Adaptive functioning in children and adolescents with Trisomy X: An exploratory analysis

Abstract

Identifying the factors related to adaptive functioning will improve the information available to families and providers of females with Trisomy X. Cognitive and behavioral features were assessed in 50 females ages 12.2 ± 3.6 years using the Behavior Assessment System for Children Second Edition (BASC-2) and Wechsler Scales of Intelligence. Executive functioning, social skills, and autistic traits were evaluated in a subset. Adaptive functioning was assessed using the BASC-2 adaptive skills composite score (ASC). Participants were classified as average adaptive skills (ASC T-score > 40) or deficits (ASC T-score < 40). Group comparisons were conducted. Multiple linear regression examined which factors contributed to ASC score. Twenty-eight females (55.6%) had adaptive skills deficits with functional communication being the most commonly affected adaptive domain. The group with ASC in the average range had higher verbal IQ (VIQ) and lower rates of numerous behavioral concerns. Internalizing behavior composite, DSM-IV inattentive symptoms score, and VIQ were significant predictors of ASC. Prenatally diagnosed females comprised over 70% of those with average adaptive skills. In this study, internalizing behaviors, inattentive ADHD symptoms, and VIQ were associated with poorer adaptive functioning. Early interventions targeting internalizing behaviors, attention/executive functioning, and communication skills may improve adaptive skills and deserve further study.

1 |. INTRODUCTION

The rise in genetic diagnoses via genomic testing in the prenatal and neonatal setting provides opportunities and challenges to clinical care and research. Early genetic diagnosis has the potential to optimize outcomes through timely intervention and surveillance. Ascertainment in the prenatal or neonatal period has allowed researchers to make important observations on the variability and natural history of many rare disorders. At the same time, clinicians are tasked with providing accurate and comprehensive counseling on a growing number of rare conditions. This becomes more challenging where there is substantial phenotypic variability or there is limited data in the medical literature to inform clinicians and families.

Trisomy X (47,XXX or Triple X syndrome) is a prime example of a genetic condition in females characterized by variable medical and psychological features ranging from asymptomatic to significant neurodevelopmental challenges. Most children and adolescents with Trisomy X are healthy, although a subset of females have congenital heart and genitourinary malformations, seizures, and increased risk for autoimmune disease and premature ovarian failure (Goswami et al., 2003; Grosso et al., 2004; Roubertie, Humbertclaude, Leydet, Lefort, & Echenne, 2006; Wigby et al., 2016). Comprehensive reports of medical problems associated with Trisomy X across the lifespan and have been summarized elsewhere (Otter, Schrander-Stumpel, & Curfs, 2010; Stochholm, Juul, & Gravholt, 2010; Tartaglia, Cordeiro, Howell, Wilson, & Janusz, 2010; Tartaglia, Howell, Sutherland, Wilson, & Wilson, 2010).

For families facing a new diagnosis of Trisomy X, questions regarding long-term adaptive functioning frequently arise. Adaptive functioning can be defined as how well a person handles the common demands in life and how independent they are compared to others of a similar age and background in skills needed for functioning in daily life. General domains of adaptive functioning include communication, self-care, home and community living skills, adaptability, and social skills. While adaptive deficits have been reported in females with Trisomy X (Tartaglia, Cordeiro, et al., 2010; Tartaglia, Howell, et al., 2010), the factors influencing and associated with differences in adaptive functioning in Trisomy X have been largely unexplored. Beginning to identify these factors will improve the information available to families of females with Trisomy X who often inquire as to the expected outcomes and independence for their daughter. Additionally, it can identify domains which may be targeted for interventions. The aim of this study is to investigate the relationship of cognitive, behavioral, and psychological features with adaptive functioning in females with Trisomy X in the first two decades of life.

The relative paucity of data regarding Trisomy X compared with other sex chromosome aneuploidies (SCAs) has been frequently attributed to low clinical ascertainment. Although 47,XXX has a relatively common incidence estimated at 1 in 1,000 live female births (Jacobs, 1979), prior studies have suggested clinical ascertainment may be as low as 10% (Nielsen, 1990). A recent population-based study in Denmark using national cytogenetic registry data confirmed that non-diagnosis remains a pervasive challenge as they estimated that 87% of females with Trisomy X remained undiagnosed (Berglund et al., 2019). Low clinical ascertainment has been proposed to be due to many 47,XXX females being asymptomatic or showing only mild physical or neurodevelopmental concerns that are under-recognized as features of the condition and do not meet threshold of severity for current genetic testing practice recommendations (Moeschler et al., 2014; Wigby et al., 2016).

Clinical ascertainment rates are expected to increase due to widespread noninvasive prenatal genetic screening (NIPS) using cell free fetal DNA as well as expanded postnatal clinical cytogenetic testing including chromosomal microarray (Committee on Practice Bulletins-Obstetrics, 2016; Konman et al., 2018). Of note, Trisomy X has a low positive predictive value compared with other aneuploidies screened by NIPS with estimates ranging from 18 to 50% (Deng et al., 2019; Peterson et al., 2017). This underscores the importance of confirmatory prenatal or postnatal testing for positive NIPS Trisomy X cases.

Longitudinal studies of prospectively identified children with SCAs by Arthur Robinson and international collaborators contributed important early descriptions of developmental and behavioral manifestations of Trisomy X (Evans, Hamerton, & Robinson, 1990; Ratcliffe & Paul, 1986; Robinson, Puck, Pennington, Borelli, & Hudson, 1979; Stewart, Netley, & Park, 1982). Investigators screened 200,000 consecutive births with Barr body analysis of placental membranes and identified approximately 200 children with SCAs who were subsequently studied longitudinally, including 43 females with 47,XXX. Speech and language delays in addition to learning disabilities were common, although a subset of females were reported to have no academic difficulties. Shy personality was described in many cases. Performance on intellectual testing typically fell in the normal range; however, females with Trisomy X on average scored 10–20 points lower when compared to their siblings. With small sample sizes and a range of neurodevelopmental abilities, the investigators cautiously concluded that a precise prognosis was difficult. Linden and Bender (2002) subsequently described a separate prenatally ascertained cohort of children with SCAs and compared the findings to Robinson et al.’ s newborn screening cohort. Among 17 females with nonmosaic 47,XXX (ages 7–18 years) diagnosed prenatally described by Linden and Bender, six (35%) were reported to have no academic difficulties. An additional eight females had mild academic difficulties such as requiring additional time to complete tests, using a tutor or additional classroom support. Only three females required special education classes. No females in the study were described to carry clinical mental health diagnoses.

