Phenylketonuria (PKU) is an autosomal recessive condition characterized by a deficiency in the phenylalanine hydroxylase enzyme that is necessary for the metabolism of the amino acid phenylalanine (phe) to tyrosine. PKU is associated with higher-than-normal levels of phe as well reductions in tyrosine and its metabolites such as dopamine and other catecholamines, in both the central nervous system and peripheral tissues (de Groot et al., Reference de Groot, Hoeksma, Blau, Reijngoud and van Spronsen2010). In addition, the excess phe competes with other available large neutral amino acids (e.g., tryptophan, valine, leucine) to cross the blood-brain barrier (Pardridge, Reference Pardridge1977; Shulkin et al., Reference Shulkin, Betz, Koeppe and Agranoff1995; de Groot et al., Reference de Groot, Hoeksma, Blau, Reijngoud and van Spronsen2010). As such, central levels of other noncatecholaminergic neurotransmitters (e.g., serotonin) are also impacted (Smith et al., Reference Smith, Momma, Aoyagi and Rapoport1987; Welsh et al., Reference Welsh, Pennington, Ozonoff, Rouse and McCabe1990). Elevated brain phe may also directly impact processes such as protein synthesis and myelin formation (Pardridge, Reference Pardridge1998; de Groot et al., Reference de Groot, Hoeksma, Blau, Reijngoud and van Spronsen2010).
If left untreated, PKU is associated with severe neurologic and cognitive impairment (Paine, Reference Paine1957; Bechar et al., Reference Bechar, Bornstein, Elian and Sandbank1965). Current practices encourage individuals with PKU to initiate a phe-restricted diet shortly after birth and to maintain these restrictions across one’s lifetime (Vockley et al., Reference Vockley, Andersson, Antshel, Braverman, Burton, Frazier, Mitchell, Smith, Thompson and Berry2014). Although individuals with early-treated PKU (ETPKU) experience much improved outcomes relative to those who do not undergo treatment, many still experience neurologic and cognitive disruptions.
The most widely reported neurologic finding among individuals with ETPKU is cortical white matter (WM) abnormalities. Indeed, a review by Anderson and Leuzzi (Reference Anderson and Leuzzi2010) estimated that WM abnormalities are detectable in over 90% of adults with ETPKU. These abnormalities are typically observed proximal to the posterior aspects of the lateral ventricles and extend more anteriorly with increased severity. The extent and severity of the WM abnormalities are positively correlated with age and blood phe levels (Hood et al., Reference Hood, Antenor-Dorsey, Rutlin, Hershey, Shimony, McKinstry, Grange, Christ, Steiner and White2015; Wesonga et al., Reference Wesonga, Shimony, Rutlin, Grange and White2016; Clocksin et al., Reference Clocksin, Hawks, White and Christ2021). More recent research suggests that gray matter structures (e.g., posterior cortex, putamen, cerebellum) may be affected in PKU as well (e.g., Bodner et al., Reference Bodner, Aldridge, Moffitt, Peck, White and Christ2012; Christ et al., Reference Christ, Price, Bodner, Saville, Moffitt and Peck2016; Aldridge et al., Reference Aldridge, Cole, Moffitt Gunn, Peck, White and Christ2020).
The majority of past neuropsychological studies of ETPKU have focused on children and adolescents. This literature supports a neurocognitive profile characterized by slight decrease in overall intellectual ability coupled with circumscribed impairments in particular domains including executive function, attention, processing speed, and fine motor control (for review, see Canton et al., Reference Canton, Gall, Feillet, Bonnemains and Roy2019). For example, Anderson et al. (Reference Anderson, Wood, Francis, Coleman, Anderson and Boneh2007) administered a neuropsychological battery to a sample of children with and without ETPKU. They found that the ETPKU group performed more poorly than the non-PKU group on full scale IQ as well as measures of processing speed, sustained attention, and memory. The ETPKU group also showed delays in their acquisition of achievement such as reading and arithmetic. Consistent with other studies, they found that task performance was negatively correlated with blood phe levels (an indicator of metabolic control) in many cases.
