Neuropsychological functions, sleep, and mental health in adults with Klinefelter syndrome

A few studies have examined neuropsychological functions, sleep, and mental health combined in Klinefelter syndrome (KS; 47,XXY). We investigated neuropsychological functions with standard tests, sleep with actigraphy, and self‐reported mental health in 30 men with KS (Mean age = 36.7 years) compared to 21 controls (Mean age = 36.8 years). Men with KS scored significantly lower on mental speed, attention span, working memory, inhibition, and set‐shifting tests, as well as overall IQ (mean effect size difference Cohen's d = 0.79). Men with KS had significantly longer night wakes, with no differences in other sleep variables (mean d = 0.34). Men with KS reported poorer mental health than controls (mean d = 1.16). Regression analyses showed neuropsychological functions explained variance in some sleep domains for men with KS but not for controls. Neuropsychological functions explained variance in some mental health domains for controls. For men with KS, however, verbal IQ was the only significant predictor of mental health. Altogether, men with KS display problems in neuropsychological functions and mental health but do not appear different from controls on most sleep parameters. Our findings indicate that relations between neuropsychological functions, sleep, and mental health differ between men with KS and controls.

facing men with KS include increased risk of developing metabolic syndrome and/or diabetes Type 2 Ishikawa, Yamaguchi, Kondo, Takenaka, & Fujisawa, 2008).
In terms of cognitive functioning, full-scale IQ scores are generally within the normal range, with considerable individual variation (Leggett et al., 2010). However, the IQ profile is commonly skewed toward significantly lower verbal IQ relative to performance IQ . Compared to controls, specific cognitive challenges that have been identified more frequently among men with KS include receptive and expressive language abilities (Ross et al., 2012;Ross, Zeger, Kushner, Zinn, & Roeltgen, 2009). The increased prevalence of reading and writing difficulties for men with KS is likely associated with the broader language problems (Boone et al., 2001;Rovet, Netley, Keenan, Bailey, & Stewart, 1996;Stewart, Bailey, Netley, Rovet, & Park, 1986). A specific area of concern among men with KS is problems with executive functions, which refers to cognitive control processes necessary for goal-directed behavior and problem solving (e.g., organization, planning, judgment, decision-making; Gravholt et al., 2018). A common division of core executive functions encompasses three theoretically derived components: Inhibitory control, working memory, and mental set shifting, which are all found to play important roles in learning/memory and educational achievements (e.g., Miyake et al., 2000;Miyake & Friedman, 2012). All are viewed as basic and moderately correlated control functions that are critical for higher-order executive functions.
The current study utilized the three-component model, as it seems highly relevant for understanding the cognitive difficulties associated with KS. We also included tests of mental efficiency/speed and attention, because studies have reported that men with KS have problems with these cognitive domains as well (Fales et al., 2003;Kompus et al., 2011;Ross et al., 2008;Temple & Sanfilippo, 2003; van Rijn & Swaab, 2015).
With regard to psychological functioning, men with KS have an increased risk of experiencing general psychological distress, as well as higher prevalence rates of depression, anxiety disorders, attention deficit hyperactivity disorder, autism spectrum disorder, and schizophrenia (Giagulli et al., 2019;Skakkebaek, Wallentin, & Gravholt, 2015). The risk of being admitted to a psychiatric ward has been estimated as 3.65 the risk of the general population (Bojesen, Juul, Birkebaek, & Gravholt, 2006). Up to 45% of KS samples have shown psychotic symptoms (Bruining, Swaab, Kas, & van Engeland, 2009). Depression rates among men with KS have ranged from 19 to 69% (Boks et al., 2007;Turriff, Levy, & Biesecker, 2011).
In terms of socioeconomic functioning, Danish Registry data have shown poorer socioeconomic functioning for men with KS compared to other men, among others via low education and low lifetime income . In a larger Australian study of 87 adult males with one or more surplus X chromosomes, 22% reported to be unemployed or on benefit pensions (Herlihy et al., 2011). A Danish registry study showed that the mean retirement age among 903 men with KS was 46 years, compared to 60 years in the control group.
Over a third of men with KS were retired, compared to 20% of controls (Bojesen et al., 2011).
Despite the growing documentation of challenges in multiple areas of functioning among men with KS, there are considerable knowledge gaps. Sleep is an area with practically no systematic knowledge regarding men with KS. This is surprising, as many of the documented challenges for men with KS, including neuromuscular problems, endocrinological problems, and low socioeconomic status, are all associated with poor sleep in the general population (Joiner, 2016;Walker & Stickgold, 2005;Walker & van der Helm, 2009). A study of 53 adults with KS showed that these men had considerably poorer self-reported sleep compared to normative data (Fjermestad & Stokke, 2018). Although this finding is concerning, it is important to note that there is considerable evidence of limited overlap between subjectively and objectively measured sleep (Mezick, Wing, & McCaffery, 2014

