with ASD (mean age = 15.19; SD = 5.50; range
= 9yrs to 45 yrs) fulfilled all inclusion criteria and were included in the analysis. Table 1 describes
the participant characteristics of both groups and results of statistical comparisons. There was no
significant group difference on chronological age. The CdLS group was significantly less able than
both the FXS and ASD groups on self help skills, mobility and speech and had poorer vision and
hearing than these groups. Individuals with FXS were less mobile and had poorer vision than the
Demographic Questionnaire. The demographic questionnaire detailed age, gender, mobility, verbal
ability, diagnostic status.
ASD in CdLS and FXS
Wessex Scale (Kushlick, Blunden, & Cox., 1973). The Wessex Scale is an informant questionnaire
designed to assess social and physical abilities in children and adults with intellectual disabilities.
The scale comprises 15 items which rate the frequency of occurrence as frequent (more than once a
week), occasionally or never. Subscales include continence, mobility, self help skills, speech and
literacy and information on vision and hearing is also included. Items regarding the individual's
ability to feed, wash and dress himself comprise the self-help skills subscale. The Wessex Scale has
good inter-rater reliability at subscale level for both children and adults. Percentage agreement of
responses at subscale level is reported to range from 78-92% (Kushlick et al., 1973) and Kappa
scores at item level range from .37 to .89 (Palmer & Jenkins, 1982). The measure has been used to
characterize levels of ability in a range of syndrome groups (i.e. anonymised for blind review 2011;
anonymised for blind review et al., 2010; anonymised for blind review et al., 2009; anonymised for
blind review et al., 2011)
Social Communication Questionnaire (SCQ; Rutter, Bailey, Lord & Berument, 2003) The SCQ
(originally called the Autism Screening Questionnaire; Berument, Rutter, Lord, Pickles, & Bailey,
1999) was employed in the current study1. This 40 item screening tool was derived from the ADI-R
and was designed for use as a screening tool for epidemiological research and for describing ASD
symptomatology across various clinical populations. The authors suggest a cut-off point for ASD of
fifteen. This score was found to differentiate individuals with Pervasive Developmental Disorders
from other diagnoses (excluding those with ID) with a specificity of .80 and a sensitivity of .96 and
differentiated individuals with autism from individuals with ID with a specificity of .67 and a
sensitivity of .96. A higher cut-off point of 22 or more is required to differentiate individuals with
autism from other Pervasive Developmental Disorders with a sensitivity of .75 and a specificity of
.60. Internal consistency is good (α = .90 for the total scale; Berument et al., 1999). Item level
validity is reported by the authors to be good with 33 out of 39 items significantly differentiating
individuals with ASD from those without. Findings consistent with these were identified by Bölte,
Holtmann and Poustka (2008). A study of pre-school children suggested that this figure may be
lower (15 of 39; Eaves, Wingert, Ho, & Mickelson, 2006) but this discrepancy may be accounted for
by differences in age between the two samples and by the smaller sample of the latter study.
1 The Autism Screening Questionnaire (ASQ) was used in the CdLS and FXS groups. The published version (now
known as the Social Communication Questionnaire; SCQ) was used in the ASD reference group. One item (item 20;
social chat) differed for nonverbal individuals between the ASQ and the SCQ. For consistency across the groups, this
item was treated as missing data and was prorated for all nonverbal participants. The prorated score was calculated as the
mean item score, based on other completed items within the communication domain. Analysis indicates that the use of
this prorated item does not impact on the significance or direction of the findings in the communication domain. This
item was not included in the item level analysis.
ASD in CdLS and FXS
Scoring the SCQ:
The SCQ consists of forty items which are scored to indicate the presence (score=1) or absence
(score =0) of autistic like impairments. Items 2 to 7 are only scored for individuals who are reported
to be verbal. Items 2 to 40 are used in the SCQ scoring algorithm (with the exception of items 17, 18
& 38) to create three domains including: Communication, Repetitive Behavior and Restricted
Interests and Reciprocal Social Interaction. Nonverbal individuals are able to achieve a score on only
8 of the 13 communication domain items of the SCQ. This introduces a scoring bias when comparing
across groups of different ability levels. In the current study, comparison of domain scores across the
groups was conducted using a mean item domain score, in order to account for both the scoring bias
on the communication domain and the differences between SCQ domains regarding the number of
items that they comprise (total domain scores are provided for reference in Table 2).