In comparison to the aforementioned childhood literature, much less is known regarding the neurocognitive and psychological profile of adults with ETPKU. Relatively recent changes in treatment guidelines (e.g., consensus support for treatment and maintenance of metabolic control throughout life – not just until person reaches young adulthood), a preliminary focus within the literature on the effects of phe level exposure in childhood, and the large number of adults lost to follow-up have contributed to a discrepancy in the number of studies focused on adult neurocognitive outcomes relative to those focused on pediatric outcomes (for further discussion, see Berry et al., Reference Berry, Brown, Grant, Greene, Jurecki, Koch, Moseley, Suter, van Calcar, Wiles and Cederbaum2013; Vockley et al., Reference Vockley, Andersson, Antshel, Braverman, Burton, Frazier, Mitchell, Smith, Thompson and Berry2014). In addition, research suggests potential developmental changes in the neurocognitive profile as children with ETPKU age. For example, White et al., (Reference White, Nortz, Mandernach, Huntington and Steiner2002) found that impairments in working memory performance were more evident in older as compared to younger children with ETPKU. Many other past studies (e.g., Moyle et al., Reference Moyle, Fox, Arthur, Bynevelt and Burnett2007; Nardecchia et al., Reference Nardecchia, Manti, Chiarotti, Carducci, Carducci and Leuzzi2015) have included small sample sizes coupled with wide age ranges, encompassing both adolescents and adults, thus making it difficult to disentangle potential developmental contributions to the observed neurocognitive profile. In terms of the available studies on adults with ETPKU, most have focused on a few select cognitive functions rather than representing a comprehensive evaluation across domains (e.g., Dawson et al., Reference Dawson, Murphy, Maritz, Chan, Ellerton, Carpenter and Lachmann2011; Weglage et al., Reference Weglage, Fromm, van Teeffelen-Heithoff, Möller, Koletzko, Marquardt, Rutsch and Feldmann2013). Taken together, these factors have contributed to a lack of consensus regarding the neurocognitive profile of adults with ETPKU and the impact of historic and concurrent metabolic control on cognitive functioning (for additional discussion of these issues, see Hofman et al., Reference Hofman, Champ, Lawton, Henderson and Dye2018). For example, based on data from nine studies, a meta-analysis by Bilder et al. (Reference Bilder, Noel, Baker, Irish, Chen, Merilainen, Prasad and Winslow2016) found that working memory was not significantly affected in adults with ETPKU. In contrast, a systematic review by Hofman et al. (Reference Hofman, Champ, Lawton, Henderson and Dye2018) covering data from sixteen studies concluded that working memory represented one of the most consistent deficits in adults with ETPKU.
In light of these challenges, more recent work with adults has endeavored to include more comprehensive neuropsychological batteries. For example, Palermo et al. (Reference Palermo, Geberhiwot, MacDonald, Limback, Hall and Romani2017) included 28 tasks spanning several cognitive domains including executive function, sustained attention, visuomotor coordination, language, and memory in a study of 37 adults with ETPKU. Similarly to the aforementioned childhood literature, the adult ETPKU group had significantly lower FSIQ than the non-PKU comparison group, suggesting that differences in overall intellectual ability may persist across development. In addition, adults with ETPKU performed worse on measures of sustained attention, processing speed, visuomotor skills, and higher-order executive function requiring flexibility and planning.
The researchers conducted additional analyses to examine the relationship between the cognitive data and phe levels during different developmental epochs (Romani et al., Reference Romani, Palermo, MacDonald, Limback, Hall and Geberhiwot2017). Although several cognitive areas (e.g., sustained attention) showed the strongest correlation with recent phe levels, others revealed vulnerability across multiple developmental epochs. For example, visual-motor coordination was correlated with recent, lifetime, adolescent, and adult phe levels. Yet still other domains such as complex executive function were most strongly related to phe levels during adolescence. These findings are consistent with the notion that the impact of phe levels on neuropsychological outcome varies based on developmental timing and cognitive domain.
The present study
The present study was designed to further this line of research and advance our understanding of the neurocognitive and psychological profile of adults with ETPKU. Within this context, we utilized a relatively new measure with emerging popularity in the cognitive assessment community, namely the NIH Toolbox (for full description, see Heaton et al., Reference Heaton, Akshoomoff, Tulsky, Mungas, Weintraub, Dikmen, Beaumont, Casaletto, Conway, Slotkin and Gershon2014). Briefly, the NIH Toolbox is a compilation of computerized assessments that can be flexibly administered either singly or as part of a larger battery to evaluate functional domains such as overall cognition or motor functions. Specific areas of cognition assessed by the NIH Toolbox include attention, episodic memory, processing speed, inhibitory control, and working memory. The NIH Toolbox has been used to successfully evaluate neurocognitive functioning among individuals with a diverse array of acute and chronic conditions (e.g., stroke, Carlozzi et al., Reference Carlozzi, Tulsky, Wolf, Goodnight, Heaton, Casaletto, Wong, Baum, Gershon and Heinemann2017; sickle cell disease, Hood et al., Reference Hood, King, Fields, Ford, Guilliams, Hulbert, Lee and White2019; alcohol use disorder, Meredith et al., Reference Meredith, Lim and Ray2020; among others). The present study represents one of the first to utilize the NIH Toolbox with individuals with ETPKU.
So as to provide a full comprehensive picture of neurocognitive functioning, we coupled the NIH toolbox with additional standardized measures designed to further assess overall intellectual ability, sustained attention, processing speed, verbal ability, executive functioning, visuospatial skills, and fine motor control. In order to capture psychological and day-to-day aspects of functioning, we also included report measures of executive function, attention, anxiety, depression, and social communication symptomatology.