| METHODS
The study was approved by the Regional committees for medical and health research ethics-South-Eastern Norway. All participants provided written informed consent prior to participation. All participants took part in a draw for a universal gift certificate (≈100 USD), one for men with KS and one for controls. Participants were not compensated in other ways, but travel and accommodation costs were covered and food and snacks were served during test days.

| Sample and recruitment
The KS sample comprised 30 men with KS aged 18-60 years (M [mean] age = 36.7 years, SD = 10.6). They were recruited from multiple nonclinical settings, that is, the user registry of Frambu resource center for rare disorders, a national (nonclinical) advisory center for rare disorders, the annual Klinefelter syndrome user association meeting, and an online video ad posted on various websites, including the Klinefelter syndrome user association website, and various rare disorders-oriented online forums.
The control sample comprised 21 men without KS aged 18-65 years (M age = 36.8 years, SD = 14.4). Controls were recruited from multiple settings, that is, ads in local newspapers; an online video ad posted on various websites, including the Klinefelter syndrome user association website and various rare disorders-oriented online forums; and the social network of male KS participants (note that there was only one family relation, i.e., a cousin). See Table 1 for background information on both samples.

| Procedures
Neuropsychological testing took place at the Department of Psychology at the University of Oslo, Norway. Test administers were a team of clinical-program psychology students trained by a specialist in clinical neuropsychology with >20 years of clinical neuropsychological experience. Participants completed the mental health questionnaire on site. They were given an actigraphy watch toward the end of the test appointment, along with verbal instructions about how to use it.
The actigraphy watches were returned by regular mail in envelopes with prepaid postage. All participants received a written report summarizing their neuropsychological profile (percentiles), mental health profile, and sleep data. The reports were followed up with a telephone consultation with the PI, who is a Clinical Psychologist. Men with KS who showed elevated mental health scores were assisted with clinical referrals when needed. The current study is part of a larger trial and the neuropsychological data were gathered after participants had been through resting-state structural MR imaging and performed auditive tasks while undergoing electroencephalogram recordings (data not reported here).

| Digit span
We used the Digit Span from WAIS-IV to measure auditory attention span and working memory capacity. The participant listens while the test administrator reads a series of digits aloud; the first set starting with only two digits. Two lists are presented at each set size. If the participant accurately reproduces at least one digit list, then the set size is increased by one digit. Testing terminates when the participant fails to reproduce both digit lists in a set. We ran all three conditions of the Digit Span task, that is, forward span (e.g., 4-2-6 repeated as 4-2-6), backward span (e.g., 4-2-6 repeated as 6-2-4), and sequencing, in which the participant is to repeat the digits in increasing order (e.g., 4-2-6 repeated as 2-4-6). The backward and sequencing conditions place concurrent demands on short-term memory storage and manipulation of information with increasing working memory load. The three tasks are scored separately and combined to a total score.