The profile of scores on the SCQ at cut off, domain and total score levels was assessed across the
three participant groups. In order to evaluate the first research question regarding the profile of ASD
symptomatology in CdLS and FXS, differences between the groups on the proportion meeting ASD
and Autism criteria on the SCQ were evaluated using Chi squared analysis with pair-wise post hoc
comparisons. One Way ANOVA with Scheffé post hoc tests were used to examine group differences
in total SCQ scores and Mixed ANOVA were used for domain level analysis. Item level scores for
individuals with CdLS and FXS were compared to the ASD reference group using the same strategy
as that employed by Hall et al. (2010). Odds ratio analysis, using 99% confidence intervals, was used
in order to evaluate the likelihood of individuals with either CdLS or FXS scoring (score of 1) on
each SCQ item compared to the ASD reference group. An odds ratio of less than 1 indicates that the
syndrome group is less likely score on a given item relative to the ASD reference group, while an
odds ratio of more than 1 indicates that the syndrome group is more likely to score on a given item
relative to the normative ASD reference group. In order to consider the third research question
regarding the reason for reaching threshold on the SCQ and to reduce heterogeneity within the
groups, analyses were repeated with a subsample of each syndrome group, including only those
individuals who scored above the cut off for ASD on the SCQ (CdLS>SCQ cut off and FXS>SCQ
2 For conciseness, the terms ‘CdLS>SCQ cut off’ and ‘FXS>SCQ cut off’ are used here and throughout the paper to
describe the subgroups of participants with CdLS and FXS scoring above the cut off on the SCQ for ASD, it does not
imply that these individuals have a confirmed diagnosis of ASD.
ASD in CdLS and FXS
For item level analyses a score of 0 was assigned on items 2-7 of the SCQ for individuals who were
nonverbal in order to reflect the approach taken in the SCQ manual and in the Hall et al., 2010
analysis. Due to the differences between the groups regarding verbal ability and the high proportion
of individuals with CdLS who were non-verbal in the study, analysis of items 2 to 7 was also
repeated excluding individuals from all groups who were not able to achieve a score due to lack of
In order to evaluate the impact of adaptive behavior skills (as measured by the Wessex) on severity
of ASD symptomatology in CdLS, FXS and ASD, multiple regression models were conducted in
each group with SCQ total score as the outcome variable and Wessex self help and chronological age
as the predictor variables. These were repeated in the CdLS>SCQ cut off and FXS>SCQ cut off
groups in order to identify whether the impact of adaptive behavior on ASD symptomatology was
different in these more refined groups.
Prevalence of ASD symptomatology in CdLS and FXS:
78.6% of individuals with CdLS (N=81), 83.6% of individuals with FXS (N=148) and 100% of
individuals with ASD (N=142) scored above the cut off for ASD on the SCQ (χ2 (2) = .30.94; p
<.001; ASD>CdLS, FXS). These figures were 45.6% (N=47), 48.6% (N=86) and 79.6% (N=113)
respectively for those scoring above the Autism cut off (χ2 (2) = .40.11; p <.001; ASD>CdLS, FXS).
Profile of domain and total scores in the total sample groups and >SCQ cut off groups:
+++Insert table 2 here+++
Table 2 describes the domain and total SCQ scores for the total sample CdLS, FXS and ASD groups
and for the CdLS>SCQ cut off and FXS>SCQ cut off subgroups. Comparisons between the total
CdLS and FXS groups and the ASD reference group identified significant differences between the
three groups on the total score of the SCQ (F (2, 421) = 43.19; p <.001; CdLS, FXS<ASD). This
pattern of findings was also identified for comparisons between the CdLS>SCQ cut off and
FXS>SCQ cut off subgroups and the ASD reference group (F (2,370) = 23.44; p <.001; CdLS,
ASD in CdLS and FXS
++++Insert Figure 1 about here++++
A Mixed ANOVA identified a significant interaction between syndrome group and SCQ domain in
both the total sample (F3.37,706.72 = 7.95, p <.001) and the >SCQ cut off samples (F3.40, 625.46 = 12.14, p
<.001). Figure 1 describes these interactions and results of post hoc tests. No statistical difference
was identified between the CdLS>SCQ cut off sample and the ASD reference group on the
communication domain or the social interaction domain (both groups scored significantly higher than
the FXS>SCQ cut off sample), while the CdLS>SCQ cut off sample scored significantly lower than
both the ASD reference group and the FXS>SCQ cut off group on the repetitive behavior domain. In
the total sample, the CdLS group scored significantly lower than the ASD reference group in all
domains, significantly higher than the FXS group on the communication domain and showed no
statistical difference to the FXS group on the social interaction domain.