More recent studies have expanded on the range of neurobehavioral features in females with Trisomy X in different cohorts recruited primarily for research studies or from clinical samples. Several studies have reported an increased frequency of several psychological disorders including ADHD, autism spectrum disorders (ASD), learning disabilities, language disorders, and anxiety disorders at a variable frequency depending on the sample studied (Bender, Harmon, Linden, & Robinson, 1995; Bishop et al., 2019; Linden et al., 1988; Tartaglia, Ayari, Hutaff-Lee, & Boada, 2012; van Rijn, Stockmann, van Buggenhout, van Ravenswaaij-Arts, & Swaab, 2014; Wigby et al., 2016). Executive functioning and social cognition deficits have also been described (Lee et al., 2012; van Rijn et al., 2014; van Rijn & Swaab, 2015). Freilinger et al. (2018) characterized emotional and behavioral challenges in girls and women with Trisomy X from Germany and reported that compared to age matched controls, a significantly greater proportion of females with Trisomy X exhibited scores within the clinical range on the Child Behavior Checklist for attention, school performance, and internalizing problems. A survey of psychological symptoms (Symptom Checklist-90-revised) of 32 adult women with Trisomy X identified elevated mean scores for several subscales compared to controls including interpersonal sensitivity, depression, phobic anxiety, and paranoid ideation. However, mean scores for the Trisomy X group remained below the clinically significant cut-offs for the scale (Freilinger et al., 2018). In total, these findings have characterized both clinically significant and more subtle neurobehavioral features that can impact adaptive and social functioning.

One study in adult women with Trisomy X highlights the long-term consequences of deficits in social and adaptive skills on socio-economic status and social functioning. In a population-based study in Demark that included 108 women with a median age of diagnosis of 19 years, Stochholm et al. reported poorer socioeconomic status among women with Trisomy X compared when compared with age-matched controls including reduced income and educational attainment (Stochholm, Juul, & Gravholt, 2013). Women with Trisomy X in the study also had lower proportions of partnership, cohabitation, and motherhood compared to controls. The authors hypothesized that common behavioral features of Trisomy X such as increased shyness and sensitivity may help account for the differences in social functioning. These findings underscore the need for increased understanding of the developmental trajectory of adaptive functioning outcomes in Trisomy X. They also highlight the need for further studies examining if diagnosis and earlier recognition of risk factors for adaptive impairments may improve adult outcomes.

1.1 |. Study aims

Understanding the factors that influence adaptive functioning would provide important prognostic information as well as identify potential behavioral targets for intervention. In this study, we first characterize adaptive functioning in a cohort of school-aged females with Trisomy X and compare the neurobehavioral profiles of Trisomy X females with adaptive skills deficits to those with age appropriate adaptive functioning. We then examine which neurobehavioral features contribute the most to adaptive skills composite (ASC) scores. Our clinical experience and the prior literature support variable neurodevelopmental phenotypes in females with Trisomy X, including females who perform in the average range on cognitive testing but still manifest deficits in adaptive functioning. Therefore, we hypothesize that other factors associated with the Trisomy X phenotype including deficits in language/verbal abilities, executive functioning, attentional deficits, and anxiety/depression are associated with adaptive impairments in Trisomy X. Finally, our previous work in Trisomy X highlighted the phenotypic differences based on timing of diagnosis with females who receive a postnatal diagnosis tending to have more significant neurobehavioral impairments, although variability exists among prenatally diagnosed females as well (Wigby et al., 2016). Therefore, we evaluated differences in neurobehavioral profiles by timing of diagnosis with an emphasis on females with prenatal diagnoses as this group is expected to become a growing proportion of Trisomy X diagnoses.

2 |. MATERIALS AND METHODS

2.1 |. Recruitment

Females with nonmosaic Trisomy X were evaluated at one of two sites: Children’ s Hospital of Colorado (CHCO) eXtraordinarY Kids Clinic and at the University of California-Davis Medical Center, Medical Investigation of Neurodevelopmental Disorders (MIND) Institute. All participants or their parents provided informed consent and assent, as appropriate, prior to participation in research. The study was approved by the institutional review board at each site (COMIRB 08–0513; MIND Institute IRB 220354).

The participants evaluated at the CHCO participated in a larger study on health and development of children and adults with SCAs. As part of this study, participants underwent comprehensive phenotyping including physical examination, parent reported and directly administered assessments of cognitive, behavioral, and psychological functioning, and medical record review. Some participants were previously reported in the cohort by Wigby et al. (2016), however, the present study includes an additional eight participants evaluated at CHCO as well as unique participants from the MIND Institute who participated as members of a control comparison group (see further description below). Participants were recruited from multiple sources including national advocacy and support organizations for SCAs and an interdisciplinary clinic for children and adolescents with SCA (eXtraordinary Kids Clinic). Additional participants were identified by referral through genetics and pediatrics clinics. Although multiple recruitment strategies were employed, most Trisomy X females in the CHCO sample were identified by physician or self-referral and therefore represent a clinically ascertained sample.

Ascertainment bias has been a significant challenge in the SCA literature and can result in the over-representation of individuals with more significant medical and neurobehavioral challenges. To improve the diversity of participants, a second Trisomy X sample recruited purely for research purposes was included from the University of California MIND Institute. This group was recruited as a comparison group for a study of visual–spatial skills in individuals with 22q11.2 deletion syndrome. At both sites, genetic testing results were reviewed to confirm a nonmosaic 47,XXX karyotype.

We applied shared inclusion criteria which are (a) nonmosaic 47,XXX karyotype, (b) school age participant (6–20 years), and (c) completion of the Behavior Assessment System for Children Second Edition (BASC-2) Parent Rating Scales (PRS). This resulted in a Trisomy X sample of 27 participants from CHCO and 23 participants from MIND. Participants were compared across the two sites to verify that there were no duplicate participants. Assessment of the participants at the two sites revealed no significant differences in age, race/ethnicity, proportion of prenatal diagnosis, or level of maternal education (Table 1). Therefore, we combined data from the two sites to comprise a more heterogeneous Trisomy X sample that is more representative of the variability present in this condition among females with a known Trisomy X diagnosis.

TABLE 1.