| Trail making test
We used three conditions of the Trail Making Test (TMT) to measure processing speed and cognitive set-shifting (Delis et al., 2001). The participant is instructed to draw a line between circles containing numbers and letters distributed across an A3-sized sheet of paper, as quickly as possible without lifting the pencil from the paper. In Divorced/separated 2 (6.7) 1 (4.7) Note: KS, Klinefelter syndrome. *Frequency distribution difference is significant at the p < .05 level (chi-square).
draw the line from the first to the last letter in alphabetical order (A-P), ignoring numbers. The Number-Letter Switching condition (TMT-4) requires cognitive set shifting in addition to processing speed. It follows the same principles as the Number and Letter Sequencing conditions (TMT-2 and 3), but the participant is now to draw lines to connect the circles in an ascending pattern alternating between the numbers and letters (i.e., 1-A-2-B-3-C and so on). Time in seconds is recorded, but there is no maximum time limit. If the participant makes a mistake, the test administrator points out the mistake and the participant is allowed to correct the mistake and continue without stopping time taking.

| Color-word interference test
The Color-Word Interference Test (CWIT) was employed to measure processing speed, inhibitory control, and set-shifting (Delis et al., 2001).
It contains four conditions with different demands on executive control. In the Color Naming condition (CWIT-1) the participant is to name different color patches, whereas the Word Reading condition (CWIT-2) requires reading words in black ink that denote colors. Color Naming and Word Reading (CWIT-1 and 2) are believed to index language-mediated processing speed that is demanded for the following conditions placing additional demands on cognitive control processes.
The subsequent Inhibition condition (CWIT-3), which is an extension of the classic Stroop test (Stroop, 1935), measures interference control and response inhibition. Finally, the Inhibition/Switching condition (CWIT-4) has the added requirement of set shifting. As in the Inhibition condition (CWIT-3), color words are printed in an inconsistent color ink, and some of the color words are boxed in. The participant is instructed to name the color of the ink instead of reading the word if the word is not boxed in, but to read the word instead of name the color of the ink if the word is boxed in (i.e., alternating between two task rules). All four task conditions are timed and the participant is instructed to work as fast as possible without making mistakes. Task completion time in seconds, as well as the number of errors, is recorded.

| Sleep assessment
We assessed sleep and circadian rhythm with an actigraph, a device used to measure and record motion over a period of time. The actigraph (Actiwatch Spectrum Plus, Phillips, the Netherlands) is integrated into a small wristwatch. The actigraphy raw data are reviewed and predefined algorithms are used to analyze the recorded data.
Actigraphy is commonly used to predict sleep/wake patterns and cir-

| Mental health
We used the Hopkins Symptom Checklist-90-Revised (SCL-90-R; Derogatis, 1994) as a screening measure of mental health. The SCL-90-R is a 90-item self-report questionnaire where items describing various mental health symptoms experienced during the last 7 days are rated 0 (not at all) to 4 (very much). The SCL-90-R comprises a global severity index as well as nine subscale scores (i.e., somatization; obsessive-compulsive symptoms; interpersonal sensitivity, depression, anxiety; hostility; phobic anxiety; paranoid ideation; and psychoticism). The SCL-90-R has documented psychometric properties and is widely used in adult mental health services (Siqveland & Leiknes, 2016). In the current sample, internal consistency (Cronbach's α) was excellent for men with KS (α = .97) and controls (α = .92).

| Data analytic plan
In addition to descriptive analyses, we ran independent samples ttests to examine if variables related to cognitive functions, sleep, and mental health were different between men with KS and controls. We investigated potential differences in subdomains of IQ in men with KS with paired samples t-tests Effect size differences were calculated as Cohen's d using the formula (M Group1 -M Group2 )/SD pooled (Cohen, 1992). We interpreted effect sizes using the following criteria: Small effect 0.2 ≤ d < 0.5, medium: 0.5 ≤ d < 0.8, large: d ≥ 0.8 (Cohen, 1988

| Group differences in cognitive functions
In terms of information processing speed, men with KS scored significantly lower than controls on all measures, and irrespective of oral

| Group differences in sleep domains
Men with KS had significantly longer night wakes than controls (Tables 2 and 3). There were no other significant differences in any of the other sleep variables. The pattern was the same when we excluded males with KS with IQ scores <70.

