PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Am J Intellect Dev Disabil. Author manuscript; available in PMC 2011 January 1.
Published in final edited form as:
PMCID: PMC2887668
NIHMSID: NIHMS173223

Autism Spectrum Disorder in Children and Adolescents with Fragile X Syndrome: Within-Syndrome Differences and Age-Related Changes

Abstract

The Autism Diagnostic Interview-Revised was used to examine diagnostic profiles and age-related changes in autism symptoms for a group of verbal children and adolescents with FXS, with and without autism. After controlling for nonverbal IQ, statistically significant between-group differences for lifetime and current autism symptoms were found for the Communication and Restricted Interests/Repetitive Behaviors domains, but not the Reciprocal Social Interaction domain. Effect sizes for differences in Reciprocal Social Interaction also were smaller than effect sizes for the other domains with one exception. Overall, severity of autism symptoms improved with age for all participants, with the least improvement noted for Restricted Interests and Repetitive Behaviors. FMRP did not account for unique variance in autism symptoms over and above nonverbal IQ.

Keywords: Fragile X syndrome, autism, diagnosis, intellectual disability, behavioral phenotype

Autism Spectrum Disorder in Children and Adolescents with Fragile X Syndrome: Within-Syndrome Differences and Age-Related Changes

Fragile X syndrome (FXS) is the most common inherited cause of intellectual disability. In general, prevalence rates have been reported as 1 in 4000 males and 1 in 8000 females (Crawford et al., 2001); however, recent studies have yielded rates closer to 1 in 2500 (Fernandez-Carvajal., Walichiewicz, Xiaosen, Pan, Hagerman, & Tassone, 2009; Hagerman, 2008). FXS is caused by an expansion of a repetitive CGG nucleotide sequence in the FMR1 gene, located on the X chromosome (Kaufmann & Reiss, 1999). This expansion causes methylation and transcriptional silencing of the gene, resulting in a reduction or absence of the protein that is normally produced (Kaufmann, Cortell, Kau, et al., 2004). This protein (FMRP) is critical for the regulation of processes involved in synaptic maturation, axonal guidance, and experience-dependent learning (Darnell, Warren, & Darnell, 2004; Hagerman, Ono, & Hagerman, 2005). Behaviors characteristic of autism are frequent in FXS, with many individuals meeting diagnostic criteria for autism (Hagerman, 1999). Most estimates place the prevalence of autism in FXS at 25%, although recent studies utilizing diagnostic criteria for the full spectrum of autism disorders (i.e., inclusive of PDD-NOS) have reported rates approaching 50% (Demark, Feldman, & Holden, 2003; Kaufmann et al, 2004; Philofsky, Hepburn, Hayes, Hagerman, & Rogers, 2004). In the present study, we were interested in within-syndrome differences in, and the age-related trajectories of, autism symptoms in FXS.

Although the association between FXS and autism has been well documented, the developmental course of autism within FXS is not well understood. It is not known whether the symptoms of autism are stable across development in FXS or whether there is an abatement of symptoms with age as has been observed for idiopathic autism (Seltzer, Krauss, Shattuck, Orsmond, Swe, & Lord, 2003; Shattuck et al., 2007). The behavioral characteristics that distinguish between individuals with FXS with and without a diagnosis of autism are not fully understood. We also lack data on the role of FMRP in autism symptomatology in FXS. The present study was designed to address these gaps.

Autism Spectrum Disorder

Autism is a spectrum of behavioral characteristics observable before three years of age and characterized by impairments in reciprocal social interactions, language and communication, and repetitive and stereotyped behaviors and restricted interests (American Psychiatric Association, 1994). In contrast to FXS, which is diagnosed through DNA testing, a diagnosis of autism (AUT) is based on behavioral testing and clinical judgment relative to criteria specified in the Diagnostic and Statistical Manual-fourth edition (DSM-IV; American Psychiatric Association, 1994) and the tenth edition of the International Classification of Diseases and Related Health Problems (ICD-10; World Health Organization, 1992).

Although the characteristics of autism vary widely across individuals, poor modulation of eye contact in social interaction is the most widely reported behavioral feature of the disorder (Lord & Spence, 2006). In fact, the avoidance of eye gaze also is characteristic of individuals with FXS and may contribute to the perceived similarity between the two disorders. Other commonalities between FXS and autism include insistence on sameness, hand stereotypies, self-injurious behavior, and inappropriate and repetitive use of objects (Levitas, Hagerman, Braden, Rimland, McBogg, & Matus, 1983).

Diagnostic instruments

The recent development of standardized instruments for the diagnosis of autism spectrum disorders has allowed increased precision and consistency in characterizing the differences between individuals with FXS only and individuals with comorbid FXS and autism (hereafter, FXS+AUT). The Autism Diagnostic Interview-Revised (ADI-R; LeCouteur, Lord, & Rutter, 2003) is widely regarded as one of the gold standards for the diagnosis of autism. The ADI-R closely follows the DSM-IV (American Psychiatric Association, 1994) and ICD-10 (World Health Organization, 1992) diagnostic criteria for autism and elicits information, through parent interview, relative to the three domains that define the disorder.

For most ADI-R items, responses are scored for a) the current degree of impairment, and b) the greatest impairment noted between ages 4 to 5 or ever in the individual's lifetime. Diagnostic classification is based on computation of an algorithm, composed of a subset of lifetime items (i.e., in reference to ages 4 to 5 or ever). When administered to parents of older children and adolescents, the ADI-R offers the potential for examining symptom change retrospectively by comparing behaviors queried in the diagnostic items to queries about current behaviors. In fact, use of the ADI-R in this manner (Seltzer et al., 2003; Shattuck et al., 2007) has yielded results similar to those of prospective longitudinal studies of individuals with idiopathic autism (Howlin, Goode, Hutton, & Rutter, 2004). For the present study, we used the ADI-R to examine age-related symptom change for older children and adolescents with FXS by comparing lifetime scores to scores reported for current behaviors.

Age-related changes in symptoms of autism

Although, by definition, autism is a lifelong disorder, and social impairments persist in virtually all individuals diagnosed early in life, there is evidence of significant improvement in symptoms with age in idiopathic autism (Boelte & Poustka, 2000; Gilchrist, Green, Cox, Burton, Rutter, & LeCouteur, 2001; Piven, Harper, Palmer, & Arndt, 1996; Seltzer, et al., 2003). Seltzer et al. (2003) examined diagnostic stability and patterns of symptom change over time by comparing current symptoms of autism with lifetime ratings as measured by the ADI-R in 405 individuals diagnosed with an autism spectrum disorder. Participants were between the ages of 10 and 53 years (M = 22 years) and did not have FXS. Significantly fewer participants met diagnostic cutoffs based upon current behavior than based upon lifetime ratings of past behavior (55% vs. 97%) and a general pattern of symptom abatement in all three ADI-R domains was observed. In a follow-up study, Seltzer and colleagues (Shattuck et al., 2007) conducted a prospective analysis of the current ADI-R profiles of 241 individuals at two time points across four years. The results were consistent with the retrospective analyses conducted by Seltzer et al. (2003), suggesting that retrospective analysis of ADI-R scores can be used to obtain a reliable index of age-related symptom change.

Cognitive impairment and symptoms of autism

Cognitive ability is a robust correlate of symptom severity and a predictor of later social functioning and independent living status for adults with idiopathic autism (Howlin, Goode, Hutton, & Rutter, 2004; Howlin, Mawhood, & Rutter, 2000). Individuals with autism who have IQs in the range of intellectual disability display slower growth in many cognitive domains (Sigman & McGowan, 2005) and continue to manifest difficulty as adults in areas such as friendship, work, and independent living (Howlin et al., 2004). Such findings suggest the need to examine the association between autism symptoms and cognitive ability in FXS because a high proportion of this population has intellectual disabilities and because their rate of cognitive development appears to decrease in later childhood and adolescence, as reflected in age-related declines in IQ (Kover, Abbeduto, Kim, & Brown, 2008; Hall, Burns, Lightbody, & Reiss, 2008).

Autism in Fragile X Syndrome

Profile of symptoms

Results across several studies suggest that children with comorbid FXS and autism differ in their symptom profiles from children with FXS only, with the former being more similar to individuals with idiopathic autism than to individuals with FXS only (Bailey, Mesibov, Hatton, Clark, Roberts, & Mayhew, 1998; Demark et. al., 2003; Lewis et al., 2006; Rogers, Wehner, & Hagerman, 2001). For example, Rogers et al. (2001) examined profiles of autistic behavior in children with FXS, ages 21 to 48 months of age. Participants received a diagnosis of autism if they met criteria on two of three diagnostic measures: the ADI-R, the ADOS, and a clinical diagnosis. Rogers et al. found no significant differences in ADI-R and ADOS total scores or individual domain scores between the autism only group and the group with comorbid FXS and autism. Both groups with autism, however, differed from the FXS only group in two ADI-R domains and all three ADOS domains.