Demographic features of Trisomy X participants and site comparisons

	Total N = 50	CHCO N = 27	MIND N = 23	p-Value a
% Caucasian	84	85	83	1.000
% Maternal college degreeb	57	76	60	.336
% Prenatal diagnosis	57	48	68	.246
	Mean (SD), range	Mean (SD)	Mean (SD)
Age (years)	12.4 (3.4), 7.7–20.9	13.1 (4.0)	11.6 (2.4)	.125
VIQ	88.9(14.2), 57–117	84.5 (12.9)	93.8 (14.2)	.024
GAI	91.9(14.3), 58–126	87.5 (13.2)	96.8 (14.2)	.026

Abbreviations: CHCO, Children’s Hospital of Colorado; GAI, Global abilities index; MIND, University of California, Davis MIND Institute; VIQ, verbal IQ.

^aStudent’s t tests used for comparisons of mean age, VIQ, GAI. Fisher's exact two-tailed tests used for comparisons of proportions of non-Hispanic white participants, maternal college degree, and prenatal diagnosis.

^bCompletion of college or advanced degree.

2.2 |. Evaluation

Participants received neurobehavioral assessments to characterize their cognitive, adaptive, and behavioral functioning and included individually administered assessments (e.g., Wechsler Scales of Intelligence) as well as parent completed measures. All participants received the BASC-2 PRS, which evaluates for a variety of childhood behavioral and emotional problems and adaptive functioning domains (Reynolds & Kamphaus, 2004). The measure is validated for ages 2–21 years and is divided into three age-group forms: preschool (ages 2–5 years), child (ages 6–11 years), and adolescent (ages 12–21 years). Depending on the age-group form, there are 134 to 160 items that are rated on a 4-point scale of frequency ranging from “Never” to “Almost Always.” For the present study, the child or adolescent BASC-2 PRS forms were utilized and completed by the parent or guardian.

Adaptive functioning was assessed using the BASC-2 ASC score, which measures adaptive functioning via five subdomains: adaptability, social skills, leadership, activities of daily living, and functional communication (see Table 2). A summary score, the ASC score, is generated from the subscales. Scores are reported as T-scores with a test mean of 50 and a SD of 10. For the ASC score and subdomains, a T-score of 30 to 39 indicates risk for adaptive skills deficits, while T-scores below 30 indicate clinically significant deficits. The BASC-2 PRS has a reported test–retest reliability across clinical samples and age-matched general samples ranging from 0.81 to 0.87 and interrater reliability of 0.7 to 0.83 (Reynolds & Kamphaus, 2004).

TABLE 2.

Description of BASC-2 PRS adaptive skills subdomains

Subdomain name	Abr	Description	Item examples
Activities of daily living	ADL	Ability to independently perform daily tasks including dressing, feeding, toileting, household chores	Needs help tying shoes, acts in a safe manner, needs to be reminded to brush teeth
Adaptability	ADP	Ability to adapt to a variety of situations and changes in plans	Gets upset when plans are changed, shares with other children
Leadership	LDR	Ability to initiate and work in a group	Other children look to them for help, will speak up if the situation calls for it
Functional communication	FC	Ability to clearly communicate requests, thoughts, and emotions	Able to describe feelings, answers the telephone properly
Social skills	SS	Ability to interact with others	Offers to help other children; shows interest in others' ideas
Adaptive skills composite	ASC	Composite score of all adaptive skills domains

Abbreviations: BASC-2, Behavior Assessment System for Children Second Edition; PRS, Parent Rating Scales.

Previous studies have found the BASC-2 ASC score to be strongly correlated with other gold standard assessments of adaptive behavior such as the Adaptive Behavior Assessment System (ABAS; Harrison & Oakland, 2003; Papazoglou, Jacobson, & Zabel, 2013) and the Vineland Adaptive Behavior Scales (Reynolds & Kamphaus, 2004). The same was found in our sample such that BASC-2 ASC score was strongly correlated with the Vineland composite score in a subset of participants (n = 19) for whom both measures were administered (r_s(19) = .79, p < .001) and with the ABAS composite score (r_s(24) = .57, p = .003) in another subset that had received the ABAS. Based on the degree of correlation of the BASC-2 ASC score with other gold standard measures of adaptive functioning in both previous studies and our current dataset, we concluded that the BASC-2 was a valid measurement of adaptive functioning in this population for use in this analysis.

Participants were also evaluated for a variety of childhood behavioral problems using the BASC-2 clinical scales which includes nine clinical subdomains (hyperactivity, aggression, conduct problems, anxiety, depression, somatization, attention problems, atypicality, and withdrawal), as well as composite scores in internalizing behaviors (anxiety, depression, and somatization), externalizing behaviors (hyperactivity, aggression, conduct problems) and a behavioral symptoms index (BSI, combines all behavioral symptoms). Attention and hyperactivity subdomains were not individually analyzed as these subdomains were better represented by the Conners’ Revised PRS (Conners’; described below). The clinical subscales and composites of the BASC-2 also have a mean of 50 and a SD of 10; however, higher scores indicate more problems. For these clinical scales, a T-score of 60–69 is classified as at-risk and T-scores greater than 69 fall in the clinically significant range. Scores of 59 and below on a given BASC-2 clinical subdomain are considered within the nonclinical range.

Cognitive evaluation was performed using one of the Wechsler Scales of Intelligence including WASI (Wechsler, 1999), WISC-III (Wechsler, 1991), WISC-IV (Wechsler, 2004), WAIS-III (Wechsler, 1997), or WAIS-IV (Wechsler, 2008). Most participants from CHCO (n = 24, 88.9%) and MIND (n = 22, 95.6%) completed one of the Wechsler Scales with variability of measures due to the date and age of participant at the time of evaluation. General Abilities Index (GAI) was calculated for all scales based on verbal IQ (VIQ) and performance IQ (PIQ) scores to allow for comparison between tests (Tulsky, Saklofske, Wilkins, & Weiss, 2001). Measures performed at the same evaluation as the BASC-2 or within 1 year of assessment were included in the analysis.