| Group differences in mental health domains
Men with KS reported significantly higher levels of mental health problems on the SCL-90 Global Severity Index and all SCL-90-R subscales compared to controls (Figure 2; Table 2). The pattern was the same when we excluded participants with IQ scores <70.
The highest scores for men with KS were on the domains somatization, obsessive-compulsive symptoms, and anxiety, which also represent the areas with the largest difference compared to controls.

| Executive functions and sleep
We ran a series of multiple regression models to examine if the neuro-

| Executive functions and mental health
We ran a series of multiple regression models to examine if the neuro- Men with KS scored poorer than controls across several of the tested neuropsychological domains, including information processing speed, auditory attention span, and the key executive functions working memory (one out of two tests), interference control/inhibition, and set-shifting. The mean scores were, with a few exceptions, not larger than 1 SD from the normative mean. The findings are in line with previous studies (e.g., Kompus et al., 2011;Ross et al., 2008;van Rijn & Swaab, 2015). The overall effect size differences were borderline medium to large. In terms of overall IQ, the effect size difference between the samples was large, however, with a particularly large discrepancy for verbal IQ. This finding concurs with previous studies, but the 17-point difference in mean total IQ points between men with KS and controls in our sample is more pronounced than the average finding of a 10-point difference across studies . It is important to note that the men with KS in our sample had a total IQ close to the population average of 100, and performance IQ slightly above average. The latter finding indicates that nonverbal analysis and reasoning abilities are not impacted in our KS sample. Taken together, our results align with existing evidence demonstrating that adult men with KS have lower results than controls in several neuropsychological domains, including verbal IQ, despite close to average total and performance IQ. It is important to note, however, that the controls in the current study were in the upper range on many domains, and that men with KS, although they scored lower than controls, were within the normal range on many tests. The large difference between verbal and performance IQ for the KS group is also important to note. There is accumulating evidence that the typical gap between verbal and T A B L E 2 Effect size (d) differences in neuropsychological functions, sleep, and self-reported psychological health between 30 men with Klinefelter syndrome and 21 controls performance IQ observed in the KS population emerges in childhood, and that the difference sometimes disappears in adulthood . In our sample, there was a significant difference in adults with KS, indicating specific language problems.
A novel aspect of the current study was the inclusion of objectively measured sleep over 7 consecutive days. We found no difference between men with KS and controls on any sleep domain, except the duration of the night wakes. This is in contrast to a previous study showing that men with KS self-reported poorer sleep than male controls (Fjermestad & Stokke, 2018). Generally, studies tend to find limited overlap between subjectively and objectively measured sleep (Mezick et al., 2014). Our findings may reflect that although men with KS think they sleep poorer than other men, whereas objective data does not corroborate their subjective experience. However, it may also be that the subjective experience of sleep is different for men with KS. It is also important to note that all males with KS, except two, in the current study received testosterone supplementation, which positively impacts sleep, reducing sleep length and improving quality of sleep (Shigehara et al., 2018). Note, however, that a recent review found little evidence that testosterone positively impacts cognitive functions in men (Buskbjerg, Gravholt, Dalby, Amidi, & Zachariae, 2019). Previous studies have demonstrated increased negative effect and neuroticism traits among men with KS (Skakkebaek et al., 2017(Skakkebaek et al., , 2018a(Skakkebaek et al., , 2018b In terms of self-reported mental health, men with KS had significantly elevated scores on all domains compared to controls, with large effect sizes in all domains except the two most severe (paranoid ideation and psychoticism), on which effect size differences were low. Previous research has demonstrated an increased prevalence of also the most severe mental health problems, for example, psychosis (Bojesen, Kristensen, et al., 2006;Bruining et al., 2009).