Given the prior success with other clinical populations, we expect that the NIH Toolbox will provide a valid assessment of neurocognitive functioning among individuals with ETPKU, with findings in line with the more established measures included in our test battery. Specifically, based on past work (e.g., Palermo et al., Reference Palermo, Geberhiwot, MacDonald, Limback, Hall and Romani2017; Hofman et al., Reference Hofman, Champ, Lawton, Henderson and Dye2018), we hypothesize that adults with ETPKU will demonstrate particular difficulties in executive function and attention, motor skills, and processing speed. Conversely, we anticipate that verbal ability will be relatively intact/spared. We also anticipate that the magnitude of cognitive and psychological disruption will be correlated with blood phe levels (a marker of metabolic control).
Methods
Participants
A sample of 40 adults with ETPKU (18–36 years of age) and a demographically matched comparison group of 32 adults without PKU participated in the study. Demographic information for the participant sample is included in Table 1.
Demographic information
All participants with ETPKU were diagnosed at birth (33 classic PKU, 1 moderate PKU, 3 mild PKU; 2 hyperphenylalaninemia; 1 unspecified). [Note that PKU categorization is based on blood phe levels at the time of diagnosis (Classic: > 1200 μmol/L; Moderate: 900–1200 μmol/L; Mild: 600–900 μmol/L; Hyperphenylalaninemia: 120–600 μmol/L).] All ETPKU participants began a phe-restricted diet shortly after birth as indicated by medical records and/or patient report. Individuals with major medical conditions unrelated to PKU (e.g., diabetes) were excluded. Participants in the non-PKU group were recruited from the local community.
For participants with ETPKU, data on historic and recent plasma phe levels was compiled from available medical records and is included in Table 1. Lifetime phe level data was unavailable for three participants, so these individuals were only included in the recent and past year phe level analyses. Most recent phe level on record as well as mean phe level for the year preceding study participation were calculated. As described previously (Christ et al., Reference Christ, Abbene, Clocksin and Wegrzyn2021), the index of dietary control (the mean of all half-year median phe levels) was computed for early childhood (0–5 years), middle childhood (6–11 years), adolescence (12–17 years), adulthood (18+ years), and lifetime (0-present). See Supplementary materials for more details.
Materials and procedure
The present study was approved by the University of Missouri Internal Review Board (Review ID# 2011602) and was carried out in accordance with the provisions of the World Medical Association Declaration of Helsinki. Informed consent was obtained for all individuals prior to participation. Participants completed all tasks in a small, quiet room with adequate overhead lighting. Participants were given opportunities to take breaks in between tests.
A full list of administered measures and associated cognitive domains are included in Table 2 and more detailed task descriptions are included in the online Supplementary materials. In brief, the test battery consisted of both performance-based measures including the Wechsler Abbreviated Scale of Intelligence – Second Edition (WASI-2; Wechsler, Reference Wechsler1999), cognitive and motor batteries from the NIH Toolbox, select subtests from the Wechsler Adult Intelligence Scale – Fourth Edition (WAIS-IV; Wechsler, Reference Wechsler2008), Connors’ Continuous Performance Test – Third Edition (CPT-3), as well as self-report measures, including Behavior Rating Inventory of Executive Function – Adult Version (BRIEF-A), Connors’ Adult ADHD Rating Scales (CAARS), Social Responsiveness Scale – Second Edition (SRS-II), Beck Anxiety Inventory (BAI), and Beck Depression Inventory – Second Edition (BDI-2).
Tasks administered and neurocognitive domains assessed
Note. WASI-II = Wechsler Abbreviated Scale of Intelligence – Second Edition; WAIS-IV = Wechsler Adult Intelligence Scale – Fourth Edition; CPT-3 = Conners Continuous Performance Test – Third Edition.
Data processing & analysis
Performance-based measures
Each of the performance-based measures (see Table 2 for full list) yielded age-normed standard scores. [Note that normative scores are not generated automatically for the Oral Symbol Digit and Walk Gait subtests of the NIH Toolbox. For these, scores were calculated manually using the normative data reported in the NIH Toolbox Technical Manual.] So as to allow for easy comparison as well as calculation of composite domain scores (see below), normed scores for all performance-based measures were converted to a standard score scale with a mean of 100 and a standard deviation of 15. Also, where applicable, the performance-based scores were adjusted such that a higher score reflected better performance.
In a procedure similar to Romani et al. (Reference Romani, Palermo, MacDonald, Limback, Hall and Geberhiwot2017), the standard scores of contributing measures were averaged to generate composite scores in the areas of verbal ability, visuospatial skills, executive functioning, motor skills, and processing speed. Table 2 includes a list of measures included in each composite. For both the performance-based and self-report measures, to protect against an increase in the likelihood of a Type I error related to multiple comparisons, we confirmed a statistically significant finding on one of the broader composite scores before proceeding to analyze the contributing individual subtest scores.