Most studies examining autism symptoms in FXS have dichotomized their samples into those with and without autism, without making distinctions within the broader autism spectrum. In an attempt to make a more fine-grained distinction, Kaufmann et al. (2006) used the ADI-R to characterize 56 mostly nonverbal young boys with FXS (ages 3 to 8 years), as FXS+AUT, FXS+ASD, or FXS only. Kaufmann et al., found that the FXS+ASD group and the FXS only groups differed significantly on Reciprocal Social Interaction domain scores, whereas the FXS+ASD group and the FXS+AUT group differed significantly on Communication domain scores. This profile suggests that the diagnosis of an autism spectrum disorder in FXS results largely from differences in social reciprocity and is not merely an artifact of communication challenges in FXS. Kaufmann and colleagues have confirmed these results in a follow-up longitudinal study (Hernandez, Feinberg, Vaurio, Passanante, Thompson, & Kaufmann, 2009).

Although the Kaufmann et al. (2004) and Hernandez et al. (2009) studies add important information to the question of how characteristics of autism are distributed within the population of individuals with FXS, the young ages of their participants, their primarily nonverbal status, and the inclusion of only males limits the generalizability of these results. The present study provides an opportunity to extend previous findings by examining ADI-R profiles for a group of older children and adolescents with FXS, including both males and females who met the criterion for having verbal language status.

Age-related changes in autism symptoms in FXS

We do not yet have a clear picture of symptom change with age in individuals with FXS. Some researchers suggest that, similar to the trajectory observed for those with idiopathic autism, the symptoms of autism in FXS abate with age (Hagerman, Jackson, Levitas, Rimland, & Braden, 1986; Baumgardner, Reiss, Freund, Abrams, 1995; Reiss & Freund, 1992; Rogers et al., 2001). Most studies, however, have used cross-sectional comparisons and thus, are subject to the confounding effect of cohort differences (Hatton et al., 2006). Moreover, there is some inconsistency even in the cross-sectional findings.

In a longitudinal study covering the age span of 1 to 14 years, Hatton et al. (2006) found small but significant age-related increases in symptom severity, as measured by the Childhood Autism Rating Scale (CARS; Schopler, Reichler, & Reynolds, 1988), for a sample of 116 children with FXS. These investigators also found that of those 39 participants who exceeded the CARS cutoff for a diagnosis of autism at one or more measurement periods, only13% improved with age such that they no longer exceeded the CARS cutoff. It is important to note that although the Hatton et al., (2006) cohort extended into early adolescence, most participants were quite young, with the mean age under 5 years at the initial visit.

In fact, there are few data regarding the age-related course of autism symptoms in FXS during adolescence and into adulthood. In the only longitudinal study involving older individuals with FXS, Sabaratnam, Murthy, Wijeratne, Payne, and Buckingham (2003), suggested that the symptoms of autism remain stable during adulthood. The mean age of participants in Sabaratnam et al. study, however, was over 46 years. In addition, Sabaratnam et al. employed a brief caretaker interview not specifically designed to assess autism. It is possible that an examination of autism symptoms during an earlier developmental period and using a more appropriate and better validated measure, such as the ADI-R, would reveal a more nuanced picture of longitudinal symptom change.

Cognitive impairment and symptoms of autism in FXS

On average, cognitive abilities in individuals with comorbid FXS and autism are lower relative to individuals with FXS only during childhood and this profile continues into adolescence and early adulthood (Bailey, Hatton, Mesibov, Ament, & Skinner, 2000; Bailey, Hatton, Skinner, & Mesibov, 2001; Demark, Feldman, & Holden, 2003; Hernandez et al., 2009; Kau et al., 2004; Kaufmann et al., 2004; Kover et al., 2008; Lewis et al., 2006; Philofsky et al., 2004; Rogers et al., 2001). In turn, cognitive ability has been found to be positively correlated with FMRP levels (Dyer Friedman et al., 2002; Kover et al., 2008; Loesch, Huggins, & Hagerman, 2004; Reiss, Freund, Baumgardner, Abrams, & Denckla, 1995; Tassone et al., 1999), although some studies have failed to confirm this relationship (Skinner et al., 2005). Hall and colleagues (2008) recently demonstrated the relationship between IQ and FMRP maintains even after controlling for age and gender, suggesting a unique contribution of FMRP to cognitive ability.

The association between FMRP levels and symptoms of autism is unclear. Although Hatton et al. (2006) found a positive association between FMRP and autism symptoms as measured by the CARS, they did not include IQ as a covariate in their analyses, leaving unresolved the question of whether FMRP accounts for unique variance in autism symptoms over and above the contribution of IQ. Other studies have failed to detect a positive association between FMRP and autism severity (Bailey et al., 2000; Harris et al., 2008; Hessl et al., 2001). Loesch et al. (2007), however, found that both FMRP and FSIQ were significant predictors of ADOS Communication domain scores for full mutation females and that FMRP was a significant predictor for full mutation males.

Given the high degree of shared variance between FMRP and IQ, Hatton et al. (2006) suggested that null findings in measuring the relationship between FMRP and autism symptom severity may depend, in part, upon whether IQ is included as a covariate. In addition, the choice of instrument used to measure the presence of autism symptoms may be critical for detecting such an association. Finally, it may be that the association between FMRP and autism symptoms cannot be detected within samples of only male participants due to the truncated range of FMRP in males. In the present study, we investigated whether FMRP accounts for unique variance in predicting current symptoms of autism, over and above the contribution of nonverbal IQ, when using the ADI-R to measure symptoms.

Research Questions

In summary, this study addressed the following questions: (1) Which lifetime symptoms of autism distinguish participants with FXS only from those with FXS who meet diagnostic criteria for autism? (2) Which current symptoms of autism distinguish participants with FXS only from those who meet diagnostic criteria for autism? (3) Within each diagnostic group, which symptoms of autism show significant improvement over time? (4) Which symptoms of autism best predict group membership? (5) Do FMRP levels account for unique variance in predicting current symptoms of autism, after controlling for nonverbal cognitive ability?

Method

Participants

Participants were 51 children and adolescents with FXS, who were part of a larger, longitudinal examination of language development in FXS and Down Syndrome. All participants were between the ages of 10 and 16 years at the time of data collection for the current study. They were recruited nationally through mailings to professionals, attendance at national and regional parent meetings, postings to internet listservs and websites, advertisements on nationally syndicated radio shows and in newspapers in selected urban areas, and though a university registry of families with children with developmental disabilities.

The current sample included 35 males and 15 females, all of whom were native English speakers. Six families had two children participate; however, there were no sibling pairs in the FXS+AUT group. According to parent report, all participants used spoken language as their primary means of communication and produced three-word phrases on a daily basis. Participants were identified as having autism based on the diagnostic algorithm of the ADI-R (Rutter et al., 2003) as described in detail below. The participants with FXS only (n = 26; 13 females) and FXS+AUT (n = 24; 2 females) did not differ significantly in age (see Table 1). As expected, however, individuals with FXS+AUT had significantly lower nonverbal mental ages and IQ's, and lower receptive and expressive language age-equivalent scores on standardized measures.

Table 1
Participant Characteristics (Means and Standard Deviations)

All participants had a diagnosis of FXS confirmed through molecular genetic testing prior to entry into the study. We also confirmed the diagnosis through use of Southern Analysis and polymerase chain reaction (PCR) testing (Brown et al., 1993; Nolin et al., 2003) conducted on peripheral blood samples for all but 9 participants (2 declined to be retested and blood samples were not available for the other 7 participants for logistical reasons). All participants had the FMR1 full mutation, although 11 males were mosaic (in methylation status or repeat size). Using a sample of 200 cells for males and 400 cells for females, and employing the method of Willemson et al. (1997) and Tassone et al. (1999), the average proportion of cells that expressed the protein FMRP in participants with FXS only was .04 (SD = .09, Range .00 −.30) for males and .48 (SD = .05, Range .34 − .51) for females. For those with FXS+AUT, average FMRP expression was .04 (SD = .07, Range .00 − .20) for males and .50 (SD = .00, Range .50 −.50) for females.