Profile of item level scores in the total sample groups and the >SCQ cut off groups:
+++Insert Table 3, Figure 2 and Figure 3 here+++
Table 3 shows that the proportion of individuals in the total CdLS and FXS groups, the CdLS>SCQ
cut off and FXS>SCQ cut off subgroups and the ASD reference group who score as impaired (item
score = 1) on each item of the SCQ. Figure 2 shows the odds ratio analysis comparing the total CdLS
and FXS groups to the ASD reference group. Figure 3 shows the odds ratio analysis comparing the
CdLS>SCQ cut off and FXS>SCQ cut off groups to the ASD reference group. For verbal items,
odds ratios are reported in two ways, firstly with nonverbal participants included with a score of 0 (to
reflect SCQ manual approach & Hall et al., 2010 analysis) and secondly with nonverbal participants
from all groups excluded from the analysis. Significant results are summarized below.
Communication Domain: The total CdLS group were less likely than the ASD reference group to
score on the following items ‘conversation’, ‘stereotyped utterances’, ‘inappropriate questions’,
‘pronoun reversal’ and ‘neologisms’. Four of these differences became non-significant when
nonverbal individuals were removed from the analysis (‘conversation’, ‘stereotyped utterances’,
‘pronoun reversal’ and ‘neologisms’). The same pattern of findings was identified in the CdLS>SCQ
cut off group.
ASD in CdLS and FXS
The total FXS group was less likely to score on the following items: ‘neologisms’, ‘imitation’,
‘nodding to mean yes’ and 'head shaking to mean no'. Removing nonverbal individuals from the
analysis did not impact on these findings. Findings were slightly different for the FXS>SCQ cut off
group, who were less likely to score on ‘imitation’ and ‘pointing’. When verbal ability was taken into
account by removing nonverbal individuals from the analysis, the FXS>SCQ cut off group were
more likely than the ASD group to score on ‘stereotyped utterances’.
Repetitive Behavior and Restricted Interests Domain:
The total CdLS group was less likely to score on four out of eight items including: ‘verbal rituals’,
‘unusual preoccupations’, ‘circumscribed interests’ and ‘sensory interests’. The difference on ‘verbal
rituals’ remained significant when nonverbal individuals were removed from the analysis. A slightly
different pattern of findings was identified for the CdLS>SCQ cut off group who were less likely to
score on ‘verbal rituals’, ‘circumscribed interests’ and ‘sensory interests’ while the difference on
‘verbal rituals’ was no longer significant when verbal ability was accounted for.
The total FXS group was less likely to score on the following items: ‘unusual preoccupations’,
'repetitive use of objects’, ‘unusual sensory interests’, and ‘complex body mannerisms’. The
differences on ‘repetitive use of objects’ and ‘unusual sensory interests’ were also identified in the
FXS>SCQ cut off group.
Reciprocal Social Interaction Domain:
On the social interaction domain, the total CdLS group were less likely to score on four out of 15
items including: ‘eye gaze’, ‘social smiling’, ‘interest in other children’, ‘response to others’ and
more likely to score on ‘range of facial expression’. In contrast, while the CdLS>SCQ cut off group
were also less likely to score on ‘eye gaze’, they were more likely to score on the following items:
‘offering comfort’ and ‘range of facial expression’.
The total FXS group were less likely to score on ‘inappropriate facial expression’, ‘use of other’s
body to communicate’, ‘social smiling’, ‘offering to share’, ‘seeking to share enjoyment’, ‘offering
comfort’, ‘quality of social overtures’, and ‘interest in children’. The FXS>SCQ cut off group were
less likely to score on ‘social smiling’, ‘seeking to share enjoyment’ and ‘quality of social overtures’
and more likely to score on ‘range of facial expression’.
ASD in CdLS and FXS
Association between adaptive behavior skills and ASD symptomatology:
In the total CdLS group, both chronological age (β = .32; p<.001) and self help skills (β = -.61;
p<.001) significantly predicted SCQ total scores. Findings in the CdLS>SCQ cut off group were
consistent with this (CA: β = .31; p=.003; self help: β = -.49; p<.001). The findings indicate that
lower levels of self help skills (when controlling for chronological age) and higher chronological age
(when controlling for self help skills) are significantly associated with higher scores on the SCQ. The
r 2 values for the overall models were 39.4% (total CdLS) and 25.4% (CdLS>SCQ cut off) with self
help skills accounting for 35.6% and 21.9% of the variance respectively.