Self-report measures
All of the self-report measures generated normed t scores (M = 50, SD = 10) in which a higher score reflected worse/more symptomatology (Beck & Steer, Reference Beck and Steer1993; Beck et al., Reference Beck, Steer and Brown1996; Roth et al., Reference Roth, Isquith and Gioia2005; Constantino, Reference Constantino2012). Scores generated from self-report measures were not modified due to their common scale. Similar to the approach taken with the performance-based measures, a composite score reflecting everyday difficulties with executive function and attention was computed based on overall scores from the BRIEF-A (Global Executive Composite; GEC) and CAARS (Overall ADHD Index). Scores from the BAI and BDI-2 were averaged to create a composite score representing symptomatology related to internalizing problems. Given that the SRS-2 captures a unique aspect of everyday functioning relative to the other measures, it was analyzed separately.
For both the performance-based and self-report measures, to protect against an increase in the likelihood of a Type I error related to multiple comparisons, we confirmed a statistically significant finding on one of the broader composite scores before proceeding to analyze the contributing individual subtest scores. A false discovery rate (FDR; q < .05) approach was used to correct for multiple comparisons associated with the correlation analyses between scores and phe levels.
Statistical approach
The primary statistical approach was an independent samples t-test with Hedges g as a measure of effect size. The normative scores for the reported measures were already corrected for age, and the majority corrected for sex as well. (Scores from the WASI-II and WAIS-IV were corrected for age only.) As such, these factors were not included as covariates. Correlations between scores and phe level measurements were also calculated.
Results
Mean composite scores are listed in Table 3 and illustrated in Figures 1 and 2. Two participants (one from each group) did not complete the CPT-3 due to technical problems. NIH Toolbox motor data were unavailable for three ETPKU participants. As such, they were excluded from analysis of the associated composite score.
Means and standard deviations for composite neurocognitive and psychological measures
Note. SRS-II = Social Responsiveness Scale – Second Edition.
* Degrees of Freedom: Visuospatial Skills and Motor Skills = 67; all others = 70.
Mean composite scores shown separately for cognitive domain and group (ETPKU and non-PKU). Higher composite scores are indicative of better performance. A dashed line is included at a standard score of 100 for reference. *p < .05; **p < .005.
Mean t scores shown separately for each self-report measure and group (ETPKU and non-PKU). Higher t scores are indicative of increased difficulties and symptomatology. A dashed line is included at a T score of 50 for reference. *p < .05; **p < .00.
Performance-based neurocognitive measures
Overall intellectual functioning
No group differences were observed in overall intellectual ability as estimated using the WASI-2 [M PKU = 101; M non-PKU = 104; t(70) = 1.16, p = .250, g = .27], although individuals with ETPKU obtained a significantly lower Fluid Cognition Composite score on the NIH Toolbox than individuals without PKU [M PKU = 98; M non-PKU = 108; t(70) = 2.64, p = .010, g = .62].
Verbal ability
The groups performed similarly on the verbal ability composite [M PKU = 99; M non-PKU = 98; t(70) = .27, p = .789, g = .06].
Visuospatial skills
Individuals with ETPKU demonstrated worse performance on the visuospatial composite as compared to individuals without PKU [M PKU = 101; M non-PKU = 110; t(67) = 3.44, p = .001, g = .82]. Further analysis confirmed that both tests (Block Design and Visual Puzzles) contributed to this finding [t(67) > 2.35, p < .022, g > .56, in both instances].
In order to rule out that the finding was driven by secondary task demands related to processing speed and executive function, a post hoc hierarchical linear regression was conducted. Scores on the visuospatial skills composite served as the dependent variable. Scores on the executive function and processing speed composites were entered in the first step of the model, followed by group (PKU and non-PKU) in the second step. Importantly, after accounting for variance related to executive function and processing speed, there continued to be a significant effect of group [F(1,65) = 5.49, p = .022, pr 2 = .08].
Executive function
The ETPKU group scored significantly lower on the executive function composite as compared to the non-PKU group [M PKU = 98; M non-PKU = 102; t(70) = 2.38, p = .020, g = .56]. Subsequent analysis of contributing tests revealed that the finding was primarily driven by group differences on measures of working memory [List Sort Working Memory; t(70) = 3.47, p = .001, g = .82] and sustained attention [CPT Omissions; t(39) = 2.09, p = .043, g = .46]. No statistically significant group differences were found on the remaining tests [t(66) < 1.23, p > .22, g < .29, in all instances].