Assessments and Measures

Autism Status

Autism status for the current study was based on the diagnostic algorithm of the ADI-R. ADI-R items elicit information relevant to early development, as well as the domains of Reciprocal Social Interaction, Communication, and Restricted Interests and Stereotyped Behaviors. Individual ADI-R items are scored according to the examiner's judgment of the presence/absence or extent of a given behavior using a scale of 0 (behavior of the type specified not present), 1 (behavior present but not sufficiently severe or frequent to meet criteria for a score of 2), 2 (definite abnormality of the type specified) or 3 (definite abnormality of the type specified and marked in severity). Items are scored for the presence/extent of the behavior a) during the three months immediately preceding the interview (Current), or b) between the ages of 4 to 5 years or ever in the targeted individual's lifetime (Lifetime). Some items provide specific age periods for coding (e.g., Friendships from 10 to 15 years of age).

Summary scores for current and lifetime ratings are calculated for each domain and provide a quantitative measure of the presence of autism symptoms. Items inquiring about developmental history are used to reflect the examiner's judgment as to age of onset. In the standard version of the ADI-R, a diagnostic algorithm, based upon domain cutoff scores for lifetime ratings, is calculated using 37 items that have been shown to have the best discriminative validity between individuals with and without autism (Lecavalier et al., 2006). In order to prevent a few items from contributing excessively to the calculation of the algorithm, scores of 3 are converted to 2's. Given the eligibility criteria for the current study, all participants were considered to be verbal for purposes of ADI-R scoring. Domain cutoffs used in the current study to determine a classification of autism/no autism were those provided in the manual for participants with verbal language status.

The biological mothers of participants in the current study were interviewed using an abbreviated version of the ADI-R comprised of 42 items (Seltzer, Krauss, Shattuck, Orsmond, Swe, & Lord, 2003; Shattuck et al., 2007) by one of two research-reliable examiners. The protocol included three items designed to gather background and age of onset information and 37 items used to compute domain scores for Reciprocal Social Interaction (16 items), Communication (13 items) and Restricted Interests and Repetitive Behaviors (8 items). These 37 items were queried for lifetime ratings and constituted the diagnostic algorithm.

In this abbreviated version of the ADI-R, current behavior also was queried for 29 items (11, 10, and 8 of the items in the Reciprocal Social Interaction, Communication, Restricted Interests and Repetitive Behaviors domains, respectively). Although an algorithm to determine diagnostic status based on current symptoms is not available, items for which both lifetime and current ratings were available were used to examine, retrospectively, age-related changes in autism symptoms and to examine between-group differences in current symptoms of autism. Mean scores for ADI-R domains are presented in Table 2 and scores for individual lifetime and current items within each domain are presented in Tables 3, ,4,4, and and55.

Table 2
Lifetime and Current ADI-R Domain Scores: Means (and Standard Deviations)
Table 3
ADI-R Reciprocal Social Interaction Domain Lifetime and Current Scores: Adjusted Means (and Standard Errors)
Table 4
ADI-R Communication Domain Lifetime and Current Scores: Adjusted Means (and Standard Errors)
Table 5
ADI-R Restricted Interests and Repetitive Behaviors Domain Lifetime and Current Scores: Adjusted Means (and Standard Errors)

Internal consistency

Because the ADI-R was not developed for the FXS population, we examined internal consistency of the ADI-R diagnostic algorithm items for the current sample. Chronbach's α values were .87, .81, and .67 for Reciprocal Social Interaction, Communication, and Restricted Interests and Repetitive Behaviors, respectively. The latter domain may have had slightly lower internal consistency because it contains the fewest items. The coefficients for this sample are similar to those obtained for the same domains by Lecavalier et al. (2006) for individuals with idiopathic autism (i.e., .84, .76, and .54, respectively).

Inter-rater Agreement

Inter-rater agreement was computed by having a second research-reliable examiner recode 25% (13) of the ADI-R protocols. We used Cohen's kappa, which corrects for chance agreement by considering base rates of the coded variables, and evaluated the resultant kappas using the benchmarks provided by Landis and Koch (1977): 0.81-1.00 (very good), 0.61-0.80 (substantial), 0.41-0.60 (moderate), and 0.21-0.40 (fair). The kappa for the diagnosis of autism was .58, which was the maximum value possible given the distribution of the rater's marginal scores (Cohen, 1969, p. 42). Kappas for lifetime ratings were 1.00, .70, and .68 for the Reciprocal Social Interaction, Communication, and Restricted Interests/Repetitive Behaviors domains, respectively. Kappa was at the maximum possible value for the latter two domains. Kappa for the ADI-R section assessing whether symptoms of autism were evident before 36 months of age was .75, also the maximum possible value. Average percentage agreement exceeded 85% for the diagnosis of autism and for each lifetime domain score. Kappa was at the maximum possible value for 33 of the 40 individual lifetime items and for 17 of the 29 items queried for current behavior. The remaining kappa values ranged between .52 and .83; thus, none of the kappa values for the individual items fell below the moderate level of agreement. Average percentage agreement was 88% for individual lifetime items (range 69 to 100%, and 89% for individual current items (range 69 to 100%).

Nonverbal Cognition

The Brief IQ Screener (i.e., the Figure Ground, Form Completion, Sequential Order, and Repeated Patterns subtests) of the Leiter-R Visualization and Reasoning Battery (Roid & Miller, 1997) was administered. For each participant, age-equivalent scores across the four subtests were averaged to provide a mean age-equivalent and scaled scores were summed to calculate a deviation IQ.

Procedures

All participants were administered a battery of measures focusing on their language and cognition. The ADI-R was administered to each participant's biological mother. The ADI-R was completed at the first annual visit for 47 participants and at the second annual visit for 4 participants. Nonverbal cognitive and language scores were obtained concurrently with the time period in which the ADI-R was administered. Participants were classified as FXS+AUT if they met the designated lifetime cutoff scores for all ADI-R domains, including age of onset. Lifetime and current ADI-R domain scores obtained for each domain are presented in Table 2. Means are reported separately for female and males; however, all analyses collapsed across gender.

Results

Diagnostic Symptoms of Autism

Between-group differences for each ADI-R domain were examined using separate multivariate analyses of covariance. Group (FXS only, FXS+AUT) was a fixed factor and diagnostic algorithm (lifetime) item scores for the individual ADI-R domain under consideration were the dependent variables. Nonverbal IQ was a covariate. Significant multivariate tests were followed by simple effects testing, using Holm's sequentially rejective procedure (Holm, 1979). One-tailed significance levels were used as participants with FXS+AUT were expected to attain higher scores on all ADI-R items than participants with FXS only.

Reciprocal Social Interaction

Between-group differences in diagnostic algorithm scores from the Reciprocal Social Interaction domain of the ADI-R failed to reach significance when controlling for differences in nonverbal IQ, F(16,33) = 1.49, p = .16, partial η2 = .42. The covariate-adjusted mean scores for algorithm items in this domain are presented in Table 3.

Communication

There was a significant effect of Group on diagnostic algorithm scores for the Communication domain, F(13,36) = 3.63, p < .001, partial η2 = .57. This represents a large effect size and indicates that 57% of the variance in algorithm scores for this domain was accounted for by group membership. The univariate tests revealed significant differences for six symptoms of autism: Pointing to Express Interest (C7), Stereotyped Utterances/Delayed Echolalia (C1), Nodding (C8), Head Shaking (C9), Spontaneous Imitation of Actions (C11), and Imitative Social Play (C13). The covariate-adjusted means for all Communication algorithm items are presented in Table 4.

Restricted Interests and Repetitive Behaviors

There was a significant effect of Group on diagnostic algorithm scores for the Restricted Interests and Repetitive Behaviors domain, F(8,41) = 3.19, p = .007, partial η2 = .38. Again, this represents a large effect size. The univariate tests revealed significant differences for three symptoms of autism: Repetitive Object Use/Interest in Parts of Objects (R4), Compulsions and Rituals (R5), and Circumscribed Interests (R3). The covariate-adjusted means for all algorithm items within this domain are presented in Table 5.

Current Symptoms of Autism

The same analytic strategy described above was followed to make between-group comparisons for current behavior items within each ADI-R domain.

Reciprocal Social Interaction

Between-group differences in current autism symptoms in the Reciprocal Social Interaction domain failed to reach significance, F(11,38) = 1.57, p = .15, partial η2 = .31,. See Table 3 for presentation of the covariate-adjusted mean scores for current items in the Reciprocal Social Interaction domain.