Attention deficit and hyperactivity disorder symptomatology was examined using the Conners’ Revised PRS, from which the DSM-IV Inattentive and DSM-IV Hyperactivity/Impulsivity subscales were utilized, which has a reported test–retest reliability of 0.76 and interrater reliability of 0.94 (Swanson et al., 2001). The DSM-IV Inattentive domain includes items related to sustaining attention, distractibility, making careless mistakes, not listening when spoken to directly, not following through on instructions, failing to finish or having difficulty organizing tasks/activities in school and home settings, losing materials, avoiding or disliking tasks requiring sustained mental effort, and being forgetful in daily activities. The Hyperactivity/Impulsivity subscale covers behaviors such as fidgeting, excessive movement or excessive talking, leaving seat when expected to sit, restlessness, impulsive behaviors, and interrupting or intruding on others.

Additional evaluations of other neurobehavioral features associated with Trisomy X were obtained in a subset of participants as there were differences in the study protocol questionnaire battery at each site. Other additional assessments of executive functioning including the Behavior Rating Inventory of Executive Functioning, Second Edition (BRIEF-2, n = 22) (Gioia, Isquith, Guy, & Kenworthy, 2015) and social behaviors associated with ASD using the Social Responsiveness Scales, Second Edition (SRS-2) (n = 29) (Constantino & Gruber, 2012). Test–retest reliability and interrater reliability for the BRIEF-2 was 0.79 and 0.73–0.95 in a Dutch sample (Huizinga & Smidts, 2011) and for the SRS-2 was 0.88–0.95 and 0.61 (Constantino & Gruber, 2012).

2.3 |. Analysis of results

All analyses were conducted with SPSS, v26 (IBM, 2019). Descriptive statistics were calculated by adaptive skills group, timing of diagnosis and for the total sample. Participants were classified as average skills (T-score ≥ 40) or at risk (T-score < 40) based on BASC-2 ASC score. Independent sample two-tailed t tests and Fisher’ s exact tests were used with continuous and categorical variables, respectively, to examine differences between average and at-risk adaptive groups, as well as prenatal and postnatal groups on demographic, cognitive, and behavioral measures. Statistical significance was set at <.05, and additional Bonferroni correction to account for the number of comparisons (including 5 comparisons for the BASC-2 subdomains and 17 comparisons for the demographic and individual cognitive/behavioral subdomains) is also reported. All analyses were performed with the available data without imputation.

Univariate correlations between different aspects of the clinical phenotype (i.e., cognitive skills, executive functioning, behavior symptoms, etc.) and adaptive skills (ASC) were examined. Correlations greater than .5, with p-value <.05 and of potential clinical significance were considered for inclusion in a multiple linear regression. Multiple linear regression analysis was done to identify factors associated with adaptive skills. Model fit was examined using Q–Q plots, and the variance inflation factor was used to assess multicollinearity. In a stepwise regression, predictor variables are entered into the regression equation one at a time. At each step in the analysis, the predictor variable that contributes the most to the prediction equation (i.e., increasing the multiple correlation) is entered first. This process continues by adding variables that contribute to the regression equation. The regression analysis ends when additional predictor variables do not add anything statistically meaningful to the regression equation. In the final multiple linear regression, variables with a p-value <.05 were declared significantly associated with adaptive skills.

3 |. RESULTS

Fifty females with Trisomy X met inclusion criteria. Table 1 presents demographic characteristics for all participants, as well as comparisons by site. Mean age at assessment was 12.2 ± 3.6 years (range 7.7–20.9 years). There were no significant differences between sites for age, race/ethnicity, maternal education level, or proportion of females with a prenatal Trisomy X diagnosis. Participants from the MIND Institute had higher Wechsler VIQ (p = .024) and GAI scores (p = .026) consistent with the expected differences in sample ascertainment between the two sites.

Overall, 44% (n = 22) of participants had ASC scores in the average range. The remaining 28 females had adaptive skills deficits (ASC T-score < 40, 37.4 ± 10.4, range 17–39) including 17 (34%) scoring in the at-risk range (ASC T-score 30–39) indicating more mild adaptive skills impairments and 11 (22%) with scores in the range of clinically significant deficits (ASC T-score < 30). Among the subdomains of adaptive skills (Figure 1), participants had the highest mean scores in social skills (T-score 42.8 ± 10.2) and adaptability (T-score 42.7 ± 12.3), indicating areas of relative strength in adaptive functioning for most of the participants. Conversely, functional communication was a concern for nearly two-thirds (65.9%) of participants (T-score 34.9 ± 11.6, range 10–63) with 34.1% of parents reporting concerns with functional communication that T-scored in the clinically significant range.

FIGURE 1.

Behavior Assessment System for Children Second Edition (BASC-2) adaptive skill subdomain scores: proportion average, at-risk, and clinically significant (total n = 50)

Comparisons of Trisomy X females with age-appropriate adaptive skills (n = 22) to those with deficits (n = 28) showed that there were no significant differences in age (p = .677), race/ethnicity (p = .718), or maternal education (p = .528). Females with prenatal diagnoses accounted for over 73% of all females with age-appropriate adaptive skill scores (p = .046). Females with average adaptive skills had higher scores across all adaptive skills subdomains. Several key differences between the two groups were identified that remained significant after correction for multiple comparisons (Table 3). Females with age appropriate adaptive skills had higher mean VIQ and GAI scores (p < .001). Behaviorally, females with adaptive skills deficits had higher scores (more challenges) in aggression, depression, withdrawal and atypicality on the clinical subdomains of the BASC-2 with mean T-scores in the depression and withdrawal subdomains that fell in clinically significant range. Females with adaptive skills deficits also had higher Conners’ DSM-IV inattention scores. The degree of social deficits and autism symptoms, as measured by the SRS-2 total T-score (n = 29) also revealed higher T-scores in Trisomy X females with adaptive skills deficits (p < .001). Further, the group with age-appropriate adaptive skills had fewer anxiety symptoms and higher executive functioning skills (both p < .05), although these did not withstand Bonferroni correction.

TABLE 3.