Motor skills
Individuals with ETPKU performed worse on the motor skills composite than individuals without PKU [M PKU = 96; M non-PKU = 100; t(67) = 2.42, p = .018, g = .58]. Further analysis of contributing subtests showed that this difference was driven primarily by the grooved pegboard task with participant’s dominant hand as well as the simple pegboard task with participant’s nondominant hand [t(70) > 2.52, p < .014, g > .59, in both instances]. A trend toward significance was found on the grooved pegboard task with participant’s nondominant hand [t(60) = 1.89, p = 0.06, g = .42]. Group differences for the simple pegboard task with participant’s dominant hand did not reach significance [t(70) < 1, p = 0.49, g = .16]. A trend toward significance was also observed on a few measures of gross motor skills, including walk endurance and walk gait, with individuals with ETPKU performing worse than individuals without PKU [t(67) > 1.76, p < 0.082, g > .42, in both instances]. No statistically significant group differences were found for any of the other contributing tests [t(67) < 1.01, p > .31, g < .24, in all instances].
Processing speed
The ETPKU group demonstrated significantly lower scores on the processing speed composite as compared to the non-PKU group [M PKU = 106; M non-PKU = 111; t(70) = 2.23, p = .029, g = .52]. Group differences were most apparent on the WAIS-IV Symbol Search subtest that contributed to this composite [t(70) = 3.21, p = .002, g = .75]. Participants with ETPKU also demonstrated lower scores on the NIH Toolbox Pattern Comparison subtest, although this difference did not reach significance [t(70) = 1.81, p = .074, g = .43]. No other significant findings were observed for contributing tests [t(66) < 1.46, p > .14, g < .35, in all instances].
Self-report measures
The ETPKU group scored significantly higher (poorer) on the executive function and attention composite measure as compared to the non-PKU group [M PKU = 49; M non-PKU = 43; t(70) = 3.15, p = .002, g = .74]. Subsequent analysis confirmed that scores from both the BRIEF-A and CAARS contributed to this finding. The ETPKU group demonstrated significantly elevated GEC, MI, and BRI composite scores on the BRIEF as compared to the non-PKU group [t(69) > 3.03, p < .003, g > .71, in all instances]. With regards to contributing BRIEF clinical scales, group differences were most apparent on the Working Memory clinical scale [M PKU = 56; M non-PKU = 47; t(67) = 4.10, p < .001, g = .92]. Significant group differences were also observed on the Inhibition, Shifting, Emotional Control, and Organization of Materials clinical scales [t(65) > 2.45, p < .017, g > .57 in all instances]. A trend toward significance was observed on the remaining clinical scales [t(70) > 1.76, p < .083, g > .41, in all instances].
On the CAARS, the ETPKU group reported significantly more attention-related symptomatology in general as evidenced by group differences on the Overall ADHD Index [M PKU = 47; M non-PKU = 42; t(70) = 2.46, p = .016, g = .58]. Group differences were also observed on Total DSM-IV ADHD Symptoms, DSM-IV Inattentive Symptoms, DSM-IV Hyperactive/Impulsive Symptoms, Inattention/Memory Problems, Impulsivity/Emotional Lability, and the Hyperactivity/Restlessness subscales [t(65) > 2.13, p < .037, g > .48, in all instances]. Group differences approached significance on the Problems with Self-Concept subscale [t(70) = 1.92, p = .059, g = .45].
No evidence of group differences were observed for the internalizing symptom composite based on the BAI and BDI-2 [M PKU = 52; M non-PKU = 51; t(70) < 1, p = .46, g = .17]. The ETPKU group also did not report significantly more symptoms of social communication problems as compared to the non-PKU group on the SRS-2 [M PKU = 53; M non-PKU = 50; t(70) = 1.55, p = .13, g = .36].
Clinical significance
Despite statistically significant groups differences across multiple performance-based and self-report measures, the associated mean scores for the ETPKU group were only modestly elevated relative to the available normative data. To investigate this issue further, we conducted additional analyses to examine the potential clinical significance of the observed elevations. Using the established criteria of 1.5 standard deviations above the normative mean as an indication of a “clinically elevated” score, we conducted chi-squared tests to compare the observed frequency of individuals with and without ETPKU who demonstrated a clinically elevated score on one or more of the performance-based composite scores. The frequency of individuals with elevated scores was significantly greater for the ETPKU group (18%) as compared to the non-PKU group (6%) [χ2(1, 40) = 8.64, p = .003]. Analysis of the self-report data from the executive function (ETPKU = 10%; non-PKU = 3%) and internalizing symptom composite scores (ETPKU = 15%; non-PKU = 3%) yielded similar results [χ2(1, 40) > 6.41, p < .011 in both instances].
Relationship with blood phe levels
Correlations between composite scores and phe level measurements are listed in Tables 4 and 5 and illustrated in supplemental Figures S1 and S2. Scatter plots of the data are also included in supplemental Figures S3 and S4. All p values noted below have been corrected for multiple comparisons using a FDR approach (q < .05).
Correlations between metabolic markers and task performance
Note: IDC = Index of Dietary Control (mean of all half-year median phe levels).
*p < .05 FDR corrected; **p < .005 FDR corrected; p < .05 uncorrected indicated in bold.
Correlations between metabolic markers and self-report measures
Note: IDC = Index of Dietary Control (mean of all half-year median phe levels).