Communication

There was a significant effect of Group, F(10,39) = 3.57, p = .002, partial η2 = .48, for current items in the Communication domain reflecting a large effect size. Univariate tests revealed significant between-group differences for two items: Stereotyped Utterances/Delayed Echolalia (C1) and Reciprocal Conversation (C3). See Table 4 for presentation of the covariate-adjusted mean scores for current items in the Communication domain.

Restricted Interests and Repetitive Behaviors

There was a significant effect of Group, F(8,41) = 3.13, p = .007, partial η2 = .38, for current items in the Restricted Interests and Repetitive Behaviors domain reflecting a large effect size. Univariate tests revealed a significant between-group difference for three items: Circumscribed Interests (R3), Compulsions and Rituals (R5), and Unusual Preoccupations (R6). See Table 5 for the covariate-adjusted mean scores for current items in the Restricted Interests and Repetitive Behaviors domain.

Age-Related Change in Autism Symptoms

Twenty-nine items are included in both the lifetime and the current item pool of the ADI-R. For these individual items, difference scores were computed by subtracting current scores from lifetime scores; thus, a difference score greater than zero indicated age-related improvement for the behavior under consideration while a difference score less than zero indicated worsening of the behavior. One sample t-tests were used to detect significant differences from zero. Two-tailed p-values were used and a Bonferroni correction was employed to adjust for multiple significance tests within each domain. Prior to conducting these analyses, we examined the bivariate correlations between chronological age and lifetime scores, current scores, and difference scores. None of these correlations were significant, indicating that chronological age of the participants did not contribute to the profile of age-related changes in autism symptoms.

FXS only Group

Of the 11 items in the Reciprocal Social Interaction domain queried for both the lifetime and current behavior, scores for two items improved significantly with child age: Offering to Share (S6), t(25) = 3.73, p < .001, d = .73, and Quality of Social Overtures (S9), t(25) = 3.43, p = .002, d = .67. These effect sizes were in the medium range. Examination of the means for the other 9 items in this domain indicated that scores for these items were all at or below 1 for both the lifetime and current ratings.

Of the 10 items in the Communication domain queried for both the lifetime and current ratings, two showed significant age-related improvement: Social Verbalization/Chat (C2), t(25) = 4.50, p < .001, d = .88, and Reciprocal Conversation (C3), t(25) = 5.28, p < .001, d = 1.03. These effects sizes are large. Examination of the means for the other items in this domain indicated that scores were all at or below 1 for both lifetime and current ratings, with the following four items at or close to zero for both ratings: Neologisms/Idiosyncratic Language (C6), Nodding (C8), Head Shaking (C9), and Conventional/Instrumental Gestures (C10).

None of the 8 items in the Restricted Interests and Repetitive Behaviors domain queried for both lifetime and current ratings showed age-related improvement; however, means for these items were all at or below 1 for both ratings.

FXS+AUT Group

Of the 11 items in the Reciprocal Social Interaction domain queried for both for both lifetime and current ratings, seven showed significant age-related improvement, all with medium to large effect sizes: Social Smiling (S4), t(25) = 3.36, p = .003, d = .67, Showing and Directing Attention (S5), t(25) = 3.66, p < .001, d = .73, Offering to Share (S6), t(25) = 5.28, p < .001, d = 1.05, Seeking to Share Enjoyment (S7), t(25) = 3.53, p = .002, d = .70, Offering Comfort (S8), t(25)= 3.38, p = .002, d = .68, Quality of social overtures (S9), t(25) = 3.84, p < .001, d = .77, and Appropriateness of Social Responses (S12), t(25) = 4.30, p < .001, d = .86.

Of the ten items in the Communication domain queried for both for both lifetime and current ratings, scores for six items showed significant age-related improvement, with medium or large effect sizes: Social Verbalization/Chat (C2), t(24) = 4.94, p<.001, d = .99, Reciprocal Conversation (C3), t(24) = 4.24, p < .001, d = .85, Pronominal Reversal (C5), t(24) = 3.76, p < .001, d = .75, Pointing (C7), t(24) = 4.93, p < .001, d = .98, Nodding (C8), t(24) = 3.12, p < .005, d = .62 and Conventional/Instrumental Gestures (C10), t(24) = 3.16, p = .004, d = .63. Examination of the remaining four items in this domain revealed means that were at or below 1 for both lifetime and current ratings, with the exception of Stereotyped Utterances/Delayed Echolalia. This latter item had a mean lifetime score of 1.79 (.15) and a mean current score of 1.44 (.14). Although the ability to engage in reciprocal conversations did improve significantly with age, the mean current score for this item remained elevated at 1.30 (.15).

For the eight items in the Restricted Interests and Repetitive Behaviors domain queried for both lifetime and current ratings, two showed significant age-related improvement with medium effect sizes: Repetitive Object Use/Interest in Parts of Objects (R4), t(24) = 3.46, p = .002, d = .69, and Hand and Finger Mannerisms (R7), t(24) = 3.17, p = .004, d = .63.

Prediction of Group Membership

Using a series of discriminant analyses, we sought to determine the combination of diagnostic (lifetime) items within each ADI-R domain that best predicted group membership. Nonverbal IQ and the diagnostic algorithm items from one domain were included in each respective analysis using a stepwise method. For each domain, the value of the squared canonical correlation reflects a measure of effect size and indicates the percent of variation in group assignment accounted for by the significant independent variables in each discriminant analysis. A significant Wilk's lambda allowed us to (a) reject the null hypothesis that the FXS only and FXS+AUT groups had the same mean discriminant function score and (b) conclude that the model for that ADI-R domain was discriminating. Finally, the standardized discriminant function coefficients indicate the relative importance of the independent variables in predicting the dependent variable, similar to the standardized betas in a regression analysis.

Reciprocal Social Interaction

The groups were significantly discriminated, canonical R2 = .45, Λ = .55, χ2(4) = 28.59, p < .001, by three items: Group play with Peers (S15), Nonverbal IQ, and Social Smiling (S4). The pooled within-group correlations between the three discriminating variables and the standardized discriminant function were .72, −.68, and .52, respectively, with algorithm items positively related and nonverbal IQ inversely related to the single discriminant function. Functions at the group centroids (i.e., the mean variate scores for each group) were −.87 for the FXS only group and .91 for the FXS+AUT group. The combination of Group Play with Peers, Nonverbal IQ, and Social Smiling correctly classified 80% of participants with FXS only and 88% of participants with FXS+AUT.

Communication

The groups were significantly discriminated, canonical R2 = .65, Λ = .35, χ2(4) = 49.83, p < .001, by four items: Stereotyped Utterances/Delayed Echolalia (C1), Pointing to Express Interest (C7), Nodding (C8), and Nonverbal IQ. The pooled within-group correlations between the four discriminating variables and the standardized discriminant function were .58, .54, .53, and −.45, respectively, with algorithm items positively related and nonverbal IQ inversely related to the discriminant function. The functions at the group centroids were -1.32 for the FXS only group and 1.37 for the FXS+AUT group. The combination of these four variables correctly classified 96% of participants with FXS only and 92% of participants with FXS+AUT.

Restricted Interests and Repetitive Behaviors

The groups were significantly discriminated, canonical R2 = .41, Λ = .59, χ2(4) = 25.19, p < .001, by three items: Repetitive Object Use (R4), Circumscribed Interests (R3), and Verbal Rituals (R1). It should be noted that noverbal IQ did not contribute significantly to group separation for this domain. The pooled within-group correlations between the three discriminating variables and the standardized discriminant function were .68, .58, and .47, respectively, with these algorithm items positively related to the discriminant function. The functions at the group centroids were −.80 for the FXS only group and .84 for the FXS+AUT group. The combination of these three algorithm items correctly classified 80% of participants in the FXS only group and 72% of participants in the FXS+AUT group.