Comparison of adaptive functioning and neurobehavioral profiles in females with Trisomy X

	Average adaptive skills N = 22 mean (SD)	Adaptive deficits N = 28 mean (SD)	p-Value
BASC-2 Adaptive Scalesa
Activities of daily living	44.8 (9.5)	29.1 (11.1)	<.001*
Adaptability	51.1 (9.6)	35.3 (9.2)	<.001*
Functional communication	43.3 (8.2)	28.0 (9.2)	<.001*
Leadership	44.4 (6.9)	32.4 (5.6)	<.001*
Social skills	49.9 (9.0)	37.3 (7.2)	<.001*
ASC	46.5 (6.7)	30.3 (6.0)	<.001*
Variable
Age (years)	12.2 (3.1)	12.6 (3.7)	.677
% Caucasian	81.8	85.7	.718
% Maternal college degreeb	63.2	73.1	.528
% Prenatal Dx	73.9	42.9	.046
Wechsler IQc
VIQ	97.4 (12.9)	81.8 (11.2)	<.001*
PIQ	100.1 (15.3)	90.8 (12.8)	.030
GAI	99.3 (14.4)	85.7 (11.2)	<.001*
Conners’a
DSM-IV inattention index	64.5 (11.8)	74.1 (11.7)	.017
DSM-IV hyperactivity/impulsivity index	55.9 (13.1)	61.6 (13.4)	.194
BASC-2 Clinical Scalesa
Aggression	46.0 (6.1)	56.8 (10.6)	<.001*
Conduct problems	52.0 (10.9)	57.1 (17.2)	.204
Anxiety	55.8 (9.3)	64.6 (16.8)	.022
Depression	52.4 (8.4)	72.2 (19.1)	<.001*
Somatization	58.0 (17.0)	62.8 (19.9)	.366
Atypicality	51.2 (9.4)	67.1 (14.7)	<.001*
Withdrawal	58.3 (11.5)	70.2 (14.4)	.003*
Externalizing symptoms composite	49.8 (8.3)	59.0 (12.4)	.003*
Internalizing symptoms composite	56.5 (10.9)	69.8 (20.2)	.005
BSI	53.0 (6.1)	69.0 (10.3)	<.001*
SRS-2 total T-scored, n = 29	59.9 (6.1)	82.9 (14.9)	<.001*
BRIEF Global Executive Compositee, n = 22	59.7 (9.0)	76.3 (11.8)	.006

Note: Original p-values reported from independent samples t tests and/or Fisher exact test.

^*p-Value significant after correction for multiple comparisons (p ≤ .003).

Abbreviations: ASC, adaptive composite; BASC-2, Behavior Assessment System for Children Second Edition; BRIEF-2; Behavior Rating Inventory of Executive Functioning, Second Edition; BSI, Behavioral Symptom Index; GAI, Global abilities index; PIQ, performance IQ; SRS-2, Social Responsiveness Scales, Second Edition.

^aConners’ and BASC-2 Clinical Scales are reported as T-scores with a mean of 50 and SD of 10. Scores 60–69 fall in the “at-risk” range and T-scores >69. Scores >70 indicate clinically significant symptoms. On the BASC-2 Adaptive Scales, Scores 30–39 fall in the “at-risk” range and T-scores <30 indicate clinically significant symptoms.

^bCompletion of undergraduate college degree or higher.

^cWechsler IQ standard scores are reported and have a mean of 100 with a SD of 15.

^dSRS-2 scales are reported as T-scores with Scores ≤59 fall within the normal range. T-scores 60–65 indicate mild social deficits, Scores 66–75 moderate social deficits, Scores ≥76 indicate severe deficits.

^eBRIEF2 scales are reported as T-scores with Scores ≤59 fall within the normal range. T-scores 60–65 indicate mild deficits, Scores 66–69 moderate deficits, Scores ≥70 indicate severe or clinically significant executive functioning deficits.

The following variables were selected for the stepwise multiple linear regression VIQ, BASC-2 Internalizing Composite score and Conners’ DSM-IV Inattentive subscale score as these factors were significantly correlated with ASC and of clinical significance. Scatterplots depict the relationship and report the correlations between each of these variables with ASC score by timing of diagnosis (see Figure S2a–c in supplementary material). As can be seen, more internalizing behaviors and inattentive symptoms, as well as lower VIQ are associated with lower adaptive skills. Multiple linear regression selected all three predictors and indicated that there was a collective significant association between BASC-2 Internalizing Composite score, Conner’ s DSM-IV Inattentive T-score, VIQ and BASC-2 ASC score, (adjusted R² = .535 (F(3, 31) = 14.06, p < .001), and together accounted for 53.5% of the variance in ASC (see Table 4). The individual predictors were examined further and indicated that internalizing behavior composite score (t = −2.435 p = .021), DSM-IV inattentive score (t = −3.278, p = .003) and VIQ (t = 2.379, p = .024) were significant predictors in the model and together accounted for 53.5% of the variance in ASC (see Table 5).

TABLE 4.

Summary of stepwise regression analysis for variables predicting adaptive skills (N = 34)

Variable	Regression model
	B	SE B	β	95% Confidence interval
Constant	49.824	11.896		25.562, 74.086
Internalizing behavior	−.163	.067	−.324*	−.299, −.026
Inattentive index	−.302	.092	−.405**	−.491, −.114
VIQ	.206	.087	.306*	.029, .383
Adjusted R2		.535
F for change in R2		5.659*

Note:B, unstandardized coefficient; SE B, standard error of the coefficient; β, standardized coefficient.

Abbreviation: VIQ, verbal IQ.

^*p < .05.

^**p < .01.

TABLE 5.