*p < .05 FDR corrected; **p < .005 FDR corrected; p < .05 uncorrected indicated in bold.
Performance-based neurocognitive measures
As can be seen in Table 4 and supplemental Figure S1, in general, task performance was strongly correlated with recent and adulthood phe levels [r = .37–73, p < .05, in all instances] but less so for earlier levels [0–5 years old: r = .14-.25, p > .200; 6–11 years old: r = .20–30, p > .131]. This was true across all of the cognitive domains assessed. Within this context, executive function had the most robust relationship to recent phe levels [r = .64, p < .001] whereas verbal ability had the least significant [r = .37, p = .037].
Self-report measures
Results for the self-report executive function composite was generally consistent with findings for the performance-based cognitive measures. Self-reported symptoms were robustly correlated with adolescent and adult phe levels [r > .55, p < .005, in both instances] but less so for levels from early and middle childhood [r < .26, p > .161, in both instances]. A different pattern was found for measures of psychoemotional functioning. Self-reported symptoms of internalizing symptoms showed robust correlations with recent phe levels as well as those from middle childhood and adolescence [r > .42, p < .023, in all instances] with no significant relationship with early childhood or adult phe levels [r < .34, p > .110, in both instances]. The self-report measure of social communication correlated robustly with adolescent and most recent phe levels [r > .54, p < .004, in all instances] and less so with earlier levels [r < .32, p > .082, in both instances].
Discussion
The majority of past studies on neuropsychological functioning and ETPKU have focused on children and adolescents. Much less is known regarding the neurocognitive and psychological profile of adults with ETPKU. The present study aimed to provide a comprehensive evaluation of neurocognitive and psychoemotional functioning among adults with ETPKU using the NIH Toolbox and other standardized measures. In addition, we examined the relationship between task performance and historic blood phe levels (i.e., a marker of metabolic control) during different developmental epochs.
In the present study, ETPKU-related impairments in executive function were apparent on both performance and self-report measures. Subsequent analysis confirmed that the areas of working memory and sustained attention appeared to be particularly affected, with group-related differences on the executive function composite being driven primarily by performance on the NIH Toolbox’s List Sort Working Memory subtest and the Connors’ Continuous Performance Test (i.e., a test of sustained attention). Whereas the broader literature on working memory in adults with ETPKU has yielded somewhat mixed findings (for review, see Bilder et al., Reference Bilder, Noel, Baker, Irish, Chen, Merilainen, Prasad and Winslow2016), the present results are consistent with findings on the Cambridge Neuropsychological Test Automated Battery (CANTAB), another computerized assessment tool. For example, Bik-Multanowski et al., (Reference Bik-Multanowski, Pietrzyk and Mozrzymas2011) reported that adults with ETPKU performed below expectations (based on normative data) on tests of spatial working memory and sustained attention from CANTAB.
The present ETPKU group also scored significantly worse than the non-PKU group on self-report measures of executive dysfunction and attentional symptomatology (i.e., the BRIEF and CAARS, respectively). Group differences on the BRIEF were most apparent on the Working Memory clinical scale. This finding is in line with the results of a recent study of BRIEF performance in a large sample (n = 286) of children and adults with ETPKU, which found that the Working Memory scale showed the largest effect size, with 19% of children and 29% of adults scoring in the “abnormally elevated” range (Christ et al., Reference Christ, Clocksin, Burton, Grant, Waisbren, Paulin, Bilder, White and Saville2020). In terms of the present CAARS data, scores for the ETPKU group were elevated relative to the non-PKU group in terms of both inattentive and hyperactive symptomatology. Antshel & Waisbren, (Reference Antshel and Waisbren2003) also reported elevated ADHD-related symptomatology in a sample of children with ETPKU. Interestingly, in contrast to the present findings, the aforementioned study found that children with ETPKU were more likely to demonstrate inattentive symptoms (as opposed to hyperactive/impulsive symptoms). Additional research is needed to better understand potential age-related changes in the manifestation of ADHD-related symptomatology in individuals with ETPKU.
ETPKU-related impairments were also observed on the motor skills composite. This finding was driven primarily by group differences on measures of fine motor skills (e.g., grooved pegboard and simple pegboard) more so than those assessing gross motor ability (e.g., grip strength, walk endurance). The presence of disruptions in fine motor control has been widely reported in past studies of both children and adults with ETPKU (for reviews, see Moyle et al., Reference Moyle, Fox, Arthur, Bynevelt and Burnett2007; Christ et al., Reference Christ, Abbene, Clocksin and Wegrzyn2021). Recent research from Christ et al., (Reference Christ, Abbene, Clocksin and Wegrzyn2021) suggests that disruptions in motor learning likely contribute to the observed fine motor difficulties. In addition, neurological studies in ETPKU have identified structural and/or functional disruptions in several brain regions (e.g., prefrontal cortex, parietal cortex, basal ganglia, cerebellum) known to play an important role in fine motor control and learning (Bodner et al., Reference Bodner, Aldridge, Moffitt, Peck, White and Christ2012; Christ et al., Reference Christ, Price, Bodner, Saville, Moffitt and Peck2016, Reference Christ, Abbene, Clocksin and Wegrzyn2021; Aldridge et al., Reference Aldridge, Cole, Moffitt Gunn, Peck, White and Christ2020). Taken together, these findings reinforce the need for future research to evaluate the relationship between neurological sequelae and observed disruptions in motor skills among individuals with ETPKU.