FMRP and Current Symptoms of Autism

A large and positive association was detected between FMRP and Nonverbal IQ, r(42) = .65, p<.01, two-tailed. The correlations between FMRP and the three current ADI-R domain scores were moderate and negative, with values ranging between −.31 and −.39 (all p's <.05), as were the correlations between nonverbal IQ and current domain scores, which ranged between −.42 and −.49 (all p's <.01). The question of whether FMRP accounted for unique variance in predicting current symptoms of autism, over and above nonverbal IQ, was examined using a separate stepwise multiple regression equation for each domain score. In each regression, the percentage of cells expressing FMRP was entered at the first step, followed by nonverbal IQ at the second step. When entered as the sole predictor, the contribution of FMRP was significant for Reciprocal Social Interaction, Communication, and Restricted Interests and Repetitive Behaviors, F(1,40) = 7.07, p < .01, F(1,40) = 4.32, p = .04, and F(1,40) = 4.48, p = .04, two-tailed, respectively. Each overall regression equation remained significant when nonverbal IQ was entered into the model, F(2,39) = 5.70, p < .01, F(2,39) = 5.95, p < .01, and F(2,39) = 5.00, p < .01, two-tailed, respectively. However, for all three analyses, the proportion of cells expressing FMRP failed to account for unique variance in current ADI-R scores over and above the contribution of nonverbal IQ. Results of these analyses are presented in Tables 6, ,7,7, and and8.8. It should be noted that there were no significant differences in FMRP level between males with FXS only and those FXS+AUT or between females with FXS only and those with FXS+AUT.

Table 6
Hierarchical Regression Predicting ADI-R Reciprocal Social Interaction Domain Current Total Score
Table 7
Hierarchical Regression Predicting ADI-R Communication Domain Current Total Score
Table 8
Hierarchical Regression Predicting ADI-R Restricted Interests and Repetitive Behaviors Domain Current Total Score

Discussion

We used the ADI-R to examine the behavioral symptoms that distinguish between older children and adolescents with FXS with and without autism. We also took advantage of the fact that the ADI-R queries the informant about the target individual at two points in the life course (i.e., through lifetime and current ratings) to examine, retrospectively, age-related changes in the symptoms of autism. Finally, we examined the contribution of FMRP to the manifestation of autism symptoms in FXS.

Deficits in social reciprocity are considered to reflect the essential and defining feature of idiopathic autism (Volkmar & Klin, 2005). After controlling for nonverbal IQ, however, we did not detect statistically significant differences between the groups in either the lifetime or current symptoms of autism that are queried in the ADI-R Reciprocal Social Interaction domain. Moreover, while effect sizes for the between group differences in this domain were large, and perhaps clinically important, they were smaller than the effect sizes obtained for lifetime and current ratings in the Communication domain as well as current ratings in the Restricted Interests and Repetitive Behaviors domain.

Based upon our findings, it appears that impairment in the Reciprocal Social Interaction domain is not the primary feature distinguishing individuals with FXS with and without a co-morbid autism diagnoses. The current findings differ, however, from those reported by Kaufmann and colleagues (Kaufmann et al., 2004; Hernandez et al., 2009), who found that the Reciprocal Social Interaction domain represented the most significant determinant of autism behavior in a group of young, largely nonverbal, males with FXS. Importantly, Kaufmann and colleagues did not covary IQ scores in the various regression models that they used to explore profiles of autism symptoms in FXS. Thus, the problems individuals with FXS display in the Reciprocal Social Interaction domain may reflect cognitive impairments that influence their ability to experience and share enjoyment and interest with a social partner rather than a social indifference or a lack of motivation to engage with others

In contrast, we did find significant between-group differences for six diagnostic items within the Communication domain after controlling for nonverbal IQ. Three of these items involve the production of communicative gestures (Pointing to Express Interest, Nodding, Head Shaking), two involve imitation (Imitation of Actions and Imitative Social Play), and one involves the use of stereotyped utterances or delayed echolalia. Of the six items that differed between the groups for lifetime ratings, however, only the use of stereotyped utterances and delayed echolalia distinguished between the groups in current ratings, after controlling for nonverbal IQ. One additional item, the ability to engage in reciprocal conversations, which did not differ for the lifetime rating, differed significantly for the current rating. It should be noted that the two lifetime algorithm items focusing on imitation are not queried in the current rating.

The use of stereotyped utterances and delayed echolalia is not unexpected for individuals with FXS, even without autism, because the FXS phenotype is associated with pragmatic difficulties, including repetitive language, tangential talk, and topic perseveration (Murphy & Abbeduto, 2007; Roberts et al., 2007; Sudhalter & Belser, 2001). That the ability to engage in reciprocal conversations emerged as significantly different between the groups in the current, but not the lifetime, rating is particularly interesting given the lack of group differences in the Reciprocal Social Interaction domain. One possibility is that, for participants with FXS+AUT, difficulties in reciprocal conversation reflect limitations in lexical and syntactic knowledge or in the specific pragmatic skills needed to engage in such conversations rather than a lack of motivation to communicate with others. Neither of the Communication items for which we observed significant between-group differences in the current ratings was included in the analyses of Hernandez et al. (2009) as these items are not queried for nonverbal participants.

Three lifetime symptoms in the Restricted Interests and Repetitive Behaviors domain differed significantly between the groups. These symptoms reflect repetitive object use, compulsions and rituals, and circumscribed interests. Compulsions and rituals as well as circumscribed interests also showed between-group differences in the current ratings. Finally, although the group difference in Unusual Preoccupations failed to reach significance for the lifetime rating, it did emerge as a significantly different for the current rating.

In a recent study examining the presence and types of repetitive behaviors in genetic syndromes, Moss, Oliver, Arron, Burbidge, and Berg (2009) found that a large sample of participants with FXS (ranging in age from 4 – 47 years) demonstrated the highest frequency and greatest number of types of repetitive behaviors relative to any other syndrome group examined. The FXS sample was especially elevated relative to the other genetic conditions in items reflecting hand stereotypies, lining up objects, preference for routines, echolalia, and restricted conversations. In the Moss et al. study, lining up objects and preference for routines were categorized together in the subdomain of compulsions, a symptom of autism which was significantly different for our participants for both lifetime and current ratings. Participants with FXS+AUT in the current study also demonstrated relatively higher levels of stereotyped utterances and delayed echolalia. Although these two characteristics are queried in the Communication domain of the ADI-R, they correspond in topography to restricted conversation and echolalia observed for verbal participants in the Moss et al. study. Importantly, Moss et al. (2009) reported that none of the repetitive behaviors identified as having high specificity for individuals with FXS was correlated with total scores on the Autism Screening Questionnaire (ASQ; Berument et al., 1999), suggesting that some language characteristics and repetitive behaviors observed in FXS may not be associated with diagnosis of autism.

In part because relative increases in cognitive impairment over time have been reported for FXS (Hall et al., 2008; Skinner et al., 2005), an additional focus of the current study was whether participants would show lessening or worsening of autism symptoms over time within each ADI-R domain. Hatton et al. (2006) found that symptom severity did increase slightly over time as indexed by CARS scores in children who were younger on average than participants in the current study. In another sample of young boys with FXS, Hernandez et al. (2009) found a general trend of worsening symptoms for young boys with FXS only and improvement of symptoms for young boys with comorbid FXS and ASD. In the current study, all participants in both groups showed improvement in autism symptoms from lifetime to current ratings.

Within the Reciprocal Social Interaction domain, Offering to Share and Quality of Social Overtures improved with age for participants with FXS only as reflected by differences between the lifetime and current ratings. In the same domain, seven symptoms of autism showed age-related improvement for participants with FXS+AUT. Use of Other's Body to Communicate and the two items related to facial expressions did not show improvement from lifetime to current ratings for participants with FXS+AUT, but these three items had adjusted mean scores that were less than 1 for the lifetime rating. Importantly, only two items, Appropriateness of Social Responses and Friendships, had adjusted current mean scores of greater than 1 for participants with FXS+AUT.

Within the Communication domain, we found that skills, such as the use of gestures, did improve with age for participants with comorbid FXS and autism. However, the use of stereotyped speech and delayed echolalia did not improve for either group, with scores for this symptom of autism remaining stable between lifetime and current ratings and differing significantly between the groups for the current rating. Although participants with FXS only improved significantly in the ability to engage in reciprocal conversations between the lifetime and current rating, this pattern of improvement was not evident for participants with FXS+AUT. Thus, despite age-related improvements in the use of gestural means of communication, the ability to engage reciprocally in conversational turn-taking remained especially challenging for participants with FXS+AUT.

In contrast to the other two symptom domains, little improvement between lifetime and current ratings was noted for the Restricted Interests and Repetitive Behaviors domain, although none of the items evidenced a worsening of symptoms. The only improvements were observed for the group with comorbid FXS and autism. Repetitive use of objects as well as hand and finger mannerisms improved significantly for participants with FXS+AUT across the two sets of ratings. In fact, repetitive object use was no longer significantly different between the groups in the current rating, with the groups not differing in hand and finger mannerisms for either set of ratings.