In the total FXS and FXS>SCQ cut off group, self help skills significantly predicted SCQ total score
(β = -.22; p=.005; β =-.19; p = .03). Chronological age did not predict SCQ total scores in either the
total FXS or FXS>SCQ cut off group. The proportion of variance accounted for by the overall model
was minimal in both groups (r2 = 4.6% and 3.2% respectively) with self help skills accounting for
4.5% and 3.2% of the variance in each case.
In the ASD reference group, there was a marginally significant effect of self help skills on SCQ total
score (β = -.18; p=.04), although the proportion of variance this accounted for was very small (3.1%).
Chronological age did not significantly predict the total SCQ score.
Both CdLS and FXS are reported within the literature to show a strong association with ASD. Both
groups have been reported to show a similar prevalence rate of ASD but have been demonstrated to
show a somewhat unusual presentation of ASD like characteristics. In the current study we aimed to
evaluate the profile of ASD symptomatology in individuals with CdLS and FXS compared to those
with idiopathic ASD who were well matched according to chronological age, using the SCQ.
Specifically, we evaluated the profile of ASD symptomatology on a number of different
measurement levels (cut off, total, domain and item levels) and in comparison to individuals with
idiopathic ASD in order to achieve a detailed picture of ASD symptomatology in the two groups and
to identify areas of overlap and difference relative to individuals with ASD who do not have a
genetic syndrome. We considered these findings both at the total sample level and in relation to the
subgroups of individuals scoring above the cut off for ASD on the SCQ (CdLS>SCQ cut off and
FXS>SCQ cut off). This enabled us to consider whether there are similarities and differences in the
presentation of ASD symptomatology when the groups are comparable with regard to the way in
ASD in CdLS and FXS
which they are behaviorally defined (i.e. score above ASD cut off on the SCQ). Finally, we
evaluated the association between adaptive behavior skills (as measured by the Wessex),
chronological age and ASD symptomatology in both syndrome groups.
The proportion of individuals with CdLS and FXS meeting the cut off for ASD was 78.6% and
83.6% respectively. These are both somewhat higher than previous reports and this may suggest that
the less conservative ASD cut off on the SCQ lacks specificity in these syndrome groups. Forty-five
percent of individuals with CdLS and 48.6% of individuals with FXS were reported to meet the
autism cut off. These figures are more consistent with previous findings (Basile, Villa, Selicorni, &
Molteni, 2007; Berney, Ireland, & Burn, 1999; Bhyuian et al., 2006; Moss & Howlin, 2009; Moss et
al., 2008; Moss et al., 2012; Oliver, Arron, Sloneem, & Hall, 2008; Oliver, Berg, Moss, Arron, &
Burbidge, 2011) and confirm suggestions that the rate of association between CdLS and ASD is
similar to that observed in FXS.
Consistent with these findings, the CdLS and FXS groups (total samples and >SCQ cut off samples)
showed a similar degree of impairment, according to the total score of the SCQ, and both groups
scored significantly lower than the ASD group on this score. The findings suggest that at the
broadest level of evaluation, individuals with CdLS and FXS may show a milder presentation of
ASD symptomatology than individuals with idiopathic ASD, even when they meet the ASD cut off
on the SCQ. This is consistent with previous findings in FXS (Bailey, Hatton, Skinner, & Mesibov,
2001; Demark, Feldman, & Holden, 2003; Moss & Howlin, 2009) but has not previously been
reported in CdLS.
A milder presentation of ASD symptomatology in CdLS and FXS was also evident at the domain
level analysis of the total sample groups in which both groups scored significantly lower than the
ASD reference group on all domains. A significant interaction between group and domain indicated
that the relative contribution of scores from each domain of the triad of impairments to the overall
presentation of ASD symptomatology differs between the FXS and CdLS groups. The CdLS group
showed greater impairments in the communication domain than the FXS group but significantly
lower levels of repetitive behavior than the FXS group. Both groups were equally impaired in the
domain of reciprocal social interaction.
Differences in the profile and severity of ASD symptomatology were more prominent in the >SCQ
cut off samples and a distinct pattern of findings was identified. Consistent with the total sample