Although the symptom levels on these severe mental health problems were higher than controls also for the current sample, it is important to note that the largest effect size differences were identified for somatization, obsessive-compulsive symptoms, and anxiety. The mental health global severity index, as well as subscales somatization, and obsessive-compulsive symptoms were above clinical cutoff. This provides direction for health professionals in which mental health problems are most essential to target. It is important to note that the SCL-90-R obsessive-compulsive subscale indicates concentration and memory problems along with compensatory behavioral strategies (e.g., controlling own work/doing things slowly to ensure correctness) more than typical obsessive-compulsive traits, and that it has been found to reflect subjectively experienced cognitive difficulties (Siqveland & Leiknes, 2016). The SCL-90 obsessive compulsive scale has been found to poorly predict obsessivecompulsive behaviors and to overlap more with general anxiety (Woody, Steketee, & Chambless, 1995). It is also important to note that the high scores on the somatization scale may reflect KSrelated physical issues and not psychosomatic difficulties. Our findings regarding mental health largely align with previous studies , and it seems clear that mental health problems in men with KS is an area in urgent need of tailored interven- found to be associated with executive functions (Skakkebaek et al., 2017). Another study of the same KS sample found that the personality trait neuroticism, which is typically inter-related with depression and anxiety, was associated with attention switching, which is an important aspect of executive function (Skakkebaek, Moore, Pedersen, et al., 2018). However, that study used self-reported attention switching and personality measures, so common-rater variance would be likely to enhance the overlap (Shirk, Reyes, & Cristotomo, 2013 The current study has limitations. The sample size is small, which is a common problem for rarely diagnosed disorders such as KS. A major limitation is that the control sample had higher education and a larger percentage was working. Although this does limit the validity of our comparison, it does also reflect the reality of the difference between men with KS and their same-age peers . An additional limitation concerns recruitment. Although recruitment from nonclinical settings has potential advantages in terms of representativeness of less severe cases, being the member of a patient advocacy user group and/or being registered at a specialist resource center for rare disorders may represent other forms of ascertainment bias. Another limitation is that we only used a self-reported measure of mental health. A previous study found that the personality trait neuroticism was the main predictor of depression and anxiety in men with KS, and that this pattern was not evident for controls . Therefore, negative personality bias may have influenced our mental health results. The inclusion of personality variables may have added to the interpretability of our results. We also lack information concerning the use of medications for mental health and/or sleep problems among our participants. In addition to potentially influencing results, medication use may also have been different between the groups, and such information should be included in future studies.
The use of actigraphy measures provided novel information about sleep domains for men with KS. In particular, the objectivity of this sleep measure represents an important advantage given the potential self-report problems for men with KS due to language problems.
However, there are also limitations with the use of actigraphy, compared to even more advanced objective sleep measures such as polysomnography. For instance, actigraphy cannot be used to measure sleep stages, which would provide important information regarding sleep disturbance (Smith et al., 2018). That being said, the use of actigraphy provides insights into "everyday life" sleep patterns while being cost-effective. There is no need for participants to travel, and the only costs involved are personnel costs in calibrating the watches and extracting data. The current study shows using actigraphy is feasible when studying men with KS. Future studies should consider longer registration periods (e.g., 14 days) to obtain an even broader assessment.
The main practice implications from the current article are: (a) Health professionals should assess neuropsychological functioning and mental health in men with KS, or refer men with KS to such services. A thorough case-based neuropsychiatric profile is essential to tailor current and future care plans, given the multiple challenges experienced by men with KS . (b) When assessing sleep in men with KS, which should be done due to previous evidence of poor subjective sleep (Fjermestad & Stokke, 2018), and general documentation of the importance of sleep quality for other areas of functioning (Joiner, 2016), it seems important that professionals add an objective sleep measure if possible. There are several sleep registration apps available that will provide some frame of reference for objective sleep, if actigraphy equipment is not an option.
Objectively measuring sleep is important because men with KS may subjectively report poorer sleep than what is evident from the objective measures. Thus, a (potential) discrepancy can be used in clinical consultations to address sleep expectations, beliefs, and habits. (c) As there may be associations between neuropsychological functions, sleep, and mental health for men with KS at the group level, and of course for the individual with KS, it is extremely important that health services are not compartmentalized but coordinated as well as target the many domains that add up to determine overall functioning and quality of life functioning for men with KS.