It is interesting to compare patterns of change in autism symptoms detected in the current study with those reported by Shattuck et al. (2007) for a large group of older children, adolescents and young adults with idiopathic autism. Overall, Shattuck et al. observed longitudinal improvements in social reciprocity and verbal communication, as well as in restricted interests and repetitive behaviors. As was the case for the current study, none of the individual ADI-R items queried by Shattuck et al. significantly worsened over time. Unlike participants with FXS+AUT in the current study, however, participants with idiopathic autism did not improve in nonverbal communication behaviors over the 4.5-year study period.

When examining age-related symptom change in the current study, it is important to note that the nonverbal IQ-adjusted mean item scores for the FXS only group hovered around 1 for all three ADI-R domains, indicating impairments that were, on average, not marked enough to be scored as clearly characteristic of autism. There were, however, two exceptions to this pattern. First, both FXS only and FXS+AUT participants had scores exceeding 1 for Friendships and this item did not differ significantly or improve from lifetime to current ratings for either group. In fact, for the group with FXS only, this item received the highest (i.e., most impaired) rating in the Reciprocal Social Interaction domain for both ratings. Second, participants with FXS only had a lifetime mean score of 1.4 (SD = .15) for the diagnostic item tapping the ability to engage in reciprocal conversation. Although the two participant groups in the current study did not differ on this Communication domain item for the lifetime rating, the group with FXS only showed age-related improvement such that this item distinguished the groups in the current rating.

Interestingly, Shattuck et al. (2007) reported that these same two items, impairments in friendships and impairments in reciprocal conversation, were the most prevalent autism symptoms observed for individuals with idiopathic autism over the age of 10 years. Our results suggest that impairments in reciprocal conversation are present to a significant degree in individuals with FXS+AUT, with between-group differences increasing with age, whereas impairments in friendship are symptomatic of all individuals with FXS regardless of autism status. It seems likely that delays in language learning and impairments in the ability to engage in back-and-forth conversational turn taking could have a cumulative and negative impact on the ability to establish and maintain friendships with peers, regardless of an individual's motivation to interact with others. Indeed, it is also possible that hyperarousal and anxiety experienced by males with FXS in social contexts contributes to deficits in peer relationships and friendships throughout development (Sudhalter & Belser, 2004). Of course, individuals with FXS also may experience a lack of opportunity to establish friendships in addition to any possible lack of motivation to engage in reciprocal social interaction. In addition, the ADI-R scoring for friendships includes the ability to interact reciprocally around non-stereotyped activities. Thus, the presence of circumscribed interests could also interfere with the ability to establish friendships. Regardless of the source, the establishment of reciprocal peer relationships should be targeted for treatment for all children and adolescents with FXS.

In addition to using multivariate analysis of variance to identify algorithm items that differed significantly between FXS only and FXS+AUT, we used a discriminant function analysis to identify the items within each ADI-R domain that maximally discriminated between the groups at the diagnostic time point (Bray & Maxwell, 1985). For the Reciprocal Social Interaction domain, item scores for Group Play with Peers, Social Smiling, as well as nonverbal IQ, resulted in the highest number of participants that were classified into the correct diagnostic group. For the Communication domain, algorithm scores for Stereotyped Utterances, Pointing to Express Interest, Nodding as well as nonverbal IQ best discriminated between the groups. Finally, for the Restricted Interests and Repetitive Behaviors domain, Repetitive Object Use, Circumscribed Interests and Verbal Rituals best discriminated between the groups. Interestingly, nonverbal IQ did not contribute to group discrimination for the Restricted Interests/Repetitive Behaviors domain. Taken together, results of the discriminant analyses support the proposal that behavioral symptoms of autism in the areas of reciprocal social interaction and communication may be secondary to cognitive impairments in individuals with FXS. However, cognitive impairments seem to influence the expression of repetitive and restricted behaviors equally in all individuals with FXS. As other authors suggest, it is possible that a psychological mechanism such as arousal might better describe the relationship between FXS and autism symptoms in the domain of restricted interests and repetitive behaviors (Roberts, Boccia, Bailey, Hatton, & Skinner, 2001; Miller et al., 1999).

Finally, we found that FMRP was not related to ADI-R domain scores after controlling for nonverbal IQ. In contrast, Hatton et al. (2006) found that FMRP was a significantly and negatively predictive of scores on the Childhood Autism Rating Scales (CARS: Schopler, Reischler, & Renner, 1988) for children with FXS, whereas Loesch et al. (2007) found a similar relationship using the Autism Diagnostic Observation Schedule (ADOS; Lord, Rutter, DiLavore, & Risi, 1999) for individuals with FXS ranging from childhood to adulthood. Hatton et al. (2006), however, did not control for nonverbal IQ. In the Loesch et al. (2007) study, although both FMRP and FSIQ were significant predictors of ADOS Communication algorithm score, only FSIQ accounted for unique variance in the ADOS algorithm score for Reciprocal Social Interaction or the aggregated variable representing the sum of the Communication and Reciprocal Social Interaction algorithm scores. Taken together, the data suggest that FMRP levels have a largely indirect effect on symptoms of autism and that this effect is likely mediated through IQ.

One issue that could not be addressed in the current study concerns changes in diagnostic classification over time. When used as intended by the developers, the ADI-R does not provide a diagnostic algorithm that can be used with current scores (i.e., an ADI-R diagnosis of autism is computed with an algorithm using behaviors reported for lifetime rating), preventing us from examining the stability of the autism diagnosis with age, as Hatton et al. (2006) did by examining CARS cutoff scores. By scoring the ADI-R according to the conventions provided in the manual, we were limited to examining changes in those individual ADI-R items that were queried at both lifetime and current time points.

The present study also was limited by the use of only one measure of autism symptoms. Hernandez et al. (2009) assigned diagnostic status to young boys with FXS by using DSM-IV criteria, the ADI-R, and the ADOS. According to these authors, the ADOS resulted in over-identification of participants as having autism, whereas complete agreement was obtained between the ADI-R and DSM-IV diagnoses. Harris and colleagues (2008) utilized the ADOS, ADI-R and DSM-IV to determine autism status for a group of males with FXS and concluded that it was the ADI-R that over-identified autism in FXS. As researchers do not know the “true” prevalence of autism in FXS, it is difficult to know how well any of these instruments is performing when used with individuals with FXS. This dilemma reinforces the importance of understanding whether autism symptoms in FXS actually reflect the operation of the same underlying pathology as in idiopathic autism.

Results of the current study also are limited by the absence of a comparison group of nonverbal IQ-matched individuals with idiopathic autism. We plan to add such a comparison group in future studies. Results also are limited by reliance on retrospective rather than longitudinal measures of change, although there is evidence from idiopathic cases of autism that the two approaches yield much the same results. Nevertheless, replication with a longitudinal design is necessary. An additional limitation involves our sample of participants, all of whom had achieved phrase speech, resulting in use of the verbal algorithm for the Communication domain of the ADI-R. Results of the current study might have differed if the participant sample had included individuals considered nonverbal according to the ADI-R scoring conventions. It is possible that nonverbal individuals with FXS differ systematically from verbal individuals with FXS not only on items included in the Reciprocal Social Interaction domain but also in ways that affect the acquisition of spoken language. That is, children with FXS who more readily demonstrate social reciprocity may also achieve more proficient spoken language status. This is an interesting issue given that most factor analytic examinations of the ADI-R have yielded a two factor structure consisting of a social-communication factor and a stereotyped language and repetitive behavior factor (Frazier, Youngstrom, Kubu, Sinclair, & Rezai, 2008; Snow, Lecavalier, & Houts, 2009). Finally, few studies on autism in FXS have included females. Of the 15 females in the current sample of older children and adolescents with FXS, eight had nonverbal IQs less than or equal to 70. Of those eight females, two met criteria for FXS+AUT. This proportion is in general agreement with Hatton et al. (2006), in which 2 of 32 females met the CARS autism cut-off. Future studies of autism symptoms in larger samples of females with FXS will be particularly helpful given the wide range of nonverbal cognitive abilities and variability in FMRP expression in females with FXS.

