Emotional and behavioral problems, quality of life and metabolic control in NTBC-treated Tyrosinemia type 1 patients

In total, 31 TT1 patients (19 males) were included in this cross-sectional study (mean age 13.9 ± 5.3 years) (Table 1). Patients <six years and/or patients who received liver transplantation were excluded. All patients were treated with NTBC (typically at a dose of 1 mg/kg body weight/day) and a phenylalanine-tyrosine-restricted diet. Eleven patients had (or still) received phenylalanine supplementation at the time of assessment because phenylalanine concentrations were below the recommended limit of 30 μmol/L [12]. Patients were included from different centers in the UK, Belgium, and the Netherlands between October 2012 and August 2018. Plasma phenylalanine and tyrosine concentrations were collected from patient records. The study was approved by the Medical Ethical Committees of the participating centers. All patients and/or parents gave informed consent to participate in this study.

To assess emotional and behavioral problems, the Achenbach System of Empirically Based Assessment (ASEBA) questionnaires were used [15, 16]. The Child Behavior Checklist (CBCL), suitable for age 6–12 years, was completed by parents. Adolescents (13–17 years) and adults (≥18 years) filled out the Youth Self Report (YSR) or the Adult Self Report (ASR). The CBCL, YSR, and ASR measure emotional and behavioral problems in seven different dimensions: withdrawn/depressed, somatic complaints, anxious/depressed, thought problems, attention problems, rule-breaking behavior, and aggressive behavior. Additionally, the CBCL and YSR assess social problems, whereas the ASR assesses intrusive behavior. The sum of scores on dimensions withdrawn/depressed, somatic complaints, and anxious/depressed forms the broadband scale ‘internalizing problems’, while the broadband scale ‘externalizing problems’ is based on the rule-breaking, aggressive, and (in case of the ASR) intrusive behavior dimensions [15, 16]. Furthermore, six Diagnostic and Statistical Manual of Mental Disorders (DSM) IV oriented scales were used. Different scales were used for the CBCL and YSR (affective, anxiety, somatic, attention deficit hyperactivity, oppositional, and conduct problems), and the ASR (depressive, anxiety, somatic, avoidant personality, attention deficit hyperactivity, and antisocial personality problems). For all ASEBA questionnaires, higher scores indicate more problems in that domain.

To assess HR-QoL, the TNO AZL Children’s Quality Of Life (TACQOL) questionnaire was used for children aged < 16 years [17]. For analysis, this questionnaire was divided into two parts: 1) for children aged 8–11 and 2) for children aged 12–15 years. Both questionnaires included 63 items and were filled out by the parents. For children aged 8–11, the TACQOL questionnaire consisted of seven different dimensions measuring physical functioning/complaints (‘body’), motor functioning (‘motor’), independent daily functioning (‘autonomy’), cognitive functioning and school performances (‘cognition’), social contacts with parents and peers (‘social’), occurrence of positive moods (‘positive emotions’), and of negative moods (‘negative emotions’). For children aged 12–15, the dimensions autonomy and social were combined, measuring interaction with peers (‘peers’). For participants aged ≥16 years, the TNO AZL Adult’s Quality Of Life (TAAQOL) questionnaire was used; this consisted of 45 items. The TAAQOL included 12 different dimensions, i.e. gross motor functioning (‘gross motor’), fine motor functioning (‘fine motor’), cognitive functioning (‘cognition’), sleep quality (‘sleep’), pain (‘pain’), social functioning (‘social’), independent daily functioning (‘daily activities’), sexuality (‘sex’), vitality (‘vitality’), positive moods (‘happiness’), depressive moods (‘depressive’), and angry moods (‘anger’). For all questionnaires, a higher score indicated a better HR-QoL.

The HR-QoL data of TT1 patients was compared to a Dutch reference population divided into 3 reference groups of participants aged 8–11 years (N = 548), aged 12–15 (N = 393), and aged 16–30 years (N = 394).