Comparison of neurobehavioral profiles by time of diagnosis

	Time of diagnosis			Prenatal diagnosis only
	Prenatal N = 28 mean (SD)	Postnatal N = 21 mean (SD)	p-Value	Average adaptive skills N = 16 mean (SD)	Adaptive deficits N = 12 mean (SD)	p-Value
BASC-2 Adaptive Scalesa
Activities of daily living	39.9 (9.6)	30.6 (15.4)	.031	43.3 (8.8)	34.8 (8.9)	.028
Adaptability	44.6 (10.7)	39.8 (14.3)	.200	50.1 (9.7)	36.7 (6.4)	.001*
Functional communication	38.9 (9.8)	28.7 (11.3)	.003*	42.4 (9.2)	33.6(8.5)	.024
Leadership	40.3 (8.4)	33.7 (7.3)	.006*	44.3 (7.5)	35.1 (6.7)	.003*
Social skills	43.7 (10.0)	41.1 (10.4)	.392	47.9 (9.3)	38.1 (8.2)	.008*
ASC	40.2 (8.7)	33.1 (10.9)	.015	45.3 (6.6)	33.5 (6.3)	<.001*
Variable
Age (years)	11.8 (3.1)	13.2 (3.7)	.155	12.3 (3.1)	11.1 (3.2)	.347
% Caucasian	82.1	85.7	1.000	87.5	75	.624
% Maternal college degreeb	76.9	55.6	.191	71.4	83.3	.652
Wechsler IQc
VIQ	93.2 (14.9)	82.6 (10.6)	.010	99.4 (11.6)	83.9 (14.8)	.007
PIQ	97.4 (13.6)	91.5 (15.7)	.183	100.2 (15.5)	93.2 (9.0)	.169
GAI	95.9 (13.7)	86.1 (13.3)	.019	100.6 (13.8)	88.9 (10.5)	.033
Conners’a
DSM-IV Inattention Index	66.5 (11.5)	73.6 (13.3)	.093	64.7 (10.9)	69.5 (12.4)	.363
DSM-IV Hyperactivity/Impulsivity Index	55.1(11.3)	62.6 (13.6)	.075	53.3 (11.1)	57.9 (11.9)	.383
BASC-2 Clinical Scalesa
Aggression	48.9 (10.0)	56.2 (9.8)	.014	45.0 (5.8)	54.2 (12.1)	.028
Conduct problems	50.3 (11.7)	61.1 (16.9)	.017	52.1 (12.0)	47.8 (11.4)	.339
Anxiety	56.9 (11.8)	66.5 (16.4)	.021	56.1 (10.4)	58.1 (13.8)	.661
Depression	58.0 (15.0)	71.7 (19.4)	.010	52.9 (9.6)	64.8 (18.4)	.035
Somatization	57.7 (19.4)	65.6 (17.0)	.144	56.4 (16.8)	59.5 (23.2)	.682
Atypicality	55.9(11.6)	66.3 (16.9)	.013	51.9 (10.2)	61.2 (11.6)	.033
Withdrawal	62.9 (15.4)	68.2 (12.6)	.204	58.8 (12.3)	68.4 (17.8)	.126
Externalizing symptoms	50.5 (8.9)	60.7 (12.8)	.004	49.2 (8.2)	52.3 (9.9)	.366
Internalizing symptoms	59.2 (16.4)	71.3 (17.6)	.017	56.1 (11.6)	63.3 (21.0)	.257
Behavioral symptom index	57.5 (8.4)	68.3 (13.1)	.002*	53.2 (6.4)	63.2 (7.4)	.001*
SRS-2 total T-scored	67.7 (15.4)	82.2 (15.0)	.017	60.3 (6.6)	77.1 (18.7)	.056
BRIEF global executive compositee	66.2 (10.5)	77.3 (13.8)	.047	59.7 (9.0)	74.0 (6.0)	.014

Note: Original p-values reported from independent samples t tests and/or Fisher exact test.

^*p-value significant after correction for multiple comparisons (p ≤ .003).

Abbreviations: BASC-2, Behavior Assessment System for Children Second Edition; BRIEF-2; Behavior Rating Inventory of Executive Functioning, Second Edition; GAI, Global abilities index; PIQ, performance IQ; SRS-2, Social Responsiveness Scales, Second Edition.

^aConners’ and BASC-2 Clinical Scales are reported as T-scores with a mean of 50 and SD of 10. Scores 60–69 fall in the “at-risk” range and T-scores >69. Scores >70 indicate clinically significant symptoms. On the BASC-2 Adaptive Scales, Scores 30–39 fall in the “at-risk” range and T-scores <30 indicate clinically significant symptoms.

^bCompletion of undergraduate college degree or higher.

^cWechsler IQ standard scores are reported and have a mean of 100 with a SD of 15.

^dSRS-2 scales are reported as T-scores with Scores ≤59 fall within the normal range. T-scores 60–65 indicate mild social deficits, Scores 66–75 moderate social deficits, Scores ≥76 indicate severe deficits. Prenatal diagnosis group n = 16 (nine average adaptive skills and seven adaptive skills deficits), postnatal diagnosis group n = 13.

^eBRIEF 2 scales are reported as T-scores with Scores ≤59 fall within the normal range. T-scores 60–65 indicate mild deficits, Scores 66–69 moderate deficits, Scores ≥70 indicate severe or clinically significant executive functioning deficits. Prenatal diagnosis group n = 11 (six average adaptive skills and five adaptive skills deficits), postnatal diagnosis group n = 11.

Interestingly, 46.4% (n = 13) of participants had deficits in adaptive skills despite average cognitive skills (defined as a GAI of 85 or greater, see scatterplot in Figure S2a in supplemental data). To further investigate the factors associated with adaptive skills deficits in individuals with average cognitive skills, we restricted our analysis to include only participants with a GAI of 85 or greater (data not shown). This yielded 13 females with adaptive skills deficits and 20 females with average adaptive skills. As with the larger group, mean age did not differ between the two groups (12.7 ± 3.9 vs. 11.8 ± 3.2 years, p = .471). Even among females with average general cognitive abilities, those with adaptive skills deficits had: (a) lower VIQ scores (88.6 ± 8.6 (n = 20) versus 100.5 ± 10.5 (n = 13), p = .002); (b) greater challenges with social deficits/autism symptoms (SRS-2 total T-score 82.7 ± 18.7 (n = 9) versus 60.5 ± 6.0 (n = 11), p = .007); (c) more executive function deficits (BRIEF-2 Global Executive Composite T-score 74.5 ± 8.0 (n = 8) versus 58.8 ± 9.8 (n = 5), p = .009); (d) more symptoms of aggression (57.8 ± 10.3 (n = 13) versus 46.5 ± 6.5 (n = 18), p = .003); and (e) more depression (72.3 ± 17.1 (n = 13) versus 51.7 ± 9.0 (n = 18), p < .001).

As females with a prenatal diagnosis of Trisomy X accounted for a significantly greater proportion of females with age-appropriate adaptive scores, we further compared the profiles between prenatal and postnatal diagnosis groups (Table 5). Females with prenatal diagnoses had higher scores in the functional communication (p = .003) and leadership (p = .006) adaptive subdomains. The prenatally diagnosed group also had higher VIQ and GAI scores as well as lower scores on several BASC-2 clinical domains including aggression, anxiety, atypicality, conduct problems, depression and composite scores of internalizing and externalizing problems, and BSI (p < .05), although only BSI scores remained significantly different after correction. Within the prenatal group, mean scores fell in the “at-risk” range only in the domains of inattention, withdrawal, social responsiveness, and executive functioning. Finally, levels of executive dysfunction and social deficits/autistic symptoms were greater in females with postnatal Trisomy X diagnoses.