In summary, findings of the current study suggest that differences in social reciprocity, the defining feature of idiopathic autism, are not observed in individuals with FXS relative to autism status when cognitive impairment is taken into account. This finding was observed for both lifetime (i.e., diagnostic) and current ratings obtained for the Autism Diagnostic Interview-Revised (Rutter et al., 2003). Differences in communication as well as restricted interests and repetitive behaviors were observed, however, even after controlling for nonverbal IQ. Symptoms of autism improved over time for individuals with FXS regardless of autism status. Controlling for intercorrelations between the ADI-R items within each domain, we found that nonverbal IQ added to group separation for the Reciprocal Social Interaction and Communication domains, but not for the Restricted Interests and Repetitive Behaviors domain. Finally, although significant and negative associations with ADI-R domain scores were detected for both FMRP and nonverbal IQ, FMRP did not account for unique variance in predicting domain scores once nonverbal IQ was added to the regression model.

Acknowledgements

This research was supported by NIH grants R01 HD024356 and P30 HD03352. We would like to thank the families who participated in this study for their time, patience and enthusiastic support.

References

  • American Psychiatric Association . Diagnostic and statistical manual of mental disorders. 4th ed. American Psychiatric Association; Washington, DC: 1994.
  • Bailey DB, Mesibov GB, Hatton DD, Clark RD, Roberts JE, Mahew L. Autistic behavior in young boys with fragile X syndrome. Journal of Autism and Developmental Disorders. 1998;28:499–507. [PubMed]
  • Bailey DB, Hatton DD, Mesibov GB, Ament N, Skinner M. Early development, temperament, and functional impairment in autism and fragile X syndrome. Journal of Autism and Developmental Disorders. 2000;30:557–567. [PubMed]
  • Bailey DB, Hatton DD, Skinner M, Mesibov G. Autism behavior, FMR1 protein, and developmental trajectories in young males with fragile X syndrome. Journal of Autism and Developmental Disorders. 2001;31:165–174. [PubMed]
  • Bailey D, Roberts J, Hooper S, Hatton D, Mirrett P, Roberts J, et al. Research on fragile X syndrome and autism: Implications for the study of genes, environments, and developmental language disorders. In: Rice M, Warren S, editors. Developmental language disorders: From phenotypes to etiologies. Erlbaum; Mahwah, NJ: 2004. pp. 121–150.
  • Baumgardner T, Reiss A, Freund L, Abrams B. Specification of the neurobehavioral phenotype in males with fragile X syndrome. Pediatrics. 1995;45:744–752. [PubMed]
  • Berument S, Rutter M, Lord C, Pickles A, Bailey A. Autism screening questionnaire: Diagnostic validity. The British Journal of Psychiatry. 1999;175:444–451. [PubMed]
  • Boelte S, Poustka F. Diagnosis of autism: The connection between current and historical information. Autism. 2000;4(4):382–390.
  • Bray J, Maxwell S. Multivariate analysis of variance. Sage Publications; London: 1985.
  • Brown WT, Houck GE, Jr., Jeziorowska A, Levinson F, Ding X-H, Dobkin C, Zhong N, Henderson J, Sklower-Brooks S, Jenkins EC. Rapid fragile X carrier screening and prenatal diagnosis using a nonradioactive PCR test. Journal American Medical Association. 1993;270:1569–1575. [PubMed]
  • Cohen I. Behavioral profiles of autistic and nonautistic fragile X males. Developmental Brain Dysfunction. 1995;8:252–269.
  • Cohen J. A coefficient of agreement for nominal scales. Educational and Psychological Measurement. 1969;20:37–46.
  • Crawford D, Acuna J, Sherman S. FMR1 and the fragile X syndrome: human genome epidemiology review. Genetics in Medicine. 2001;3:359–371. [PubMed]
  • Darnell J, Warren S, Darnell R. The fragile X mental retardation protein, FMRP, recognizes G-quartets. Mental Retardation and Developmental Disabilities Research Reviews. 2004;10:49–52. [PubMed]
  • Demark J, Feldman M, Holden J. Behavioral Relationship Between Autism and Fragile X Syndrome. American Journal on Mental Retardation. 2003 September;108(5):314–326. [PubMed]
  • Dyer-Friedman J, Glaser B, Hessl D, Huffman L, Wisbeck J, Reiss A, et al. Genetic and environmental influences on the cognitive outcomes of children with fragile X syndrome. Journal of the American Academy of Child & Adolescent Psychiatry. 2002 March;41(3):237–244. [PubMed]
  • Fernandez-Carvajal I, Walichiewicz P, Xiaosen X, Pan R, Hagerman P, Tassone F. Screening for expanded alleles of the FMR1 gene in blood spots from newborn males in a Spanish population. Journal of Molecular Diagnostics. 2009;11:324–328. [PubMed]
  • Frazier T, Youngstrom E, Kubu C, Sinclair L, Rezai A. Exploratory and confirmatory factor analysis of the Autism Diagnostic Interview-Revised. Journal of Autism and Developmental Disorders. 2008;38:474–480. [PubMed]
  • Gilchrist A, Green J, Cox A, Burton D, Rutter M, Le Couteur A. Development and current functioning in adolescents with Asperger syndrome: A comparative study. Journal of Child Psychology and Psychiatry. 2001 February;42(2):227–240. [PubMed]
  • Gotham K, Pickles A, Lord C. Standardizing ADOS scores for a measure of severity in autism. Journal of Autism and Developmental Disorders. 2009;39:693–705. [PMC free article] [PubMed]
  • Hagerman P. The fragile X prevalence paradox. Journal of Medical Genetics. 2008;45:498–499. [PMC free article] [PubMed]
  • Hagerman R. Clinical and molecular aspects of fragile X syndrome. In: Tager-Flusberg H, editor. Neurodevelopmental disorders. The MIT Press; Cambridge, MA: 1999. pp. 27–42.
  • Hagerman R, Jackson A, Levitas A, Rimland B, Braden M. An analysis of autism in fifty males with the fragile X syndrome. American Journal of Medical Genetics. 1986;64:356–361. [PubMed]
  • Hagerman R, Ono M, Hagerman P. Recent advances in fragile X: a model for autism and neurodegeneration. Current Opinion in Psychiatry. 2005;18:490–496. [PubMed]
  • Hall S, Burns D, Lightbody A, Reiss A. Longitudinal changes in intellectual development in children with fragile X syndrome. Journal of Abnormal Child Psychology. 2008 August;36(6):927–939. [PubMed]
  • Harris S, Hessl D, Goodlin-Jones B, Ferranti J, Bacalman S, Barbato I, et al. Autism profiles of males with fragile X syndrome. American Journal on Mental Retardation. 2008;113(6):427–438. [PMC free article] [PubMed]
  • Hatton D, Sideris J, Skinner M, Mankowski J, Bailey DB, Roberts J, Mirrett P. Autism behavior in children with fragile X syndrome: prevalence, stability, and the impact of FMRP. American Journal of Medical Genetics. 2006;140:1804–1813. [PubMed]
  • Hernandez R, Feinberg R, Vaurio R, Passanante N, Thompson R, Kaufmann W. Autism spectrum disorder in Fragile X syndrome: A longitudinal evaluation. American Journal of Medical Genetics. 2009 published online 13 May 2009. [PMC free article] [PubMed]
  • Hessl D, Dyer-Friedman J, Glaser B, Wisbeck J, Barajas G, Taylor A, Reiss A. The influence of environmental and genetic factors on behavior problems and autistic symptoms in boys and girls with fragile X syndrome. Pediatrics. 2001;108:88–97. [PubMed]
  • Holm S. A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics. 1979;6:65–70.
  • Holmes N, Shah A, Wing L. Disability assessment schedule: a brief screening device for use with the mentally retarded. Psychological Medicine. 1982;12:879–890. [PubMed]
  • Howlin P, Goode S, Hutton J, Rutter M. Adult outcome for children with autism. Journal of Child Psychology and Psychiatry. 2004;45(2):212–229. [PubMed]
  • Howlin P, Mawhood L, Rutter M. Autism and developmental receptive language disorder—A follow-up comparison in early adult life. II: Social, behavioural, and psychiatric outcomes. Journal of Child Psychology and Psychiatry. 2000 July;41(5):561–578. [PubMed]
  • Kau A, Tierney E, Bukelis I, Stump M, Kates W, Trescher W, et al. Social behavior profile in young males with fragile X syndrome: Characteristics and specificity. American Journal of Medical Genetics. 2004;126A:9–17. [PubMed]