We also found that the ETPKU group performed more poorly on measures of processing speed as compared to the non-PKU group, a finding that is also consistent with the existing literature (Janzen & Nguyen, Reference Janzen and Nguyen2010; Palermo et al., Reference Palermo, Geberhiwot, MacDonald, Limback, Hall and Romani2017; Aitkenhead et al., Reference Aitkenhead, Krishna, Ellerton, Moinuddin, Matcham, Shiel, Hossain, Kiffin, Foley, Skeath, Cleary, Lachmann and Murphy2021). From a theoretical perspective, it is hypothesized that disruptions in structural and functional brain connectivity likely contribute to the processing speed deficits frequently observed in individuals with ETPKU (e.g., Anderson & Leuzzi, Reference Anderson and Leuzzi2010; Christ et al., Reference Christ, Moffitt, Peck, White and Hilgard2012; Clocksin et al., Reference Clocksin, Hawks, White and Christ2021) although research directly confirming this link is still needed.
Visuospatial skills represent a relatively understudied cognitive construct in individuals with ETPKU, with the few studies examining visuospatial skills reporting discrepant results (Brumm et al., Reference Brumm, Azen, Moats, Stern, Broomand, Nelson and Koch2004; Channon et al., Reference Channon, German, Cassina and Lee2004; Janzen & Nguyen, Reference Janzen and Nguyen2010). The present study found that the ETPKU group performed significantly worse on the visuospatial skills composite than the non-PKU group. It can be difficult with many visuospatial tests to disentangle the potential contribution of secondary demands (e.g., executive function, fine motor control, processing speed) to task performance. Within this context, significant group differences remained even after statistically accounting for the potential contribution of such secondary demands. Also, it is noteworthy that we observed group differences on the WAIS-IV Visual Puzzles subtest, a motor-free visuospatial test with minimal speed demands (i.e., participants just had to respond within 30 s to each item) that requires participants to view a completed puzzle and select three pieces that together would reconstruct the puzzle. Neuroimaging studies of ETPKU have documented irregularities in parietal and occipital regions and related WM tracts (i.e., superior longitudinal fasciculus) known to play an important role in visuospatial processing (Christ et al., Reference Christ, Price, Bodner, Saville, Moffitt and Peck2016; Clocksin et al., Reference Clocksin, Hawks, White and Christ2021).
The present study also examined the relationship between task performance and historic blood phe levels (i.e., a marker of treatment adherence) during different developmental epochs (e.g., early childhood, middle childhood, adolescence, adulthood, lifetime, and recent). The results point to a similar finding across multiple cognitive domains: task performance was strongly correlated with adolescent, adulthood, and recent phe levels but less so for levels from early and middle childhood. A similar pattern was observed for the self-report measures of executive function and attention. These findings are generally in line with those of previous studies on adults with ETPKU (e.g., Channon et al., Reference Channon, German, Cassina and Lee2004; Jahja et al., Reference Jahja, Huijbregts, de Sonneville, Meere, Legemaat, Bosch, Hollak, Rubio-Gozalbo, Brouwers, Hofstede, Vries, Janssen, Ans, Langendonk and Spronsen2017; Romani et al., Reference Romani, Palermo, MacDonald, Limback, Hall and Geberhiwot2017; Christ et al., Reference Christ, Clocksin, Burton, Grant, Waisbren, Paulin, Bilder, White and Saville2020; Aitkenhead et al., Reference Aitkenhead, Krishna, Ellerton, Moinuddin, Matcham, Shiel, Hossain, Kiffin, Foley, Skeath, Cleary, Lachmann and Murphy2021). Of note, upon initial glance, the lack of a strong correlation between task performance and early childhood phe levels may seem surprising given the substantial literature linking early treatment and cognitive outcome (e.g., Anderson et al., Reference Anderson, Wood, Francis, Coleman, Anderson and Boneh2007; Waisbren et al., Reference Waisbren, Noel, Fahrbach, Cella, Frame, Dorenbaum and Levy2007; Canton et al., Reference Canton, Gall, Feillet, Bonnemains and Roy2019). We postulate that this finding (or lack thereof) is likely a consequence of insufficient variability in our sample’s early phe levels necessary for significant correlations to emerge.