Patient data were collected from clinical records to investigate differences within the TT1 group. Additionally, all phenylalanine and tyrosine concentrations were retrieved from the clinical patient records. Dried blood spot samples, when stated as such in the patient files, were excluded in order to only include venous samples in our analysis. In general, three to four venous samples were obtained every year. Based on the existing literature on associations between metabolic control and behavioral outcomes in TT1 and phenylketonuria we focused on phenylalanine and tyrosine levels during three periods in life: the first year after birth, throughout lifetime (until assessment), and the last year before assessment [5, 8, 13, 18, 19]. Since some plasma phenylalanine concentrations < 30 μmol/L were unspecified, imputation was used to allow inclusion of these concentrations in analyses. For the imputation, median phenylalanine concentrations < 30 μmol/L were calculated using available data from other patients in whom the plasma concentrations < 30 μmol/L were specified. Lifetime concentrations were calculated as the median of yearly median phenylalanine and tyrosine levels.

For the analysis of emotional and behavioral problems, raw ASEBA-scores were converted into T-scores to allow comparison to the general population and between different groups of TT1 patients. Descriptive analyses were performed using pre-defined T-scores indicating scores in the normal range (50–64), borderline range (65–69) and clinical range (≥70) for all different subdomains of the ASEBA questionnaires, and borderline range (60–63), and clinical range (64–100) for internalizing problems and externalizing problems [15, 16]. The borderline range indicates that scores are not clearly deviant from norm scores, but high enough to be of concern. Scores in the clinical range indicate a deviation from norm scores, and therefore possible clinical importance which may warrant intervention. For the assessment of HR-QoL, TT1 patients were compared with Dutch reference data using Mann-Whitney U tests. Kruskal Wallis tests and/or Mann-Whitney U tests were performed to analyze differences within the TT1 group, comparing pre-symptomatically versus symptomatically diagnosed patients, and symptomatically diagnosed patients diagnosed at different ages (< 2 months, 2–6 months or > 6 months of age [2]). Spearman correlation tests were performed to study associations between behavior problems and HR-QoL, and between behavior problems and HR-QoL-outcomes on the one hand and plasma phenylalanine and tyrosine concentrations on the other. The broadband ASEBA-scales were excluded from these last correlational analyses since these were calculated using the subdomains and therefore significant results would be expected when subdomains were statistically significant. All statistical analyses were performed using IBM SPSS Statistics 22nd version and p < 0.05 was considered statistically significant.

Table 2 summarizes the findings of the ASEBA questionnaires. When considering internalizing problems, 47% of children, 22% of adolescents, and 14% of adults with TT1 scored within the borderline or clinical range. When considering externalizing problems, 54% of children, 22% of adolescents and no adults scored within the borderline or clinical range. TT1 children showed problems in all domains, but especially with respect to attention problems, with 53% being in clinical range. Adolescents and adults especially showed thought and attention problems, respectively. When investigating differences between the different age groups using Kruskal-Wallis analysis, children showed significantly more attention problems than adults (p = 0.025). There were no significant differences in behavior problems observed between symptomatically and pre-symptomatically diagnosed patients.

Table 2 also shows the descriptive findings on the DSM-IV oriented scales of the ASEBA. TT1 patients reported problems on most scales, especially affective, somatic, attention deficit hyperactivity, and conduct issues. Although not statistically significant, scores within the borderline or clinical range were generally observed more frequently in children than in adolescents and adults.

Figure 1a shows the results of the HR-QoL questionnaires of TT1 children compared to the Dutch reference population. The TACQOL questionnaire for children aged 8–11 showed a significantly lower HR-QoL for the domains autonomy (p = 0.011), cognition (p = 0.029) and social (p = 0.039) in TT1 patients. When comparing the TACQOL scores for TT1 children aged 12–15 with the corresponding reference population, a significantly lower HR-QoL was observed on the domain body (p = 0.014) (Fig. 1b). For both age groups (8–11 and 12–15), analyses of the TACQOL showed no significant differences between pre-symptomatically and symptomatically diagnosed patients, and no differences between the different times of diagnosis.