  • Kaufmann W, Cortell R, Kau A, Bukelis I, Tierney E, Gray R, Cox C, Capone G, Stanard P. Autism spectrum disorder in fragile X syndrome: communication, social interaction and specific behaviors. American Journal of Medical Genetics. 2004;129A:225–234. [PubMed]
  • Kaufmann W, Reiss A. Molecular and cellular genetics of fragile X syndrome. American Journal of Medical Genetics. 1999;88:11–24. [PubMed]
  • Kover S, Abbeduto L, Kim J-S, Brown T. Change and stability in nonverbal cognitive abilities in girls and boys with fragile X syndrome.. Poster presented at the 11th biennial International Fragile X Conference of the National Fragile X Foundation; St. Louis, MO. Jul, 2008.
  • Krug D, Arick J, Almond P. Autism Screening Instrument for Educational Planning: An Assessment and Educational Planning System for Autism and Developmental Disabilities. 2nd Edition Pro-Ed; Austin, TX: 1993.
  • Landis J, Koch G. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159–174. [PubMed]
  • Lecavalier L, Scahill L, McDougle C, McCracken J, Vitiello B, Tierney E, et al. Validity of the Autism Diagnostic Interview-Revised. American Journal on Mental Retardation. 2006;3:199–215. [PubMed]
  • Levitas A, Hagerman R, Braden M, Rimland B, McBogg P, Matus I. Autism and the fragile X syndrome. Journal of Developmental Behavioral Pediatrics. 1983;4:151–158. [PubMed]
  • Lewis P, Abbeduto L, Murphy M, Giles N, Schroeder S, Bruno L, et al. Cognitive, language and social-cognitive skills of individuals with Fragile X Syndrome with and without autism. Journal of Intellectual Disability Research. 2006;50(7):532–545. [PubMed]
  • Loesch DZ, Bui QM, Dissanayake C, Clifford S, Gould E, Bulhak-Paterson D, et al. Molecular and cognitive predictors of the continuum of autism behaviours in fragile X. Neuroscience & Biobehavioral Reviews. 2007;31(3):315–326. [PMC free article] [PubMed]
  • Loesch D, Huggins R, Hagerman R. Phenotypic Variation and FMRP Levels in Fragile X. Mental Retardation and Developmental Disabilities Research Reviews. 2004;10(1):31–41. [PubMed]
  • Lord C, Rutter M, DiLavore P, Risi S. The Autism Diagnostic Observation Schedule. Western Psychological Services; Los Angeles, CA: 1999.
  • Lord C, Spence S. Understanding autism: From basic neuroscience to treatment. CRC Press; Boca Raton, FL: 2006. Autism Spectrum Disorders: Phenotype and Diagnosis. pp. 1–23.
  • Miller L, McIntosh D, McGrath J, Shyu V, Lampe M, Taylor A, Tassone F, Neitzel K, Stackhouse T, Hagerman R. Electrodermal responses to sensory stimuli in individuals with fragile X syndrome: A preliminary project. American Journal of Medical Genetics. 1999;83:268–279. [PubMed]
  • Moss J, Oliver C, Arron K, Burbidge Cl., Berg K. The prevalence and phenomenology of repetitive behavior in genetic syndromes. Journal of Autism and Developmental Disorders. 2009;39:572–588. [PubMed]
  • Murphy MM, Abbeduto L. Gender differences in repetitive language in fragile X syndrome. Journal of Intellectual Disability Research. 2007;51(5):387–400. [PubMed]
  • Nolin S, Dobkin C, Brown WT. Molecular Analysis of Fragile X Syndrome. In: Dracopoli NC, Haines JL, Korf BR, editors. Current Protocols in Human Genetics. Wiley; New York: 2003. pp. 9.5.1–9.5.12.
  • Philofsky A, Hepburn S, Hayes A, Rogers S, Hagerman R. Linguistic and cognitive functioning and autism symptoms in young children with Fragile X syndrome. American Journal on Mental Retardation. 2004 May;109(3):208–218. [PubMed]
  • Piven J, Harper J, Palmer P, Arndt S. Course of behavioral change in autism: A retrospective study of high-IQ adolescents and adults. Journal of the American Academy of Child & Adolescent Psychiatry. 1996 April;35(4):523–529. [PubMed]
  • Reiss A, Freund L. Behavioral phenotype of fragile X syndrome: DSM-III-R autistic behavior in male children. American Journal of Medical Genetics. 1992;43:35–46. [PubMed]
  • Reiss A, Freund L, Baumgardner, Abrams M, Denckla M. Contribution of the FMR1 gene mutation to human intellectual dysfunction. Nature Genetics. 1995;11:331–334. [PubMed]
  • Roberts J, Boccia M, Bailey D, Hatton D, Skinner M. Cardiovascular indicators of arousal in boys affected by fragile X syndrome. Developmental Psychology. 2001;39:107–123. [PubMed]
  • Roberts J, Martin G, Moskowitz L, Harris A, Foreman J, Nelson L. Discourse skills of boys with fragile X syndrome in comparison to boys with down syndrome. Journal of Speech, Language, and Hearing Research. 2007;50:475–492. [PubMed]
  • Roid G, Miller L. Leiter International Performance Scale - Revised. Stoelting Co.; Wood Dale, IL: 1997.
  • Rogers S, Wehner E, Hagerman R. The behavioral phenotype in fragile X: Symptoms of autism in very young children with fragile X syndrome, idiopathic autism, and other developmental disorders. Journal of Developmental & Behavioral Pediatrics. 2001 December;22(6):409–417. [PubMed]
  • Rutter M, LeCouteur A, Lord C. Autism Diagnostic Interview-revised (ADI-R) Western Psychological Services; Los Angeles, CA: 2003.
  • Sabaratnam M, Murthy NV, Wijeratne A, Payne S, Buckingham A. Autism-like behavior profile and psychiatric morbidity in fragile X syndrome: a prospective 10-year follow-up study. European Child and Adolescent Psychiatry. 2003;12:172–177. [PubMed]
  • Schopler E, Reichler R, Renner B. The childhood autism rating scale (CARS) Western Psychological Services; Los Angeles: 1988.
  • Seltzer M, Krauss M, Shattuck P, Orsmond G, Swe A, Lord C. The symptoms of autism spectrum disorders in adolescence and adulthood. Journal of Autism and Developmental Disorders. 2003;33(6):565–581. [PubMed]
  • Shattuck P, Seltzer M, Greenberg J, Orsmond G, Bolt D, Kring S, Lounds J, Lord C. Change in autism symptoms and maladaptive behaviors in adolescents and adults with an autism spectrum disorder. Journal of Autism and Developmental Disorders. 2007 [PMC free article] [PubMed]
  • Sigman M, McGovern C. Improvement in cognitive and language skills from preschool to adolescence in autism. Journal of Autism and Developmental Disorders. 2005;35:15–23. [PubMed]
  • Skinner M, Hooper S, Hatton D, Roberts J, Mirrett P, Schaaf J, Sullivan K, Wheeler A, Bailey D. Mapping nonverbal IQ in young boys with fragile X syndrome. American Journal of Medical Genetics A. 2005;132:25–32. [PubMed]
  • Snow A, Lecavalier L, Houts C. The structure of the Autism Diagnostic Interview-Revised: Diagnostic and phenotypic implications. Journal of Child Psychology and Psychiatry. 2009;50:734–742. [PubMed]
  • Sudhalter V, Belser R. Conversational characteristics of children with fragile X syndrome: Tangential language. American Journal on Mental Retardation. 2001;106(5):389–400. [PubMed]
  • Sudhalter V, Belser R. Atypical language production of males with Fragile X syndrome. In: Hughes D, editor. Educating children with fragile X syndrome. Routledge Falmer; London, UK: 2004. pp. 25–31.
  • Tassone F, Hagerman RJ, Ikle DN, Dyer PN, Lampe M, Willemsen R, et al. FMRP expression as a potential prognostic indicator in fragile X syndrome. American Journal of Medical Genetics. 1999b;84(3):250–261. [PubMed]
  • Turk J, Graham P. Fragile X syndrome, autism and autistic features. Autism. 1997;1:175–197.
  • Volkmar F, Klin A. Issues in the classification of autism and related conditions. In: Volkmar F, Paul R, Klin A, Cohen D, editors. Handbook of autism and pervasive developmental disorders. Wiley; Hoboken, NJ: 2005. pp. 5–41.
  • Wechsler D. Wechsler Intelligence Scales for Children – Third Edition. Manual. The Psychological Corporation; San Antonio, TX: 1991.
  • Willemson R, R., Smits A, Mohkamsing S, van Beerendonk H, de Hann M, de Vries B, et al. Rapid antibody test for diagnosing fragile X syndrome: a validation of the technique. Human Genetics. 1997;99:308–311. [PubMed]
  • World Health Organization . The international statistical classification of diseases and related health disorders. Geneva: 1992.