Given the ongoing consideration of ascertainment bias in Trisomy X research, we further explored comparisons in the subset of our sample that was diagnosed in the prenatal period, acknowledging the small sample size and the potential of recruitment bias in this group (Table 5). When we focused on analysis of prenatally diagnosed females with average adaptive skills (n = 16) compared with those who had challenges (n = 12), the findings of higher VIQ and GAI scores were replicated (p < .05). Females with prenatal ascertainment and adaptive skills challenges were found to have higher behavior symptoms index (BSI) scores on the BASC-2 (p = .001). This indicates increased parental concerns regarding overall behavior in females with adaptive skills challenges; however, no specific subdomain on the clinical scales that comprises the BSI (hyperactivity, aggression, depression, attention problems, atypicality, and withdrawal) was individually significant after Bonferroni correction.

4 |. DISCUSSION

This study characterized adaptive functioning in a school age cohort of females with Trisomy X and explored the contribution of cognitive, behavioral, and psychological factors as well as timing of diagnosis to adaptive skills outcomes. Over half (56%) of participating females with Trisomy X had difficulties in adaptive functioning including nearly two-thirds (65.9%) with functional communication challenges. Areas of relative adaptive skill strengths were seen in domains of adaptability and social skills in this cohort. Multiple linear regression analysis revealed that internalizing behaviors, inattentive symptoms, and language/verbal deficits were significant contributors to adaptive functioning. Together these three factors account for over 50% of the variability in ASC scores and internalizing behaviors was the strongest predictor of ASC score. Understanding of the different factors contributing to adaptive functioning outcomes is important for considerations of interventions and supports for the Trisomy X population.

The relationship between cognitive skills (IQ) and adaptive skills is widely recognized and was replicated here in this Trisomy X cohort as lower global cognitive abilities (GAI) scores was associated with adaptive deficits. Lower GAI scores were primarily driven by lower VIQ scores rather than non-VIQ. Lower verbal cognitive skills in Trisomy X are well established as part of the phenotype (Bender et al., 1983; Bishop et al., 2019), although these findings build on previous literature by supporting a significant relationship of verbal and language abilities with daily adaptive functioning skills in Trisomy X. This pattern was also observed among prenatally diagnosed females who had average adaptive skills compared to prenatally diagnosed females with adaptive deficits (Table 5) and lower VIQ scores remained associated with adaptive skills deficits even among females with average cognitive scores (GAI 85 or greater). Functional communication was also a concern in nearly two-thirds of Trisomy X females evaluated in this study, which highlights the role of language/verbal deficits on everyday functioning including considerations of both receptive language skills as well as self-expression in home, academic, and social settings.

Females with Trisomy X as well as the other sex chromosome trisomies (47,XXY and 47,XYY) are at risk for speech and language deficits which has been well known since the early natural history studies by Arthur Robinson and colleagues and replicated in more recent studies as well (Bender et al., 1983; Bishop et al., 2019; Legett, Jacobs, Nation, Scerif, & Bishop, 2010; Wigby et al., 2016). These disorders manifest over the course of childhood and range from delays in acquisition of receptive and expressive language milestones in infancy and early childhood to speech dyspraxia and challenges with language pragmatics. While more common and more severe in the Trisomy X patients with a postnatal diagnosis or from a high bias group, this study and others support the high risk for language difficulties within the prenatally diagnosed or low bias cohorts compared to normative samples (Bishop et al., 2019). Language and communication impairments can exacerbate psychological domains including anxiety, depression, and irritability which may manifest as behavioral displays of frustration/tantrums, shyness, or social withdrawal. Thus, early recognition and treatment of language delays may have a positive impact beyond language development in improving self-expression and overall psychological and behavioral functioning in females with Trisomy X.

Interestingly, within the adaptive skills deficits group over 50% (n = 15) had GAI IQ scores of 85 or greater (range 85–108, see scatterplot in Supplemental Figure S2d). This finding further supports the role of additional phenotypic factors contributing to adaptive functioning deficits in the Trisomy X population. Therefore, while IQ is an important contributor for adaptive functioning, as it is well recognized in psychology literature, it is important for clinicians to have a heightened awareness that adaptive and language deficits may manifest in Trisomy X females across the IQ spectrum (Table 4).

Mood and behavioral findings are also known areas of risk in Trisomy X (Otter et al., 2010; Tartaglia, Cordeiro, et al., 2010; Tartaglia, Howell, et al., 2010). Our findings further probed specific psychological domains to explore their relationship with adaptive functioning, and identified the domains of depressed mood, aggressive behavior, withdrawal, and atypicality as measured by the BASC-2 to be associated with deficits in adaptive skill outcomes. Results from the regression analysis identified the internalizing composite to be responsible for the largest amount of variance in the adaptive composite (31.5%), which includes subdomains of depression, anxiety, and somatization. The significant role of mental health symptoms such as anxiety and depression on adaptive outcomes has also been reported in the general pediatric population as well as other groups such as autism, 22q11.2 deletion syndrome, and Rett syndrome (Angkustsiri et al., 2012; Barnes et al., 2015; Gardiner & Iarocci, 2018). Thus, identification and treatment of the mental health conditions may be important beyond relief of immediate symptoms in improving overall functioning across adaptive domains. Further research is needed to determine the optimal mental health treatments in Trisomy X and how earlier identification may improve adaptive outcomes.

The atypicality domain of the BASC-2 and the SRS both contain items that measure social deficits and atypical social communication that can be associated with ASD. Deficits in social communication and increased risk for ASD have been previously reported in Trisomy X cohorts (Bishop et al., 2011; van Rijn et al., 2014; Wilson, King, & Bishop, 2019), and adaptive deficits are well known to be associated with ASD in the general population as well (Kanne et al., 2011). Given the strong relationship of these domains to adaptive outcomes in this cohort, further research in the profile of social communication and ASD symptoms in Trisomy X is needed.

The DSM-IV inattentive subscale was also identified as a significant contributor to the regression model. While this scale contains items inquiring about attentional skills and distractibility, it is important to note that many of the items on this scale inquire about skills that also include broader executive functions (such as organization of materials and tasks, following directions, etc.) that are also known to be an area of risk in Trisomy X. While only assessed in the Denver subset, the BRIEF-2 also identified differences in executive functioning between the two adaptive skills groups. Interventions for executive functioning and attention have been shown to lead to improvement in adaptive outcomes in other groups such as ASD (Kenworthy et al., 2014) and similar interventions should be evaluated in Trisomy X to determine efficacy.