In addition to cognitive difficulties, there is growing evidence that individuals with ETPKU may also experience a higher prevalence of mental health concerns including anxiety and depression (Pietz et al., Reference Pietz, Fätkenheuer, Armbruster, Esser and Schmidt1997; Manti et al., Reference Manti, Nardecchia, Chiarotti, Carducci, Carducci and Leuzzi2016; Bilder et al., Reference Bilder, Kobori, Cohen-Pfeffer, Johnson, Jurecki and Grant2017). Consistent with past research (e.g., ten Hoedt et al., Reference Ten Hoedt, de Sonneville, Francois, Ter Horst, Janssen, Rubio-Gozalbo, Wijburg, Hollak and Bosch2011; Clacy et al., Reference Clacy, Sharman and McGill2014; Jahja et al., Reference Jahja, Huijbregts, de Sonneville, Meere, Legemaat, Bosch, Hollak, Rubio-Gozalbo, Brouwers, Hofstede, Vries, Janssen, Ans, Langendonk and Spronsen2017), we found significant positive correlations between phe levels and internalizing symptom levels. The finding of robust correlations with phe levels across middle childhood and adolescence (6–17 years) underscores the importance of this epoch in the development of emotional processing and regulation (Weir et al., Reference Weir, Zakama and Rao2012). Additional work is needed to better understand not only the nature of the mental health concerns associated with ETPKU but also the potential interaction between psychological and cognitive difficulties in this population. Indeed, past research with other clinical populations (e.g., generalized anxiety disorder) suggests that the two phenomena may be linked, with increased anxiety contributing to poorer cognitive performance (Balderston et al., Reference Balderston, Vytal, O’Connell, Torrisi, Letkiewicz, Ernst and Grillon2017).
Lastly, elevated phe levels (particularly during adolescence and later ages) were also associated with increased difficulties with social communication. This is not surprising given that several of the previously discussed cognitive and psychological abilities (e.g., executive function, processing speed, emotional processing) that are adversely affected by high phe levels contribute to proficient social communication. The result is also generally consistent with past findings of poorer social-cognitive functioning and perspective taking in individuals with higher phe levels as compared to those with lower phe levels (Jahja et al., Reference Jahja, van Spronsen, de Sonneville, van der Meere, Bosch, Hollak, Rubio-Gozalbo, Brouwers, Hofstede, de Vries, Janssen, van der Ploeg, Langendonk and Huijbregts2016; Palermo et al., Reference Palermo, MacDonald, Limback, Robertson, Howe, Geberhiwot and Romani2020).
Limitations and future directions
The present study represents one of the few comprehensive studies of neuropsychological and psychological functioning in adults with ETPKU. Looking forward, it will be important for future studies to include larger sample sizes with a broader/older age range. State-sponsored early screening programs for PKU began in the mid-to-late 1960s, which means that the oldest cohort of ETPKU patients are currently approaching their early to mid-50s. (The oldest participants in the present study were in their mid-30s.) Within this context, critical questions for the field moving forward include “What does it mean to grow older with PKU in terms of physical, neurologic, and psychological health? Also, what role does metabolic control (both historic and current) play in moderating health risks in this older cohort?”
Consistent with previous studies, we found a significant relationship between current phe levels and task performance, with particularly robust correlations with the executive functioning composite. To the extent that effective treatment management and adherence is associated with cognitive/executive demands, there is a potential cyclical relationship between worsening treatment adherence and poorer executive function. Future research incorporating treatment intervention and longitudinal data collection will be critical for understanding the cause/effect nature of this relationship.
As noted earlier, the present study represents one of the first to incorporate the NIH Toolbox for the study of ETPKU. Our results are promising and support the potential utility of the NIH Toolbox as a relatively quick assessment tool for individuals with ETPKU. This represents only the first step, however, towards fully validating the NIH Toolbox for use with individuals with ETPKU. It will be important for future studies to include larger sample sizes with a broader age range (i.e., including children), to directly compare performance on the NIH Toolbox with that on more established neuropsychological tests assessing the same cognitive constructs, and to also evaluate test-retest reliability.
Summary and conclusions
The present study provides a comprehensive view of neuropsychological functioning in adults with ETPKU and its relationship to historic and recent phe levels. Between-group comparisons revealed ETPKU-related weaknesses in visuospatial skills, executive function, motor skills, and processing speed. Within the ETPKU group, phe levels (particularly recent levels) were significantly correlated with performance across most cognitive domains and aspects of psychological functioning (e.g., anxiety & depression symptoms; attentional difficulties). Taken together, these findings support the presence of a circumscribed pattern of cognitive difficulties in adults with ETPKU and reinforce the importance for continued metabolic control throughout the lifespan.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1355617722000686
Funding statement
The present research was supported by a grant from the National PKU Alliance.
Conflicts of interest
S.E.C. has received consulting fees and/or honoraria from BioMarin Pharmaceutical Inc., Synlogic Therapeutics, and Horizon Therapeutics. BioMarin has also funded past research in S.E.C’s laboratory. The other authors report no competing interests to declare.