Figure 1c shows that adult TT1 patients (N = 9) reported a significantly lower HR-QoL in the domains gross motor (0.016), fine motor (p < 0.001), cognition (p = 0.001), and vitality (p = 0.046). When comparing pre-symptomatically (N = 3) and symptomatically (N = 6) diagnosed patients, a significant difference was observed on the domain vitality (p = 0.024), indicating a lower HR-QoL in pre-symptomatically diagnosed patients. When comparing groups based on age at diagnosis, no significant differences were observed.

HR-QoL and emotional/behavioral problems were often significantly correlated (see Additional files 1 and 2).

As phenylalanine is a metabolic precursor of tyrosine, a relationship between phenylalanine and tyrosine concentrations is expected. Phenylalanine and tyrosine concentrations during the first year of life were, indeed, positively correlated (ρ = 0.532; p = 0.007), as were the concentrations during the last year (ρ = 0.631; p < 0.001). Lifetime concentrations showed a positive trend (ρ = 0.329; p = 0.094).

Significant negative correlations were observed between first year phenylalanine concentrations and somatic complaints (ρ = − 0.421; p = 0.040), social problems (ρ = − 0.457; p = 0.043), thought problems (ρ = − 0.430; p = 0.040), and attention problems (ρ = − 0.513; p = 0.010). Furthermore, negative correlations were observed between first year tyrosine concentrations and social problems (ρ = − 0.608; p = 0.004), attention problems (ρ = − 0.598; p = 0.002), delinquent behavior (ρ = − 0.417; p = 0.043), and aggressive behavior (ρ = − 0.438; p = 0.032). This indicates that low phenylalanine (and the associated lower tyrosine) concentrations during the first year of life were associated with both internalizing and externalizing behavior problems.

The DSM-IV oriented scales showed negative correlations between first year tyrosine concentrations and affective problems (ρ = − 0.449; p = 0.047), attention deficit hyperactivity problems (ρ = − 0.493; p = 0.014), and conduct problems (ρ = − 0.667; p = 0.001).

Results for the HR-QoL in children, showed a positive correlation between first year plasma phenylalanine concentrations and autonomy (ρ = 0.609; p = 0.016). In adults, a negative correlation was observed between first year phenylalanine concentrations and cognition (ρ = − 0.943; p = 0.005). Results of the correlation analyses are also summarized in Additional file 3.

No significant correlations were observed between results on the ASEBA questionnaires and lifetime plasma phenylalanine and tyrosine concentrations. Correlational analyses with the HR-QoL questionnaires showed a positive correlation between lifetime phenylalanine concentrations and autonomy (ρ = 0.517; p = 0.034) in children, indicating more autonomy with higher (normal rather than low) phenylalanine concentrations. Furthermore, negative correlations were observed between lifetime phenylalanine concentrations and cognition (ρ = − 0.829; p = 0.021) and social contacts (ρ = − 0.802; p = 0.030) in patients aged > 16 years, indicating better HR-QoL with lower phenylalanine concentrations (also see Additional file 3).

A significant correlation was observed between the last year phenylalanine concentrations and the score in the withdrawn/depressed domain (ρ = 0.411; p = 0.030) indicating more problems with higher phenylalanine concentrations. The last year tyrosine concentrations also correlated with the withdrawn/depressed score (ρ = 0.492; p = 0.008), indicating more problems with higher tyrosine concentrations. When investigating the DSM-oriented ASEBA scales, higher phenylalanine concentrations correlated with increased affective problems (ρ = 0.418; p = 0.047) and anxiety problems (ρ = 0.381; p = 0.045). These findings indicate that higher (usually normal) phenylalanine concentrations and the associated higher tyrosine concentrations are associated with more problems.

Results for the HR-QoL in children, showed a significant negative correlation between the last year tyrosine concentrations and positive emotions (ρ = − 0.505; p = 0.033), indicating more positive emotions with lower (towards normal rather than high) tyrosine levels. In adult patients, the last year phenylalanine concentrations negatively correlated with the score in the social domain (ρ = − 0.802; p = 0.030) (also see Additional file 3).