4.1 |. Timing of diagnosis: Prenatal versus postnatal ascertainment

The timing of diagnostic ascertainment (e.g., prenatal diagnosis vs. postnatal diagnosis) has been associated with differences in cognitive and adaptive outcomes in the sex chromosome trisomies. Specifically, within Trisomy X, prenatally ascertained females have higher mean IQ and adaptive functioning scores on the Vineland (Wigby et al., 2016). In the present study, we found that over 73% of females with age-appropriate adaptive skills were prenatally ascertained and the prenatal diagnosis group had higher composite scores of adaptive functioning on the BASC-2 and higher mean VIQ and GAI scores. Scatterplot figures of VIQ and GAI scores (Supplemental Figure S2c,d) were consistent with these findings and showed a degree of overlap among more impaired Trisomy X prenatal diagnosis females and the postnatal group. Females with postnatal Trisomy X diagnoses also had greater levels of impairment across multiple clinical domains, underscoring the importance of a comprehensive neuropsychological evaluation at the time of postnatal diagnosis to identify cooccurring psychiatric diagnoses.

Investigators have suggested several hypotheses for the improved outcomes in prenatally ascertained children with SCAs including environmental factors such as a higher parental education levels and higher socioeconomic status (Tartaglia, Cordeiro, et al., 2010; Tartaglia, Howell, et al., 2010), although in the present study, there were no significant differences in levels of maternal education. As many prenatally ascertained females with Trisomy X are identified incidentally (e.g., screening due to advanced maternal age), it is more likely that this group includes females who are asymptomatic or have only mild neurodevelopmental concerns. Furthermore, developmental surveillance due to a prenatally identified genetic condition likely results in earlier recognition and intervention when delays are recognized. In contrast, females who are ascertained by postnatal karyotype would have a higher a priori risk for poorer adaptive functioning as developmental or cognitive concerns are the most frequent indication for obtaining a postnatal karyotype.

It is important to note that over 40% of prenatally ascertained females in this sample had abnormal adaptive skills scores (n = 12/28), however the majority of scores fell within the “at risk” range and only 1 female had clinically significant adaptive skills impairments. Parents of these females reported higher levels of behavioral concerns (based on BSI scores). Prenatally diagnosed females with adaptive skills concerns also had more symptoms of executive functioning problems (reflected by higher BRIEF-2 global executive functioning score), however this did not withstand Bonferroni correction. These findings deserve further study in a larger prenatally ascertained cohort to better inform prenatal genetic counseling. Such a study would not only reflect more of the variability in outcomes of prenatally ascertained female but that which is due to true phenotypic heterogeneity in Trisomy X. However, the data highlight that prenatally diagnosed females with Trisomy X are also at risk for adaptive deficits and supports an association of mood with adaptive skills in the prenatal diagnosis group.

4.2 |. Limitations

We acknowledge the limitations of the present observational study. While we reported findings of the BASC-2 in 50 females with Trisomy X, only a subset had results of executive functioning (BRIEF-2) and social functioning/autistic traits (SRS-2) due to protocol differences between sites. Thus, we were limited in our ability to detect subtle differences in these domains that may influence adaptive skills. Reporting bias is also always a consideration when subjective questionnaires instruments are utilized to quantify behavioral traits and can contribute to correlations between questionnaire results. Direct measures of speech-language skills beyond verbal cognitive scores would also be an important component of a future study. Further, the data presented are cross-sectional across a large age range; a longitudinal, prospective study design would better define predictors of adaptive outcomes in adulthood.

Measures of adaptive functioning including the BASC-2 as well as gold-standard measures such as the Vineland and ABAS are all based on parent report of the child’ s symptoms and are therefore open to reporting bias. We also acknowledge that the BASC-2 ASC is not a commonly utilized measure of adaptive functioning compared to more comprehensive interview-based assessments such as the Vineland or parent-report measures such as the ABAS which have significantly more items. However, a large subset of our sample (N = 43) also had received one of these gold-standard assessments and significant correlations were found between the BASC-2 ASC and adaptive composite scores on the Vineland (r_s(19) = .79, p < .001) or ABAS (r_s(24) = .57, p = .003) versions administered. Thus, we also suggest that the BASC ASC continue to be explored as a proxy measure of overall adaptive skills considering the markedly decreased administration time and scoring efforts of the BASC compared to the more traditional measures.

There are likely other environmental or genetic factors influencing adaptive functioning in females with Trisomy X that were not directly assessed by this study. For example, phenotypic variability in other studies in SCAs has been shown to be associated with differences in copy number variations (Le Gall et al., 2017), early life stressors (Bender, Linden, & Robinson, 1987; van Rijn, Barneveld, Descheemaeker, Giltay, & Swaab, 2018), or familial learning disabilities (Samango-Sprouse, Stapleton, Sadeghin, & Gropman, 2013), although all of these studies would also benefit from validation in larger cohorts.

Due to marked phenotypic variability in Trisomy X, larger sample sizes allowing for more robust analyses are always ideal. One option for future study is the creation of an international open SCA dataset thereby allowing researchers to conduct studies using a larger sample than may be achieved at any one institution. Such a dataset would be useful to validate the preliminary findings of this study, further delineate differences based on timing, method of diagnostic ascertainment, and country of origin, and explore other phenotypic, environmental, and genetic factors contributing to adaptive outcomes in Trisomy X and other SCAs.

4.3 |. Clinical implications and future directions

Results of this exploratory study expands the current knowledge about neurobehavioral features in females with Trisomy X showing that internalizing behaviors, lower VIQ scores, deficits in functional communication, and inattentive symptoms are associated with poorer adaptive functioning. Screening for internalizing psychological symptoms (including depression, withdrawal, and anxiety), inattention, executive functioning, and social communication are important when considering treatment planning of adaptive functioning deficits in Trisomy X. Furthermore, interventions to promote the acquisition of self-expression, language skills, attention span, and social communication may improve adaptive skills development; however, further study with randomized controlled trials would be important to determine if these interventions are indeed effective. Future longitudinal studies should investigate the contributions of the child’ s neurobehavioral phenotype, genetic background, and environmental factors, as well as focus on evaluating different interventions to improve adaptive outcomes in the subset of females with Trisomy X who have adaptive challenges.